ScipyLectures Simple

SciKits Numpy Matplotlib
2017
SciPy Python EDITION
IP[y]:
Cython IPython
Scipy Edited by
Gal Varoquaux
Lecture Notes Emmanuelle Gouillart

Olaf Vahtras
www.scipy-lectures.org
Gal Varoquaux Emmanuelle Gouillart Olav Vahtras

Christopher Burns Adrian Chauve Robert Cimrman Christophe Combelles
Pierre de Buyl Ralf Gommers Andr Espaze Zbigniew Jdrzejewski-Szmek
Valentin Haenel Gert-Ludwig Ingold Fabian Pedregosa Didrik Pinte
Nicolas P. Rougier Pauli Virtanen
an d man y ot her s...
Contents
I Getting started with Python for science 2

1 Python scientific computing ecosystem 4
1.1 Why Python? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.2 The Scientific Python ecosystem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.3 Before starting: Installing a working environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.4 The workflow: interactive environments and text editors . . . . . . . . . . . . . . . . . . . . . . . . 8
2 The Python language 12

2.1 First steps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.2 Basic types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.3 Control Flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2.4 Defining functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.5 Reusing code: scripts and modules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
2.6 Input and Output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
2.7 Standard Library . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
2.8 Exception handling in Python . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
2.9 Object-oriented programming (OOP) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
3 NumPy: creating and manipulating numerical data 47

3.1 The NumPy array object . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
3.2 Numerical operations on arrays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
3.3 More elaborate arrays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
3.4 Advanced operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
3.5 Some exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
3.6 Full code examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
4 Matplotlib: plotting 97
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
4.2 Simple plot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
4.3 Figures, Subplots, Axes and Ticks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
4.4 Other Types of Plots: examples and exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
4.5 Beyond this tutorial . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
4.6 Quick references . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
5 Scipy : high-level scientific computing 184

5.1 File input/output: scipy.io . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
5.2 Special functions: scipy.special . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186
5.3 Linear algebra operations: scipy.linalg . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186
i
5.4 Interpolation: scipy.interpolate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188
5.5 Optimization and fit: scipy.optimize . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188
5.6 Statistics and random numbers: scipy.stats . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194
5.7 Numerical integration: scipy.integrate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196
5.8 Fast Fourier transforms: scipy.fftpack . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198
5.9 Signal processing: scipy.signal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201
5.10 Image manipulation: scipy.ndimage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203
5.11 Summary exercises on scientific computing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208
5.12 Full code examples for the scipy chapter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221
6 Getting help and finding documentation 259
II Advanced topics 262

7 Advanced Python Constructs 264
7.1 Iterators, generator expressions and generators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 265
7.2 Decorators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 269
7.3 Context managers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277
8 Advanced NumPy 281

8.1 Life of ndarray . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 282
8.2 Universal functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 295
8.3 Interoperability features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 304
8.4 Array siblings: chararray, maskedarray, matrix . . . . . . . . . . . . . . . . . . . . . . . . . . . . 307
8.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 310
8.6 Contributing to NumPy/Scipy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 310
9 Debugging code 314

9.1 Avoiding bugs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 315
9.2 Debugging workflow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 317
9.3 Using the Python debugger . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 318
9.4 Debugging segmentation faults using gdb . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 323
10 Optimizing code 325

10.1 Optimization workflow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 326
10.2 Profiling Python code . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 326
10.3 Making code go faster . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 328
10.4 Writing faster numerical code . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 330
11 Sparse Matrices in SciPy 332

11.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 332
11.2 Storage Schemes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 334
11.3 Linear System Solvers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 346
11.4 Other Interesting Packages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 351
12 Image manipulation and processing using Numpy and Scipy 352

12.1 Opening and writing to image files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 353
12.2 Displaying images . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 355
12.3 Basic manipulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 356
12.4 Image filtering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 359
12.5 Feature extraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 364
12.6 Measuring objects properties: ndimage.measurements . . . . . . . . . . . . . . . . . . . . . . . . 366
12.8 Examples for the image processing chapter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 371
13 Mathematical optimization: finding minima of functions 397

13.1 Knowing your problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 398
13.2 A review of the different optimizers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 400
13.4 Examples for the mathematical optimization chapter . . . . . . . . . . . . . . . . . . . . . . . . . . 406
ii
13.5 Practical guide to optimization with scipy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 432
13.6 Special case: non-linear least-squares . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 435
13.7 Optimization with constraints . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 437
13.9 Examples for the mathematical optimization chapter . . . . . . . . . . . . . . . . . . . . . . . . . . 438
14 Interfacing with C 439

14.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 439
14.2 Python-C-Api . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 440
14.3 Ctypes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 445
14.4 SWIG . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 449
14.5 Cython . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 453
14.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 458
14.7 Further Reading and References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 458
14.8 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 458
III Packages and applications 460

15 Statistics in Python 462
15.1 Data representation and interaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 463
15.2 Hypothesis testing: comparing two groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 468
15.3 Linear models, multiple factors, and analysis of variance . . . . . . . . . . . . . . . . . . . . . . . . 471
15.4 More visualization: seaborn for statistical exploration . . . . . . . . . . . . . . . . . . . . . . . . . . 476
15.5 Testing for interactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 480
15.6 Full code for the figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 481
15.7 Solutions to this chapters exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 502
16 Sympy : Symbolic Mathematics in Python 505

16.1 First Steps with SymPy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 506
16.2 Algebraic manipulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 507
16.3 Calculus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 508
16.4 Equation solving . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 510
16.5 Linear Algebra . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 511
17 Scikit-image: image processing 513

17.1 Introduction and concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 514
17.2 Input/output, data types and colorspaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 515
17.3 Image preprocessing / enhancement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 517
17.4 Image segmentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 521
17.5 Measuring regions properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 523
17.6 Data visualization and interaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 524
17.7 Feature extraction for computer vision . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 525
17.9 Examples for the scikit-image chapter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 526
18 Traits: building interactive dialogs 538

18.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 539
18.2 Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 540
18.3 What are Traits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 540
19 3D plotting with Mayavi 557

19.1 Mlab: the scripting interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 558
19.2 Interactive work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 564
19.3 Slicing and dicing data: sources, modules and filters . . . . . . . . . . . . . . . . . . . . . . . . . . . 565
19.4 Animating the data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 568
19.5 Making interactive dialogs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 569
19.6 Putting it together . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 571
20 scikit-learn: machine learning in Python 572
iii
20.1 Introduction: problem settings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 573
20.2 Basic principles of machine learning with scikit-learn . . . . . . . . . . . . . . . . . . . . . . . . . . 577
20.3 Supervised Learning: Classification of Handwritten Digits . . . . . . . . . . . . . . . . . . . . . . . 582
20.4 Supervised Learning: Regression of Housing Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . 586
20.5 Measuring prediction performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 589
20.6 Unsupervised Learning: Dimensionality Reduction and Visualization . . . . . . . . . . . . . . . . 594
20.7 The eigenfaces example: chaining PCA and SVMs . . . . . . . . . . . . . . . . . . . . . . . . . . . . 597
20.8 Parameter selection, Validation, and Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 603
20.9 Examples for the scikit-learn chapter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 610
Index 652
iv
Scipy lecture notes, Edition 2017.1
Contents 1
Part I
Getting started with Python for science
2
This part of the Scipy lecture notes is a self-contained introduction to everything that is needed to use Python
for science, from the language itself, to numerical computing or plotting.
3
CHAPTER 1
Python scientific computing ecosystem
Authors: Fernando Perez, Emmanuelle Gouillart, Gal Varoquaux, Valentin Haenel
1.1 Why Python?
1.1.1 The scientists needs
Get data (simulation, experiment control),

Manipulate and process data,
Visualize results, quickly to understand, but also with high quality figures, for reports or publications.
1.1.2 Pythons strengths
Batteries included Rich collection of already existing bricks of classic numerical methods, plotting or
data processing tools. We dont want to re-program the plotting of a curve, a Fourier transform or a fitting
algorithm. Dont reinvent the wheel!
Easy to learn Most scientists are not payed as programmers, neither have they been trained so. They
need to be able to draw a curve, smooth a signal, do a Fourier transform in a few minutes.
Easy communication To keep code alive within a lab or a company it should be as readable as a book
by collaborators, students, or maybe customers. Python syntax is simple, avoiding strange symbols or
lengthy routine specifications that would divert the reader from mathematical or scientific understand-
ing of the code.
Efficient code Python numerical modules are computationally efficient. But needless to say that a very
fast code becomes useless if too much time is spent writing it. Python aims for quick development times
and quick execution times.
4
Universal Python is a language used for many different problems. Learning Python avoids learning a
new software for each new problem.
1.1.3 How does Python compare to other solutions?
Compiled languages: C, C++, Fortran. . .
Pros
Very fast. For heavy computations, its difficult to outperform these languages.
Cons
Painful usage: no interactivity during development, mandatory compilation steps, ver-
bose syntax, manual memory management. These are difficult languages for non pro-
grammers.
Matlab scripting language
Pros
Very rich collection of libraries with numerous algorithms, for many different domains.
Fast execution because these libraries are often written in a compiled language.
Pleasant development environment: comprehensive and help, integrated editor, etc.
Commercial support is available.
Cons
Base language is quite poor and can become restrictive for advanced users.
Not free.
Julia
Pros
Fast code, yet interactive and simple.
Easily connects to Python or C.
Cons
Ecosystem limited to numerical computing.
Still young.
Other scripting languages: Scilab, Octave, R, IDL, etc.
Pros
Open-source, free, or at least cheaper than Matlab.
Some features can be very advanced (statistics in R, etc.)
Cons
Fewer available algorithms than in Matlab, and the language is not more advanced.
Some software are dedicated to one domain. Ex: Gnuplot to draw curves. These pro-
grams are very powerful, but they are restricted to a single type of usage, such as plotting.
1.1. Why Python? 5

Python
Pros
Very rich scientific computing libraries
Well thought out language, allowing to write very readable and well structured code: we
code what we think.
Many libraries beyond scientific computing (web server, serial port access, etc.)
Free and open-source software, widely spread, with a vibrant community.
A variety of powerful environments to work in, such as IPython, Spyder, Jupyter note-
books, Pycharm
Cons
Not all the algorithms that can be found in more specialized software or toolboxes.
1.2 The Scientific Python ecosystem
Unlike Matlab, or R, Python does not come with a pre-bundled set of modules for scientific computing. Below
are the basic building blocks that can be combined to obtain a scientific computing environment:
Python, a generic and modern computing language

The language: flow control, data types (string, int), data collections (lists, dictionaries), etc.
Modules of the standard library: string processing, file management, simple network protocols.
A large number of specialized modules or applications written in Python: web framework, etc. . . . and
scientific computing.
Development tools (automatic testing, documentation generation)
See also:
chapter on Python language
Core numeric libraries
1.2. The Scientific Python ecosystem 6

Numpy: numerical computing with powerful numerical arrays objects, and routines to manipulate
them. http://www.numpy.org/
See also:
chapter on numpy
Scipy : high-level numerical routines. Optimization, regression, interpolation, etc http://www.scipy.org/
See also:
chapter on scipy
Matplotlib : 2-D visualization, publication-ready plots http://matplotlib.org/
See also:
chapter on matplotlib
Advanced interactive environments:

IPython, an advanced Python console http://ipython.org/
Jupyter, notebooks in the browser http://jupyter.org/
1.2. The Scientific Python ecosystem 7

Domain-specific packages,
Mayavi for 3-D visualization
pandas, statsmodels, seaborn for statistics
sympy for symbolic computing
scikit-image for image processing
scikit-learn for machine learning
and much more packages not documented in the scipy lectures.
See also:
chapters on advanced topics
chapters on packages and applications
1.3 Before starting: Installing a working environment
Python comes in many flavors, and there are many ways to install it. However, we recommend to install a
scientific-computing distribution, that comes readily with optimized versions of scientific modules.
Under Linux
If you have a recent distribution, most of the tools are probably packaged, and it is recommended to use your
package manager.
Other systems
There are several fully-featured Scientific Python distributions:
Anaconda
EPD
WinPython
Python 3 or Python 2?
In 2008, Python 3 was released. It is a major evolution of the language that made a few changes. Some old
scientific code does not yet run under Python 3. However, this is infrequent and Python 3 comes with many
benefits. We advise that you install Python 3.
1.4 The workflow: interactive environments and text editors
Interactive work to test and understand algorithms: In this section, we describe a workflow combining inter-
active work and consolidation.
Python is a general-purpose language. As such, there is not one blessed environment to work in, and not only
one way of using it. Although this makes it harder for beginners to find their way, it makes it possible for Python
to be used for programs, in web servers, or embedded devices.
1.4.1 Interactive work
We recommend an interactive work with the IPython console, or its offspring, the Jupyter notebook. They are
handy to explore and understand algorithms.
1.3. Before starting: Installing a working environment 8

Under the notebook
To execute code, press shift enter
Start ipython:
In [1]: print('Hello world')

Hello world
Getting help by using the ? operator after an object:
In [2]: print?
Type: builtin_function_or_method
Base Class: <type 'builtin_function_or_method'>
String Form: <built-in function print>
Namespace: Python builtin
Docstring:
print(value, ..., sep=' ', end='\n', file=sys.stdout)
Prints the values to a stream, or to sys.stdout by default.

Optional keyword arguments:
file: a file-like object (stream); defaults to the current sys.stdout.
sep: string inserted between values, default a space.
end: string appended after the last value, default a newline.
See also:
IPython user manual: http://ipython.org/ipython-doc/dev/index.html
Jupyter Notebook QuickStart: http://jupyter.readthedocs.io/en/latest/content-quickstart.html
1.4.2 Elaboration of the work in an editor
As you move forward, it will be important to not only work interactively, but also to create and reuse Python
files. For this, a powerful code editor will get you far. Here are several good easy-to-use editors:
Spyder: integrates an IPython console, a debugger, a profiler. . .
PyCharm: integrates an IPython console, notebooks, a debugger. . . (freely available, but commercial)
Atom
Some of these are shipped by the various scientific Python distributions, and you can find them in the menus.
As an exercise, create a file my_file.py in a code editor, and add the following lines:
s = 'Hello world'
print(s)
Now, you can run it in IPython console or a notebook and explore the resulting variables:
In [1]: %run my_file.py

Hello world
In [2]: s
Out[2]: 'Hello world'
In [3]: %whos
Variable Type Data/Info
----------------------------
s str Hello world
1.4. The workflow: interactive environments and text editors 9

From a script to functions
While it is tempting to work only with scripts, that is a file full of instructions following each other, do plan
to progressively evolve the script to a set of functions:
A script is not reusable, functions are.
Thinking in terms of functions helps breaking the problem in small blocks.
1.4.3 IPython and Jupyter Tips and Tricks
The user manuals contain a wealth of information. Here we give a quick introduction to four useful features:
history, tab completion, magic functions, and aliases.
Command history Like a UNIX shell, the IPython console supports command history. Type up and down to
navigate previously typed commands:
In [1]: x = 10
In [2]: <UP>
In [2]: x = 10
Tab completion Tab completion, is a convenient way to explore the structure of any object youre dealing
with. Simply type object_name.<TAB> to view the objects attributes. Besides Python objects and keywords,
tab completion also works on file and directory names.*
In [1]: x = 10
In [2]: x.<TAB>
x.bit_length x.denominator x.imag x.real
x.conjugate x.from_bytes x.numerator x.to_bytes
Magic functions The console and the notebooks support so-called magic functions by prefixing a command
with the % character. For example, the run and whos functions from the previous section are magic functions.
Note that, the setting automagic, which is enabled by default, allows you to omit the preceding % sign. Thus,
you can just type the magic function and it will work.
Other useful magic functions are:
%cd to change the current directory.
In [1]: cd /tmp
/tmp

%cpaste allows you to paste code, especially code from websites which has been prefixed with the stan-
dard Python prompt (e.g. >>>) or with an ipython prompt, (e.g. in [3]):
In [2]: %cpaste
Pasting code; enter '--' alone on the line to stop or use Ctrl-D.
:>>> for i in range(3):
:... print(i)
:--
0
1
2
%timeit allows you to time the execution of short snippets using the timeit module from the standard
library:
In [3]: %timeit x = 10
10000000 loops, best of 3: 39 ns per loop
See also:
Chapter on optimizing code
%debug allows you to enter post-mortem debugging. That is to say, if the code you try to execute, raises
an exception, using %debug will enter the debugger at the point where the exception was thrown.
In [4]: x === 10
File "<ipython-input-6-12fd421b5f28>", line 1
x === 10
^
SyntaxError: invalid syntax
In [5]: %debug
> /.../IPython/core/compilerop.py (87)ast_parse()
86 and are passed to the built-in compile function."""
---> 87 return compile(source, filename, symbol, self.flags | PyCF_ONLY_AST, 1)
88
ipdb>locals()
{'source': u'x === 10\n', 'symbol': 'exec', 'self':
<IPython.core.compilerop.CachingCompiler instance at 0x2ad8ef0>,
'filename': '<ipython-input-6-12fd421b5f28>'}
See also:
Chapter on debugging
Aliases Furthermore IPython ships with various aliases which emulate common UNIX command line tools
such as ls to list files, cp to copy files and rm to remove files (a full list of aliases is shown when typing alias).
Getting help
The built-in cheat-sheet is accessible via the %quickref magic function.

A list of all available magic functions is shown when typing %magic.

CHAPTER 2
The Python language
Authors: Chris Burns, Christophe Combelles, Emmanuelle Gouillart, Gal Varoquaux
Python for scientific computing
We introduce here the Python language. Only the bare minimum necessary for getting started with Numpy
and Scipy is addressed here. To learn more about the language, consider going through the excellent tutorial
https://docs.python.org/tutorial. Dedicated books are also available, such as http://www.diveintopython.
net/.
Tip: Python is a programming language, as are C, Fortran, BASIC, PHP, etc. Some specific features of Python
are as follows:
an interpreted (as opposed to compiled) language. Contrary to e.g. C or Fortran, one does not compile
Python code before executing it. In addition, Python can be used interactively: many Python inter-
preters are available, from which commands and scripts can be executed.
a free software released under an open-source license: Python can be used and distributed free of
charge, even for building commercial software.
multi-platform: Python is available for all major operating systems, Windows, Linux/Unix, MacOS X,
most likely your mobile phone OS, etc.
a very readable language with clear non-verbose syntax
12
a language for which a large variety of high-quality packages are available for various applications, from
web frameworks to scientific computing.
a language very easy to interface with other languages, in particular C and C++.
Some other features of the language are illustrated just below. For example, Python is an object-oriented
language, with dynamic typing (the same variable can contain objects of different types during the
course of a program).
See https://www.python.org/about/ for more information about distinguishing features of Python.
2.1 First steps
Start the Ipython shell (an enhanced interactive Python shell):

by typing ipython from a Linux/Mac terminal, or from the Windows cmd shell,
or by starting the program from a menu, e.g. in the Python(x,y) or EPD menu if you have installed one
of these scientific-Python suites.
Tip: If you dont have Ipython installed on your computer, other Python shells are available, such as the plain
Python shell started by typing python in a terminal, or the Idle interpreter. However, we advise to use the
Ipython shell because of its enhanced features, especially for interactive scientific computing.
Once you have started the interpreter, type
>>> print("Hello, world!")

Hello, world!
Tip: The message Hello, world! is then displayed. You just executed your first Python instruction, congratu-
lations!
To get yourself started, type the following stack of instructions
>>> a = 3
>>> b = 2*a
>>> type(b)
<type 'int'>
>>> print(b)
6
>>> a*b
18
>>> b = 'hello'
>>> type(b)
<type 'str'>
>>> b + b
'hellohello'
>>> 2*b
'hellohello'
Tip: Two variables a and b have been defined above. Note that one does not declare the type of a variable
before assigning its value. In C, conversely, one should write:
int a = 3;
2.1. First steps 13

In addition, the type of a variable may change, in the sense that at one point in time it can be equal to a value
of a certain type, and a second point in time, it can be equal to a value of a different type. b was first equal to an
integer, but it became equal to a string when it was assigned the value hello. Operations on integers (b=2*a)
are coded natively in Python, and so are some operations on strings such as additions and multiplications,
which amount respectively to concatenation and repetition.
2.2 Basic types
2.2.1 Numerical types
Tip: Python supports the following numerical, scalar types:
Integer
>>> 1 + 1
2
>>> a = 4
>>> type(a)
<type 'int'>
Floats
>>> c = 2.1
>>> type(c)
<type 'float'>
Complex
>>> a = 1.5 + 0.5j

>>> a.real
1.5
>>> a.imag
0.5
>>> type(1. + 0j)
<type 'complex'>
Booleans
>>> 3 > 4
False
>>> test = (3 > 4)
>>> test
False
>>> type(test)
<type 'bool'>
Tip: A Python shell can therefore replace your pocket calculator, with the basic arithmetic operations +, -, *,
/, % (modulo) natively implemented
>>> 7 * 3.
21.0
>>> 2**10
1024
>>> 8 % 3
2
2.2. Basic types 14

Type conversion (casting):
>>> float(1)
1.0
Warning: Integer division

In Python 2:
>>> 3 / 2
1
In Python 3:
>>> 3 / 2
1.5
To be safe: use floats:

>>> 3 / 2.
1.5
>>> a = 3
>>> b = 2
>>> a / b # In Python 2
1
>>> a / float(b)
1.5
Future behavior: to always get the behavior of Python3

>>> from __future__ import division
>>> 3 / 2
1.5
Tip: If you explicitly want integer division use //:

>>> 3.0 // 2
1.0
Note: The behaviour of the division operator has changed in Python 3.
2.2.2 Containers
Tip: Python provides many efficient types of containers, in which collections of objects can be stored.
Lists
Tip: A list is an ordered collection of objects, that may have different types. For example:
>>> colors = ['red', 'blue', 'green', 'black', 'white']

>>> type(colors)
<type 'list'>
Indexing: accessing individual objects contained in the list:
2.2. Basic types 15

>>> colors[2]
'green'
Counting from the end with negative indices:
>>> colors[-1]
'white'
>>> colors[-2]
'black'
Warning: Indexing starts at 0 (as in C), not at 1 (as in Fortran or Matlab)!
Slicing: obtaining sublists of regularly-spaced elements:
>>> colors
['red', 'blue', 'green', 'black', 'white']
>>> colors[2:4]
['green', 'black']
Warning: Note that colors[start:stop] contains the elements with indices i such as start<= i <
stop (i ranging from start to stop-1). Therefore, colors[start:stop] has (stop - start) elements.
Slicing syntax: colors[start:stop:stride]
Tip: All slicing parameters are optional:
>>> colors
>>> colors[3:]
['black', 'white']
>>> colors[:3]
['red', 'blue', 'green']
>>> colors[::2]
['red', 'green', 'white']
Lists are mutable objects and can be modified:
>>> colors[0] = 'yellow'

>>> colors
['yellow', 'blue', 'green', 'black', 'white']
>>> colors[2:4] = ['gray', 'purple']
>>> colors
['yellow', 'blue', 'gray', 'purple', 'white']
Note: The elements of a list may have different types:
>>> colors = [3, -200, 'hello']

>>> colors
[3, -200, 'hello']
>>> colors[1], colors[2]
(-200, 'hello')
Tip: For collections of numerical data that all have the same type, it is often more efficient to use the array
type provided by the numpy module. A NumPy array is a chunk of memory containing fixed-sized items. With
2.2. Basic types 16

NumPy arrays, operations on elements can be faster because elements are regularly spaced in memory and
more operations are performed through specialized C functions instead of Python loops.
Tip: Python offers a large panel of functions to modify lists, or query them. Here are a few examples; for more
details, see https://docs.python.org/tutorial/datastructures.html#more-on-lists
Add and remove elements:
>>> colors = ['red', 'blue', 'green', 'black', 'white']

>>> colors.append('pink')
>>> colors
['red', 'blue', 'green', 'black', 'white', 'pink']
>>> colors.pop() # removes and returns the last item
'pink'
>>> colors
>>> colors.extend(['pink', 'purple']) # extend colors, in-place
>>> colors
['red', 'blue', 'green', 'black', 'white', 'pink', 'purple']
>>> colors = colors[:-2]
>>> colors
Reverse:
>>> rcolors = colors[::-1]

>>> rcolors
['white', 'black', 'green', 'blue', 'red']
>>> rcolors2 = list(colors)
>>> rcolors2
>>> rcolors2.reverse() # in-place
>>> rcolors2
Concatenate and repeat lists:
>>> rcolors + colors

['white', 'black', 'green', 'blue', 'red', 'red', 'blue', 'green', 'black', 'white']
>>> rcolors * 2
['white', 'black', 'green', 'blue', 'red', 'white', 'black', 'green', 'blue', 'red']
Tip: Sort:
>>> sorted(rcolors) # new object

['black', 'blue', 'green', 'red', 'white']
>>> rcolors
>>> rcolors.sort() # in-place
>>> rcolors
['black', 'blue', 'green', 'red', 'white']
Methods and Object-Oriented Programming
The notation rcolors.method() (e.g. rcolors.append(3) and colors.pop()) is our first example of
2.2. Basic types 17

object-oriented programming (OOP). Being a list, the object rcolors owns the method function that is
called using the notation .. No further knowledge of OOP than understanding the notation . is necessary for
going through this tutorial.
Discovering methods:
Reminder: in Ipython: tab-completion (press tab)

In [28]: rcolors.<TAB>
rcolors.append rcolors.index rcolors.remove
rcolors.count rcolors.insert rcolors.reverse
rcolors.extend rcolors.pop rcolors.sort
Strings
Different string syntaxes (simple, double or triple quotes):
s = 'Hello, how are you?'

s = "Hi, what's up"
s = '''Hello, # tripling the quotes allows the
how are you''' # string to span more than one line
s = """Hi,
what's up?"""
In [1]: 'Hi, what's up?'

------------------------------------------------------------
File "<ipython console>", line 1
'Hi, what's up?'
^
SyntaxError: invalid syntax
The newline character is \n, and the tab character is \t.
Tip: Strings are collections like lists. Hence they can be indexed and sliced, using the same syntax and rules.
Indexing:
>>> a = "hello"
>>> a[0]
'h'
>>> a[1]
'e'
>>> a[-1]
'o'
Tip: (Remember that negative indices correspond to counting from the right end.)
Slicing:
>>> a = "hello, world!"

>>> a[3:6] # 3rd to 6th (excluded) elements: elements 3, 4, 5
'lo,'
>>> a[2:10:2] # Syntax: a[start:stop:step]
'lo o'
2.2. Basic types 18

>>> a[::3] # every three characters, from beginning to end

'hl r!'
Tip: Accents and special characters can also be handled in Unicode strings (see https://docs.python.org/
tutorial/introduction.html#unicode-strings).
A string is an immutable object and it is not possible to modify its contents. One may however create new
strings from the original one.
In [53]: a = "hello, world!"

In [54]: a[2] = 'z'
---------------------------------------------------------------------------
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
TypeError: 'str' object does not support item assignment
In [55]: a.replace('l', 'z', 1)

Out[55]: 'hezlo, world!'
In [56]: a.replace('l', 'z')
Out[56]: 'hezzo, worzd!'
Tip: Strings have many useful methods, such as a.replace as seen above. Remember the a. object-oriented
notation and use tab completion or help(str) to search for new methods.
See also:
Python offers advanced possibilities for manipulating strings, looking for patterns or formatting. The inter-
ested reader is referred to https://docs.python.org/library/stdtypes.html#string-methods and https://docs.
python.org/library/string.html#new-string-formatting
String formatting:
>>> 'An integer: %i ; a float: %f ; another string: %s ' % (1, 0.1, 'string')
'An integer: 1; a float: 0.100000; another string: string'
>>> i = 102
>>> filename = 'processing_of_dataset_%d .txt' % i
>>> filename
'processing_of_dataset_102.txt'
Dictionaries
Tip: A dictionary is basically an efficient table that maps keys to values. It is an unordered container
>>> tel = {'emmanuelle': 5752, 'sebastian': 5578}

>>> tel['francis'] = 5915
>>> tel
{'sebastian': 5578, 'francis': 5915, 'emmanuelle': 5752}
>>> tel['sebastian']
5578
>>> tel.keys()
['sebastian', 'francis', 'emmanuelle']
>>> tel.values()
[5578, 5915, 5752]
>>> 'francis' in tel
True
2.2. Basic types 19

Tip: It can be used to conveniently store and retrieve values associated with a name (a string for a date, a
name, etc.). See https://docs.python.org/tutorial/datastructures.html#dictionaries for more information.
A dictionary can have keys (resp. values) with different types:
>>> d = {'a':1, 'b':2, 3:'hello'}

>>> d
{'a': 1, 3: 'hello', 'b': 2}
More container types
Tuples
Tuples are basically immutable lists. The elements of a tuple are written between parentheses, or just separated
by commas:
>>> t = 12345, 54321, 'hello!'

>>> t[0]
12345
>>> t
(12345, 54321, 'hello!')
>>> u = (0, 2)
Sets: unordered, unique items:
>>> s = set(('a', 'b', 'c', 'a'))

>>> s
set(['a', 'c', 'b'])
>>> s.difference(('a', 'b'))
set(['c'])
2.2.3 Assignment operator
Tip: Python library reference says:

Assignment statements are used to (re)bind names to values and to modify attributes or items of
mutable objects.
In short, it works as follows (simple assignment):
1. an expression on the right hand side is evaluated, the corresponding object is created/obtained
2. a name on the left hand side is assigned, or bound, to the r.h.s. object
Things to note:
a single object can have several names bound to it:
In [1]: a = [1, 2, 3]
In [2]: b = a
In [3]: a
Out[3]: [1, 2, 3]
In [4]: b
Out[4]: [1, 2, 3]
In [5]: a is b
Out[5]: True
2.2. Basic types 20

In [6]: b[1] = 'hi!'

In [7]: a
Out[7]: [1, 'hi!', 3]
to change a list in place, use indexing/slices:
In [1]: a = [1, 2, 3]
In [3]: a
Out[3]: [1, 2, 3]
In [4]: a = ['a', 'b', 'c'] # Creates another object.
In [5]: a
Out[5]: ['a', 'b', 'c']
In [6]: id(a)
Out[6]: 138641676
In [7]: a[:] = [1, 2, 3] # Modifies object in place.
In [8]: a
Out[8]: [1, 2, 3]
In [9]: id(a)
Out[9]: 138641676 # Same as in Out[6], yours will differ...
the key concept here is mutable vs. immutable

mutable objects can be changed in place
immutable objects cannot be modified once created
See also:
A very good and detailed explanation of the above issues can be found in David M. Beazleys article Types and
Objects in Python.
2.3 Control Flow
Controls the order in which the code is executed.
2.3.1 if/elif/else
>>> if 2**2 == 4:
... print('Obvious!')
...
Obvious!
Blocks are delimited by indentation
Tip: Type the following lines in your Python interpreter, and be careful to respect the indentation depth. The
Ipython shell automatically increases the indentation depth after a colon : sign; to decrease the indentation
depth, go four spaces to the left with the Backspace key. Press the Enter key twice to leave the logical block.
>>> a = 10
>>> if a == 1:
... print(1)
... elif a == 2:
... print(2)
... else:
... print('A lot')
A lot
2.3. Control Flow 21

Indentation is compulsory in scripts as well. As an exercise, re-type the previous lines with the same indenta-
tion in a script condition.py, and execute the script with run condition.py in Ipython.
2.3.2 for/range
Iterating with an index:
>>> for i in range(4):

... print(i)
0
1
2
3
But most often, it is more readable to iterate over values:
>>> for word in ('cool', 'powerful', 'readable'):

... print('Python is %s ' % word)
Python is cool
Python is powerful
Python is readable
2.3.3 while/break/continue
Typical C-style while loop (Mandelbrot problem):
>>> z = 1 + 1j
>>> while abs(z) < 100:
... z = z**2 + 1
>>> z
(-134+352j)
More advanced features

break out of enclosing for/while loop:
>>> z = 1 + 1j
>>> while abs(z) < 100:

... if z.imag == 0:
... break
... z = z**2 + 1
continue the next iteration of a loop.:
>>> a = [1, 0, 2, 4]
>>> for element in a:
... if element == 0:
... continue
... print(1. / element)
1.0
0.5
0.25
2.3.4 Conditional Expressions
if <OBJECT>
Evaluates to False:

any number equal to zero (0, 0.0, 0+0j)

an empty container (list, tuple, set, dictionary, . . . )
False, None
Evaluates to True:
everything else
a == b Tests equality, with logics:
>>> 1 == 1.
True
a is b Tests identity: both sides are the same object:
>>> 1 is 1.
False
>>> a = 1
>>> b = 1
>>> a is b
True
a in b For any collection b: b contains a
>>> b = [1, 2, 3]
>>> 2 in b
True
>>> 5 in b
False
If b is a dictionary, this tests that a is a key of b.
2.3.5 Advanced iteration
Iterate over any sequence
You can iterate over any sequence (string, list, keys in a dictionary, lines in a file, . . . ):
>>> vowels = 'aeiouy'
>>> for i in 'powerful':

... if i in vowels:
... print(i)
o
e
u
>>> message = "Hello how are you?"

>>> message.split() # returns a list
['Hello', 'how', 'are', 'you?']
>>> for word in message.split():
... print(word)
...
Hello
how
are
you?

Tip: Few languages (in particular, languages for scientific computing) allow to loop over anything but in-
tegers/indices. With Python it is possible to loop exactly over the objects of interest without bothering with
indices you often dont care about. This feature can often be used to make code more readable.
Warning: Not safe to modify the sequence you are iterating over.
Keeping track of enumeration number
Common task is to iterate over a sequence while keeping track of the item number.
Could use while loop with a counter as above. Or a for loop:
>>> words = ('cool', 'powerful', 'readable')

>>> for i in range(0, len(words)):
... print((i, words[i]))
(0, 'cool')
(1, 'powerful')
(2, 'readable')
But, Python provides a built-in function - enumerate - for this:
>>> for index, item in enumerate(words):

... print((index, item))
(0, 'cool')
(1, 'powerful')
(2, 'readable')
Looping over a dictionary
Use items:
>>> d = {'a': 1, 'b':1.2, 'c':1j}
>>> for key, val in sorted(d.items()):

... print('Key: %s has value: %s ' % (key, val))
Key: a has value: 1
Key: b has value: 1.2
Key: c has value: 1j
Note: The ordering of a dictionary in random, thus we use sorted() which will sort on the keys.
2.3.6 List Comprehensions
>>> [i**2 for i in range(4)]

[0, 1, 4, 9]

Exercise
Compute the decimals of Pi using the Wallis formula:

4i 2
=2
Y
2
i =1 4i 1
2.4 Defining functions
2.4.1 Function definition
In [56]: def test():

....: print('in test function')
....:
....:
In [57]: test()
in test function
Warning: Function blocks must be indented as other control-flow blocks.
2.4.2 Return statement
Functions can optionally return values.
In [6]: def disk_area(radius):

...: return 3.14 * radius * radius
...:
In [8]: disk_area(1.5)
Out[8]: 7.0649999999999995
Note: By default, functions return None.
Note: Note the syntax to define a function:

the def keyword;
is followed by the functions name, then
the arguments of the function are given between parentheses followed by a colon.
the function body;
and return object for optionally returning values.
2.4.3 Parameters
Mandatory parameters (positional arguments)
2.4. Defining functions 25

In [81]: def double_it(x):

....: return x * 2
....:
In [82]: double_it(3)
Out[82]: 6
In [83]: double_it()
---------------------------------------------------------------------------
TypeError: double_it() takes exactly 1 argument (0 given)
Optional parameters (keyword or named arguments)
In [84]: def double_it(x=2):

....: return x * 2
....:
Out[85]: 4
In [86]: double_it(3)
Out[86]: 6
Keyword arguments allow you to specify default values.
Warning: Default values are evaluated when the function is defined, not when it is called. This can be
problematic when using mutable types (e.g. dictionary or list) and modifying them in the function body,
since the modifications will be persistent across invocations of the function.
Using an immutable type in a keyword argument:
In [124]: bigx = 10
In [125]: def double_it(x=bigx):

.....: return x * 2
.....:
In [126]: bigx = 1e9 # Now really big
Out[128]: 20
Using an mutable type in a keyword argument (and modifying it inside the function body):
In [2]: def add_to_dict(args={'a': 1, 'b': 2}):
...: for i in args.keys():
...: args[i] += 1
...: print args
...:
In [3]: add_to_dict
Out[3]: <function __main__.add_to_dict>
In [4]: add_to_dict()
{'a': 2, 'b': 3}
{'a': 3, 'b': 4}
{'a': 4, 'b': 5}

Tip: More involved example implementing pythons slicing:
In [98]: def slicer(seq, start=None, stop=None, step=None):

....: """Implement basic python slicing."""
....: return seq[start:stop:step]
....:
In [101]: rhyme = 'one fish, two fish, red fish, blue fish'.split()
In [102]: rhyme
Out[102]: ['one', 'fish,', 'two', 'fish,', 'red', 'fish,', 'blue', 'fish']
In [103]: slicer(rhyme)
Out[103]: ['one', 'fish,', 'two', 'fish,', 'red', 'fish,', 'blue', 'fish']
In [104]: slicer(rhyme, step=2)

Out[104]: ['one', 'two', 'red', 'blue']
In [105]: slicer(rhyme, 1, step=2)

Out[105]: ['fish,', 'fish,', 'fish,', 'fish']
In [106]: slicer(rhyme, start=1, stop=4, step=2)

Out[106]: ['fish,', 'fish,']
The order of the keyword arguments does not matter:
In [107]: slicer(rhyme, step=2, start=1, stop=4)

Out[107]: ['fish,', 'fish,']
but it is good practice to use the same ordering as the functions definition.
Keyword arguments are a very convenient feature for defining functions with a variable number of arguments,
especially when default values are to be used in most calls to the function.
2.4.4 Passing by value
Tip: Can you modify the value of a variable inside a function? Most languages (C, Java, . . . ) distinguish
passing by value and passing by reference. In Python, such a distinction is somewhat artificial, and it is a
bit subtle whether your variables are going to be modified or not. Fortunately, there exist clear rules.
Parameters to functions are references to objects, which are passed by value. When you pass a variable to a
function, python passes the reference to the object to which the variable refers (the value). Not the variable
itself.
If the value passed in a function is immutable, the function does not modify the callers variable. If the value
is mutable, the function may modify the callers variable in-place:
>>> def try_to_modify(x, y, z):

... x = 23
... y.append(42)
... z = [99] # new reference
... print(x)
... print(y)
... print(z)
...

>>> a = 77 # immutable variable

>>> b = [99] # mutable variable
>>> c = [28]
>>> try_to_modify(a, b, c)
23
[99, 42]
[99]
>>> print(a)
77
>>> print(b)
[99, 42]
>>> print(c)
[28]
Functions have a local variable table called a local namespace.

The variable x only exists within the function try_to_modify.
2.4.5 Global variables
Variables declared outside the function can be referenced within the function:
In [114]: x = 5
In [115]: def addx(y):

.....: return x + y
.....:
In [116]: addx(10)
Out[116]: 15
But these global variables cannot be modified within the function, unless declared global in the function.
This doesnt work:
In [117]: def setx(y):

.....: x = y
.....: print('x is %d ' % x)
.....:
.....:
In [118]: setx(10)
x is 10
In [120]: x
Out[120]: 5
This works:
In [121]: def setx(y):

.....: global x
.....: x = y
.....: print('x is %d ' % x)
.....:
.....:
In [122]: setx(10)
x is 10
In [123]: x
Out[123]: 10

2.4.6 Variable number of parameters
Special forms of parameters:

*args: any number of positional arguments packed into a tuple
**kwargs: any number of keyword arguments packed into a dictionary
In [35]: def variable_args(*args, **kwargs):

....: print 'args is', args
....: print 'kwargs is', kwargs
....:
In [36]: variable_args('one', 'two', x=1, y=2, z=3)

args is ('one', 'two')
kwargs is {'y': 2, 'x': 1, 'z': 3}
2.4.7 Docstrings
Documentation about what the function does and its parameters. General convention:
In [67]: def funcname(params):

....: """Concise one-line sentence describing the function.
....:
....: Extended summary which can contain multiple paragraphs.
....: """
....: # function body
....: pass
....:
In [68]: funcname?
Type: function
Base Class: type 'function'>
String Form: <function funcname at 0xeaa0f0>
Namespace: Interactive
File: <ipython console>
Definition: funcname(params)
Docstring:
Concise one-line sentence describing the function.
Extended summary which can contain multiple paragraphs.
Note: Docstring guidelines

For the sake of standardization, the Docstring Conventions webpage documents the semantics and conven-
tions associated with Python docstrings.
Also, the Numpy and Scipy modules have defined a precise standard for documenting scientific func-
tions, that you may want to follow for your own functions, with a Parameters section, an Examples sec-
tion, etc. See http://projects.scipy.org/numpy/wiki/CodingStyleGuidelines#docstring-standard and http://
projects.scipy.org/numpy/browser/trunk/doc/example.py#L37
2.4.8 Functions are objects
Functions are first-class objects, which means they can be:

assigned to a variable
an item in a list (or any collection)

passed as an argument to another function.
In [38]: va = variable_args
In [39]: va('three', x=1, y=2)

args is ('three',)
kwargs is {'y': 2, 'x': 1}
2.4.9 Methods
Methods are functions attached to objects. Youve seen these in our examples on lists, dictionaries, strings,
etc. . .
2.4.10 Exercises
Exercise: Fibonacci sequence
Write a function that displays the n first terms of the Fibonacci sequence, defined by:

U0 = 0
U =1
1
Un+2 = Un+1 +Un
Exercise: Quicksort
Implement the quicksort algorithm, as defined by wikipedia
function quicksort(array)
var list less, greater
if length(array) < 2
return array
select and remove a pivot value pivot from array
for each x in array
if x < pivot + 1 then append x to less
else append x to greater
return concatenate(quicksort(less), pivot, quicksort(greater))
2.5 Reusing code: scripts and modules
For now, we have typed all instructions in the interpreter. For longer sets of instructions we need to change
track and write the code in text files (using a text editor), that we will call either scripts or modules. Use your fa-
vorite text editor (provided it offers syntax highlighting for Python), or the editor that comes with the Scientific
Python Suite you may be using.
2.5.1 Scripts
Tip: Let us first write a script, that is a file with a sequence of instructions that are executed each time the script
is called. Instructions may be e.g. copied-and-pasted from the interpreter (but take care to respect indentation
rules!).
2.5. Reusing code: scripts and modules 30

The extension for Python files is .py. Write or copy-and-paste the following lines in a file called test.py
message = "Hello how are you?"

for word in message.split():
print word
Tip: Let us now execute the script interactively, that is inside the Ipython interpreter. This is maybe the most
common use of scripts in scientific computing.
Note: in Ipython, the syntax to execute a script is %run script.py. For example,
In [1]: %run test.py

Hello
how
are
you?
In [2]: message
Out[2]: 'Hello how are you?'
The script has been executed. Moreover the variables defined in the script (such as message) are now available
inside the interpreters namespace.
Tip: Other interpreters also offer the possibility to execute scripts (e.g., execfile in the plain Python inter-
preter, etc.).
It is also possible In order to execute this script as a standalone program, by executing the script inside a shell
terminal (Linux/Mac console or cmd Windows console). For example, if we are in the same directory as the
test.py file, we can execute this in a console:
$ python test.py
Hello
how
are
you?
Tip: Standalone scripts may also take command-line arguments

In file.py:
import sys
print sys.argv
$ python file.py test arguments

['file.py', 'test', 'arguments']
Warning: Dont implement option parsing yourself. Use modules such as optparse, argparse or
:moddocopt.

2.5.2 Importing objects from modules
In [1]: import os
In [2]: os
Out[2]: <module 'os' from '/usr/lib/python2.6/os.pyc'>
In [3]: os.listdir('.')
Out[3]:
['conf.py',
'basic_types.rst',
'control_flow.rst',
'functions.rst',
'python_language.rst',
'reusing.rst',
'file_io.rst',
'exceptions.rst',
'workflow.rst',
'index.rst']
And also:
In [4]: from os import listdir
Importing shorthands:
In [5]: import numpy as np
Warning:
from os import *
This is called the star import and please, Do not use it

Makes the code harder to read and understand: where do symbols come from?
Makes it impossible to guess the functionality by the context and the name (hint: os.name is the
name of the OS), and to profit usefully from tab completion.
Restricts the variable names you can use: os.name might override name, or vise-versa.
Creates possible name clashes between modules.
Makes the code impossible to statically check for undefined symbols.
Tip: Modules are thus a good way to organize code in a hierarchical way. Actually, all the scientific computing
tools we are going to use are modules:
>>> import numpy as np # data arrays

>>> np.linspace(0, 10, 6)
array([ 0., 2., 4., 6., 8., 10.])
>>> import scipy # scientific computing
2.5.3 Creating modules
Tip: If we want to write larger and better organized programs (compared to simple scripts), where some
objects are defined, (variables, functions, classes) and that we want to reuse several times, we have to create

our own modules.
Let us create a module demo contained in the file demo.py:
"A demo module."
def print_b():
"Prints b."
print 'b'
def print_a():
"Prints a."
print 'a'
c = 2
d = 2
Tip: In this file, we defined two functions print_a and print_b. Suppose we want to call the print_a
function from the interpreter. We could execute the file as a script, but since we just want to have access to the
function print_a, we are rather going to import it as a module. The syntax is as follows.
In [1]: import demo
In [2]: demo.print_a()
a
In [3]: demo.print_b()
b
Importing the module gives access to its objects, using the module.object syntax. Dont forget to put the
modules name before the objects name, otherwise Python wont recognize the instruction.
Introspection
In [4]: demo?
Type: module
Base Class: <type 'module'>
String Form: <module 'demo' from 'demo.py'>
File: /home/varoquau/Projects/Python_talks/scipy_2009_tutorial/source/demo.py
Docstring:
A demo module.
In [5]: who
demo
In [6]: whos
------------------------------
demo module <module 'demo' from 'demo.py'>
In [7]: dir(demo)
Out[7]:
['__builtins__',
'__doc__',
'__file__',
'__name__',

'__package__',
'c',
'd',
'print_a',
'print_b']
In [8]: demo.
demo.c demo.print_a demo.py
demo.d demo.print_b demo.pyc
Importing objects from modules into the main namespace
In [9]: from demo import print_a, print_b
In [10]: whos
--------------------------------
demo module <module 'demo' from 'demo.py'>
print_a function <function print_a at 0xb7421534>
print_b function <function print_b at 0xb74214c4>
In [11]: print_a()
a
Warning: Module caching

Modules are cached: if you modify demo.py and re-import it in the old session, you will get the
old one.
Solution:
In [10]: reload(demo)
In Python3 instead reload is not builtin, so you have to import the importlib module first and then do:
In [10]: importlib.reload(demo)
2.5.4 __main__ and module loading
Tip: Sometimes we want code to be executed when a module is run directly, but not when it is imported by
another module. if __name__ == '__main__' allows us to check whether the module is being run directly.
File demo2.py:
def print_b():
"Prints b."
print 'b'
def print_a():
"Prints a."
print 'a'
# print_b() runs on import

print_b()
if __name__ == '__main__':
# print_a() is only executed when the module is run directly.
print_a()

Importing it:
In [11]: import demo2

b
In [12]: import demo2
Running it:
In [13]: %run demo2

b
a
2.5.5 Scripts or modules? How to organize your code
Note: Rule of thumb

Sets of instructions that are called several times should be written inside functions for better code
reusability.
Functions (or other bits of code) that are called from several scripts should be written inside a module,
so that only the module is imported in the different scripts (do not copy-and-paste your functions in the
different scripts!).
How modules are found and imported
When the import mymodule statement is executed, the module mymodule is searched in a given list of direc-
tories. This list includes a list of installation-dependent default path (e.g., /usr/lib/python) as well as the
list of directories specified by the environment variable PYTHONPATH.
The list of directories searched by Python is given by the sys.path variable
In [1]: import sys
In [2]: sys.path
Out[2]:
['',
'/home/varoquau/.local/bin',
'/usr/lib/python2.7',
'/home/varoquau/.local/lib/python2.7/site-packages',
'/usr/lib/python2.7/dist-packages',
'/usr/local/lib/python2.7/dist-packages',
...]
Modules must be located in the search path, therefore you can:

write your own modules within directories already defined in the search path (e.g. $HOME/.local/lib/
python2.7/dist-packages). You may use symbolic links (on Linux) to keep the code somewhere else.
modify the environment variable PYTHONPATH to include the directories containing the user-defined
modules.
Tip: On Linux/Unix, add the following line to a file read by the shell at startup (e.g. /etc/profile, .profile)
export PYTHONPATH=$PYTHONPATH:/home/emma/user_defined_modules

On Windows, http://support.microsoft.com/kb/310519 explains how to handle environment variables.
or modify the sys.path variable itself within a Python script.
Tip:
import sys
new_path = '/home/emma/user_defined_modules'
if new_path not in sys.path:
sys.path.append(new_path)
This method is not very robust, however, because it makes the code less portable (user-dependent path)
and because you have to add the directory to your sys.path each time you want to import from a module
in this directory.
See also:
See https://docs.python.org/tutorial/modules.html for more information about modules.
2.5.6 Packages
A directory that contains many modules is called a package. A package is a module with submodules (which
can have submodules themselves, etc.). A special file called __init__.py (which may be empty) tells Python
that the directory is a Python package, from which modules can be imported.
$ ls
cluster/ io/ README.txt@ stsci/
__config__.py@ LATEST.txt@ setup.py@ __svn_version__.py@
__config__.pyc lib/ setup.pyc __svn_version__.pyc
constants/ linalg/ setupscons.py@ THANKS.txt@
fftpack/ linsolve/ setupscons.pyc TOCHANGE.txt@
__init__.py@ maxentropy/ signal/ version.py@
__init__.pyc misc/ sparse/ version.pyc
INSTALL.txt@ ndimage/ spatial/ weave/
integrate/ odr/ special/
interpolate/ optimize/ stats/
$ cd ndimage
$ ls
doccer.py@ fourier.pyc interpolation.py@ morphology.pyc setup.pyc
doccer.pyc info.py@ interpolation.pyc _nd_image.so
setupscons.py@
filters.py@ info.pyc measurements.py@ _ni_support.py@
setupscons.pyc
filters.pyc __init__.py@ measurements.pyc _ni_support.pyc tests/
fourier.py@ __init__.pyc morphology.py@ setup.py@
From Ipython:
In [1]: import scipy
In [2]: scipy.__file__
Out[2]: '/usr/lib/python2.6/dist-packages/scipy/__init__.pyc'
In [3]: import scipy.version
In [4]: scipy.version.version
Out[4]: '0.7.0'
In [5]: import scipy.ndimage.morphology

In [6]: from scipy.ndimage import morphology
In [17]: morphology.binary_dilation?
Type: function
Base Class: <type 'function'>
String Form: <function binary_dilation at 0x9bedd84>
File: /usr/lib/python2.6/dist-packages/scipy/ndimage/morphology.py
Definition: morphology.binary_dilation(input, structure=None,
iterations=1, mask=None, output=None, border_value=0, origin=0,
brute_force=False)
Docstring:
Multi-dimensional binary dilation with the given structure.
An output array can optionally be provided. The origin parameter

controls the placement of the filter. If no structuring element is
provided an element is generated with a squared connectivity equal
to one. The dilation operation is repeated iterations times. If
iterations is less than 1, the dilation is repeated until the
result does not change anymore. If a mask is given, only those
elements with a true value at the corresponding mask element are
modified at each iteration.
2.5.7 Good practices
Use meaningful object names

Indentation: no choice!
Tip: Indenting is compulsory in Python! Every command block following a colon bears an additional
indentation level with respect to the previous line with a colon. One must therefore indent after def
f(): or while:. At the end of such logical blocks, one decreases the indentation depth (and re-increases
it if a new block is entered, etc.)
Strict respect of indentation is the price to pay for getting rid of { or ; characters that delineate logical
blocks in other languages. Improper indentation leads to errors such as
------------------------------------------------------------
IndentationError: unexpected indent (test.py, line 2)
All this indentation business can be a bit confusing in the beginning. However, with the clear indenta-
tion, and in the absence of extra characters, the resulting code is very nice to read compared to other
languages.
Indentation depth: Inside your text editor, you may choose to indent with any positive number of spaces
(1, 2, 3, 4, . . . ). However, it is considered good practice to indent with 4 spaces. You may configure your
editor to map the Tab key to a 4-space indentation.
Style guidelines
Long lines: you should not write very long lines that span over more than (e.g.) 80 characters. Long lines
can be broken with the \ character
>>> long_line = "Here is a very very long line \

... that we break in two parts."
Spaces
Write well-spaced code: put whitespaces after commas, around arithmetic operators, etc.:

>>> a = 1 # yes
>>> a=1 # too cramped
A certain number of rules for writing beautiful code (and more importantly using the same conven-
tions as anybody else!) are given in the Style Guide for Python Code.
Quick read
If you want to do a first quick pass through the Scipy lectures to learn the ecosystem, you can directly skip
to the next chapter: NumPy: creating and manipulating numerical data.
The remainder of this chapter is not necessary to follow the rest of the intro part. But be sure to come back
and finish this chapter later.
2.6 Input and Output
To be exhaustive, here are some information about input and output in Python. Since we will use the Numpy
methods to read and write files, you may skip this chapter at first reading.
We write or read strings to/from files (other types must be converted to strings). To write in a file:
>>> f = open('workfile', 'w') # opens the workfile file

>>> type(f)
<type 'file'>
>>> f.write('This is a test \nand another test')
>>> f.close()
To read from a file
In [1]: f = open('workfile', 'r')
In [2]: s = f.read()
In [3]: print(s)
This is a test
and another test
In [4]: f.close()
See also:
For more details: https://docs.python.org/tutorial/inputoutput.html
2.6.1 Iterating over a file
In [6]: f = open('workfile', 'r')
In [7]: for line in f:

...: print line
...:
This is a test
and another test
In [8]: f.close()
2.6. Input and Output 38

File modes
Read-only: r
Write-only: w
Note: Create a new file or overwrite existing file.
Append a file: a
Read and Write: r+
Binary mode: b
Note: Use for binary files, especially on Windows.
2.7 Standard Library
Note: Reference document for this section:

The Python Standard Library documentation: https://docs.python.org/library/index.html
Python Essential Reference, David Beazley, Addison-Wesley Professional
2.7.1 os module: operating system functionality
A portable way of using operating system dependent functionality.
Directory and file manipulation
Current directory:
In [17]: os.getcwd()
Out[17]: '/Users/cburns/src/scipy2009/scipy_2009_tutorial/source'
List a directory:
In [31]: os.listdir(os.curdir)
Out[31]:
['.index.rst.swo',
'.python_language.rst.swp',
'.view_array.py.swp',
'_static',
'_templates',
'basic_types.rst',
'conf.py',
'control_flow.rst',
'debugging.rst',
...
Make a directory:
In [32]: os.mkdir('junkdir')
In [33]: 'junkdir' in os.listdir(os.curdir)

Out[33]: True
Rename the directory:
2.7. Standard Library 39

In [36]: os.rename('junkdir', 'foodir')
In [37]: 'junkdir' in os.listdir(os.curdir)

Out[37]: False
In [38]: 'foodir' in os.listdir(os.curdir)

Out[38]: True
In [41]: os.rmdir('foodir')
In [42]: 'foodir' in os.listdir(os.curdir)

Out[42]: False
Delete a file:
In [44]: fp = open('junk.txt', 'w')
In [45]: fp.close()
In [46]: 'junk.txt' in os.listdir(os.curdir)

Out[46]: True
In [47]: os.remove('junk.txt')
In [48]: 'junk.txt' in os.listdir(os.curdir)

Out[48]: False
os.path: path manipulations
os.path provides common operations on pathnames.
In [70]: fp = open('junk.txt', 'w')
In [71]: fp.close()
In [72]: a = os.path.abspath('junk.txt')
In [73]: a
Out[73]: '/Users/cburns/src/scipy2009/scipy_2009_tutorial/source/junk.txt'
In [74]: os.path.split(a)
Out[74]: ('/Users/cburns/src/scipy2009/scipy_2009_tutorial/source',
'junk.txt')
In [78]: os.path.dirname(a)
Out[78]: '/Users/cburns/src/scipy2009/scipy_2009_tutorial/source'
In [79]: os.path.basename(a)
Out[79]: 'junk.txt'
In [80]: os.path.splitext(os.path.basename(a))
Out[80]: ('junk', '.txt')
In [84]: os.path.exists('junk.txt')
Out[84]: True
In [86]: os.path.isfile('junk.txt')
Out[86]: True
In [87]: os.path.isdir('junk.txt')
Out[87]: False

In [88]: os.path.expanduser('~/local')
Out[88]: '/Users/cburns/local'
In [92]: os.path.join(os.path.expanduser('~'), 'local', 'bin')

Out[92]: '/Users/cburns/local/bin'
Running an external command
In [8]: os.system('ls')
basic_types.rst demo.py functions.rst python_language.rst standard_library.rst
control_flow.rst exceptions.rst io.rst python-logo.png
demo2.py first_steps.rst oop.rst reusing_code.rst
Note: Alternative to os.system

A noteworthy alternative to os.system is the sh module. Which provides much more convenient ways to
obtain the output, error stream and exit code of the external command.
In [20]: import sh
In [20]: com = sh.ls()
In [21]: print com

basic_types.rst exceptions.rst oop.rst standard_library.rst
control_flow.rst first_steps.rst python_language.rst
demo2.py functions.rst python-logo.png
demo.py io.rst reusing_code.rst
In [22]: print com.exit_code

0
In [23]: type(com)
Out[23]: sh.RunningCommand
Walking a directory
os.path.walk generates a list of filenames in a directory tree.
In [10]: for dirpath, dirnames, filenames in os.walk(os.curdir):

....: for fp in filenames:
....: print os.path.abspath(fp)
....:
....:
/Users/cburns/src/scipy2009/scipy_2009_tutorial/source/.index.rst.swo
/Users/cburns/src/scipy2009/scipy_2009_tutorial/source/.view_array.py.swp
/Users/cburns/src/scipy2009/scipy_2009_tutorial/source/basic_types.rst
/Users/cburns/src/scipy2009/scipy_2009_tutorial/source/conf.py
/Users/cburns/src/scipy2009/scipy_2009_tutorial/source/control_flow.rst
...
Environment variables:
In [9]: import os
In [11]: os.environ.keys()
Out[11]:

['_',
'FSLDIR',
'TERM_PROGRAM_VERSION',
'FSLREMOTECALL',
'USER',
'HOME',
'PATH',
'PS1',
'SHELL',
'EDITOR',
'WORKON_HOME',
'PYTHONPATH',
...
In [12]: os.environ['PYTHONPATH']
Out[12]: '.:/Users/cburns/src/utils:/Users/cburns/src/nitools:
/Users/cburns/local/lib/python2.5/site-packages/:
/usr/local/lib/python2.5/site-packages/:
/Library/Frameworks/Python.framework/Versions/2.5/lib/python2.5'
In [16]: os.getenv('PYTHONPATH')
Out[16]: '.:/Users/cburns/src/utils:/Users/cburns/src/nitools:
/Users/cburns/local/lib/python2.5/site-packages/:
/usr/local/lib/python2.5/site-packages/:
/Library/Frameworks/Python.framework/Versions/2.5/lib/python2.5'
2.7.2 shutil: high-level file operations
The shutil provides useful file operations:

shutil.rmtree: Recursively delete a directory tree.
shutil.move: Recursively move a file or directory to another location.
shutil.copy: Copy files or directories.
2.7.3 glob: Pattern matching on files
The glob module provides convenient file pattern matching.

Find all files ending in .txt:
In [18]: import glob
In [19]: glob.glob('*.txt')
Out[19]: ['holy_grail.txt', 'junk.txt', 'newfile.txt']
2.7.4 sys module: system-specific information
System-specific information related to the Python interpreter.

Which version of python are you running and where is it installed:
In [117]: sys.platform
Out[117]: 'darwin'
In [118]: sys.version
Out[118]: '2.5.2 (r252:60911, Feb 22 2008, 07:57:53) \n
[GCC 4.0.1 (Apple Computer, Inc. build 5363)]'

In [119]: sys.prefix
Out[119]: '/Library/Frameworks/Python.framework/Versions/2.5'
List of command line arguments passed to a Python script:
In [100]: sys.argv
Out[100]: ['/Users/cburns/local/bin/ipython']
sys.path is a list of strings that specifies the search path for modules. Initialized from PYTHONPATH:
In [121]: sys.path
Out[121]:
['',
'/Users/cburns/local/bin',
'/Users/cburns/local/lib/python2.5/site-packages/grin-1.1-py2.5.egg',
'/Users/cburns/local/lib/python2.5/site-packages/argparse-0.8.0-py2.5.egg',
'/Users/cburns/local/lib/python2.5/site-packages/urwid-0.9.7.1-py2.5.egg',
'/Users/cburns/local/lib/python2.5/site-packages/yolk-0.4.1-py2.5.egg',
'/Users/cburns/local/lib/python2.5/site-packages/virtualenv-1.2-py2.5.egg',
...
2.7.5 pickle: easy persistence
Useful to store arbitrary objects to a file. Not safe or fast!
In [1]: import pickle
In [2]: l = [1, None, 'Stan']
In [3]: pickle.dump(l, file('test.pkl', 'w'))
In [4]: pickle.load(file('test.pkl'))
Out[4]: [1, None, 'Stan']
Exercise
Write a program to search your PYTHONPATH for the module site.py.
path_site
2.8 Exception handling in Python
It is likely that you have raised Exceptions if you have typed all the previous commands of the tutorial. For
example, you may have raised an exception if you entered a command with a typo.
Exceptions are raised by different kinds of errors arising when executing Python code. In your own code, you
may also catch errors, or define custom error types. You may want to look at the descriptions of the the built-in
Exceptions when looking for the right exception type.
2.8.1 Exceptions
Exceptions are raised by errors in Python:
2.8. Exception handling in Python 43

In [1]: 1/0
---------------------------------------------------------------------------
ZeroDivisionError: integer division or modulo by zero
In [2]: 1 + 'e'
---------------------------------------------------------------------------
TypeError: unsupported operand type(s) for +: 'int' and 'str'
In [3]: d = {1:1, 2:2}
In [4]: d[3]
---------------------------------------------------------------------------
KeyError: 3
In [5]: l = [1, 2, 3]
In [6]: l[4]
---------------------------------------------------------------------------
IndexError: list index out of range
In [7]: l.foobar
---------------------------------------------------------------------------
AttributeError: 'list' object has no attribute 'foobar'
As you can see, there are different types of exceptions for different errors.
2.8.2 Catching exceptions
try/except
In [10]: while True:

....: try:
....: x = int(raw_input('Please enter a number: '))
....: break
....: except ValueError:
....: print('That was no valid number. Try again...')
....:
Please enter a number: a
That was no valid number. Try again...
Please enter a number: 1
In [9]: x
Out[9]: 1
try/finally
In [10]: try:
....: x = int(raw_input('Please enter a number: '))
....: finally:
....: print('Thank you for your input')
....:
....:
Please enter a number: a
Thank you for your input
---------------------------------------------------------------------------
ValueError: invalid literal for int() with base 10: 'a'
Important for resource management (e.g. closing a file)

Easier to ask for forgiveness than for permission
In [11]: def print_sorted(collection):

....: try:
....: collection.sort()
....: except AttributeError:
....: pass
....: print(collection)
....:
....:
In [12]: print_sorted([1, 3, 2])

[1, 2, 3]
In [13]: print_sorted(set((1, 3, 2)))

set([1, 2, 3])
In [14]: print_sorted('132')
132
2.8.3 Raising exceptions
Capturing and reraising an exception:
In [15]: def filter_name(name):

....: try:
....: name = name.encode('ascii')
....: except UnicodeError as e:
....: if name == 'Gal':
....: print('OK, Gal')
....: else:
....: raise e
....: return name
....:
In [16]: filter_name('Gal')
OK, Gal
Out[16]: 'Ga\xc3\xabl'
In [17]: filter_name('Stfan')
---------------------------------------------------------------------------
UnicodeDecodeError: 'ascii' codec can't decode byte 0xc3 in position 2: ordinal not in
,range(128)
Exceptions to pass messages between parts of the code:
In [17]: def achilles_arrow(x):

....: if abs(x - 1) < 1e-3:
....: raise StopIteration
....: x = 1 - (1-x)/2.
....: return x
....:
In [18]: x = 0
In [19]: while True:

....: try:
....: x = achilles_arrow(x)
....: except StopIteration:
....: break
....:

....:
In [20]: x
Out[20]: 0.9990234375
Use exceptions to notify certain conditions are met (e.g. StopIteration) or not (e.g. custom error raising)
2.9 Object-oriented programming (OOP)
Python supports object-oriented programming (OOP). The goals of OOP are:

to organize the code, and
to re-use code in similar contexts.
Here is a small example: we create a Student class, which is an object gathering several custom functions
(methods) and variables (attributes), we will be able to use:
>>> class Student(object):

... def __init__(self, name):
... self.name = name
... def set_age(self, age):
... self.age = age
... def set_major(self, major):
... self.major = major
...
>>> anna = Student('anna')
>>> anna.set_age(21)
>>> anna.set_major('physics')
In the previous example, the Student class has __init__, set_age and set_major methods. Its at-
tributes are name, age and major. We can call these methods and attributes with the following notation:
classinstance.method or classinstance.attribute. The __init__ constructor is a special method we
call with: MyClass(init parameters if any).
Now, suppose we want to create a new class MasterStudent with the same methods and attributes as the pre-
vious one, but with an additional internship attribute. We wont copy the previous class, but inherit from
it:
>>> class MasterStudent(Student):

... internship = 'mandatory, from March to June'
...
>>> james = MasterStudent('james')
>>> james.internship
'mandatory, from March to June'
>>> james.set_age(23)
>>> james.age
23
The MasterStudent class inherited from the Student attributes and methods.
Thanks to classes and object-oriented programming, we can organize code with different classes correspond-
ing to different objects we encounter (an Experiment class, an Image class, a Flow class, etc.), with their own
methods and attributes. Then we can use inheritance to consider variations around a base class and re-use
code. Ex : from a Flow base class, we can create derived StokesFlow, TurbulentFlow, PotentialFlow, etc.
2.9. Object-oriented programming (OOP) 46

CHAPTER 3
NumPy: creating and manipulating
numerical data
Authors: Emmanuelle Gouillart, Didrik Pinte, Gal Varoquaux, and Pauli Virtanen
This chapter gives an overview of NumPy, the core tool for performant numerical computing with Python.
3.1 The NumPy array object
Section contents
What are NumPy and NumPy arrays?

Creating arrays
Basic data types
Basic visualization
Indexing and slicing
Copies and views
Fancy indexing
47
3.1.1 What are NumPy and NumPy arrays?
NumPy arrays
Python objects
high-level number objects: integers, floating point
containers: lists (costless insertion and append), dictionaries (fast lookup)
NumPy provides
extension package to Python for multi-dimensional arrays
closer to hardware (efficiency)
designed for scientific computation (convenience)
Also known as array oriented computing
>>> import numpy as np

>>> a = np.array([0, 1, 2, 3])
>>> a
array([0, 1, 2, 3])
Tip: For example, An array containing:

values of an experiment/simulation at discrete time steps
signal recorded by a measurement device, e.g. sound wave
pixels of an image, grey-level or colour
3-D data measured at different X-Y-Z positions, e.g. MRI scan
...
Why it is useful: Memory-efficient container that provides fast numerical operations.
In [1]: L = range(1000)
In [2]: %timeit [i**2 for i in L]

1000 loops, best of 3: 403 us per loop
In [3]: a = np.arange(1000)
In [4]: %timeit a**2

100000 loops, best of 3: 12.7 us per loop
NumPy Reference documentation
On the web: http://docs.scipy.org/

Interactive help:
In [5]: np.array?
String Form:<built-in function array>
Docstring:
array(object, dtype=None, copy=True, order=None, subok=False, ndmin=0, ...
3.1. The NumPy array object 48

Looking for something:
>>> np.lookfor('create array')

Search results for 'create array'
---------------------------------
numpy.array
Create an array.
numpy.memmap
Create a memory-map to an array stored in a *binary* file on disk.
In [6]: np.con*?
np.concatenate
np.conj
np.conjugate
np.convolve
Import conventions
The recommended convention to import numpy is:
3.1.2 Creating arrays
Manual construction of arrays
1-D:
>>> a = np.array([0, 1, 2, 3])

>>> a
array([0, 1, 2, 3])
>>> a.ndim
1
>>> a.shape
(4,)
>>> len(a)
4
2-D, 3-D, . . . :
>>> b = np.array([[0, 1, 2], [3, 4, 5]]) # 2 x 3 array

>>> b
array([[0, 1, 2],
[3, 4, 5]])
>>> b.ndim
2
>>> b.shape
(2, 3)
>>> len(b) # returns the size of the first dimension
2
>>> c = np.array([[[1], [2]], [[3], [4]]])

>>> c
array([[[1],
[2]],
[[3],
[4]]])

>>> c.shape
(2, 2, 1)
Exercise: Simple arrays
Create a simple two dimensional array. First, redo the examples from above. And then create your
own: how about odd numbers counting backwards on the first row, and even numbers on the second?
Use the functions len(), numpy.shape() on these arrays. How do they relate to each other? And to
the ndim attribute of the arrays?
Functions for creating arrays
Tip: In practice, we rarely enter items one by one. . .
Evenly spaced:
>>> a = np.arange(10) # 0 .. n-1 (!)

>>> a
array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9])
>>> b = np.arange(1, 9, 2) # start, end (exclusive), step
>>> b
array([1, 3, 5, 7])
or by number of points:
>>> c = np.linspace(0, 1, 6) # start, end, num-points

>>> c
array([ 0. , 0.2, 0.4, 0.6, 0.8, 1. ])
>>> d = np.linspace(0, 1, 5, endpoint=False)
>>> d
array([ 0. , 0.2, 0.4, 0.6, 0.8])
Common arrays:
>>> a = np.ones((3, 3)) # reminder: (3, 3) is a tuple

>>> a
array([[ 1., 1., 1.],
[ 1., 1., 1.],
[ 1., 1., 1.]])
>>> b = np.zeros((2, 2))
>>> b
array([[ 0., 0.],
[ 0., 0.]])
>>> c = np.eye(3)
>>> c
array([[ 1., 0., 0.],
[ 0., 1., 0.],
[ 0., 0., 1.]])
>>> d = np.diag(np.array([1, 2, 3, 4]))
>>> d
array([[1, 0, 0, 0],
[0, 2, 0, 0],
[0, 0, 3, 0],
[0, 0, 0, 4]])
np.random: random numbers (Mersenne Twister PRNG):

>>> a = np.random.rand(4) # uniform in [0, 1]

>>> a
array([ 0.95799151, 0.14222247, 0.08777354, 0.51887998])
>>> b = np.random.randn(4) # Gaussian

>>> b
array([ 0.37544699, -0.11425369, -0.47616538, 1.79664113])
>>> np.random.seed(1234) # Setting the random seed
Exercise: Creating arrays using functions
Experiment with arange, linspace, ones, zeros, eye and diag.

Create different kinds of arrays with random numbers.
Try setting the seed before creating an array with random values.
Look at the function np.empty. What does it do? When might this be useful?
3.1.3 Basic data types
You may have noticed that, in some instances, array elements are displayed with a trailing dot (e.g. 2. vs 2).
This is due to a difference in the data-type used:
>>> a = np.array([1, 2, 3])

>>> a.dtype
dtype('int64')
>>> b = np.array([1., 2., 3.])

>>> b.dtype
dtype('float64')
Tip: Different data-types allow us to store data more compactly in memory, but most of the time we simply
work with floating point numbers. Note that, in the example above, NumPy auto-detects the data-type from
the input.
You can explicitly specify which data-type you want:
>>> c = np.array([1, 2, 3], dtype=float)

>>> c.dtype
dtype('float64')
The default data type is floating point:
>>> a = np.ones((3, 3))

>>> a.dtype
dtype('float64')
There are also other types:

Complex
>>> d = np.array([1+2j, 3+4j, 5+6*1j])

>>> d.dtype
dtype('complex128')

Bool
>>> e = np.array([True, False, False, True])

>>> e.dtype
dtype('bool')
Strings
>>> f = np.array(['Bonjour', 'Hello', 'Hallo',])

>>> f.dtype # <--- strings containing max. 7 letters
dtype('S7')
Much more
int32
int64
uint32
uint64
3.1.4 Basic visualization
Now that we have our first data arrays, we are going to visualize them.
Start by launching IPython:
$ ipython
Or the notebook:
$ ipython notebook
Once IPython has started, enable interactive plots:
>>> %matplotlib
Or, from the notebook, enable plots in the notebook:
>>> %matplotlib inline
The inline is important for the notebook, so that plots are displayed in the notebook and not in a new window.
Matplotlib is a 2D plotting package. We can import its functions as below:
>>> import matplotlib.pyplot as plt # the tidy way
And then use (note that you have to use show explicitly if you have not enabled interactive plots with
%matplotlib):
>>> plt.plot(x, y) # line plot

>>> plt.show() # <-- shows the plot (not needed with interactive plots)
Or, if you have enabled interactive plots with %matplotlib:
1D plotting:
>>> x = np.linspace(0, 3, 20)

>>> y = np.linspace(0, 9, 20)

[<matplotlib.lines.Line2D object at ...>]

>>> plt.plot(x, y, 'o') # dot plot
2D arrays (such as images):
>>> x = np.linspace(0, 3, 20)

>>> y = np.linspace(0, 9, 20)
>>> plt.plot(x, y, 'o') # dot plot
See also:
More in the: matplotlib chapter
Exercise: Simple visualizations
Plot some simple arrays: a cosine as a function of time and a 2D matrix.

Try using the gray colormap on the 2D matrix.
3.1.5 Indexing and slicing
The items of an array can be accessed and assigned to the same way as other Python sequences (e.g. lists):
>>> a = np.arange(10)
>>> a
array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9])

>>> a[0], a[2], a[-1]

(0, 2, 9)
Warning: Indices begin at 0, like other Python sequences (and C/C++). In contrast, in Fortran or Matlab,
indices begin at 1.
The usual python idiom for reversing a sequence is supported:
>>> a[::-1]
array([9, 8, 7, 6, 5, 4, 3, 2, 1, 0])
For multidimensional arrays, indexes are tuples of integers:
>>> a = np.diag(np.arange(3))
>>> a
array([[0, 0, 0],
[0, 1, 0],
[0, 0, 2]])
>>> a[1, 1]
1
>>> a[2, 1] = 10 # third line, second column
>>> a
array([[ 0, 0, 0],
[ 0, 1, 0],
[ 0, 10, 2]])
>>> a[1]
array([0, 1, 0])
Note:
In 2D, the first dimension corresponds to rows, the second to columns.
for multidimensional a, a[0] is interpreted by taking all elements in the unspecified dimensions.
Slicing: Arrays, like other Python sequences can also be sliced:
>>> a
array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9])
>>> a[2:9:3] # [start:end:step]
array([2, 5, 8])
Note that the last index is not included! :
>>> a[:4]
array([0, 1, 2, 3])
All three slice components are not required: by default, start is 0, end is the last and step is 1:
>>> a[1:3]
array([1, 2])
>>> a[::2]
array([0, 2, 4, 6, 8])
>>> a[3:]
array([3, 4, 5, 6, 7, 8, 9])
A small illustrated summary of NumPy indexing and slicing. . .

You can also combine assignment and slicing:
>>> a[5:] = 10
>>> a
array([ 0, 1, 2, 3, 4, 10, 10, 10, 10, 10])
>>> b = np.arange(5)
>>> a[5:] = b[::-1]
>>> a
array([0, 1, 2, 3, 4, 4, 3, 2, 1, 0])
Exercise: Indexing and slicing
Try the different flavours of slicing, using start, end and step: starting from a linspace, try to obtain
odd numbers counting backwards, and even numbers counting forwards.
Reproduce the slices in the diagram above. You may use the following expression to create the array:
>>> np.arange(6) + np.arange(0, 51, 10)[:, np.newaxis]
array([[ 0, 1, 2, 3, 4, 5],
[10, 11, 12, 13, 14, 15],
[20, 21, 22, 23, 24, 25],
[30, 31, 32, 33, 34, 35],
[40, 41, 42, 43, 44, 45],
[50, 51, 52, 53, 54, 55]])
Exercise: Array creation
Create the following arrays (with correct data types):

[[1, 1, 1, 1],
[1, 1, 1, 1],
[1, 1, 1, 2],
[1, 6, 1, 1]]
[[0., 0., 0., 0., 0.],

[2., 0., 0., 0., 0.],
[0., 3., 0., 0., 0.],
[0., 0., 4., 0., 0.],
[0.,The
3.1. 0.,NumPy
0., 5.,array
0.], object 55
[0., 0., 0., 0., 6.]]
Par on course: 3 statements for each

Hint: Individual array elements can be accessed similarly to a list, e.g. a[1] or a[1, 2].
Hint: Examine the docstring for diag.
Exercise: Tiling for array creation
Skim through the documentation for np.tile, and use this function to construct the array:
[[4, 3, 4, 3, 4, 3],
[2, 1, 2, 1, 2, 1],
[4, 3, 4, 3, 4, 3],
[2, 1, 2, 1, 2, 1]]
3.1.6 Copies and views
A slicing operation creates a view on the original array, which is just a way of accessing array data. Thus the
original array is not copied in memory. You can use np.may_share_memory() to check if two arrays share the
same memory block. Note however, that this uses heuristics and may give you false positives.
When modifying the view, the original array is modified as well:
>>> a
array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9])
>>> b = a[::2]
>>> b
array([0, 2, 4, 6, 8])
>>> np.may_share_memory(a, b)
True
>>> b[0] = 12
>>> b
array([12, 2, 4, 6, 8])
>>> a # (!)
array([12, 1, 2, 3, 4, 5, 6, 7, 8, 9])
>>> c = a[::2].copy() # force a copy
>>> c[0] = 12
>>> a
array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9])
>>> np.may_share_memory(a, c)
False
This behavior can be surprising at first sight. . . but it allows to save both memory and time.
Worked example: Prime number sieve

Compute prime numbers in 099, with a sieve

Construct a shape (100,) boolean array is_prime, filled with True in the beginning:
>>> is_prime = np.ones((100,), dtype=bool)
Cross out 0 and 1 which are not primes:

>>> is_prime[:2] = 0
For each integer j starting from 2, cross out its higher multiples:
>>> N_max = int(np.sqrt(len(is_prime) - 1))
>>> for j in range(2, N_max + 1):
... is_prime[2*j::j] = False
Skim through help(np.nonzero), and print the prime numbers

Follow-up:
Move the above code into a script file named prime_sieve.py
Run it to check it works
Use the optimization suggested in the sieve of Eratosthenes:
1. Skip j which are already known to not be primes
2. The first number to cross out is j 2
3.1.7 Fancy indexing
Tip: NumPy arrays can be indexed with slices, but also with boolean or integer arrays (masks). This method
is called fancy indexing. It creates copies not views.
Using boolean masks
>>> np.random.seed(3)
>>> a = np.random.randint(0, 21, 15)
>>> a
array([10, 3, 8, 0, 19, 10, 11, 9, 10, 6, 0, 20, 12, 7, 14])

>>> (a % 3 == 0)
array([False, True, False, True, False, False, False, True, False,
True, True, False, True, False, False], dtype=bool)
>>> mask = (a % 3 == 0)
>>> extract_from_a = a[mask] # or, a[a%3==0]
>>> extract_from_a # extract a sub-array with the mask
array([ 3, 0, 9, 6, 0, 12])
Indexing with a mask can be very useful to assign a new value to a sub-array:
>>> a[a % 3 == 0] = -1
>>> a
array([10, -1, 8, -1, 19, 10, 11, -1, 10, -1, -1, 20, -1, 7, 14])
Indexing with an array of integers
>>> a = np.arange(0, 100, 10)

>>> a
array([ 0, 10, 20, 30, 40, 50, 60, 70, 80, 90])
Indexing can be done with an array of integers, where the same index is repeated several time:
>>> a[[2, 3, 2, 4, 2]] # note: [2, 3, 2, 4, 2] is a Python list

array([20, 30, 20, 40, 20])
New values can be assigned with this kind of indexing:
>>> a[[9, 7]] = -100

>>> a
array([ 0, 10, 20, 30, 40, 50, 60, -100, 80, -100])
Tip: When a new array is created by indexing with an array of integers, the new array has the same shape than
the array of integers:
>>> idx = np.array([[3, 4], [9, 7]])
>>> idx.shape
(2, 2)
>>> a[idx]
array([[3, 4],
[9, 7]])
The image below illustrates various fancy indexing applications

Exercise: Fancy indexing
Again, reproduce the fancy indexing shown in the diagram above.

Use fancy indexing on the left and array creation on the right to assign values into an array, for instance
by setting parts of the array in the diagram above to zero.
3.2 Numerical operations on arrays
Section contents
Elementwise operations
Basic reductions
Broadcasting
Array shape manipulation
Sorting data
Summary
3.2.1 Elementwise operations
Basic operations
With scalars:
>>> a = np.array([1, 2, 3, 4])

>>> a + 1
array([2, 3, 4, 5])
>>> 2**a
array([ 2, 4, 8, 16])
All arithmetic operates elementwise:
>>> b = np.ones(4) + 1
>>> a - b
array([-1., 0., 1., 2.])
3.2. Numerical operations on arrays 59

>>> a * b
array([ 2., 4., 6., 8.])
>>> j = np.arange(5)
>>> 2**(j + 1) - j
array([ 2, 3, 6, 13, 28])
These operations are of course much faster than if you did them in pure python:
>>> a = np.arange(10000)
>>> %timeit a + 1
>>> l = range(10000)
>>> %timeit [i+1 for i in l]
Warning: Array multiplication is not matrix multiplication:

>>> c = np.ones((3, 3))
>>> c * c # NOT matrix multiplication!
array([[ 1., 1., 1.],
[ 1., 1., 1.],
[ 1., 1., 1.]])
Note: Matrix multiplication:
>>> c.dot(c)
array([[ 3., 3., 3.],
[ 3., 3., 3.],
[ 3., 3., 3.]])
Exercise: Elementwise operations
Try simple arithmetic elementwise operations: add even elements with odd elements
Time them against their pure python counterparts using %timeit.
Generate:
[2**0, 2**1, 2**2, 2**3, 2**4]
a_j = 2^(3*j) - j
Other operations
Comparisons:
>>> a = np.array([1, 2, 3, 4])

>>> b = np.array([4, 2, 2, 4])
>>> a == b
array([False, True, False, True], dtype=bool)
>>> a > b
array([False, False, True, False], dtype=bool)
Tip: Array-wise comparisons:

>>> a = np.array([1, 2, 3, 4])

>>> b = np.array([4, 2, 2, 4])
>>> c = np.array([1, 2, 3, 4])
>>> np.array_equal(a, b)
False
>>> np.array_equal(a, c)
True
Logical operations:
>>> a = np.array([1, 1, 0, 0], dtype=bool)

>>> b = np.array([1, 0, 1, 0], dtype=bool)
>>> np.logical_or(a, b)
array([ True, True, True, False], dtype=bool)
>>> np.logical_and(a, b)
array([ True, False, False, False], dtype=bool)
Transcendental functions:
>>> np.sin(a)
array([ 0. , 0.84147098, 0.90929743, 0.14112001, -0.7568025 ])
>>> np.log(a)
array([ -inf, 0. , 0.69314718, 1.09861229, 1.38629436])
>>> np.exp(a)
array([ 1. , 2.71828183, 7.3890561 , 20.08553692, 54.59815003])
Shape mismatches
>>> a + np.array([1, 2])
ValueError: operands could not be broadcast together with shapes (4) (2)
Broadcasting? Well return to that later.

Transposition:
>>> a = np.triu(np.ones((3, 3)), 1) # see help(np.triu)

>>> a
array([[ 0., 1., 1.],
[ 0., 0., 1.],
[ 0., 0., 0.]])
>>> a.T
array([[ 0., 0., 0.],
[ 1., 0., 0.],
[ 1., 1., 0.]])
Warning: The transposition is a view

As a results, the following code is wrong and will not make a matrix symmetric:
>>> a += a.T
It will work for small arrays (because of buffering) but fail for large one, in unpredictable ways.
Note: Linear algebra

The sub-module numpy.linalg implements basic linear algebra, such as solving linear systems, singular
value decomposition, etc. However, it is not guaranteed to be compiled using efficient routines, and thus
we recommend the use of scipy.linalg, as detailed in section Linear algebra operations: scipy.linalg
Exercise other operations
Look at the help for np.allclose. When might this be useful?

Look at the help for np.triu and np.tril.
3.2.2 Basic reductions
Computing sums
>>> x = np.array([1, 2, 3, 4])

>>> np.sum(x)
10
>>> x.sum()
10
Sum by rows and by columns:
>>> x = np.array([[1, 1], [2, 2]])

>>> x
array([[1, 1],
[2, 2]])
>>> x.sum(axis=0) # columns (first dimension)
array([3, 3])
>>> x[:, 0].sum(), x[:, 1].sum()
(3, 3)
>>> x.sum(axis=1) # rows (second dimension)
array([2, 4])
>>> x[0, :].sum(), x[1, :].sum()
(2, 4)
Tip: Same idea in higher dimensions:
>>> x = np.random.rand(2, 2, 2)
>>> x.sum(axis=2)[0, 1]
1.14764...

>>> x[0, 1, :].sum()

1.14764...
Other reductions
works the same way (and take axis=)

Extrema:
>>> x = np.array([1, 3, 2])

>>> x.min()
1
>>> x.max()
3
>>> x.argmin() # index of minimum

0
>>> x.argmax() # index of maximum
1
Logical operations:
>>> np.all([True, True, False])

False
>>> np.any([True, True, False])
True
Note: Can be used for array comparisons:
>>> a = np.zeros((100, 100))

>>> np.any(a != 0)
False
>>> np.all(a == a)
True
>>> a = np.array([1, 2, 3, 2])

>>> b = np.array([2, 2, 3, 2])
>>> c = np.array([6, 4, 4, 5])
>>> ((a <= b) & (b <= c)).all()
True
Statistics:
>>> x = np.array([1, 2, 3, 1])

>>> y = np.array([[1, 2, 3], [5, 6, 1]])
>>> x.mean()
1.75
>>> np.median(x)
1.5
>>> np.median(y, axis=-1) # last axis
array([ 2., 5.])
>>> x.std() # full population standard dev.

0.82915619758884995
. . . and many more (best to learn as you go).

Exercise: Reductions
Given there is a sum, what other function might you expect to see?
What is the difference between sum and cumsum?
Worked Example: data statistics
Data in populations.txt describes the populations of hares and lynxes (and carrots) in northern Canada
during 20 years.
You can view the data in an editor, or alternatively in IPython (both shell and notebook):
In [1]: !cat data/populations.txt
First, load the data into a NumPy array:

>>> data = np.loadtxt('data/populations.txt')
>>> year, hares, lynxes, carrots = data.T # trick: columns to variables
Then plot it:

>>> from matplotlib import pyplot as plt
>>> plt.axes([0.2, 0.1, 0.5, 0.8])
>>> plt.plot(year, hares, year, lynxes, year, carrots)
>>> plt.legend(('Hare', 'Lynx', 'Carrot'), loc=(1.05, 0.5))
The mean populations over time:

>>> populations = data[:, 1:]
>>> populations.mean(axis=0)
array([ 34080.95238095, 20166.66666667, 42400. ])
The sample standard deviations:

>>> populations.std(axis=0)
array([ 20897.90645809, 16254.59153691, 3322.50622558])
Which species has the highest population each year?:

>>> np.argmax(populations, axis=1)
array([2, 2, 0, 0, 1, 1, 2, 2, 2, 2, 2, 2, 0, 0, 0, 1, 2, 2, 2, 2, 2])
Worked Example: diffusion using a random walk algorithm

Tip: Let us consider a simple 1D random walk process: at each time step a walker jumps right or left with
equal probability.
We are interested in finding the typical distance from the origin of a random walker after t left or right
jumps? We are going to simulate many walkers to find this law, and we are going to do so using array com-
puting tricks: we are going to create a 2D array with the stories (each walker has a story) in one direction,
and the time in the other:
>>> n_stories = 1000 # number of walkers

>>> t_max = 200 # time during which we follow the walker
We randomly choose all the steps 1 or -1 of the walk:
>>> t = np.arange(t_max)
>>> steps = 2 * np.random.randint(0, 1 + 1, (n_stories, t_max)) - 1 # +1 because the high
,value is exclusive
>>> np.unique(steps) # Verification: all steps are 1 or -1
array([-1, 1])
We build the walks by summing steps along the time:
>>> positions = np.cumsum(steps, axis=1) # axis = 1: dimension of time

>>> sq_distance = positions**2
We get the mean in the axis of the stories:
>>> mean_sq_distance = np.mean(sq_distance, axis=0)
Plot the results:
>>> plt.figure(figsize=(4, 3))

<matplotlib.figure.Figure object at ...>
>>> plt.plot(t, np.sqrt(mean_sq_distance), 'g.', t, np.sqrt(t), 'y-')
[<matplotlib.lines.Line2D object at ...>, <matplotlib.lines.Line2D object at ...>]
>>> plt.xlabel(r"$t$")
<matplotlib.text.Text object at ...>
>>> plt.ylabel(r"$\sqrt{\langle (\delta x)^2 \rangle}$")
>>> plt.tight_layout() # provide sufficient space for labels

We find a well-known result in physics: the RMS dis-

tance grows as the square root of the time!
3.2.3 Broadcasting
Basic operations on numpy arrays (addition, etc.) are elementwise

This works on arrays of the same size.
Nevertheless, Its also possible to do operations on arrays of different
sizes if NumPy can transform these arrays so that they all have
the same size: this conversion is called broadcasting.
The image below gives an example of broadcasting:
Lets verify:

>>> a = np.tile(np.arange(0, 40, 10), (3, 1)).T

>>> a
array([[ 0, 0, 0],
[10, 10, 10],
[20, 20, 20],
[30, 30, 30]])
>>> b = np.array([0, 1, 2])
>>> a + b
array([[ 0, 1, 2],
[10, 11, 12],
[20, 21, 22],
[30, 31, 32]])
We have already used broadcasting without knowing it!:
>>> a = np.ones((4, 5))

>>> a[0] = 2 # we assign an array of dimension 0 to an array of dimension 1
>>> a
array([[ 2., 2., 2., 2., 2.],
[ 1., 1., 1., 1., 1.],
[ 1., 1., 1., 1., 1.],
[ 1., 1., 1., 1., 1.]])
An useful trick:
>>> a = np.arange(0, 40, 10)

>>> a.shape
(4,)
>>> a = a[:, np.newaxis] # adds a new axis -> 2D array
>>> a.shape
(4, 1)
>>> a
array([[ 0],
[10],
[20],
[30]])
>>> a + b
array([[ 0, 1, 2],
[10, 11, 12],
[20, 21, 22],
[30, 31, 32]])
Tip: Broadcasting seems a bit magical, but it is actually quite natural to use it when we want to solve a problem
whose output data is an array with more dimensions than input data.
Worked Example: Broadcasting
Lets construct an array of distances (in miles) between cities of Route 66: Chicago, Springfield, Saint-Louis,
Tulsa, Oklahoma City, Amarillo, Santa Fe, Albuquerque, Flagstaff and Los Angeles.
>>> mileposts = np.array([0, 198, 303, 736, 871, 1175, 1475, 1544,
... 1913, 2448])
>>> distance_array = np.abs(mileposts - mileposts[:, np.newaxis])
>>> distance_array
array([[ 0, 198, 303, 736, 871, 1175, 1475, 1544, 1913, 2448],
[ 198, 0, 105, 538, 673, 977, 1277, 1346, 1715, 2250],
[ 303, 105, 0, 433, 568, 872, 1172, 1241, 1610, 2145],
[ 736, 538, 433, 0, 135, 439, 739, 808, 1177, 1712],
[ 871, 673, 568, 135, 0, 304, 604, 673, 1042, 1577],
[1175, 977, 872, 439, 304, 0, 300, 369, 738, 1273],
[1475, 1277, 1172, 739, 604, 300, 0, 69, 438, 973],
[1544, 1346, 1241, 808, 673, 369, 69, 0, 369, 904],
[1913, 1715,
3.2. Numerical 1610, on
operations 1177, 1042, 738, 438, 369,
arrays 0, 535], 67
[2448, 2250, 2145, 1712, 1577, 1273, 973, 904, 535, 0]])
A lot of grid-based or network-based problems can also use broadcasting. For instance, if we want to compute
the distance from the origin of points on a 10x10 grid, we can do
>>> x, y = np.arange(5), np.arange(5)[:, np.newaxis]

>>> distance = np.sqrt(x ** 2 + y ** 2)
>>> distance
array([[ 0. , 1. , 2. , 3. , 4. ],
[ 1. , 1.41421356, 2.23606798, 3.16227766, 4.12310563],
[ 2. , 2.23606798, 2.82842712, 3.60555128, 4.47213595],
[ 3. , 3.16227766, 3.60555128, 4.24264069, 5. ],
[ 4. , 4.12310563, 4.47213595, 5. , 5.65685425]])
Or in color:
>>> plt.pcolor(distance)
>>> plt.colorbar()
Remark : the numpy.ogrid() function allows to di-

rectly create vectors x and y of the previous example, with two significant dimensions:
>>> x, y = np.ogrid[0:5, 0:5]

>>> x, y
(array([[0],
[1],
[2],
[3],
[4]]), array([[0, 1, 2, 3, 4]]))

>>> x.shape, y.shape

((5, 1), (1, 5))
>>> distance = np.sqrt(x ** 2 + y ** 2)
Tip: So, np.ogrid is very useful as soon as we have to handle computations on a grid. On the other hand,
np.mgrid directly provides matrices full of indices for cases where we cant (or dont want to) benefit from
broadcasting:
>>> x, y = np.mgrid[0:4, 0:4]

>>> x
array([[0, 0, 0, 0],
[1, 1, 1, 1],
[2, 2, 2, 2],
[3, 3, 3, 3]])
>>> y
array([[0, 1, 2, 3],
[0, 1, 2, 3],
[0, 1, 2, 3],
[0, 1, 2, 3]])
See also:
Broadcasting: discussion of broadcasting in the Advanced NumPy chapter.
3.2.4 Array shape manipulation
Flattening
>>> a = np.array([[1, 2, 3], [4, 5, 6]])

>>> a.ravel()
array([1, 2, 3, 4, 5, 6])
>>> a.T
array([[1, 4],
[2, 5],
[3, 6]])
>>> a.T.ravel()
array([1, 4, 2, 5, 3, 6])
Higher dimensions: last dimensions ravel out first.
Reshaping
The inverse operation to flattening:
>>> a.shape
(2, 3)
>>> b = a.ravel()
>>> b = b.reshape((2, 3))
>>> b
array([[1, 2, 3],
[4, 5, 6]])
Or,
>>> a.reshape((2, -1)) # unspecified (-1) value is inferred

array([[1, 2, 3],
[4, 5, 6]])

Warning: ndarray.reshape may return a view (cf help(np.reshape))), or copy
Tip:
>>> b[0, 0] = 99
>>> a
array([[99, 2, 3],
[ 4, 5, 6]])
Beware: reshape may also return a copy!:
>>> a = np.zeros((3, 2))

>>> b = a.T.reshape(3*2)
>>> b[0] = 9
>>> a
array([[ 0., 0.],
[ 0., 0.],
[ 0., 0.]])
To understand this you need to learn more about the memory layout of a numpy array.
Adding a dimension
Indexing with the np.newaxis object allows us to add an axis to an array (you have seen this already above in
the broadcasting section):
>>> z = np.array([1, 2, 3])

>>> z
array([1, 2, 3])
>>> z[:, np.newaxis]

array([[1],
[2],
[3]])
>>> z[np.newaxis, :]
array([[1, 2, 3]])
Dimension shuffling
>>> a = np.arange(4*3*2).reshape(4, 3, 2)
>>> a.shape
(4, 3, 2)
>>> a[0, 2, 1]
5
>>> b = a.transpose(1, 2, 0)
>>> b.shape
(3, 2, 4)
>>> b[2, 1, 0]
5
Also creates a view:
>>> b[2, 1, 0] = -1
>>> a[0, 2, 1]
-1

Resizing
Size of an array can be changed with ndarray.resize:
>>> a.resize((8,))
>>> a
array([0, 1, 2, 3, 0, 0, 0, 0])
However, it must not be referred to somewhere else:
>>> b = a
>>> a.resize((4,))
ValueError: cannot resize an array that has been referenced or is
referencing another array in this way. Use the resize function
Exercise: Shape manipulations
Look at the docstring for reshape, especially the notes section which has some more information
about copies and views.
Use flatten as an alternative to ravel. What is the difference? (Hint: check which one returns a view
and which a copy)
Experiment with transpose for dimension shuffling.
3.2.5 Sorting data
Sorting along an axis:
>>> a = np.array([[4, 3, 5], [1, 2, 1]])

>>> b = np.sort(a, axis=1)
>>> b
array([[3, 4, 5],
[1, 1, 2]])
Note: Sorts each row separately!
In-place sort:
>>> a.sort(axis=1)
>>> a
array([[3, 4, 5],
[1, 1, 2]])
Sorting with fancy indexing:
>>> a = np.array([4, 3, 1, 2])

>>> j = np.argsort(a)
>>> j
array([2, 3, 1, 0])
>>> a[j]
array([1, 2, 3, 4])
Finding minima and maxima:

>>> a = np.array([4, 3, 1, 2])

>>> j_max = np.argmax(a)
>>> j_min = np.argmin(a)
>>> j_max, j_min
(0, 2)
Exercise: Sorting
Try both in-place and out-of-place sorting.

Try creating arrays with different dtypes and sorting them.
Use all or array_equal to check the results.
Look at np.random.shuffle for a way to create sortable input quicker.
Combine ravel, sort and reshape.
Look at the axis keyword for sort and rewrite the previous exercise.
3.2.6 Summary
What do you need to know to get started?

Know how to create arrays : array, arange, ones, zeros.
Know the shape of the array with array.shape, then use slicing to obtain different views of the array:
array[::2], etc. Adjust the shape of the array using reshape or flatten it with ravel.
Obtain a subset of the elements of an array and/or modify their values with masks
>>> a[a < 0] = 0
Know miscellaneous operations on arrays, such as finding the mean or max (array.max(), array.
mean()). No need to retain everything, but have the reflex to search in the documentation (online docs,
help(), lookfor())!!
For advanced use: master the indexing with arrays of integers, as well as broadcasting. Know more
NumPy functions to handle various array operations.
Quick read
to the next chapter: Matplotlib: plotting.
and finish this chapter, as well as to do some more exercices.
3.3 More elaborate arrays
Section contents
More data types

Structured data types
maskedarray: dealing with (propagation of ) missing data
3.3. More elaborate arrays 72

3.3.1 More data types
Casting
Bigger type wins in mixed-type operations:
>>> np.array([1, 2, 3]) + 1.5

array([ 2.5, 3.5, 4.5])
Assignment never changes the type!
>>> a = np.array([1, 2, 3])

>>> a.dtype
dtype('int64')
>>> a[0] = 1.9 # <-- float is truncated to integer
>>> a
array([1, 2, 3])
Forced casts:
>>> a = np.array([1.7, 1.2, 1.6])

>>> b = a.astype(int) # <-- truncates to integer
>>> b
array([1, 1, 1])
Rounding:
>>> a = np.array([1.2, 1.5, 1.6, 2.5, 3.5, 4.5])

>>> b = np.around(a)
>>> b # still floating-point
array([ 1., 2., 2., 2., 4., 4.])
>>> c = np.around(a).astype(int)
>>> c
array([1, 2, 2, 2, 4, 4])
Different data type sizes
Integers (signed):
int8 8 bits
int16 16 bits
int32 32 bits (same as int on 32-bit platform)
int64 64 bits (same as int on 64-bit platform)
>>> np.array([1], dtype=int).dtype

dtype('int64')
>>> np.iinfo(np.int32).max, 2**31 - 1
(2147483647, 2147483647)
Unsigned integers:
uint8 8 bits
uint16 16 bits
uint32 32 bits
uint64 64 bits

>>> np.iinfo(np.uint32).max, 2**32 - 1

(4294967295, 4294967295)
Long integers
Python 2 has a specific type for long integers, that cannot overflow, represented with an L at the end. In
Python 3, all integers are long, and thus cannot overflow.
>>> np.iinfo(np.int64).max, 2**63 - 1
(9223372036854775807, 9223372036854775807L)
Floating-point numbers:
float16 16 bits
float32 32 bits
float64 64 bits (same as float)
float96 96 bits, platform-dependent (same as np.longdouble)
float128 128 bits, platform-dependent (same as np.longdouble)
>>> np.finfo(np.float32).eps
1.1920929e-07
>>> np.finfo(np.float64).eps
2.2204460492503131e-16
>>> np.float32(1e-8) + np.float32(1) == 1

True
>>> np.float64(1e-8) + np.float64(1) == 1
False
Complex floating-point numbers:
complex64 two 32-bit floats

complex128 two 64-bit floats
complex192 two 96-bit floats, platform-dependent
complex256 two 128-bit floats, platform-dependent
Smaller data types
If you dont know you need special data types, then you probably dont.
Comparison on using float32 instead of float64:
Half the size in memory and on disk
Half the memory bandwidth required (may be a bit faster in some operations)
In [1]: a = np.zeros((1e6,), dtype=np.float64)
In [2]: b = np.zeros((1e6,), dtype=np.float32)
In [3]: %timeit a*a

1000 loops, best of 3: 1.78 ms per loop
In [4]: %timeit b*b

But: bigger rounding errors sometimes in surprising places (i.e., dont use them unless you really
need them)

3.3.2 Structured data types
sensor_code (4-character string)

position (float)
value (float)
>>> samples = np.zeros((6,), dtype=[('sensor_code', 'S4'),

... ('position', float), ('value', float)])
>>> samples.ndim
1
>>> samples.shape
(6,)
>>> samples.dtype.names
('sensor_code', 'position', 'value')
>>> samples[:] = [('ALFA', 1, 0.37), ('BETA', 1, 0.11), ('TAU', 1, 0.13),

... ('ALFA', 1.5, 0.37), ('ALFA', 3, 0.11), ('TAU', 1.2, 0.13)]
>>> samples
array([('ALFA', 1.0, 0.37), ('BETA', 1.0, 0.11), ('TAU', 1.0, 0.13),
('ALFA', 1.5, 0.37), ('ALFA', 3.0, 0.11), ('TAU', 1.2, 0.13)],
dtype=[('sensor_code', 'S4'), ('position', '<f8'), ('value', '<f8')])
Field access works by indexing with field names:
>>> samples['sensor_code']
array(['ALFA', 'BETA', 'TAU', 'ALFA', 'ALFA', 'TAU'],
dtype='|S4')
>>> samples['value']
array([ 0.37, 0.11, 0.13, 0.37, 0.11, 0.13])
>>> samples[0]
('ALFA', 1.0, 0.37)
>>> samples[0]['sensor_code'] = 'TAU'

>>> samples[0]
('TAU', 1.0, 0.37)
Multiple fields at once:
>>> samples[['position', 'value']]

array([(1.0, 0.37), (1.0, 0.11), (1.0, 0.13), (1.5, 0.37), (3.0, 0.11),
(1.2, 0.13)],
dtype=[('position', '<f8'), ('value', '<f8')])
Fancy indexing works, as usual:
>>> samples[samples['sensor_code'] == 'ALFA']

array([('ALFA', 1.5, 0.37), ('ALFA', 3.0, 0.11)],
dtype=[('sensor_code', 'S4'), ('position', '<f8'), ('value', '<f8')])
Note: There are a bunch of other syntaxes for constructing structured arrays, see here and here.
3.3.3 maskedarray: dealing with (propagation of) missing data
For floats one could use NaNs, but masks work for all types:
>>> x = np.ma.array([1, 2, 3, 4], mask=[0, 1, 0, 1])

>>> x

masked_array(data = [1 -- 3 --],
mask = [False True False True],
fill_value = 999999)
>>> y = np.ma.array([1, 2, 3, 4], mask=[0, 1, 1, 1])

>>> x + y
masked_array(data = [2 -- -- --],
mask = [False True True True],
Masking versions of common functions:
>>> np.ma.sqrt([1, -1, 2, -2])

masked_array(data = [1.0 -- 1.41421356237... --],
mask = [False True False True],
fill_value = 1e+20)
Note: There are other useful array siblings
While it is off topic in a chapter on numpy, lets take a moment to recall good coding practice, which really do
pay off in the long run:
Good practices
Explicit variable names (no need of a comment to explain what is in the variable)
Style: spaces after commas, around =, etc.
A certain number of rules for writing beautiful code (and, more importantly, using the same conven-
tions as everybody else!) are given in the Style Guide for Python Code and the Docstring Conventions
page (to manage help strings).
Except some rare cases, variable names and comments in English.
3.4 Advanced operations
Section contents
Polynomials
Loading data files
3.4.1 Polynomials
NumPy also contains polynomials in different bases:

For example, 3x 2 + 2x 1:
>>> p = np.poly1d([3, 2, -1])

>>> p(0)
-1
>>> p.roots
3.4. Advanced operations 76

array([-1. , 0.33333333])
>>> p.order
2
>>> x = np.linspace(0, 1, 20)

>>> y = np.cos(x) + 0.3*np.random.rand(20)
>>> p = np.poly1d(np.polyfit(x, y, 3))
>>> t = np.linspace(0, 1, 200)

>>> plt.plot(x, y, 'o', t, p(t), '-')
See http://docs.scipy.org/doc/numpy/reference/
routines.polynomials.poly1d.html for more.
More polynomials (with more bases)
NumPy also has a more sophisticated polynomial interface, which supports e.g. the Chebyshev basis.
3x 2 + 2x 1:
>>> p = np.polynomial.Polynomial([-1, 2, 3]) # coefs in different order!

>>> p(0)
-1.0
>>> p.roots()
array([-1. , 0.33333333])
>>> p.degree() # In general polynomials do not always expose 'order'
2
Example using polynomials in Chebyshev basis, for polynomials in range [-1, 1]:
>>> x = np.linspace(-1, 1, 2000)

>>> y = np.cos(x) + 0.3*np.random.rand(2000)
>>> p = np.polynomial.Chebyshev.fit(x, y, 90)
>>> t = np.linspace(-1, 1, 200)

>>> plt.plot(x, y, 'r.')
>>> plt.plot(t, p(t), 'k-', lw=3)

The Chebyshev polynomials have some advantages in

interpolation.
3.4.2 Loading data files
Text files
Example: populations.txt:
# year hare lynx carrot
1900 30e3 4e3 48300
1901 47.2e3 6.1e3 48200
1902 70.2e3 9.8e3 41500
1903 77.4e3 35.2e3 38200

>>> data
array([[ 1900., 30000., 4000., 48300.],
[ 1901., 47200., 6100., 48200.],
[ 1902., 70200., 9800., 41500.],
...
>>> np.savetxt('pop2.txt', data)

>>> data2 = np.loadtxt('pop2.txt')
Note: If you have a complicated text file, what you can try are:
np.genfromtxt
Using Pythons I/O functions and e.g. regexps for parsing (Python is quite well suited for this)
Reminder: Navigating the filesystem with IPython
In [1]: pwd # show current directory

'/home/user/stuff/2011-numpy-tutorial'
In [2]: cd ex
'/home/user/stuff/2011-numpy-tutorial/ex'
In [3]: ls
populations.txt species.txt

Images
Using Matplotlib:
>>> img = plt.imread('data/elephant.png')

>>> img.shape, img.dtype
((200, 300, 3), dtype('float32'))
>>> plt.imshow(img)
<matplotlib.image.AxesImage object at ...>
>>> plt.savefig('plot.png')
>>> plt.imsave('red_elephant.png', img[:,:,0], cmap=plt.cm.gray)
This saved only one channel (of RGB):
>>> plt.imshow(plt.imread('red_elephant.png'))
Other libraries:
>>> from scipy.misc import imsave

>>> imsave('tiny_elephant.png', img[::6,::6])
>>> plt.imshow(plt.imread('tiny_elephant.png'), interpolation='nearest')

NumPys own format
NumPy has its own binary format, not portable but with efficient I/O:
>>> data = np.ones((3, 3))

>>> np.save('pop.npy', data)
>>> data3 = np.load('pop.npy')
Well-known (& more obscure) file formats
HDF5: h5py, PyTables

NetCDF: scipy.io.netcdf_file, netcdf4-python, . . .
Matlab: scipy.io.loadmat, scipy.io.savemat
MatrixMarket: scipy.io.mmread, scipy.io.mmwrite
IDL: scipy.io.readsav
. . . if somebody uses it, theres probably also a Python library for it.
Exercise: Text data files
Write a Python script that loads data from populations.txt:: and drop the last column and the first 5 rows.
Save the smaller dataset to pop2.txt.
NumPy internals
If you are interested in the NumPy internals, there is a good discussion in Advanced NumPy.
3.5 Some exercises
3.5.1 Array manipulations
1. Form the 2-D array (without typing it in explicitly):
3.5. Some exercises 80

[[1, 6, 11],
[2, 7, 12],
[3, 8, 13],
[4, 9, 14],
[5, 10, 15]]
and generate a new array containing its 2nd and 4th rows.
2. Divide each column of the array:

>>> a = np.arange(25).reshape(5, 5)
elementwise with the array b = np.array([1., 5, 10, 15, 20]). (Hint: np.newaxis).
3. Harder one: Generate a 10 x 3 array of random numbers (in range [0,1]). For each row, pick the number
closest to 0.5.
Use abs and argsort to find the column j closest for each row.
Use fancy indexing to extract the numbers. (Hint: a[i,j] the array i must contain the row num-
bers corresponding to stuff in j.)
3.5.2 Picture manipulation: Framing a Face
Lets do some manipulations on numpy arrays by starting with an image of a racoon. scipy provides a 2D
array of this image with the scipy.misc.face function:
>>> from scipy import misc

>>> face = misc.face(gray=True) # 2D grayscale image
Here are a few images we will be able to obtain with our manipulations: use different colormaps, crop the
image, change some parts of the image.
Lets use the imshow function of pylab to display the image.
>>> import pylab as plt

>>> face = misc.face(gray=True)
>>> plt.imshow(face)
<matplotlib.image.AxesImage object at 0x...>
The face is displayed in false colors. A colormap must be specified for it to be displayed in grey.
>>> plt.imshow(face, cmap=plt.cm.gray)

Create an array of the image with a narrower centering [for example,] remove 100 pixels from all the
borders of the image. To check the result, display this new array with imshow.
>>> crop_face = face[100:-100, 100:-100]
We will now frame the face with a black locket. For this, we need to create a mask corresponding to
the pixels we want to be black. The center of the face is around (660, 330), so we defined the mask
by this condition (y-300)**2 + (x-660)**2

>>> sy, sx = face.shape

>>> y, x = np.ogrid[0:sy, 0:sx] # x and y indices of pixels
>>> y.shape, x.shape
((768, 1), (1, 1024))
>>> centerx, centery = (660, 300) # center of the image
>>> mask = ((y - centery)**2 + (x - centerx)**2) > 230**2 # circle
then we assign the value 0 to the pixels of the image corresponding to the mask. The syntax is
extremely simple and intuitive:
>>> face[mask] = 0
>>> plt.imshow(face)
Follow-up: copy all instructions of this exercise in a script called face_locket.py then execute this
script in IPython with %run face_locket.py.
Change the circle to an ellipsoid.
3.5.3 Data statistics
The data in populations.txt describes the populations of hares and lynxes (and carrots) in northern Canada
during 20 years:

>>> year, hares, lynxes, carrots = data.T # trick: columns to variables
>>> import matplotlib.pyplot as plt

>>> plt.axes([0.2, 0.1, 0.5, 0.8])
<matplotlib.axes...Axes object at ...>
>>> plt.plot(year, hares, year, lynxes, year, carrots)
[<matplotlib.lines.Line2D object at ...>, ...]
>>> plt.legend(('Hare', 'Lynx', 'Carrot'), loc=(1.05, 0.5))
<matplotlib.legend.Legend object at ...>
Computes and print, based on the data in

populations.txt. . .
1. The mean and std of the populations of each species for the years in the period.
2. Which year each species had the largest population.
3. Which species has the largest population for each year. (Hint: argsort & fancy indexing of np.
array(['H', 'L', 'C']))
4. Which years any of the populations is above 50000. (Hint: comparisons and np.any)
5. The top 2 years for each species when they had the lowest populations. (Hint: argsort, fancy indexing)

6. Compare (plot) the change in hare population (see help(np.gradient)) and the number of lynxes.
Check correlation (see help(np.corrcoef)).
. . . all without for-loops.
Solution: Python source file
3.5.4 Crude integral approximations
Write a function f(a, b, c) that returns a b c. Form a 24x12x6 array containing its values in parameter
ranges [0,1] x [0,1] x [0,1].
Approximate the 3-d integral
Z 1Z 1Z 1
(a b c)d a d b d c
0 0 0
over this volume with the mean. The exact result is: ln 2 12 0.1931 . . . what is your relative error?
(Hints: use elementwise operations and broadcasting. You can make np.ogrid give a number of points in
given range with np.ogrid[0:1:20j].)
Reminder Python functions:
def f(a, b, c):

return some_result
3.5.5 Mandelbrot set
Write a script that computes the Mandelbrot fractal.

The Mandelbrot iteration:
N_max = 50
some_threshold = 50
c = x + 1j*y
z = 0
for j in range(N_max):
z = z**2 + c
Point (x, y) belongs to the Mandelbrot set if |z| < some_threshold.

Do this computation by:
1. Construct a grid of c = x + 1j*y values in range [-2, 1] x [-1.5, 1.5]

2. Do the iteration
3. Form the 2-d boolean mask indicating which points are in the set
4. Save the result to an image with:

>>> plt.imshow(mask.T, extent=[-2, 1, -1.5, 1.5])
>>> plt.gray()
>>> plt.savefig('mandelbrot.png')
3.5.6 Markov chain
Markov chain transition matrix P, and probability distribution on the states p:

1. 0 <= P[i,j] <= 1: probability to go from state i to state j
2. Transition rule: p new = P T p ol d
3. all(sum(P, axis=1) == 1), p.sum() == 1: normalization
Write a script that works with 5 states, and:
Constructs a random matrix, and normalizes each row so that it is a transition matrix.
Starts from a random (normalized) probability distribution p and takes 50 steps => p_50
Computes the stationary distribution: the eigenvector of P.T with eigenvalue 1 (numerically: closest to
1) => p_stationary
Remember to normalize the eigenvector I didnt. . .
Checks if p_50 and p_stationary are equal to tolerance 1e-5
Toolbox: np.random.rand, .dot(), np.linalg.eig, reductions, abs(), argmin, comparisons, all, np.
linalg.norm, etc.

3.6 Full code examples
3.6.1 Full code examples for the numpy chapter
2D plotting
Plot a basic 2D figure
import numpy as np
import matplotlib.pyplot as plt
image = np.random.rand(30, 30)

plt.imshow(image, cmap=plt.cm.hot)
plt.colorbar()
plt.show()
Total running time of the script: ( 0 minutes 0.125 seconds)

Download Python source code: plot_basic2dplot.py
Download Jupyter notebook: plot_basic2dplot.ipynb
Generated by Sphinx-Gallery
1D plotting
Plot a basic 1D figure
3.6. Full code examples 85

import numpy as np
x = np.linspace(0, 3, 20)
y = np.linspace(0, 9, 20)
plt.plot(x, y)
plt.plot(x, y, 'o')
plt.show()

Download Python source code: plot_basic1dplot.py
Download Jupyter notebook: plot_basic1dplot.ipynb
Distances exercise
Plot distances in a grid

import numpy as np
x, y = np.arange(5), np.arange(5)[:, np.newaxis]

distance = np.sqrt(x ** 2 + y ** 2)
plt.pcolor(distance)
plt.colorbar()
plt.show()

Download Python source code: plot_distances.py
Download Jupyter notebook: plot_distances.ipynb
Fitting to polynomial
Plot noisy data and their polynomial fit

import numpy as np
np.random.seed(12)
y = np.cos(x) + 0.3*np.random.rand(20)
p = np.poly1d(np.polyfit(x, y, 3))
t = np.linspace(0, 1, 200)
plt.plot(x, y, 'o', t, p(t), '-')
plt.show()

Download Python source code: plot_polyfit.py
Download Jupyter notebook: plot_polyfit.ipynb
Fitting in Chebyshev basis
Plot noisy data and their polynomial fit in a Chebyshev basis

import numpy as np
np.random.seed(0)
x = np.linspace(-1, 1, 2000)
y = np.cos(x) + 0.3*np.random.rand(2000)
p = np.polynomial.Chebyshev.fit(x, y, 90)
t = np.linspace(-1, 1, 200)
plt.plot(x, y, 'r.')
plt.plot(t, p(t), 'k-', lw=3)
plt.show()

Download Python source code: plot_chebyfit.py
Download Jupyter notebook: plot_chebyfit.ipynb
Population exercise
Plot populations of hares, lynxes, and carrots

import numpy as np
data = np.loadtxt('../data/populations.txt')
year, hares, lynxes, carrots = data.T
plt.axes([0.2, 0.1, 0.5, 0.8])

plt.plot(year, hares, year, lynxes, year, carrots)
plt.legend(('Hare', 'Lynx', 'Carrot'), loc=(1.05, 0.5))
plt.show()

Download Python source code: plot_populations.py
Download Jupyter notebook: plot_populations.ipynb
Reading and writing an elephant
Read and write images
import numpy as np

original figure
plt.figure()
img = plt.imread('../data/elephant.png')
plt.imshow(img)
red channel displayed in grey
plt.figure()
img_red = img[:, :, 0]
plt.imshow(img_red, cmap=plt.cm.gray)

lower resolution
plt.figure()
img_tiny = img[::6, ::6]
plt.imshow(img_tiny, interpolation='nearest')
plt.show()


Download Python source code: plot_elephant.py
Download Jupyter notebook: plot_elephant.ipynb
Mandelbrot set
Compute the Mandelbrot fractal and plot it

import numpy as np
from numpy import newaxis
def compute_mandelbrot(N_max, some_threshold, nx, ny):

# A grid of c-values
x = np.linspace(-2, 1, nx)
y = np.linspace(-1.5, 1.5, ny)
c = x[:,newaxis] + 1j*y[newaxis,:]
# Mandelbrot iteration
z = c
for j in range(N_max):
z = z**2 + c
mandelbrot_set = (abs(z) < some_threshold)
return mandelbrot_set
mandelbrot_set = compute_mandelbrot(50, 50., 601, 401)
plt.imshow(mandelbrot_set.T, extent=[-2, 1, -1.5, 1.5])

plt.gray()
plt.show()

Download Python source code: plot_mandelbrot.py
Download Jupyter notebook: plot_mandelbrot.ipynb

Random walk exercise
Plot distance as a function of time for a random walk together with the theoretical result
import numpy as np
# We create 1000 realizations with 200 steps each

n_stories = 1000
t_max = 200
t = np.arange(t_max)
# Steps can be -1 or 1 (note that randint excludes the upper limit)
steps = 2 * np.random.randint(0, 1 + 1, (n_stories, t_max)) - 1
# The time evolution of the position is obtained by successively

# summing up individual steps. This is done for each of the
# realizations, i.e. along axis 1.
positions = np.cumsum(steps, axis=1)
# Determine the time evolution of the mean square distance.

sq_distance = positions**2
mean_sq_distance = np.mean(sq_distance, axis=0)
# Plot the distance d from the origin as a function of time and

# compare with the theoretically expected result where d(t)
# grows as a square root of time t.
plt.figure(figsize=(4, 3))
plt.plot(t, np.sqrt(mean_sq_distance), 'g.', t, np.sqrt(t), 'y-')
plt.xlabel(r"$t$")
plt.ylabel(r"$\sqrt{\langle (\delta x)^2 \rangle}$")
plt.tight_layout()
plt.show()

Download Python source code: plot_randomwalk.py
Download Jupyter notebook: plot_randomwalk.ipynb

Download all examples in Python source code: auto_examples_python.zip

Download all examples in Jupyter notebooks: auto_examples_jupyter.zip

CHAPTER 4
Matplotlib: plotting
Thanks
Many thanks to Bill Wing and Christoph Deil for review and corrections.
Authors: Nicolas Rougier, Mike Mller, Gal Varoquaux
Chapter contents
Introduction
Simple plot
Figures, Subplots, Axes and Ticks
Other Types of Plots: examples and exercises
Beyond this tutorial
Quick references
Full code examples
4.1 Introduction
Tip: Matplotlib is probably the most used Python package for 2D-graphics. It provides both a quick way to
visualize data from Python and publication-quality figures in many formats. We are going to explore matplotlib
97
in interactive mode covering most common cases.
4.1.1 IPython, Jupyter, and matplotlib modes
Tip: The Jupyter notebook and the IPython enhanced interactive Python, are tuned for the scientific-
computing workflow in Python, in combination with Matplotlib:
For interactive matplotlib sessions, turn on the matplotlib mode

IPython console When using the IPython console, use:
In [1]: %matplotlib
Jupyter notebook In the notebook, insert, at the beginning of the notebook the following magic:
%matplotlib inline
4.1.2 pyplot
Tip: pyplot provides a procedural interface to the matplotlib object-oriented plotting library. It is modeled
closely after Matlab. Therefore, the majority of plotting commands in pyplot have Matlab analogs with
similar arguments. Important commands are explained with interactive examples.
from matplotlib import pyplot as plt
4.2 Simple plot
Tip: In this section, we want to draw the cosine and sine functions on the same plot. Starting from the default
settings, well enrich the figure step by step to make it nicer.
First step is to get the data for the sine and cosine functions:
import numpy as np
X = np.linspace(-np.pi, np.pi, 256, endpoint=True)

C, S = np.cos(X), np.sin(X)
X is now a numpy array with 256 values ranging from - to + (included). C is the cosine (256 values) and S is
the sine (256 values).
To run the example, you can type them in an IPython interactive session:
$ ipython --pylab
This brings us to the IPython prompt:
IPython 0.13 -- An enhanced Interactive Python.

? -> Introduction to IPython's features.
%magic -> Information about IPython's 'magic' % functions.
help -> Python's own help system.
object? -> Details about 'object'. ?object also works, ?? prints more.
4.2. Simple plot 98

Welcome to pylab, a matplotlib-based Python environment.

For more information, type 'help(pylab)'.
Tip: You can also download each of the examples and run it using regular python, but you will lose interactive
data manipulation:
$ python plot_exercise_1.py
You can get source for each step by clicking on the corresponding figure.
4.2.1 Plotting with default settings
Hint: Documentation
plot tutorial
plot() command
Tip: Matplotlib comes with a set of default settings that allow customizing all kinds of properties. You can
control the defaults of almost every property in matplotlib: figure size and dpi, line width, color and style,
axes, axis and grid properties, text and font properties and so on.
import numpy as np

plt.plot(X, C)
plt.plot(X, S)
plt.show()
4.2. Simple plot 99

4.2.2 Instantiating defaults
Hint: Documentation
Customizing matplotlib
In the script below, weve instantiated (and commented) all the figure settings that influence the appearance
of the plot.
Tip: The settings have been explicitly set to their default values, but now you can interactively play with the
values to explore their affect (see Line properties and Line styles below).
import numpy as np
# Create a figure of size 8x6 inches, 80 dots per inch

plt.figure(figsize=(8, 6), dpi=80)
# Create a new subplot from a grid of 1x1

plt.subplot(1, 1, 1)

# Plot cosine with a blue continuous line of width 1 (pixels)

plt.plot(X, C, color="blue", linewidth=1.0, linestyle="-")
# Plot sine with a green continuous line of width 1 (pixels)

plt.plot(X, S, color="green", linewidth=1.0, linestyle="-")
# Set x limits
plt.xlim(-4.0, 4.0)
# Set x ticks
plt.xticks(np.linspace(-4, 4, 9, endpoint=True))
# Set y limits
plt.ylim(-1.0, 1.0)
# Set y ticks
plt.yticks(np.linspace(-1, 1, 5, endpoint=True))
# Save figure using 72 dots per inch

# plt.savefig("exercise_2.png", dpi=72)
4.2. Simple plot 100

# Show result on screen

plt.show()
4.2.3 Changing colors and line widths
Hint: Documentation
Controlling line properties
Line API
Tip: First step, we want to have the cosine in blue and the sine in red and a slighty thicker line for both of
them. Well also slightly alter the figure size to make it more horizontal.
...
plt.plot(X, S, color="red", linewidth=2.5, linestyle="-")
...
4.2.4 Setting limits
Hint: Documentation
xlim() command
ylim() command

Tip: Current limits of the figure are a bit too tight and we want to make some space in order to clearly see all
data points.
...
plt.xlim(X.min() * 1.1, X.max() * 1.1)
plt.ylim(C.min() * 1.1, C.max() * 1.1)
...
4.2.5 Setting ticks
Hint: Documentation
xticks() command
yticks() command
Tick container
Tick locating and formatting
Tip: Current ticks are not ideal because they do not show the interesting values (+/-,+/-/2) for sine and
cosine. Well change them such that they show only these values.
...
plt.xticks([-np.pi, -np.pi/2, 0, np.pi/2, np.pi])
plt.yticks([-1, 0, +1])
...
4.2.6 Setting tick labels

Hint: Documentation
Working with text
xticks() command
yticks() command
set_xticklabels()
set_yticklabels()
Tip: Ticks are now properly placed but their label is not very explicit. We could guess that 3.142 is but it
would be better to make it explicit. When we set tick values, we can also provide a corresponding label in the
second argument list. Note that well use latex to allow for nice rendering of the label.
...
plt.xticks([-np.pi, -np.pi/2, 0, np.pi/2, np.pi],
[r'$-\pi$', r'$-\pi/2$', r'$0$', r'$+\pi/2$', r'$+\pi$'])
plt.yticks([-1, 0, +1],
[r'$-1$', r'$0$', r'$+1$'])
...
4.2.7 Moving spines
Hint: Documentation
Spines
Axis container
Transformations tutorial
Tip: Spines are the lines connecting the axis tick marks and noting the boundaries of the data area. They can
be placed at arbitrary positions and until now, they were on the border of the axis. Well change that since we
want to have them in the middle. Since there are four of them (top/bottom/left/right), well discard the top
and right by setting their color to none and well move the bottom and left ones to coordinate 0 in data space
coordinates.
...
ax = plt.gca() # gca stands for 'get current axis'
ax.spines['right'].set_color('none')

ax.spines['top'].set_color('none')
ax.xaxis.set_ticks_position('bottom')
ax.spines['bottom'].set_position(('data',0))
ax.yaxis.set_ticks_position('left')
ax.spines['left'].set_position(('data',0))
...
4.2.8 Adding a legend
Hint: Documentation
Legend guide
legend() command
Legend API
Tip: Lets add a legend in the upper left corner. This only requires adding the keyword argument label (that
will be used in the legend box) to the plot commands.
...
plt.plot(X, C, color="blue", linewidth=2.5, linestyle="-", label="cosine")
plt.plot(X, S, color="red", linewidth=2.5, linestyle="-", label="sine")
plt.legend(loc='upper left')
...
4.2.9 Annotate some points
Hint: Documentation

Annotating axis
annotate() command
Tip: Lets annotate some interesting points using the annotate command. We chose the 2/3 value and we
want to annotate both the sine and the cosine. Well first draw a marker on the curve as well as a straight dotted
line. Then, well use the annotate command to display some text with an arrow.
...
t = 2 * np.pi / 3
plt.plot([t, t], [0, np.cos(t)], color='blue', linewidth=2.5, linestyle="--")
plt.scatter([t, ], [np.cos(t), ], 50, color='blue')
plt.annotate(r'$cos(\frac{2\pi}{3} )=-\frac{1} {2} $',

xy=(t, np.cos(t)), xycoords='data',
xytext=(-90, -50), textcoords='offset points', fontsize=16,
arrowprops=dict(arrowstyle="->", connectionstyle="arc3,rad=.2"))
plt.plot([t, t],[0, np.sin(t)], color='red', linewidth=2.5, linestyle="--")

plt.scatter([t, ],[np.sin(t), ], 50, color='red')
plt.annotate(r'$sin(\frac{2\pi}{3} )=\frac{\sqrt{3} }{2} $',

xy=(t, np.sin(t)), xycoords='data',
xytext=(+10, +30), textcoords='offset points', fontsize=16,
...
4.2.10 Devil is in the details
Hint: Documentation
Artists
BBox
Tip: The tick labels are now hardly visible because of the blue and red lines. We can make them bigger and we
can also adjust their properties such that theyll be rendered on a semi-transparent white background. This
will allow us to see both the data and the labels.
...
for label in ax.get_xticklabels() + ax.get_yticklabels():
label.set_fontsize(16)

label.set_bbox(dict(facecolor='white', edgecolor='None', alpha=0.65))

...
4.3 Figures, Subplots, Axes and Ticks
A figure in matplotlib means the whole window in the user interface. Within this figure there can be sub-
plots.
Tip: So far we have used implicit figure and axes creation. This is handy for fast plots. We can have more
control over the display using figure, subplot, and axes explicitly. While subplot positions the plots in a regular
grid, axes allows free placement within the figure. Both can be useful depending on your intention. Weve
already worked with figures and subplots without explicitly calling them. When we call plot, matplotlib calls
gca() to get the current axes and gca in turn calls gcf() to get the current figure. If there is none it calls
figure() to make one, strictly speaking, to make a subplot(111). Lets look at the details.
4.3.1 Figures
Tip: A figure is the windows in the GUI that has Figure # as title. Figures are numbered starting from 1 as
opposed to the normal Python way starting from 0. This is clearly MATLAB-style. There are several parameters
that determine what the figure looks like:
Argument Default Description

num 1 number of figure
figsize figure.figsize figure size in inches (width, height)
dpi figure.dpi resolution in dots per inch
facecolor figure.facecolor color of the drawing background
edgecolor figure.edgecolor color of edge around the drawing background
frameon True draw figure frame or not
Tip: The defaults can be specified in the resource file and will be used most of the time. Only the number of
the figure is frequently changed.
As with other objects, you can set figure properties also setp or with the set_something methods.
When you work with the GUI you can close a figure by clicking on the x in the upper right corner. But you can
close a figure programmatically by calling close. Depending on the argument it closes (1) the current figure
(no argument), (2) a specific figure (figure number or figure instance as argument), or (3) all figures ("all" as
argument).
plt.close(1) # Closes figure 1
4.3.2 Subplots
Tip: With subplot you can arrange plots in a regular grid. You need to specify the number of rows and columns
and the number of the plot. Note that the gridspec command is a more powerful alternative.
4.3. Figures, Subplots, Axes and Ticks 106

4.3.3 Axes
Axes are very similar to subplots but allow placement of plots at any location in the fig-
ure. So if we want to put a smaller plot inside a bigger one we do so with axes.
4.3.4 Ticks
Well formatted ticks are an important part of publishing-ready figures. Matplotlib provides a totally config-
urable system for ticks. There are tick locators to specify where ticks should appear and tick formatters to
give ticks the appearance you want. Major and minor ticks can be located and formatted independently from
each other. Per default minor ticks are not shown, i.e. there is only an empty list for them because it is as
NullLocator (see below).
Tick Locators
Tick locators control the positions of the ticks. They are set as follows:
ax = plt.gca()
ax.xaxis.set_major_locator(eval(locator))
There are several locators for different kind of requirements:

All of these locators derive from the base class matplotlib.ticker.Locator. You can make your own lo-
4.3. Figures, Subplots, Axes and Ticks 107

cator deriving from it. Handling dates as ticks can be especially tricky. Therefore, matplotlib provides special
locators in matplotlib.dates.
4.4 Other Types of Plots: examples and exercises
4.4.1 Regular Plots
Starting from the code below, try to reproduce the graphic

taking care of filled areas:
Hint: You need to use the fill_between command.
n = 256
X = np.linspace(-np.pi, np.pi, n, endpoint=True)
Y = np.sin(2 * X)
plt.plot(X, Y + 1, color='blue', alpha=1.00)

plt.plot(X, Y - 1, color='blue', alpha=1.00)
Click on the figure for solution.
4.4. Other Types of Plots: examples and exercises 108

4.4.2 Scatter Plots

taking care of marker size, color and transparency.
Hint: Color is given by angle of (X,Y).
n = 1024
X = np.random.normal(0,1,n)
Y = np.random.normal(0,1,n)
plt.scatter(X,Y)
Click on figure for solution.
4.4.3 Bar Plots
Starting from the code below, try to reproduce the graphic by

adding labels for red bars.
Hint: You need to take care of text alignment.
n = 12
X = np.arange(n)
Y1 = (1 - X / float(n)) * np.random.uniform(0.5, 1.0, n)
plt.bar(X, +Y1, facecolor='#9999ff', edgecolor='white')

plt.bar(X, -Y2, facecolor='#ff9999', edgecolor='white')
for x, y in zip(X, Y1):

plt.text(x + 0.4, y + 0.05, '%.2f ' % y, ha='center', va='bottom')
plt.ylim(-1.25, +1.25)
4.4.4 Contour Plots

taking care of the colormap (see Colormaps below).
Hint: You need to use the clabel command.
def f(x, y):

return (1 - x / 2 + x ** 5 + y ** 3) * np.exp(-x ** 2 -y ** 2)
n = 256
x = np.linspace(-3, 3, n)
y = np.linspace(-3, 3, n)
X, Y = np.meshgrid(x, y)
plt.contourf(X, Y, f(X, Y), 8, alpha=.75, cmap='jet')

C = plt.contour(X, Y, f(X, Y), 8, colors='black', linewidth=.5)
4.4.5 Imshow

taking care of colormap, image interpolation and origin.

Hint: You need to take care of the origin of the image in the imshow command and use a colorbar
def f(x, y):

return (1 - x / 2 + x ** 5 + y ** 3) * np.exp(-x ** 2 - y ** 2)
n = 10
x = np.linspace(-3, 3, 4 * n)
y = np.linspace(-3, 3, 3 * n)
plt.imshow(f(X, Y))
Click on the figure for the solution.
4.4.6 Pie Charts

taking care of colors and slices size.
Hint: You need to modify Z.
Z = np.random.uniform(0, 1, 20)
plt.pie(Z)
Click on the figure for the solution.
4.4.7 Quiver Plots
Starting from the code above, try to reproduce the graphic

taking care of colors and orientations.

Hint: You need to draw arrows twice.
n = 8
X, Y = np.mgrid[0:n, 0:n]
plt.quiver(X, Y)
4.4.8 Grids

taking care of line styles.
axes = plt.gca()
axes.set_xlim(0, 4)
axes.set_ylim(0, 3)
axes.set_xticklabels([])
axes.set_yticklabels([])
4.4.9 Multi Plots
Starting from the code below, try to reproduce the graphic.
Hint: You can use several subplots with different partition.

4.4.10 Polar Axis
Hint: You only need to modify the axes line
plt.axes([0, 0, 1, 1])
N = 20
theta = np.arange(0., 2 * np.pi, 2 * np.pi / N)
radii = 10 * np.random.rand(N)
width = np.pi / 4 * np.random.rand(N)
bars = plt.bar(theta, radii, width=width, bottom=0.0)
for r, bar in zip(radii, bars):

bar.set_facecolor(cm.jet(r / 10.))
bar.set_alpha(0.5)
4.4.11 3D Plots
Hint: You need to use contourf

from mpl_toolkits.mplot3d import Axes3D
fig = plt.figure()
ax = Axes3D(fig)
X = np.arange(-4, 4, 0.25)
Y = np.arange(-4, 4, 0.25)
X, Y = np.meshgrid(X, Y)
R = np.sqrt(X**2 + Y**2)
Z = np.sin(R)
ax.plot_surface(X, Y, Z, rstride=1, cstride=1, cmap='hot')

See also:
3D plotting with Mayavi
4.4.12 Text
Try to do the same from scratch !
Hint: Have a look at the matplotlib logo.
Quick read
to the next chapter: Scipy : high-level scientific computing.
and finish this chapter later.

4.5 Beyond this tutorial
Matplotlib benefits from extensive documentation as well as a large community of users and developers. Here
are some links of interest:
4.5.1 Tutorials
Pyplot tutorial
Introduction
Controlling line properties
Working with multiple figures and axes
Working with text
Image tutorial
Startup commands
Importing image data into Numpy arrays
Plotting numpy arrays as images
Text tutorial
Text introduction
Basic text commands
Text properties and layout
Writing mathematical expressions
Text rendering With LaTeX
Annotating text
Artist tutorial
Introduction
Customizing your objects
Object containers
Figure container
Axes container
Axis containers
Tick containers
Path tutorial
Introduction
Bzier example
Compound paths
Transforms tutorial
Introduction
Data coordinates
Axes coordinates
Blended transformations
Using offset transforms to create a shadow effect
The transformation pipeline
4.5.2 Matplotlib documentation
User guide
FAQ
Installation
Usage
How-To
Troubleshooting
Environment Variables
Screenshots
4.5. Beyond this tutorial 115

4.5.3 Code documentation
The code is well documented and you can quickly access a specific command from within a python session:

>>> help(plt.plot)
Help on function plot in module matplotlib.pyplot:
plot(*args, **kwargs)
Plot lines and/or markers to the
:class:`~matplotlib.axes.Axes`. *args* is a variable length
argument, allowing for multiple *x*, *y* pairs with an
optional format string. For example, each of the following is
legal::
plot(x, y) # plot x and y using default line style and color

plot(x, y, 'bo') # plot x and y using blue circle markers
plot(y) # plot y using x as index array 0..N-1
plot(y, 'r+') # ditto, but with red plusses
If *x* and/or *y* is 2-dimensional, then the corresponding columns

will be plotted.
...
4.5.4 Galleries
The matplotlib gallery is also incredibly useful when you search how to render a given graphic. Each example
comes with its source.
4.5.5 Mailing lists
Finally, there is a user mailing list where you can ask for help and a developers mailing list that is more techni-
cal.
4.6 Quick references
Here is a set of tables that show main properties and styles.
4.6. Quick references 116

4.6.1 Line properties
Property Description Appearance

alpha (or a) alpha transparency on 0-1
scale
antialiased True or False - use antialised

rendering
color (or c) matplotlib color arg

linestyle (or ls) see Line properties
linewidth (or lw) float, the line width in points
solid_capstyle Cap style for solid lines
solid_joinstyle Join style for solid lines
dash_capstyle Cap style for dashes
dash_joinstyle Join style for dashes
marker see Markers
markeredgewidth line width around the marker
(mew) symbol
markeredgecolor edge color if a marker is used
(mec)
markerfacecolor face color if a marker is used
(mfc)
markersize (ms) size of the marker in points
4.6.2 Line styles
4.6. Quick references 117

4.6.3 Markers
4.6.4 Colormaps
All colormaps can be reversed by appending _r. For instance, gray_r is the reverse of gray.
If you want to know more about colormaps, checks Documenting the matplotlib colormaps.
4.7.1 Code samples for Matplotlib
The examples here are only examples relevant to the points raised in this chapter. The matplotlib documenta-
tion comes with a much more exhaustive gallery.

Pie chart
A simple pie chart example with matplotlib.
import numpy as np
n = 20
Z = np.ones(n)
Z[-1] *= 2
plt.axes([0.025, 0.025, 0.95, 0.95])
plt.pie(Z, explode=Z*.05, colors = ['%f ' % (i/float(n)) for i in range(n)])

plt.axis('equal')
plt.xticks(())
plt.yticks()
plt.show()

Download Python source code: plot_pie.py
Download Jupyter notebook: plot_pie.ipynb
Plotting a scatter of points
A simple example showing how to plot a scatter of points with matplotlib.

import numpy as np
n = 1024
X = np.random.normal(0, 1, n)
Y = np.random.normal(0, 1, n)
T = np.arctan2(Y, X)
plt.axes([0.025, 0.025, 0.95, 0.95])

plt.scatter(X, Y, s=75, c=T, alpha=.5)
plt.xlim(-1.5, 1.5)
plt.xticks(())
plt.ylim(-1.5, 1.5)
plt.yticks(())
plt.show()

Download Python source code: plot_scatter.py
Download Jupyter notebook: plot_scatter.ipynb
A simple, good-looking plot
Demoing some simple features of matplotlib

import numpy as np
import matplotlib
matplotlib.use('Agg')
fig = plt.figure(figsize=(5, 4), dpi=72)

axes = fig.add_axes([0.01, 0.01, .98, 0.98])
X = np.linspace(0, 2, 200, endpoint=True)
Y = np.sin(2*np.pi*X)
plt.plot(X, Y, lw=2)
plt.ylim(-1.1, 1.1)
plt.grid()
plt.show()

Download Python source code: plot_good.py
Download Jupyter notebook: plot_good.ipynb
Subplots
Show multiple subplots in matplotlib.

fig = plt.figure()
fig.subplots_adjust(bottom=0.025, left=0.025, top = 0.975, right=0.975)
plt.xticks(()), plt.yticks(())
plt.xticks(())
plt.yticks(())
plt.xticks(())
plt.yticks(())
plt.xticks(())
plt.yticks(())
plt.show()

Download Python source code: plot_multiplot.py
Download Jupyter notebook: plot_multiplot.ipynb

Simple axes example
This example shows a couple of simple usage of axes.
plt.axes([.1, .1, .8, .8])

plt.xticks(())
plt.yticks(())
plt.text(.6, .6, 'axes([0.1, 0.1, .8, .8])', ha='center', va='center',
size=20, alpha=.5)
plt.axes([.2, .2, .3, .3])

plt.xticks(())
plt.yticks(())
plt.text(.5, .5, 'axes([0.2, 0.2, .3, .3])', ha='center', va='center',
size=16, alpha=.5)
plt.show()

Download Python source code: plot_axes.py
Download Jupyter notebook: plot_axes.ipynb
A simple plotting example
A plotting example with a few simple tweaks

import numpy as np
import matplotlib
fig = plt.figure(figsize=(5, 4), dpi=72)

axes = fig.add_axes([0.01, 0.01, .98, 0.98])
x = np.linspace(0, 2, 200, endpoint=True)
y = np.sin(2 * np.pi * x)
plt.plot(x, y, lw=.25, c='k')
plt.xticks(np.arange(0.0, 2.0, 0.1))
plt.yticks(np.arange(-1.0, 1.0, 0.1))
plt.grid()
plt.show()

Download Python source code: plot_bad.py
Download Jupyter notebook: plot_bad.ipynb
Horizontal arrangement of subplots
An example showing horizontal arrangement of subplots with matplotlib.

plt.xticks(())
plt.yticks(())
plt.text(0.5, 0.5, 'subplot(2,1,1)', ha='center', va='center',
size=24, alpha=.5)
plt.xticks(())
plt.yticks(())
size=24, alpha=.5)
plt.tight_layout()
plt.show()

Download Python source code: plot_subplot-horizontal.py
Download Jupyter notebook: plot_subplot-horizontal.ipynb
Subplot plot arrangement vertical
An example showing vertical arrangement of subplots with matplotlib.

plt.xticks(())
plt.yticks(())
size=24, alpha=.5)
plt.xticks(())
plt.yticks(())
size=24, alpha=.5)
plt.tight_layout()
plt.show()

Download Python source code: plot_subplot-vertical.py
Download Jupyter notebook: plot_subplot-vertical.ipynb
3D plotting
A simple example of 3D plotting.

import numpy as np
fig = plt.figure()
ax = Axes3D(fig)
X = np.arange(-4, 4, 0.25)
Y = np.arange(-4, 4, 0.25)
R = np.sqrt(X ** 2 + Y ** 2)
Z = np.sin(R)
ax.plot_surface(X, Y, Z, rstride=1, cstride=1, cmap=plt.cm.hot)

ax.contourf(X, Y, Z, zdir='z', offset=-2, cmap=plt.cm.hot)
ax.set_zlim(-2, 2)
plt.show()

Download Python source code: plot_plot3d.py
Download Jupyter notebook: plot_plot3d.ipynb
Imshow elaborate
An example demoing imshow and styling the figure.

import numpy as np
def f(x, y):

return (1 - x / 2 + x ** 5 + y ** 3 ) * np.exp(-x ** 2 - y ** 2)
n = 10
x = np.linspace(-3, 3, 3.5 * n)
y = np.linspace(-3, 3, 3.0 * n)
Z = f(X, Y)
plt.axes([0.025, 0.025, 0.95, 0.95])

plt.imshow(Z, interpolation='nearest', cmap='bone', origin='lower')
plt.colorbar(shrink=.92)
plt.xticks(())
plt.yticks(())
plt.show()

Download Python source code: plot_imshow.py
Download Jupyter notebook: plot_imshow.ipynb
Plotting a vector field: quiver
A simple example showing how to plot a vector field (quiver) with matplotlib.

import numpy as np
n = 8
T = np.arctan2(Y - n / 2., X - n/2.)
R = 10 + np.sqrt((Y - n / 2.0) ** 2 + (X - n / 2.0) ** 2)
U, V = R * np.cos(T), R * np.sin(T)
plt.axes([0.025, 0.025, 0.95, 0.95])

plt.quiver(X, Y, U, V, R, alpha=.5)
plt.quiver(X, Y, U, V, edgecolor='k', facecolor='None', linewidth=.5)
plt.xlim(-1, n)
plt.xticks(())
plt.ylim(-1, n)
plt.yticks(())
plt.show()

Download Python source code: plot_quiver.py
Download Jupyter notebook: plot_quiver.ipynb
Plotting in polar coordinnates
A simple example showing how to plot in polar coordinnates with matplotlib.

import numpy as np
ax = plt.axes([0.025, 0.025, 0.95, 0.95], polar=True)
N = 20
theta = np.arange(0.0, 2 * np.pi, 2 * np.pi / N)
for r,bar in zip(radii, bars):

bar.set_facecolor(plt.cm.jet(r/10.))
bar.set_alpha(0.5)
ax.set_xticklabels([])
ax.set_yticklabels([])
plt.show()

Download Python source code: plot_polar.py
Download Jupyter notebook: plot_polar.ipynb
Displaying the contours of a function
An example showing how to display the contours of a function with matplotlib.

import numpy as np
def f(x,y):
return (1 - x / 2 + x**5 + y**3) * np.exp(-x**2 -y**2)
n = 256
X,Y = np.meshgrid(x, y)
plt.axes([0.025, 0.025, 0.95, 0.95])
plt.contourf(X, Y, f(X, Y), 8, alpha=.75, cmap=plt.cm.hot)

C = plt.contour(X, Y, f(X, Y), 8, colors='black', linewidth=.5)
plt.clabel(C, inline=1, fontsize=10)
plt.xticks(())
plt.yticks(())
plt.show()

Download Python source code: plot_contour.py
Download Jupyter notebook: plot_contour.ipynb

A example of plotting not quite right
An ugly example of plotting.
import numpy as np
import matplotlib
matplotlib.rc('grid', color='black', linestyle='-', linewidth=1)
fig = plt.figure(figsize=(5,4),dpi=72)
axes = fig.add_axes([0.01, 0.01, .98, 0.98], axisbg='.75')
X = np.linspace(0, 2, 40, endpoint=True)
Y = np.sin(2 * np.pi * X)
plt.plot(X, Y, lw=.05, c='b', antialiased=False)
plt.xticks(())
plt.yticks(np.arange(-1., 1., 0.2))
plt.grid()
ax = plt.gca()
plt.show()

Download Python source code: plot_ugly.py
Download Jupyter notebook: plot_ugly.ipynb
Plot and filled plots
Simple example of plots and filling between them with matplotlib.

import numpy as np
n = 256
X = np.linspace(-np.pi, np.pi, n, endpoint=True)
Y = np.sin(2 * X)
plt.axes([0.025, 0.025, 0.95, 0.95])
plt.plot(X, Y + 1, color='blue', alpha=1.00)

plt.fill_between(X, 1, Y + 1, color='blue', alpha=.25)
plt.plot(X, Y - 1, color='blue', alpha=1.00)

plt.fill_between(X, -1, Y - 1, (Y - 1) > -1, color='blue', alpha=.25)
plt.fill_between(X, -1, Y - 1, (Y - 1) < -1, color='red', alpha=.25)
plt.xlim(-np.pi, np.pi)
plt.xticks(())
plt.ylim(-2.5, 2.5)
plt.yticks(())
plt.show()

Download Python source code: plot_plot.py
Download Jupyter notebook: plot_plot.ipynb

Bar plots
An example of bar plots with matplotlib.
import numpy as np
n = 12
X = np.arange(n)
plt.axes([0.025, 0.025, 0.95, 0.95])

plt.bar(X, +Y1, facecolor='#9999ff', edgecolor='white')

plt.text(x + 0.4, y + 0.05, '%.2f ' % y, ha='center', va= 'bottom')

plt.text(x + 0.4, -y - 0.05, '%.2f ' % y, ha='center', va= 'top')
plt.xlim(-.5, n)
plt.xticks(())
plt.ylim(-1.25, 1.25)
plt.yticks(())
plt.show()

Download Python source code: plot_bar.py

Download Jupyter notebook: plot_bar.ipynb

Subplot grid
An example showing the subplot grid in matplotlib.
plt.xticks(())
plt.yticks(())
size=20, alpha=.5)
plt.xticks(())
plt.yticks(())
size=20, alpha=.5)
plt.xticks(())
plt.yticks(())

size=20, alpha=.5)
plt.xticks(())
plt.yticks(())
size=20, alpha=.5)

plt.tight_layout()
plt.show()

Download Python source code: plot_subplot-grid.py
Download Jupyter notebook: plot_subplot-grid.ipynb
Axes
This example shows various axes command to position matplotlib axes.
plt.axes([.1, .1, .5, .5])

plt.xticks(())
plt.yticks(())
plt.text(0.1, 0.1, 'axes([0.1, 0.1, .8, .8])', ha='left', va='center',
size=16, alpha=.5)
plt.axes([.2, .2, .5, .5])

plt.xticks(())
plt.yticks(())
size=16, alpha=.5)
plt.axes([0.3, 0.3, .5, .5])

plt.xticks(())
plt.yticks(())
size=16, alpha=.5)
plt.axes([.4, .4, .5, .5])

plt.xticks(())
plt.yticks(())
size=16, alpha=.5)
plt.show()

Download Python source code: plot_axes-2.py
Download Jupyter notebook: plot_axes-2.ipynb
Grid
Displaying a grid on the axes in matploblib.
ax = plt.axes([0.025, 0.025, 0.95, 0.95])
ax.set_xlim(0,4)
ax.set_ylim(0,3)

ax.xaxis.set_major_locator(plt.MultipleLocator(1.0))
ax.xaxis.set_minor_locator(plt.MultipleLocator(0.1))
ax.yaxis.set_major_locator(plt.MultipleLocator(1.0))
ax.yaxis.set_minor_locator(plt.MultipleLocator(0.1))
ax.grid(which='major', axis='x', linewidth=0.75, linestyle='-', color='0.75')
ax.grid(which='minor', axis='x', linewidth=0.25, linestyle='-', color='0.75')
ax.grid(which='major', axis='y', linewidth=0.75, linestyle='-', color='0.75')
ax.grid(which='minor', axis='y', linewidth=0.25, linestyle='-', color='0.75')
plt.show()

Download Python source code: plot_grid.py
Download Jupyter notebook: plot_grid.ipynb
3D plotting
Demo 3D plotting with matplotlib and style the figure.

from mpl_toolkits.mplot3d import axes3d
ax = plt.gca(projection='3d')
X, Y, Z = axes3d.get_test_data(0.05)
cset = ax.contourf(X, Y, Z)

ax.clabel(cset, fontsize=9, inline=1)
plt.xticks(())
plt.yticks(())
ax.set_zticks(())
ax.text2D(-0.05, 1.05, " 3D plots \n",

horizontalalignment='left',
verticalalignment='top',
bbox=dict(facecolor='white', alpha=1.0),
family='Lint McCree Intl BB',
size='x-large',
transform=plt.gca().transAxes)
ax.text2D(-0.05, .975, " Plot 2D or 3D data",

size='medium',
plt.show()

Download Python source code: plot_plot3d-2.py
Download Jupyter notebook: plot_plot3d-2.ipynb
GridSpec
An example demoing gridspec


import matplotlib.gridspec as gridspec
G = gridspec.GridSpec(3, 3)
axes_1 = plt.subplot(G[0, :])

plt.xticks(())
plt.yticks(())
plt.text(0.5, 0.5, 'Axes 1', ha='center', va='center', size=24, alpha=.5)
axes_2 = plt.subplot(G[1, :-1])

plt.xticks(())
plt.yticks(())
axes_3 = plt.subplot(G[1:, -1])

plt.xticks(())
plt.yticks(())
axes_4 = plt.subplot(G[-1, 0])

plt.xticks(())
plt.yticks(())
axes_5 = plt.subplot(G[-1, -2])

plt.xticks(())
plt.yticks(())
plt.tight_layout()
plt.show()

Download Python source code: plot_gridspec.py
Download Jupyter notebook: plot_gridspec.ipynb
Demo text printing
A example showing off elaborate text printing with matplotlib.

import numpy as np
eqs = []
eqs.append((r"$W^{3\beta}_{\delta_1 \rho_1 \sigma_2} = U^{3\beta}_{\delta_1 \rho_1} + \frac{1}
,{8 \pi 2} \int^{\alpha_2}_{\alpha_2} d \alpha^\prime_2 \left[\frac{ U^{2\beta}_{\delta_1
,\rho_1} - \alpha^\prime_2U^{1\beta}_{\rho_1 \sigma_2} }{U^{0\beta}_{\rho_1 \sigma_2}}\right]$
,"))
eqs.append((r"$\frac{d\rho}{d t} + \rho \vec{v}\cdot\nabla\vec{v} = -\nabla p + \mu\nabla^2
,\vec{v} + \rho \vec{g}$"))
eqs.append((r"$\int_{-\infty}^\infty e^{-x^2}dx=\sqrt{\pi}$"))
eqs.append((r"$E = mc^2 = \sqrt{{m_0}^2c^4 + p^2c^2}$"))
eqs.append((r"$F_G = G\frac{m_1m_2}{r^2}$"))
plt.axes([0.025, 0.025, 0.95, 0.95])
for i in range(24):
index = np.random.randint(0, len(eqs))
eq = eqs[index]
size = np.random.uniform(12, 32)
x,y = np.random.uniform(0, 1, 2)
alpha = np.random.uniform(0.25, .75)
plt.text(x, y, eq, ha='center', va='center', color="#11557c", alpha=alpha,
transform=plt.gca().transAxes, fontsize=size, clip_on=True)
plt.xticks(())
plt.yticks(())
plt.show()

Download Python source code: plot_text.py

Download Jupyter notebook: plot_text.ipynb
4.7.2 Code for the chapters exercises
Excercise 1
Solution of the excercise 1 with matplotlib.
import numpy as np
n = 256
C,S = np.cos(X), np.sin(X)
plt.plot(X, C)
plt.plot(X,S)
plt.show()

Download Python source code: plot_exercise_1.py
Download Jupyter notebook: plot_exercise_1.ipynb

Exercise 4
Exercise 4 with matplotlib.
import numpy as np

plt.subplot(111)

S = np.sin(X)
C = np.cos(X)


plt.show()

Exercise 3

import numpy as np

plt.subplot(111)


plt.xlim(-4.0, 4.0)
plt.ylim(-1.0, 1.0)
plt.show()

Exercise 5

import numpy as np

plt.subplot(111)

S = np.sin(X)
C = np.cos(X)


plt.xticks([-np.pi, -np.pi/2, 0, np.pi/2, np.pi])

plt.yticks([-1, 0, +1])
plt.show()

Exercise 6

import numpy as np

plt.subplot(111)

C = np.cos(X)
S = np.sin(X)


[r'$-\pi$', r'$-\pi/2$', r'$0$', r'$+\pi/2$', r'$+\pi$'])

plt.yticks([-1, 0, +1],
[r'$-1$', r'$0$', r'$+1$'])
plt.show()

Exercise 2

import numpy as np
# Create a new figure of size 8x6 points, using 100 dots per inch
# Create a new subplot from a grid of 1x1

plt.subplot(111)

# Plot cosine using blue color with a continuous line of width 1 (pixels)
# Plot sine using green color with a continuous line of width 1 (pixels)
plt.plot(X, S, color="green", linewidth=1.0, linestyle="-")
# Set x limits
plt.xlim(-4., 4.)
# Set x ticks
# Set y limits
plt.ylim(-1.0, 1.0)
# Set y ticks
# Show result on screen

plt.show()

Exercise 7
Exercise 7 with matplotlib
import numpy as np
plt.figure(figsize=(8,5), dpi=80)
plt.subplot(111)
X = np.linspace(-np.pi, np.pi, 256,endpoint=True)

C = np.cos(X)
S = np.sin(X)

ax = plt.gca()


[r'$-\pi$', r'$-\pi/2$', r'$0$', r'$+\pi/2$', r'$+\pi$'])

plt.yticks([-1, 0, +1],
[r'$-1$', r'$0$', r'$+1$'])
plt.show()

Exercise 8
import numpy as np
plt.figure(figsize=(8,5), dpi=80)
plt.subplot(111)

C = np.cos(X)
S = np.sin(X)

ax = plt.gca()


[r'$-\pi$', r'$-\pi/2$', r'$0$', r'$+\pi/2$', r'$+\pi$'])

plt.yticks([-1, +1],
[r'$-1$', r'$+1$'])
plt.show()

Exercise 9
import numpy as np

plt.subplot(111)


C = np.cos(X)
S = np.sin(X)

ax = plt.gca()

[r'$-\pi$', r'$-\pi/2$', r'$0$', r'$+\pi/2$', r'$+\pi$'])

plt.yticks([-1, +1],
[r'$-1$', r'$+1$'])
t = 2*np.pi/3
plt.plot([t, t], [0, np.cos(t)],
color='blue', linewidth=1.5, linestyle="--")
plt.annotate(r'$sin(\frac{2\pi}{3})=\frac{\sqrt{3}}{2}$',
xytext=(+10, +30), textcoords='offset points', fontsize=16,
plt.plot([t, t], [0, np.sin(t)],

color='red', linewidth=1.5, linestyle="--")
plt.scatter([t, ], [np.sin(t), ], 50, color='red')
plt.annotate(r'$cos(\frac{2\pi}{3})=-\frac{1}{2}$', xy=(t, np.cos(t)),
xycoords='data', xytext=(-90, -50), textcoords='offset points',
fontsize=16,
plt.show()

Exercise
Exercises with matplotlib.

import numpy as np

plt.subplot(111)


ax = plt.gca()
ax.spines['bottom'].set_position(('data', 0))
ax.spines['left'].set_position(('data', 0))

plt.xticks([-np.pi, -np.pi / 2, 0, np.pi / 2, np.pi],
[r'$-\pi$', r'$-\pi/2$', r'$0$', r'$+\pi/2$', r'$+\pi$'])

plt.yticks([-1, 1],
[r'$-1$', r'$+1$'])
t = 2*np.pi/3
plt.plot([t, t], [0, np.cos(t)],
color='blue', linewidth=1.5, linestyle="--")
plt.annotate(r'$sin(\frac{2\pi}{3})=\frac{\sqrt{3}}{2}$',

xytext=(10, 30), textcoords='offset points', fontsize=16,

plt.plot([t, t], [0, np.sin(t)],

color='red', linewidth=1.5, linestyle="--")
plt.scatter([t, ], [np.sin(t), ], 50, color ='red')
plt.annotate(r'$cos(\frac{2\pi}{3})=-\frac{1}{2}$', xy=(t, np.cos(t)),
xycoords='data', xytext=(-90, -50),
textcoords='offset points', fontsize=16,
for label in ax.get_xticklabels() + ax.get_yticklabels():

label.set_fontsize(16)
label.set_bbox(dict(facecolor='white', edgecolor='None', alpha=0.65 ))
plt.show()

4.7.3 Example demoing choices for an option
The colors matplotlib line plots
An example demoing the various colors taken by matplotlibs plot.
size = 256, 16
dpi = 72.0
figsize = size[0] / float(dpi), size[1] / float(dpi)
fig = plt.figure(figsize=figsize, dpi=dpi)
fig.patch.set_alpha(0)
plt.axes([0, 0.1, 1, .8], frameon=False)
for i in range(1,11):
plt.plot([i, i], [0, 1], lw=1.5)
plt.xlim(0, 11)
plt.xticks(())
plt.yticks(())
plt.show()

Download Python source code: plot_color.py
Download Jupyter notebook: plot_color.ipynb
Linewidth
Plot various linewidth with matplotlib.

size = 256, 16
dpi = 72.0
plt.axes([0, .1, 1, .8], frameon=False)
for i in range(1, 11):

plt.plot([i, i], [0, 1], color='b', lw=i/2.)
plt.xlim(0, 11)
plt.ylim(0, 1)
plt.xticks(())
plt.yticks(())
plt.show()

Download Python source code: plot_linewidth.py
Download Jupyter notebook: plot_linewidth.ipynb
Alpha: transparency
This example demonstrates using alpha for transparency.
size = 256,16
dpi = 72.0
figsize= size[0] / float(dpi), size[1] / float(dpi)

plt.axvline(i, linewidth=1, color='blue', alpha= .25 + .75 * i / 10.)
plt.xlim(0, 11)
plt.xticks(())
plt.yticks(())
plt.show()

Download Python source code: plot_alpha.py
Download Jupyter notebook: plot_alpha.ipynb
Aliased versus anti-aliased
This example demonstrates aliased versus anti-aliased text.

size = 128, 16
dpi = 72.0
plt.axes([0, 0, 1, 1], frameon=False)
plt.rcParams['text.antialiased'] = False
plt.text(0.5, 0.5, "Aliased", ha='center', va='center')
plt.xlim(0, 1)
plt.ylim(0, 1)
plt.xticks(())
plt.yticks(())
plt.show()

Download Python source code: plot_aliased.py
Download Jupyter notebook: plot_aliased.ipynb
Aliased versus anti-aliased
The example shows aliased versus anti-aliased text.
size = 128, 16
dpi = 72.0
plt.rcParams['text.antialiased'] = True
plt.text(0.5, 0.5, "Anti-aliased", ha='center', va='center')
plt.xlim(0, 1)
plt.ylim(0, 1)
plt.xticks(())
plt.yticks(())
plt.show()

Download Python source code: plot_antialiased.py
Download Jupyter notebook: plot_antialiased.ipynb

Marker size
Demo the marker size control in matplotlib.
size = 256, 16
dpi = 72.0

plt.plot([i, ], [1, ], 's', markersize=i, markerfacecolor='w',
markeredgewidth=.5, markeredgecolor='k')
plt.xlim(0, 11)
plt.xticks(())
plt.yticks(())
plt.show()

Download Python source code: plot_ms.py
Download Jupyter notebook: plot_ms.ipynb
Marker edge width
Demo the marker edge widths of matplotlibs markers.
size = 256, 16
dpi = 72.0
for i in range(1,11):
plt.plot([i, ], [1, ], 's', markersize=5,
markeredgewidth=1 + i/10., markeredgecolor='k', markerfacecolor='w')
plt.xlim(0, 11)
plt.xticks(())
plt.yticks(())
plt.show()

Download Python source code: plot_mew.py
Download Jupyter notebook: plot_mew.ipynb

Marker edge color
Demo the marker edge color of matplotlibs markers.
import numpy as np
size = 256,16
dpi = 72.0

r, g, b = np.random.uniform(0, 1, 3)
plt.plot([i, ], [1, ], 's', markersize=5, markerfacecolor='w',
markeredgewidth=1.5, markeredgecolor=(r, g, b, 1))
plt.xlim(0, 11)
plt.xticks(())
plt.yticks(())
plt.show()

Download Python source code: plot_mec.py
Download Jupyter notebook: plot_mec.ipynb
Marker face color
Demo the marker face color of matplotlibs markers.
import numpy as np
size = 256, 16
dpi = 72.0

r, g, b = np.random.uniform(0, 1, 3)
plt.plot([i, ], [1, ], 's', markersize=8, markerfacecolor=(r, g, b, 1),
markeredgewidth=.1, markeredgecolor=(0, 0, 0, .5))
plt.xlim(0, 11)
plt.xticks(())
plt.yticks(())
plt.show()

Download Python source code: plot_mfc.py

Download Jupyter notebook: plot_mfc.ipynb

Colormaps
An example plotting the matplotlib colormaps.
import numpy as np
plt.rc('text', usetex=False)
a = np.outer(np.arange(0, 1, 0.01), np.ones(10))
plt.subplots_adjust(top=0.8, bottom=0.05, left=0.01, right=0.99)
maps = [m for m in plt.cm.datad if not m.endswith("_r")]
maps.sort()
l = len(maps) + 1
for i, m in enumerate(maps):
plt.subplot(1, l, i+1)
plt.axis("off")
plt.imshow(a, aspect='auto', cmap=plt.get_cmap(m), origin="lower")
plt.title(m, rotation=90, fontsize=10, va='bottom')
plt.show()

Download Python source code: plot_colormaps.py
Download Jupyter notebook: plot_colormaps.ipynb
Solid cap style
An example demoing the solide cap style in matplotlib.

import numpy as np
size = 256, 16
dpi = 72.0
plt.plot(np.arange(4), np.ones(4), color="blue", linewidth=8,

solid_capstyle='butt')
plt.plot(5 + np.arange(4), np.ones(4), color="blue", linewidth=8,

solid_capstyle='round')
plt.plot(10 + np.arange(4), np.ones(4), color="blue", linewidth=8,

solid_capstyle='projecting')
plt.xlim(0, 14)
plt.xticks(())
plt.yticks(())
plt.show()

Download Python source code: plot_solid_capstyle.py
Download Jupyter notebook: plot_solid_capstyle.ipynb
Solid joint style
An example showing the differen solid joint styles in matplotlib.
import numpy as np
size = 256, 16
dpi = 72.0
plt.plot(np.arange(3), [0, 1, 0], color="blue", linewidth=8,

solid_joinstyle='miter')
plt.plot(4 + np.arange(3), [0, 1, 0], color="blue", linewidth=8,
solid_joinstyle='bevel')
plt.plot(8 + np.arange(3), [0, 1, 0], color="blue", linewidth=8,
solid_joinstyle='round')
plt.xlim(0, 12)
plt.ylim(-1, 2)
plt.xticks(())
plt.yticks(())
plt.show()


Download Python source code: plot_solid_joinstyle.py
Download Jupyter notebook: plot_solid_joinstyle.ipynb
Dash capstyle
An example demoing the dash capstyle.
import numpy as np
size = 256, 16
dpi = 72.0
plt.plot(np.arange(4), np.ones(4), color="blue", dashes=[15, 15],

linewidth=8, dash_capstyle='butt')
plt.plot(5 + np.arange(4), np.ones(4), color="blue", dashes=[15, 15],

linewidth=8, dash_capstyle='round')
plt.plot(10 + np.arange(4), np.ones(4), color="blue", dashes=[15, 15],

linewidth=8, dash_capstyle='projecting')
plt.xlim(0, 14)
plt.xticks(())
plt.yticks(())
plt.show()

Download Python source code: plot_dash_capstyle.py
Download Jupyter notebook: plot_dash_capstyle.ipynb
Dash join style
Example demoing the dash join style.
import numpy as np
size = 256, 16
dpi = 72.0

plt.plot(np.arange(3), [0, 1, 0], color="blue", dashes=[12, 5], linewidth=8,

dash_joinstyle='miter')
plt.plot(4 + np.arange(3), [0, 1, 0], color="blue", dashes=[12, 5],
linewidth=8, dash_joinstyle='bevel')
plt.plot(8 + np.arange(3), [0, 1, 0], color="blue", dashes=[12, 5],
linewidth=8, dash_joinstyle='round')
plt.xlim(0, 12)
plt.ylim(-1, 2)
plt.xticks(())
plt.yticks(())
plt.show()

Download Python source code: plot_dash_joinstyle.py
Download Jupyter notebook: plot_dash_joinstyle.ipynb
Linestyles
Plot the different line styles.
import numpy as np
def linestyle(ls, i):

X = i * .5 * np.ones(11)
Y = np.arange(11)
plt.plot(X, Y, ls, color=(.0, .0, 1, 1), lw=3, ms=8,
mfc=(.75, .75, 1, 1), mec=(0, 0, 1, 1))
plt.text(.5 * i, 10.25, ls, rotation=90, fontsize=15, va='bottom')

linestyles = ['-', '--', ':', '-.', '.', ',', 'o', '^', 'v', '<', '>', 's',
'+', 'x', 'd', '1', '2', '3', '4', 'h', 'p', '|', '_', 'D', 'H']
n_lines = len(linestyles)
size = 20 * n_lines, 300

dpi = 72.0
for i, ls in enumerate(linestyles):
linestyle(ls, i)
plt.xlim(-.2, .2 + .5*n_lines)
plt.xticks(())
plt.yticks(())
plt.show()

Download Python source code: plot_linestyles.py
Download Jupyter notebook: plot_linestyles.ipynb
Markers
Show the different markers of matplotlib.
import numpy as np
def marker(m, i):

X = i * .5 * np.ones(11)
Y = np.arange(11)
plt.plot(X, Y, lw=1, marker=m, ms=10, mfc=(.75, .75, 1, 1),

mec=(0, 0, 1, 1))
plt.text(.5 * i, 10.25, repr(m), rotation=90, fontsize=15, va='bottom')

markers = [0, 1, 2, 3, 4, 5, 6, 7, 'o', 'h', '_', '1', '2', '3', '4',

'8', 'p', '^', 'v', '<', '>', '|', 'd', ',', '+', 's', '*',
'|', 'x', 'D', 'H', '.']
n_markers = len(markers)
size = 20 * n_markers, 300

dpi = 72.0
for i, m in enumerate(markers):
marker(m, i)
plt.xlim(-.2, .2 + .5 * n_markers)
plt.xticks(())
plt.yticks(())
plt.show()

Download Python source code: plot_markers.py
Download Jupyter notebook: plot_markers.ipynb
Locators for tick on axis
An example demoing different locators to position ticks on axis for matplotlib.
import numpy as np
def tickline():
plt.xlim(0, 10), plt.ylim(-1, 1), plt.yticks([])

ax = plt.gca()
ax.spines['left'].set_color('none')
ax.yaxis.set_ticks_position('none')
ax.xaxis.set_minor_locator(plt.MultipleLocator(0.1))
ax.plot(np.arange(11), np.zeros(11))
return ax
locators = [
'plt.NullLocator()',
'plt.MultipleLocator(1.0)',
'plt.FixedLocator([0, 2, 8, 9, 10])',
'plt.IndexLocator(3, 1)',
'plt.LinearLocator(5)',
'plt.LogLocator(2, [1.0])',
'plt.AutoLocator()',
]
n_locators = len(locators)
size = 512, 40 * n_locators

dpi = 72.0
for i, locator in enumerate(locators):

plt.subplot(n_locators, 1, i + 1)
ax = tickline()
ax.xaxis.set_major_locator(eval(locator))
plt.text(5, 0.3, locator[3:], ha='center')
plt.subplots_adjust(bottom=.01, top=.99, left=.01, right=.99)

plt.show()

Download Python source code: plot_ticks.py
Download Jupyter notebook: plot_ticks.ipynb
4.7.4 Code generating the summary figures with a title
Plotting in polar, decorated
An example showing how to plot in polar coordinnate, and some decorations.

import numpy as np
plt.subplot(1, 1, 1, polar=True)
N = 20
theta = np.arange(0.0, 2 * np.pi, 2 * np.pi / N)
for r, bar in zip(radii, bars):
bar.set_facecolor(plt.cm.jet(r / 10.))
bar.set_alpha(0.5)
plt.gca().set_xticklabels([])
plt.gca().set_yticklabels([])
plt.text(-0.2, 1.02, " Polar Axis \n",

size='xx-large',
plt.text(-0.2, 1.01, "\n\n Plot anything using polar axis ",

size='large',
plt.show()


Download Python source code: plot_polar_ext.py
Download Jupyter notebook: plot_polar_ext.ipynb
3D plotting vignette
Demo 3D plotting with matplotlib and decorate the figure.
import numpy as np
fig = plt.figure()
ax = Axes3D(fig)
X = np.arange(-4, 4, 0.25)
Y = np.arange(-4, 4, 0.25)
R = np.sqrt(X ** 2 + Y ** 2)
Z = np.sin(R)
ax.plot_surface(X, Y, Z, rstride=1, cstride=1, cmap=plt.cm.hot)

ax.contourf(X, Y, Z, zdir='z', offset=-2, cmap=plt.cm.hot)
ax.set_zlim(-2, 2)
plt.xticks(())
plt.yticks(())
ax.set_zticks(())

ax.text2D(0.05, .93, " 3D plots \n",

size='xx-large',
ax.text2D(0.05, .87, " Plot 2D or 3D data",

size='large',
plt.show()

Download Python source code: plot_plot3d_ext.py
Download Jupyter notebook: plot_plot3d_ext.ipynb
Plot example vignette
An example of plots with matplotlib, and added annotations.
import numpy as np
n = 256

X = np.linspace(0, 2, n)
Y = np.sin(2 * np.pi * X)
plt.plot (X, Y, lw=2, color='violet')

plt.xlim(-0.2, 2.2)
plt.xticks(())
plt.ylim(-1.2, 1.2)
plt.yticks(())
# Add a title and a box around it

from matplotlib.patches import FancyBboxPatch
ax = plt.gca()
ax.add_patch(FancyBboxPatch((-0.05, .87),
width=.66, height=.165, clip_on=False,
boxstyle="square,pad=0", zorder=3,
facecolor='white', alpha=1.0,
transform=plt.gca().transAxes))
plt.text(-0.05, 1.02, " Regular Plot: plt.plot(...)\n",

size='xx-large',
plt.text(-0.05, 1.01, "\n\n Plot lines and/or markers ",

size='large',
plt.show()

Download Python source code: plot_plot_ext.py
Download Jupyter notebook: plot_plot_ext.ipynb
Multiple plots vignette
Demo multiple plots and style the figure.

ax = plt.subplot(2, 1, 1)

ax = plt.gca()
plt.text(-0.05, 1.02, " Multiplot: plt.subplot(...)\n",

size='xx-large',
transform=ax.transAxes)
plt.text(-0.05, 1.01, "\n\n Plot several plots at once ",
size='large',
transform=ax.transAxes)

plt.show()

Download Python source code: plot_multiplot_ext.py
Download Jupyter notebook: plot_multiplot_ext.ipynb
Boxplot with matplotlib
An example of doing box plots with matplotlib
import numpy as np
fig = plt.figure(figsize=(8, 5))

axes = plt.subplot(111)
n = 5
Z = np.zeros((n, 4))
X = np.linspace(0, 2, n, endpoint=True)
Y = np.random.random((n, 4))
plt.boxplot(Y)
plt.xticks(())
plt.yticks(())


ax = plt.gca()
plt.text(-0.05, 1.02, " Box Plot: plt.boxplot(...)\n ",

size='xx-large',
transform=axes.transAxes)
plt.text(-0.04, .98, "\n Make a box and whisker plot ",

size='large',
plt.show()

Download Python source code: plot_boxplot_ext.py
Download Jupyter notebook: plot_boxplot_ext.ipynb
Plot scatter decorated
An example showing the scatter function, with decorations.

import numpy as np
n = 1024
X = np.random.normal(0, 1, n)
Y = np.random.normal(0, 1, n)
T = np.arctan2(Y,X)
plt.scatter(X, Y, s=75, c=T, alpha=.5)

plt.xlim(-1.5, 1.5)
plt.xticks(())
plt.ylim(-1.5, 1.5)
plt.yticks(())

ax = plt.gca()
plt.text(-0.05, 1.02, " Scatter Plot: plt.scatter(...)\n",

size='xx-large',

plt.text(-0.05, 1.01, "\n\n Make a scatter plot of x versus y ",

size='large',
plt.show()

Download Python source code: plot_scatter_ext.py
Download Jupyter notebook: plot_scatter_ext.ipynb
Pie chart vignette
Demo pie chart with matplotlib and style the figure.
import numpy as np
n = 20
X = np.ones(n)
X[-1] *= 2
plt.pie(X, explode=X*.05, colors = ['%f ' % (i/float(n)) for i in range(n)])
fig = plt.gcf()
w, h = fig.get_figwidth(), fig.get_figheight()
r = h / float(w)

plt.xlim(-1.5, 1.5)
plt.ylim(-1.5 * r, 1.5 * r)
plt.xticks(())
plt.yticks(())

ax = plt.gca()
plt.text(-0.05, 1.02, " Pie Chart: plt.pie(...)\n",

size='xx-large',
plt.text(-0.05, 1.01, "\n\n Make a pie chart of an array ",

size='large',
plt.show()

Download Python source code: plot_pie_ext.py
Download Jupyter notebook: plot_pie_ext.ipynb
Bar plot advanced
An more elaborate bar plot example

import numpy as np
n = 16
X = np.arange(n)
plt.bar(X, Y1, facecolor='#9999ff', edgecolor='white')
plt.xlim(-.5, n)
plt.xticks(())
plt.ylim(-1, 1)
plt.yticks(())

ax = plt.gca()
plt.text(-0.05, 1.02, " Bar Plot: plt.bar(...)\n",

size='xx-large',
plt.text(-0.05, 1.01, "\n\n Make a bar plot with rectangles ",

size='large',
plt.show()

Download Python source code: plot_bar_ext.py
Download Jupyter notebook: plot_bar_ext.ipynb
Plotting quiver decorated
An example showing quiver with decorations.
import numpy as np
n = 8
T = np.arctan2(Y - n/ 2., X - n / 2.)
R = 10 + np.sqrt((Y - n / 2.) ** 2 + (X - n / 2.) ** 2)
U, V = R * np.cos(T), R * np.sin(T)
plt.quiver(X, Y, U, V, R, alpha=.5)
plt.quiver(X, Y, U, V, edgecolor='k', facecolor='None', linewidth=.5)

plt.xlim(-1, n)
plt.xticks(())
plt.ylim(-1, n)
plt.yticks(())

ax = plt.gca()
plt.text(-0.05, 1.02, " Quiver Plot: plt.quiver(...)\n",

size='xx-large',
plt.text(-0.05, 1.01, "\n\n Plot a 2-D field of arrows ",

size='large',
plt.show()

Download Python source code: plot_quiver_ext.py
Download Jupyter notebook: plot_quiver_ext.ipynb
Imshow demo
Demoing imshow

import numpy as np
def f(x, y):

n = 10
x = np.linspace(-3, 3, 8 * n)
y = np.linspace(-3, 3, 6 * n)
Z = f(X, Y)
plt.imshow(Z, interpolation='nearest', cmap='bone', origin='lower')
plt.xticks(())
plt.yticks(())

ax = plt.gca()
plt.text(-0.05, 1.02, " Imshow: plt.imshow(...)\n",

size='xx-large',

plt.text(-0.05, 1.01, "\n\n Display an image to current axes ",

size='large',
plt.show()

Download Python source code: plot_imshow_ext.py
Download Jupyter notebook: plot_imshow_ext.ipynb
Display the contours of a function
An example demoing how to plot the contours of a function, with additional layout tweeks.
import numpy as np
def f(x,y):
n = 256

plt.contourf(X, Y, f(X, Y), 8, alpha=.75, cmap=plt.cm.hot)

C = plt.contour(X, Y, f(X,Y), 8, colors='black', linewidth=.5)
plt.clabel(C, inline=1, fontsize=10)
plt.xticks(())
plt.yticks(())

ax = plt.gca()
plt.text(-0.05, 1.02, " Contour Plot: plt.contour(..)\n",

size='xx-large',
plt.text(-0.05, 1.01, "\n\n Draw contour lines and filled contours ",
size='large',
plt.show()

Download Python source code: plot_contour_ext.py
Download Jupyter notebook: plot_contour_ext.ipynb
Grid elaborate
An example displaying a grid on the axes and tweaking the layout.


from matplotlib.ticker import MultipleLocator
fig = plt.figure(figsize=(8, 6), dpi=72, facecolor="white")

axes = plt.subplot(111)
axes.set_xlim(0, 4)
axes.set_ylim(0, 3)
axes.xaxis.set_major_locator(MultipleLocator(1.0))
axes.xaxis.set_minor_locator(MultipleLocator(0.1))
axes.yaxis.set_major_locator(MultipleLocator(1.0))
axes.yaxis.set_minor_locator(MultipleLocator(0.1))
axes.grid(which='major', axis='x', linewidth=0.75, linestyle='-', color='0.75')
axes.grid(which='minor', axis='x', linewidth=0.25, linestyle='-', color='0.75')
axes.grid(which='major', axis='y', linewidth=0.75, linestyle='-', color='0.75')
axes.grid(which='minor', axis='y', linewidth=0.25, linestyle='-', color='0.75')
axes.set_xticklabels([])
axes.set_yticklabels([])

ax = plt.gca()
plt.text(-0.05, 1.02, " Grid: plt.grid(...)\n",


size='xx-large',
plt.text(-0.05, 1.01, "\n\n Draw ticks and grid ",

size='large',

Download Python source code: plot_grid_ext.py
Download Jupyter notebook: plot_grid_ext.ipynb
Text printing decorated
An example showing text printing and decorating the resulting figure.
import numpy as np
fig = plt.figure()
plt.xticks(())
plt.yticks(())
eqs = []
eqs.append((r"$W^{3\beta}_{\delta_1 \rho_1 \sigma_2} = U^{3\beta}_{\delta_1 \rho_1} + \frac{1}
,{8 \pi 2} \int^{\alpha_2}_{\alpha_2} d \alpha^\prime_2 \left[\frac{ U^{2\beta}_{\delta_1
,\rho_1} - \alpha^\prime_2U^{1\beta}_{\rho_1 \sigma_2} }{U^{0\beta}_{\rho_1 \sigma_2}}\right]$
,"))
eqs.append((r"$\frac{d\rho}{d t} + \rho \vec{v}\cdot\nabla\vec{v} = -\nabla p + \mu\nabla^2

,\vec{v} + \rho \vec{g}$"))
eqs.append((r"$\int_{-\infty}^\infty e^{-x^2}dx=\sqrt{\pi}$"))
eqs.append((r"$E = mc^2 = \sqrt{{m_0}^2c^4 + p^2c^2}$"))
eqs.append((r"$F_G = G\frac{m_1m_2}{r^2}$"))
for i in range(24):
index = np.random.randint(0,len(eqs))
eq = eqs[index]
size = np.random.uniform(12,32)
x,y = np.random.uniform(0,1,2)
alpha = np.random.uniform(0.25,.75)
plt.text(x, y, eq, ha='center', va='center', color="#11557c", alpha=alpha,
transform=plt.gca().transAxes, fontsize=size, clip_on=True)

ax = plt.gca()
plt.text(-0.05, 1.02, " Text: plt.text(...)\n",

size='xx-large',
plt.text(-0.05, 1.01, "\n\n Draw any kind of text ",

size='large',
plt.show()

Download Python source code: plot_text_ext.py
Download Jupyter notebook: plot_text_ext.ipynb

CHAPTER 5
Scipy : high-level scientific computing
Authors: Gal Varoquaux, Adrien Chauve, Andre Espaze, Emmanuelle Gouillart, Ralf Gommers
Scipy
The scipy package contains various toolboxes dedicated to common issues in scientific computing. Its
different submodules correspond to different applications, such as interpolation, integration, optimization,
image processing, statistics, special functions, etc.
Tip: scipy can be compared to other standard scientific-computing libraries, such as the GSL (GNU Scientific
Library for C and C++), or Matlabs toolboxes. scipy is the core package for scientific routines in Python; it is
meant to operate efficiently on numpy arrays, so that numpy and scipy work hand in hand.
Before implementing a routine, it is worth checking if the desired data processing is not already implemented
in Scipy. As non-professional programmers, scientists often tend to re-invent the wheel, which leads to buggy,
non-optimal, difficult-to-share and unmaintainable code. By contrast, Scipys routines are optimized and
tested, and should therefore be used when possible.
Chapters contents
File input/output: scipy.io

Special functions: scipy.special
Linear algebra operations: scipy.linalg
Interpolation: scipy.interpolate
184
Optimization and fit: scipy.optimize

Statistics and random numbers: scipy.stats
Numerical integration: scipy.integrate
Fast Fourier transforms: scipy.fftpack
Signal processing: scipy.signal
Image manipulation: scipy.ndimage
Summary exercises on scientific computing
Full code examples for the scipy chapter
Warning: This tutorial is far from an introduction to numerical computing. As enumerating the different
submodules and functions in scipy would be very boring, we concentrate instead on a few examples to give
a general idea of how to use scipy for scientific computing.
scipy is composed of task-specific sub-modules:
scipy.cluster Vector quantization / Kmeans

scipy.constants Physical and mathematical constants
scipy.fftpack Fourier transform
scipy.integrate Integration routines
scipy.interpolate Interpolation
scipy.io Data input and output
scipy.linalg Linear algebra routines
scipy.ndimage n-dimensional image package
scipy.odr Orthogonal distance regression
scipy.optimize Optimization
scipy.signal Signal processing
scipy.sparse Sparse matrices
scipy.spatial Spatial data structures and algorithms
scipy.special Any special mathematical functions
scipy.stats Statistics
Tip: They all depend on numpy, but are mostly independent of each other. The standard way of importing
Numpy and these Scipy modules is:

>>> from scipy import stats # same for other sub-modules
The main scipy namespace mostly contains functions that are really numpy functions (try scipy.cos is
np.cos). Those are exposed for historical reasons; theres no reason to use import scipy in your code.
5.1 File input/output: scipy.io
Matlab files: Loading and saving:
>>> from scipy import io as spio

>>> a = np.ones((3, 3))
>>> spio.savemat('file.mat', {'a': a}) # savemat expects a dictionary
>>> data = spio.loadmat('file.mat')
5.1. File input/output: scipy.io 185

>>> data['a']
array([[ 1., 1., 1.],
[ 1., 1., 1.],
[ 1., 1., 1.]])
Warning: Python / Matlab mismatches, eg matlab does not represent 1D arrays

>>> a = np.ones(3)
>>> a
array([ 1., 1., 1.])
>>> spio.savemat('file.mat', {'a': a})
>>> spio.loadmat('file.mat')['a']
array([[ 1., 1., 1.]])
Notice the difference?
Image files: Reading images:

>>> misc.imread('fname.png')
array(...)
>>> # Matplotlib also has a similar function
>>> plt.imread('fname.png')
array(...)
See also:
Load text files: numpy.loadtxt()/numpy.savetxt()
Clever loading of text/csv files: numpy.genfromtxt()/numpy.recfromcsv()
Fast and efficient, but numpy-specific, binary format: numpy.save()/numpy.load()
More advanced input/output of images in scikit-image: skimage.io
5.2 Special functions: scipy.special
Special functions are transcendental functions. The docstring of the scipy.special module is well-written,
so we wont list all functions here. Frequently used ones are:
Bessel function, such as scipy.special.jn() (nth integer order Bessel function)
Elliptic function (scipy.special.ellipj() for the Jacobian elliptic function, . . . )
Gamma function: scipy.special.gamma(), also note scipy.special.gammaln() which will give the
log of Gamma to a higher numerical precision.
Erf, the area under a Gaussian curve: scipy.special.erf()
5.3 Linear algebra operations: scipy.linalg
5.2. Special functions: scipy.special 186

Tip: The scipy.linalg module provides standard linear algebra operations, relying on an underlying effi-
cient implementation (BLAS, LAPACK).
The scipy.linalg.det() function computes the determinant of a square matrix:
>>> from scipy import linalg

>>> arr = np.array([[1, 2],
... [3, 4]])
>>> linalg.det(arr)
-2.0
>>> arr = np.array([[3, 2],
... [6, 4]])
>>> linalg.det(arr)
0.0
>>> linalg.det(np.ones((3, 4)))
...
ValueError: expected square matrix
The scipy.linalg.inv() function computes the inverse of a square matrix:
>>> arr = np.array([[1, 2],

... [3, 4]])
>>> iarr = linalg.inv(arr)
>>> iarr
array([[-2. , 1. ],
[ 1.5, -0.5]])
>>> np.allclose(np.dot(arr, iarr), np.eye(2))
True
Finally computing the inverse of a singular matrix (its determinant is zero) will raise LinAlgError:
>>> arr = np.array([[3, 2],

... [6, 4]])
>>> linalg.inv(arr)
...
...LinAlgError: singular matrix
More advanced operations are available, for example singular-value decomposition (SVD):
>>> arr = np.arange(9).reshape((3, 3)) + np.diag([1, 0, 1])

>>> uarr, spec, vharr = linalg.svd(arr)
The resulting array spectrum is:
>>> spec
array([ 14.88982544, 0.45294236, 0.29654967])
The original matrix can be re-composed by matrix multiplication of the outputs of svd with np.dot:
>>> sarr = np.diag(spec)

>>> svd_mat = uarr.dot(sarr).dot(vharr)
>>> np.allclose(svd_mat, arr)
True
SVD is commonly used in statistics and signal processing. Many other standard decompositions (QR,
LU, Cholesky, Schur), as well as solvers for linear systems, are available in scipy.linalg.
5.3. Linear algebra operations: scipy.linalg 187

5.4 Interpolation: scipy.interpolate
scipy.interpolate is useful for fitting a function from experimental data and thus evaluating points where
no measure exists. The module is based on the FITPACK Fortran subroutines.
By imagining experimental data close to a sine function:
>>> measured_time = np.linspace(0, 1, 10)

>>> noise = (np.random.random(10)*2 - 1) * 1e-1
>>> measures = np.sin(2 * np.pi * measured_time) + noise
scipy.interpolate.interp1d can build a linear interpolation function:
>>> from scipy.interpolate import interp1d

>>> linear_interp = interp1d(measured_time, measures)
Then the result can be evaluated at the time of

interest:
>>> interpolation_time = np.linspace(0, 1, 50)

>>> linear_results = linear_interp(interpolation_time)
A cubic interpolation can also be selected by providing the kind optional keyword argument:
>>> cubic_interp = interp1d(measured_time, measures, kind='cubic')

>>> cubic_results = cubic_interp(interpolation_time)
scipy.interpolate.interp2d is similar to scipy.interpolate.interp1d, but for 2-D arrays. Note that

for the interp family, the interpolation points must stay within the range of given data points. See the summary
exercise on Maximum wind speed prediction at the Sprog station for a more advanced spline interpolation
example.
5.5 Optimization and fit: scipy.optimize
Optimization is the problem of finding a numerical solution to a minimization or equality.
Tip: The scipy.optimize module provides algorithms for function minimization (scalar or multi-
dimensional), curve fitting and root finding.
>>> from scipy import optimize
5.4. Interpolation: scipy.interpolate 188

5.5.1 Curve fitting
Suppose we have data on a sine wave, with some noise:
>>> x_data = np.linspace(-5, 5, num=50)

>>> y_data = 2.9 * np.sin(1.5 * x_data) + np.random.normal(size=50)
If we know that the data lies on a sine wave, but not the amplitudes or the period, we can find those by least
squares curve fitting. First we have to define the test function to fit, here a sine with unknown amplitude and
period:
>>> def test_func(x, a, b):

... return a * np.sin(b * x)
We then use scipy.optimize.curve_fit() to find a

and b:
>>> params, params_covariance = optimize.curve_fit(test_func, x_data, y_data, p0=[2, 2])

>>> print(params)
[ 3.05931973 1.45754553]
Exercise: Curve fitting of temperature data
The temperature extremes in Alaska for each month, starting in January, are given by (in degrees
Celcius):
max: 17, 19, 21, 28, 33, 38, 37, 37, 31, 23, 19, 18
min: -62, -59, -56, -46, -32, -18, -9, -13, -25, -46, -52, -58
1. Plot these temperature extremes.

2. Define a function that can describe min and max temperatures. Hint: this function has to
have a period of 1 year. Hint: include a time offset.
3. Fit this function to the data with scipy.optimize.curve_fit().
4. Plot the result. Is the fit reasonable? If not, why?
5.5. Optimization and fit: scipy.optimize 189

5. Is the time offset for min and max temperatures the same within the fit accuracy?
solution
5.5.2 Finding the minimum of a scalar function
Lets define the following function:
>>> def f(x):

... return x**2 + 10*np.sin(x)
and plot it:
>>> x = np.arange(-10, 10, 0.1)

>>> plt.plot(x, f(x))
>>> plt.show()
This function has a global minimum around -1.3 and a local minimum around 3.8.
Searching for minimum can be done with scipy.optimize.minimize(), given a starting point x0, it returns
the location of the minimum that it has found:
result type
The result of scipy.optimize.minimize() is a compound object comprising all information on the con-
vergence
>>> result = optimize.minimize(f, x0=0)

>>> result
fun: -7.9458233756...
hess_inv: array([[ 0.0858...]])
jac: array([ -1.19209...e-06])
message: 'Optimization terminated successfully.'
nfev: 18
nit: 5
njev: 6
status: 0
success: True
x: array([-1.30644...])

>>> result.x # The coordinate of the minimum

array([-1.30644...])
Methods: As the function is a smooth function, gradient-descent based methods are good options. The lBFGS
algorithm is a good choice in general:
>>> optimize.minimize(f, x0=0, method="L-BFGS-B")

fun: array([-7.94582338])
hess_inv: <1x1 LbfgsInvHessProduct with dtype=float64>
jac: array([ -1.42108547e-06])
message: ...'CONVERGENCE: NORM_OF_PROJECTED_GRADIENT_<=_PGTOL'
nfev: 12
nit: 5
status: 0
success: True
x: array([-1.30644013])
Note how it cost only 12 functions evaluation above to find a good value for the minimum.
Global minimum: A possible issue with this approach is that, if the function has local minima, the algorithm
may find these local minima instead of the global minimum depending on the initial point x0:
>>> res = optimize.minimize(f, x0=3, method="L-BFGS-B")

>>> res.x
array([ 3.83746709])
If we dont know the neighborhood of the global minimum to choose the initial point, we need to resort to
costlier global optimization. To find the global minimum, we use scipy.optimize.basinhopping() (added
in version 0.12.0 of Scipy). It combines a local optimizer with sampling of starting points:
>>> optimize.basinhopping(f, 0)
nfev: 1725
minimization_failures: 0
fun: -7.9458233756152845
x: array([-1.30644001])
message: ['requested number of basinhopping iterations completed successfully']
njev: 575
nit: 100
Note: scipy used to contain the routine anneal, it has been removed in SciPy 0.16.0.
Constraints: We can constrain the variable to the interval (0, 10) using the bounds argument:

A list of bounds
As minimize() works in general with x multidimensionsal, the bounds argument is a list of bound on
each dimension.
>>> res = optimize.minimize(f, x0=1,

... bounds=((0, 10), ))
>>> res.x
array([ 0.])
Tip: What has happened? Why are we finding 0, which is not a mimimum of our function.
Minimizing functions of several variables
To minimize over several variables, the trick is to turn them into a function of a multi-dimensional variable
(a vector). See for instance the exercise on 2D minimization below.
Note: scipy.optimize.minimize_scalar() is a function with dedicated methods to minimize functions

of only one variable.
See also:
Finding minima of function is discussed in more details in the advanced chapter: Mathematical optimization:
finding minima of functions.
Exercise: 2-D minimization
The six-hump camel-

back function
x4 2
f (x, y) = (4 2.1x 2 + )x + x y + (4y 2 4)y 2
3
has multiple global and local minima. Find the global minima of this function.
Hints:
Variables can be restricted to 2 < x < 2 and 1 < y < 1.

Use numpy.meshgrid() and pylab.imshow() to find visually the regions.

Use scipy.optimize.minimize(), optionally trying out several of its methods.
How many global minima are there, and what is the function value at those points? What hap-
pens for an initial guess of (x, y) = (0, 0) ?
solution
5.5.3 Finding the roots of a scalar function
To find a root, i.e. a point where f (x) = 0, of the function f above we can use scipy.optimize.root():
>>> root = optimize.root(f, x0=1) # our initial guess is 1

>>> root # The full result
fjac: array([[-1.]])
fun: array([ 0.])
message: 'The solution converged.'
nfev: 10
qtf: array([ 1.33310463e-32])
r: array([-10.])
status: 1
success: True
x: array([ 0.])
>>> root.x # Only the root found
array([ 0.])
Note that only one root is found. Inspecting the plot of f reveals that there is a second root around -2.5. We
find the exact value of it by adjusting our initial guess:
>>> root2 = optimize.root(f, x0=-2.5)

>>> root2.x
array([-2.47948183])
Note: scipy.optimize.root() also comes with a variety of algorithms, set via the method argument.
Now that we have found the minima and roots of f and used curve fitting on it, we put all those results together
in a single plot:
See also:

You can find all algorithms and functions with similar functionalities in the documentation of scipy.
optimize.
See the summary exercise on Non linear least squares curve fitting: application to point extraction in topo-
graphical lidar data for another, more advanced example.
5.6 Statistics and random numbers: scipy.stats
The module scipy.stats contains statistical tools and probabilistic descriptions of random processes. Ran-
dom number generators for various random process can be found in numpy.random.
5.6.1 Distributions: histogram and probability density function
Given observations of a random process, their histogram is an estimator of the random processs PDF (proba-
bility density function):
>>> samples = np.random.normal(size=1000)

>>> bins = np.arange(-4, 5)
>>> bins
array([-4, -3, -2, -1, 0, 1, 2, 3, 4])
>>> histogram = np.histogram(samples, bins=bins, normed=True)[0]
>>> bins = 0.5*(bins[1:] + bins[:-1])
>>> bins
array([-3.5, -2.5, -1.5, -0.5, 0.5, 1.5, 2.5, 3.5])
>>> from scipy import stats
>>> pdf = stats.norm.pdf(bins) # norm is a distribution object
>>> plt.plot(bins, histogram)

>>> plt.plot(bins, pdf)
The distribution objects
scipy.stats.norm is a distribution object: each distribution in scipy.stats is represented as an object.

Here its the normal distribution, and it comes with a PDF, a CDF, and much more.
5.6. Statistics and random numbers: scipy.stats 194

If we know that the random process belongs to a given family of random processes, such as normal processes,
we can do a maximum-likelihood fit of the observations to estimate the parameters of the underlying distri-
bution. Here we fit a normal process to the observed data:
>>> loc, std = stats.norm.fit(samples)

>>> loc
-0.045256707...
>>> std
0.9870331586...
Exercise: Probability distributions
Generate 1000 random variates from a gamma distribution with a shape parameter of 1, then plot a his-
togram from those samples. Can you plot the pdf on top (it should match)?
Extra: the distributions have many useful methods. Explore them by reading the docstring or by using tab
completion. Can you recover the shape parameter 1 by using the fit method on your random variates?
5.6.2 Mean, median and percentiles
The mean is an estimator of the center of the distribution:
>>> np.mean(samples)
-0.0452567074...
The median another estimator of the center. It is the value with half of the observations below, and half above:
>>> np.median(samples)
-0.0580280347...
Tip: Unlike the mean, the median is not sensitive to the tails of the distribution. It is robust.
Exercise: Compare mean and median on samples of a Gamma distribution
Which one seems to be the best estimator of the center for the Gamma distribution?
The median is also the percentile 50, because 50% of the observation are below it:
>>> stats.scoreatpercentile(samples, 50)

-0.0580280347...
Similarly, we can calculate the percentile 90:
>>> stats.scoreatpercentile(samples, 90)

1.2315935511...
Tip: The percentile is an estimator of the CDF: cumulative distribution function.
5.6. Statistics and random numbers: scipy.stats 195

5.6.3 Statistical tests
A statistical test is a decision indicator. For in-

stance, if we have two sets of observations, that we assume are generated from Gaussian processes, we can use
a T-test to decide whether the means of two sets of observations are significantly different:
>>> a = np.random.normal(0, 1, size=100)

>>> b = np.random.normal(1, 1, size=10)
>>> stats.ttest_ind(a, b)
(array(-3.177574054...), 0.0019370639...)
Tip: The resulting output is composed of:

The T statistic value: it is a number the sign of which is proportional to the difference between the two
random processes and the magnitude is related to the significance of this difference.
the p value: the probability of both processes being identical. If it is close to 1, the two process are almost
certainly identical. The closer it is to zero, the more likely it is that the processes have different means.
See also:
The chapter on statistics introduces much more elaborate tools for statistical testing and statistical data load-
ing and visualization outside of scipy.
5.7 Numerical integration: scipy.integrate
5.7.1 Function integrals

R /2
The most generic integration routine is scipy.integrate.quad(). To compute 0 si n(t )d t :
>>> from scipy.integrate import quad

>>> res, err = quad(np.sin, 0, np.pi/2)
>>> np.allclose(res, 1) # res is the result, is should be close to 1
True
>>> np.allclose(err, 1 - res) # err is an estimate of the err
True
Other integration schemes are available: scipy.integrate.fixed_quad(), scipy.integrate.

quadrature(), scipy.integrate.romberg(). . .
5.7.2 Integrating differential equations
scipy.integrate also features routines for integrating Ordinary Differential Equations (ODE). In particular,
5.7. Numerical integration: scipy.integrate 196

scipy.integrate.odeint() solves ODE of the form:
dy/dt = rhs(y1, y2, .., t0,...)
dy
As an introduction, let us solve the ODE d t = 2y between t = 0 . . . 4, with the initial condition y(t = 0) = 1.
First the function computing the derivative of the position needs to be defined:
>>> def calc_derivative(ypos, time):

... return -2 * ypos
Then, to compute y as a function of time:
>>> from scipy.integrate import odeint

>>> time_vec = np.linspace(0, 4, 40)
>>> y = odeint(calc_derivative, y0=1, t=time_vec)
Let us integrate a more complex ODE: a damped spring-mass oscillator. The position of a mass attached to a
spring obeys the 2nd order ODE y 00 + 20 y 0 + 20 y = 0 with 20 = k/m with k the spring constant, m the mass
and = c/(2m0 ) with c the damping coefficient. We set:
>>> mass = 0.5 # kg

>>> kspring = 4 # N/m
>>> cviscous = 0.4 # N s/m
Hence:
>>> eps = cviscous / (2 * mass * np.sqrt(kspring/mass))

>>> omega = np.sqrt(kspring / mass)
The system is underdamped, as:
>>> eps < 1

True
For odeint(), the 2nd order equation needs to be transformed in a system of two first-order equations for the
vector Y = (y, y 0 ): the function computes the velocity and acceleration:
>>> def calc_deri(yvec, time, eps, omega):

... return (yvec[1], -eps * omega * yvec[1] - omega **2 * yvec[0])
5.7. Numerical integration: scipy.integrate 197

Integration of the system follows:
>>> time_vec = np.linspace(0, 10, 100)

>>> yinit = (1, 0)
>>> yarr = odeint(calc_deri, yinit, time_vec, args=(eps, omega))
Tip: scipy.integrate.odeint() uses the LSODA (Livermore Solver for Ordinary Differential equations
with Automatic method switching for stiff and non-stiff problems), see the ODEPACK Fortran library for more
details.
See also:
Partial Differental Equations
There is no Partial Differential Equations (PDE) solver in Scipy. Some Python packages for solving PDEs are
available, such as fipy or SfePy.
5.8 Fast Fourier transforms: scipy.fftpack
The scipy.fftpack module computes fast Fourier transforms (FFTs) and offers utilities to handle them. The
main functions are:
scipy.fftpack.fft() to compute the FFT
scipy.fftpack.fftfreq() to generate the sampling frequencies
scipy.fftpack.ifft() computes the inverse FFT, from frequency space to signal space
As an illustration, a (noisy) input signal (sig), and its FFT:
>>> from scipy import fftpack

>>> sig_fft = fftpack.fft(sig)
>>> freqs = fftpack.fftfreq(sig.size, d=time_step)
5.8. Fast Fourier transforms: scipy.fftpack 198

Signal FFT
As the signal comes from a real function, the Fourier transform is symmetric.
The peak signal frequency can be found with freqs[power.argmax()]
Setting the Fourrier component above

this frequency to zero and inverting the FFT with scipy.fftpack.ifft(), gives a filtered signal.
Note: The code of this example can be found here
numpy.fft
Numpy also has an implementation of FFT (numpy.fft). However, the scipy one should be preferred, as it
uses more efficient underlying implementations.
Fully worked examples:

Crude periodicity finding (link) Gaussian image blur (link)
Exercise: Denoise moon landing image
1. Examine the provided image moonlanding.png, which is heavily contaminated with periodic noise.
In this exercise, we aim to clean up the noise using the Fast Fourier Transform.
2. Load the image using pylab.imread().
3. Find and use the 2-D FFT function in scipy.fftpack, and plot the spectrum (Fourier transform of )
the image. Do you have any trouble visualising the spectrum? If so, why?

4. The spectrum consists of high and low frequency components. The noise is contained in the high-
frequency part of the spectrum, so set some of those components to zero (use array slicing).
5. Apply the inverse Fourier transform to see the resulting image.
Solution
5.9 Signal processing: scipy.signal
Tip: scipy.signal is for typical signal processing: 1D, regularly-sampled signals.
Resampling scipy.signal.resample(): resample

a signal to n points using FFT.
>>> t = np.linspace(0, 5, 100)

>>> x = np.sin(t)
>>> from scipy import signal

>>> x_resampled = signal.resample(x, 25)
>>> plt.plot(t, x)
>>> plt.plot(t[::4], x_resampled, 'ko')
Tip: Notice how on the side of the window the resampling is less accurate and has a rippling effect.
This resampling is different from the interpolation provided by scipy.interpolate as it only applies to reg-
ularly sampled data.
5.9. Signal processing: scipy.signal 201

Detrending scipy.signal.detrend(): remove

linear trend from signal:
>>> t = np.linspace(0, 5, 100)

>>> x = t + np.random.normal(size=100)

>>> x_detrended = signal.detrend(x)
>>> plt.plot(t, x)
>>> plt.plot(t, x_detrended)
Filtering: For non-linear filtering, scipy.signal has filtering (median filter scipy.signal.medfilt(),
Wiener scipy.signal.wiener()), but we will discuss this in the image section.
Tip: scipy.signal also has a full-blown set of tools for the design of linear filter (finite and infinite response
filters), but this is out of the scope of this tutorial.
Spectral analysis: scipy.signal.spectrogram() compute a spectrogram frequency spectrums over con-

secutive time windows, while scipy.signal.welch() comptes a power spectrum density (PSD).
5.9. Signal processing: scipy.signal 202

5.10 Image manipulation: scipy.ndimage
scipy.ndimage provides manipulation of n-dimensional arrays as images.
5.10.1 Geometrical transformations on images
Changing orientation, resolution, ..
>>> from scipy import misc # Load an image

>>> from scipy import ndimage # Shift, roate and zoom it

>>> shifted_face = ndimage.shift(face, (50, 50))
>>> shifted_face2 = ndimage.shift(face, (50, 50), mode='nearest')
>>> rotated_face = ndimage.rotate(face, 30)
>>> cropped_face = face[50:-50, 50:-50]
>>> zoomed_face = ndimage.zoom(face, 2)
>>> zoomed_face.shape
(1536, 2048)
>>> plt.subplot(151)
<matplotlib.axes._subplots.AxesSubplot object at 0x...>
>>> plt.imshow(shifted_face, cmap=plt.cm.gray)

>>> plt.axis('off')
(-0.5, 1023.5, 767.5, -0.5)
>>> # etc.
5.10.2 Image filtering
Generate a noisy face:
5.10. Image manipulation: scipy.ndimage 203


>>> face = face[:512, -512:] # crop out square on right
>>> noisy_face = np.copy(face).astype(np.float)
>>> noisy_face += face.std() * 0.5 * np.random.standard_normal(face.shape)
Apply a variety of filters on it:
>>> blurred_face = ndimage.gaussian_filter(noisy_face, sigma=3)

>>> median_face = ndimage.median_filter(noisy_face, size=5)
>>> wiener_face = signal.wiener(noisy_face, (5, 5))
Other filters in scipy.ndimage.filters and scipy.signal can be applied to images.
Exercise
Compare histograms for the different filtered images.
5.10.3 Mathematical morphology
Tip: Mathematical morphology stems from set theory. It characterizes and transforms geometrical structures.
Binary (black and white) images, in particular, can be transformed using this theory: the sets to be transformed
are the sets of neighboring non-zero-valued pixels. The theory was also extended to gray-valued images.
Mathematical-morphology operations use a structuring element in order to modify geometrical structures.

Let us first generate a structuring element:
>>> el = ndimage.generate_binary_structure(2, 1)
>>> el

array([[False, True, False],

[...True, True, True],
[False, True, False]], dtype=bool)
>>> el.astype(np.int)
array([[0, 1, 0],
[1, 1, 1],
[0, 1, 0]])
Erosion scipy.ndimage.binary_erosion()
>>> a = np.zeros((7, 7), dtype=np.int)

>>> a[1:6, 2:5] = 1
>>> a
array([[0, 0, 0, 0, 0, 0, 0],
[0, 0, 1, 1, 1, 0, 0],
[0, 0, 1, 1, 1, 0, 0],
[0, 0, 1, 1, 1, 0, 0],
[0, 0, 1, 1, 1, 0, 0],
[0, 0, 1, 1, 1, 0, 0],
[0, 0, 0, 0, 0, 0, 0]])
>>> ndimage.binary_erosion(a).astype(a.dtype)
array([[0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 1, 0, 0, 0],
[0, 0, 0, 1, 0, 0, 0],
[0, 0, 0, 1, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0]])
>>> #Erosion removes objects smaller than the structure
>>> ndimage.binary_erosion(a, structure=np.ones((5,5))).astype(a.dtype)
array([[0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0]])
Dilation scipy.ndimage.binary_dilation()
>>> a = np.zeros((5, 5))

>>> a[2, 2] = 1
>>> a
array([[ 0., 0., 0., 0., 0.],
[ 0., 0., 0., 0., 0.],
[ 0., 0., 1., 0., 0.],
[ 0., 0., 0., 0., 0.],
[ 0., 0., 0., 0., 0.]])
>>> ndimage.binary_dilation(a).astype(a.dtype)
array([[ 0., 0., 0., 0., 0.],
[ 0., 0., 1., 0., 0.],
[ 0., 1., 1., 1., 0.],
[ 0., 0., 1., 0., 0.],
[ 0., 0., 0., 0., 0.]])
Opening scipy.ndimage.binary_opening()

>>> a[1:4, 1:4] = 1
>>> a[4, 4] = 1
>>> a
array([[0, 0, 0, 0, 0],

[0, 1, 1, 1, 0],
[0, 1, 1, 1, 0],
[0, 1, 1, 1, 0],
[0, 0, 0, 0, 1]])
>>> # Opening removes small objects
>>> ndimage.binary_opening(a, structure=np.ones((3, 3))).astype(np.int)
array([[0, 0, 0, 0, 0],
[0, 1, 1, 1, 0],
[0, 1, 1, 1, 0],
[0, 1, 1, 1, 0],
[0, 0, 0, 0, 0]])
>>> # Opening can also smooth corners
>>> ndimage.binary_opening(a).astype(np.int)
array([[0, 0, 0, 0, 0],
[0, 0, 1, 0, 0],
[0, 1, 1, 1, 0],
[0, 0, 1, 0, 0],
[0, 0, 0, 0, 0]])
Closing: scipy.ndimage.binary_closing()
Exercise
Check that opening amounts to eroding, then dilating.
An opening operation removes small structures, while a closing operation fills small holes. Such operations
can therefore be used to clean an image.
>>> a = np.zeros((50, 50))

>>> a[10:-10, 10:-10] = 1
>>> a += 0.25 * np.random.standard_normal(a.shape)
>>> mask = a>=0.5
>>> opened_mask = ndimage.binary_opening(mask)
>>> closed_mask = ndimage.binary_closing(opened_mask)
Exercise
Check that the area of the reconstructed square is smaller than the area of the initial square.
(The opposite would occur if the closing step was performed before the opening).
For gray-valued images, eroding (resp. dilating) amounts to replacing a pixel by the minimal (resp. maximal)
value among pixels covered by the structuring element centered on the pixel of interest.

>>> a[1:6, 1:6] = 3
>>> a[4, 4] = 2; a[2, 3] = 1

>>> a
array([[0, 0, 0, 0, 0, 0, 0],
[0, 3, 3, 3, 3, 3, 0],
[0, 3, 3, 1, 3, 3, 0],
[0, 3, 3, 3, 3, 3, 0],
[0, 3, 3, 3, 2, 3, 0],
[0, 3, 3, 3, 3, 3, 0],
[0, 0, 0, 0, 0, 0, 0]])
>>> ndimage.grey_erosion(a, size=(3, 3))
array([[0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0],
[0, 0, 1, 1, 1, 0, 0],
[0, 0, 1, 1, 1, 0, 0],
[0, 0, 3, 2, 2, 0, 0],
[0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0]])
5.10.4 Connected components and measurements on images
Let us first generate a nice synthetic binary image.
>>> x, y = np.indices((100, 100))

>>> sig = np.sin(2*np.pi*x/50.) * np.sin(2*np.pi*y/50.) * (1+x*y/50.**2)**2
>>> mask = sig > 1
scipy.ndimage.label() assigns a different label to each connected component:
>>> labels, nb = ndimage.label(mask)

>>> nb
8
Now compute measurements on each connected component:
>>> areas = ndimage.sum(mask, labels, range(1, labels.max()+1))

>>> areas # The number of pixels in each connected component
array([ 190., 45., 424., 278., 459., 190., 549., 424.])
>>> maxima = ndimage.maximum(sig, labels, range(1, labels.max()+1))
>>> maxima # The maximum signal in each connected component
array([ 1.80238238, 1.13527605, 5.51954079, 2.49611818,
6.71673619, 1.80238238, 16.76547217, 5.51954079])

Extract the 4th connected component, and crop the array around it:
>>> ndimage.find_objects(labels==4)
[(slice(30L, 48L, None), slice(30L, 48L, None))]
>>> sl = ndimage.find_objects(labels==4)
>>> plt.imshow(sig[sl[0]])
See the summary exercise on Image processing application: counting bubbles and unmolten grains for a more
advanced example.
5.11 Summary exercises on scientific computing
The summary exercises use mainly Numpy, Scipy and Matplotlib. They provide some real-life examples of sci-
entific computing with Python. Now that the basics of working with Numpy and Scipy have been introduced,
the interested user is invited to try these exercises.
5.11.1 Maximum wind speed prediction at the Sprog station
The exercise goal is to predict the maximum wind speed occurring every 50 years even if no measure exists
for such a period. The available data are only measured over 21 years at the Sprog meteorological station
located in Denmark. First, the statistical steps will be given and then illustrated with functions from the
scipy.interpolate module. At the end the interested readers are invited to compute results from raw data and
in a slightly different approach.
Statistical approach
The annual maxima are supposed to fit a normal probability density function. However such function is not
going to be estimated because it gives a probability from a wind speed maxima. Finding the maximum wind
speed occurring every 50 years requires the opposite approach, the result needs to be found from a defined
probability. That is the quantile function role and the exercise goal will be to find it. In the current model, it is
supposed that the maximum wind speed occurring every 50 years is defined as the upper 2% quantile.
By definition, the quantile function is the inverse of the cumulative distribution function. The latter describes
the probability distribution of an annual maxima. In the exercise, the cumulative probability p_i for a given
year i is defined as p_i = i/(N+1) with N = 21, the number of measured years. Thus it will be possible to
calculate the cumulative probability of every measured wind speed maxima. From those experimental points,
the scipy.interpolate module will be very useful for fitting the quantile function. Finally the 50 years maxima is
going to be evaluated from the cumulative probability of the 2% quantile.
5.11. Summary exercises on scientific computing 208

Computing the cumulative probabilities
The annual wind speeds maxima have already been computed and saved in the numpy format in the file
examples/max-speeds.npy, thus they will be loaded by using numpy:

>>> max_speeds = np.load('intro/summary-exercises/examples/max-speeds.npy')
>>> years_nb = max_speeds.shape[0]
Following the cumulative probability definition p_i from the previous section, the corresponding values will
be:
>>> cprob = (np.arange(years_nb, dtype=np.float32) + 1)/(years_nb + 1)
and they are assumed to fit the given wind speeds:
>>> sorted_max_speeds = np.sort(max_speeds)
Prediction with UnivariateSpline
In this section the quantile function will be estimated by using the UnivariateSpline class which can
represent a spline from points. The default behavior is to build a spline of degree 3 and points can
have different weights according to their reliability. Variants are InterpolatedUnivariateSpline and
LSQUnivariateSpline on which errors checking is going to change. In case a 2D spline is wanted, the
BivariateSpline class family is provided. All those classes for 1D and 2D splines use the FITPACK For-
tran subroutines, thats why a lower library access is available through the splrep and splev functions for
respectively representing and evaluating a spline. Moreover interpolation functions without the use of FIT-
PACK parameters are also provided for simpler use (see interp1d, interp2d, barycentric_interpolate
and so on).
For the Sprog maxima wind speeds, the UnivariateSpline will be used because a spline of degree 3 seems
to correctly fit the data:
>>> from scipy.interpolate import UnivariateSpline

>>> quantile_func = UnivariateSpline(cprob, sorted_max_speeds)
The quantile function is now going to be evaluated from the full range of probabilities:
>>> nprob = np.linspace(0, 1, 1e2)

>>> fitted_max_speeds = quantile_func(nprob)
In the current model, the maximum wind speed occurring every 50 years is defined as the upper 2% quantile.
As a result, the cumulative probability value will be:
>>> fifty_prob = 1. - 0.02
So the storm wind speed occurring every 50 years can be guessed by:
>>> fifty_wind = quantile_func(fifty_prob)

>>> fifty_wind
array(32.97989825...)
The results are now gathered on a Matplotlib figure:
Exercise with the Gumbell distribution
The interested readers are now invited to make an exercise by using the wind speeds measured over 21 years.
The measurement period is around 90 minutes (the original period was around 10 minutes but the file size
has been reduced for making the exercise setup easier). The data are stored in numpy format inside the file

Fig. 5.1: Solution: Python source file

examples/sprog-windspeeds.npy. Do not look at the source code for the plots until you have completed
the exercise.
The first step will be to find the annual maxima by using numpy and plot them as a matplotlib bar figure.
The second step will be to use the Gumbell distribution on cumulative probabilities p_i defined as
-log( -log(p_i) ) for fitting a linear quantile function (remember that you can define the degree
of the UnivariateSpline). Plotting the annual maxima versus the Gumbell distribution should give
you the following figure.
The last step will be to find 34.23 m/s for the maximum wind speed occurring every 50 years.
5.11.2 Non linear least squares curve fitting: application to point extraction in
topographical lidar data
The goal of this exercise is to fit a model to some data. The data used in this tutorial are lidar data and are
described in details in the following introductory paragraph. If youre impatient and want to practice now,
please skip it and go directly to Loading and visualization.
Introduction
Lidars systems are optical rangefinders that analyze property of scattered light to measure distances. Most of
them emit a short light impulsion towards a target and record the reflected signal. This signal is then processed
to extract the distance between the lidar system and the target.


Topographical lidar systems are such systems embedded in airborne platforms. They measure distances be-
tween the platform and the Earth, so as to deliver information on the Earths topography (see1 for more details).
In this tutorial, the goal is to analyze the waveform recorded by the lidar system2 . Such a signal contains peaks
whose center and amplitude permit to compute the position and some characteristics of the hit target. When
the footprint of the laser beam is around 1m on the Earth surface, the beam can hit multiple targets during the
two-way propagation (for example the ground and the top of a tree or building). The sum of the contributions
of each target hit by the laser beam then produces a complex signal with multiple peaks, each one containing
information about one target.
One state of the art method to extract information from these data is to decompose them in a sum of Gaussian
functions where each function represents the contribution of a target hit by the laser beam.
Therefore, we use the scipy.optimize module to fit a waveform to one or a sum of Gaussian functions.
Loading and visualization
Load the first waveform using:

>>> waveform_1 = np.load('data/waveform_1.npy')
and visualize it:

>>> t = np.arange(len(waveform_1))
>>> plt.plot(t, waveform_1)
>>> plt.show()
1 Mallet, C. and Bretar, F. Full-Waveform Topographic Lidar: State-of-the-Art. ISPRS Journal of Photogrammetry and Remote Sensing
64(1), pp.1-16, January 2009 http://dx.doi.org/10.1016/j.isprsjprs.2008.09.007
2 The data used for this tutorial are part of the demonstration data available for the FullAnalyze software and were kindly provided by
the GIS DRAIX.

As you can notice, this waveform is a 80-bin-length signal with a single peak.
Fitting a waveform with a simple Gaussian model
The signal is very simple and can be modeled as a single Gaussian function and an offset corresponding to the
background noise. To fit the signal with the function, we must:
define the model
propose an initial solution
call scipy.optimize.leastsq
Model
A Gaussian function defined by
t 2

B + A exp

can be defined in python by:
>>> def model(t, coeffs):

... return coeffs[0] + coeffs[1] * np.exp( - ((t-coeffs[2])/coeffs[3])**2 )
where
coeffs[0] is B (noise)
coeffs[1] is A (amplitude)
coeffs[2] is (center)

coeffs[3] is (width)
Initial solution
An approximative initial solution that we can find from looking at the graph is for instance:
>>> x0 = np.array([3, 30, 15, 1], dtype=float)
Fit
scipy.optimize.leastsq minimizes the sum of squares of the function given as an argument. Basically, the
function to minimize is the residuals (the difference between the data and the model):
>>> def residuals(coeffs, y, t):

... return y - model(t, coeffs)
So lets get our solution by calling scipy.optimize.leastsq() with the following arguments:
the function to minimize
an initial solution
the additional arguments to pass to the function
>>> from scipy.optimize import leastsq

>>> x, flag = leastsq(residuals, x0, args=(waveform_1, t))
>>> print(x)
[ 2.70363341 27.82020742 15.47924562 3.05636228]
And visualize the solution:
>>> plt.plot(t, waveform_1, t, model(t, x))

>>> plt.legend(['waveform', 'model'])
>>> plt.show()
Remark: from scipy v0.8 and above, you should rather use scipy.optimize.curve_fit() which takes the
model and the data as arguments, so you dont need to define the residuals any more.
Going further
Try with a more complex waveform (for instance data/waveform_2.npy) that contains three significant
peaks. You must adapt the model which is now a sum of Gaussian functions instead of only one Gaussian
peak.

In some cases, writing an explicit function to compute the Jacobian is faster than letting leastsq esti-
mate it numerically. Create a function to compute the Jacobian of the residuals and use it as an input for
leastsq.
When we want to detect very small peaks in the signal, or when the initial guess is too far from a good
solution, the result given by the algorithm is often not satisfying. Adding constraints to the parameters
of the model enables to overcome such limitations. An example of a priori knowledge we can add is the
sign of our variables (which are all positive).
With the following initial solution:
>>> x0 = np.array([3, 50, 20, 1], dtype=float)
compare the result of scipy.optimize.leastsq() and what you can get with scipy.optimize.
fmin_slsqp() when adding boundary constraints.

5.11.3 Image processing application: counting bubbles and unmolten grains
Statement of the problem
1. Open the image file MV_HFV_012.jpg and display it. Browse through the keyword arguments in the doc-
string of imshow to display the image with the right orientation (origin in the bottom left corner, and not the
upper left corner as for standard arrays).
This Scanning Element Microscopy image shows a glass sample (light gray matrix) with some bubbles (on
black) and unmolten sand grains (dark gray). We wish to determine the fraction of the sample covered by
these three phases, and to estimate the typical size of sand grains and bubbles, their sizes, etc.
2. Crop the image to remove the lower panel with measure information.
3. Slightly filter the image with a median filter in order to refine its histogram. Check how the histogram
changes.
4. Using the histogram of the filtered image, determine thresholds that allow to define masks for sand pixels,
glass pixels and bubble pixels. Other option (homework): write a function that determines automatically the
thresholds from the minima of the histogram.
5. Display an image in which the three phases are colored with three different colors.
6. Use mathematical morphology to clean the different phases.

7. Attribute labels to all bubbles and sand grains, and remove from the sand mask grains that are smaller than
10 pixels. To do so, use ndimage.sum or np.bincount to compute the grain sizes.
8. Compute the mean size of bubbles.
Proposed solution
5.11.4 Example of solution for the image processing exercise: unmolten grains in
glass
1. Open the image file MV_HFV_012.jpg and display it. Browse through the keyword arguments in the
docstring of imshow to display the image with the right orientation (origin in the bottom left corner,
and not the upper left corner as for standard arrays).
>>> dat = pl.imread('data/MV_HFV_012.jpg')
2. Crop the image to remove the lower panel with measure information.
>>> dat = dat[:-60]
3. Slightly filter the image with a median filter in order to refine its histogram. Check how the histogram
changes.

>>> filtdat = ndimage.median_filter(dat, size=(7,7))

>>> hi_dat = np.histogram(dat, bins=np.arange(256))
>>> hi_filtdat = np.histogram(filtdat, bins=np.arange(256))
4. Using the histogram of the filtered image, determine thresholds that allow to define masks for sand pix-
els, glass pixels and bubble pixels. Other option (homework): write a function that determines automat-
ically the thresholds from the minima of the histogram.
>>> void = filtdat <= 50

>>> sand = np.logical_and(filtdat > 50, filtdat <= 114)
>>> glass = filtdat > 114
5. Display an image in which the three phases are colored with three different colors.
>>> phases = void.astype(np.int) + 2*glass.astype(np.int) + 3*sand.astype(np.int)

6. Use mathematical morphology to clean the different phases.
>>> sand_op = ndimage.binary_opening(sand, iterations=2)
7. Attribute labels to all bubbles and sand grains, and remove from the sand mask grains that are smaller
than 10 pixels. To do so, use ndimage.sum or np.bincount to compute the grain sizes.
>>> sand_labels, sand_nb = ndimage.label(sand_op)

>>> sand_areas = np.array(ndimage.sum(sand_op, sand_labels, np.arange(sand_labels.
,max()+1)))
>>> mask = sand_areas > 100
>>> remove_small_sand = mask[sand_labels.ravel()].reshape(sand_labels.shape)
8. Compute the mean size of bubbles.
>>> bubbles_labels, bubbles_nb = ndimage.label(void)

>>> bubbles_areas = np.bincount(bubbles_labels.ravel())[1:]
>>> mean_bubble_size = bubbles_areas.mean()
>>> median_bubble_size = np.median(bubbles_areas)
>>> mean_bubble_size, median_bubble_size
(1699.875, 65.0)

5.12 Full code examples for the scipy chapter
5.12.1 Finding the minimum of a smooth function
Demos various methods to find the minimum of a function.
import numpy as np
def f(x):
return x**2 + 10*np.sin(x)
x = np.arange(-10, 10, 0.1)

plt.plot(x, f(x))
Now find the minimum with a few methods
from scipy import optimize
# The default (Nelder Mead)

print(optimize.minimize(f, x0=0))
Out:
fun: -7.945823375615215
hess_inv: array([[ 0.08589237]])
5.12. Full code examples for the scipy chapter 221

jac: array([ -1.19209290e-06])

nfev: 18
nit: 5
njev: 6
status: 0
success: True
x: array([-1.30644012])
print(optimize.minimize(f, x0=0, method="L-BFGS-B"))
Out:
fun: array([-7.94582338])
jac: array([ -1.42108547e-06])
message: b'CONVERGENCE: NORM_OF_PROJECTED_GRADIENT_<=_PGTOL'
nfev: 12
nit: 5
status: 0
success: True
x: array([-1.30644013])
plt.show()

Download Python source code: plot_optimize_example1.py
Download Jupyter notebook: plot_optimize_example1.ipynb
5.12.2 Detrending a signal
scipy.signal.detrend() removes a linear trend.

Generate a random signal with a trend
import numpy as np
x = t + np.random.normal(size=100)
Detrend
from scipy import signal

x_detrended = signal.detrend(x)
Plot

plt.plot(t, x, label="x")
plt.plot(t, x_detrended, label="x_detrended")
plt.legend(loc='best')
plt.show()


Download Python source code: plot_detrend.py
Download Jupyter notebook: plot_detrend.ipynb
5.12.3 Resample a signal with scipy.signal.resample
scipy.signal.resample() uses FFT to resample a 1D signal.

Generate a signal with 100 data point
import numpy as np
x = np.sin(t)
Downsample it by a factor of 4

x_resampled = signal.resample(x, 25)
Plot

plt.plot(t, x, label='Original signal')
plt.plot(t[::4], x_resampled, 'ko', label='Resampled signal')
plt.show()


Download Python source code: plot_resample.py
Download Jupyter notebook: plot_resample.ipynb
5.12.4 Integrating a simple ODE
Solve the ODE dy/dt = -2y between t = 0..4, with the initial condition y(t=0) = 1.
import numpy as np
from scipy.integrate import odeint

def calc_derivative(ypos, time):

return -2*ypos
time_vec = np.linspace(0, 4, 40)

yvec = odeint(calc_derivative, 1, time_vec)
plt.plot(time_vec, yvec)
plt.xlabel('t: Time')
plt.ylabel('y: Position')
plt.tight_layout()

Download Python source code: plot_odeint_simple.py
Download Jupyter notebook: plot_odeint_simple.ipynb
5.12.5 Comparing 2 sets of samples from Gaussians
import numpy as np
# Generates 2 sets of observations

samples1 = np.random.normal(0, size=1000)
samples2 = np.random.normal(1, size=1000)
# Compute a histogram of the sample

bins = np.linspace(-4, 4, 30)
histogram1, bins = np.histogram(samples1, bins=bins, normed=True)
histogram2, bins = np.histogram(samples2, bins=bins, normed=True)

plt.hist(samples1, bins=bins, normed=True, label="Samples 1")
plt.hist(samples2, bins=bins, normed=True, label="Samples 2")
plt.show()

Download Python source code: plot_t_test.py
Download Jupyter notebook: plot_t_test.ipynb
5.12.6 Integrate the Damped spring-mass oscillator
import numpy as np
from scipy.integrate import odeint
mass = 0.5 # kg
kspring = 4 # N/m
cviscous = 0.4 # N s/m
eps = cviscous / (2 * mass * np.sqrt(kspring/mass))

omega = np.sqrt(kspring / mass)
def calc_deri(yvec, time, eps, omega):

return (yvec[1], -eps * omega * yvec[1] - omega **2 * yvec[0])
time_vec = np.linspace(0, 10, 100)

yinit = (1, 0)
yarr = odeint(calc_deri, yinit, time_vec, args=(eps, omega))
plt.plot(time_vec, yarr[:, 0], label='y')
plt.plot(time_vec, yarr[:, 1], label="y'")
plt.show()


Download Python source code: plot_odeint_damped_spring_mass.py
Download Jupyter notebook: plot_odeint_damped_spring_mass.ipynb
5.12.7 Normal distribution: histogram and PDF
Explore the normal distribution: a histogram built from samples and the PDF (probability density function).
import numpy as np
# Sample from a normal distribution using numpy's random number generator

samples = np.random.normal(size=10000)
# Compute a histogram of the sample

bins = np.linspace(-5, 5, 30)
histogram, bins = np.histogram(samples, bins=bins, normed=True)
bin_centers = 0.5*(bins[1:] + bins[:-1])
# Compute the PDF on the bin centers from scipy distribution object
from scipy import stats
pdf = stats.norm.pdf(bin_centers)

plt.plot(bin_centers, histogram, label="Histogram of samples")
plt.plot(bin_centers, pdf, label="PDF")
plt.legend()
plt.show()

Download Python source code: plot_normal_distribution.py

Download Jupyter notebook: plot_normal_distribution.ipynb
5.12.8 Curve fitting
Demos a simple curve fitting

First generate some data
import numpy as np
# Seed the random number generator for reproducibility

np.random.seed(0)
x_data = np.linspace(-5, 5, num=50)

y_data = 2.9 * np.sin(1.5 * x_data) + np.random.normal(size=50)
# And plot it
plt.scatter(x_data, y_data)
Now fit a simple sine function to the data
def test_func(x, a, b):

return a * np.sin(b * x)
params, params_covariance = optimize.curve_fit(test_func, x_data, y_data,

p0=[2, 2])
print(params)

Out:
[ 3.05931973 1.45754553]
And plot the resulting curve on the data
plt.scatter(x_data, y_data, label='Data')
plt.plot(x_data, test_func(x_data, params[0], params[1]),
label='Fitted function')
plt.show()

Download Python source code: plot_curve_fit.py
Download Jupyter notebook: plot_curve_fit.ipynb
5.12.9 Spectrogram, power spectral density
Demo spectrogram and power spectral density on a frequency chirp.
import numpy as np
Generate a chirp signal
# Seed the random number generator

np.random.seed(0)

time_step = .01
time_vec = np.arange(0, 70, time_step)
# A signal with a small frequency chirp

sig = np.sin(0.5 * np.pi * time_vec * (1 + .1 * time_vec))
plt.plot(time_vec, sig)
Compute and plot the spectrogram
The spectrum of the signal on consecutive time windows

freqs, times, spectrogram = signal.spectrogram(sig)
plt.imshow(spectrogram, aspect='auto', cmap='hot_r', origin='lower')
plt.title('Spectrogram')
plt.ylabel('Frequency band')
plt.xlabel('Time window')
plt.tight_layout()

Compute and plot the power spectral density (PSD)
The power of the signal per frequency band
freqs, psd = signal.welch(sig)
plt.semilogx(freqs, psd)
plt.title('PSD: power spectral density')
plt.xlabel('Frequency')
plt.ylabel('Power')
plt.tight_layout()

plt.show()

Download Python source code: plot_spectrogram.py
Download Jupyter notebook: plot_spectrogram.ipynb

5.12.10 A demo of 1D interpolation
# Generate data
import numpy as np
np.random.seed(0)
measured_time = np.linspace(0, 1, 10)
noise = 1e-1 * (np.random.random(10)*2 - 1)
measures = np.sin(2 * np.pi * measured_time) + noise
# Interpolate it to new time points

from scipy.interpolate import interp1d
linear_interp = interp1d(measured_time, measures)
interpolation_time = np.linspace(0, 1, 50)
linear_results = linear_interp(interpolation_time)
cubic_interp = interp1d(measured_time, measures, kind='cubic')
cubic_results = cubic_interp(interpolation_time)
# Plot the data and the interpolation

plt.plot(measured_time, measures, 'o', ms=6, label='measures')
plt.plot(interpolation_time, linear_results, label='linear interp')
plt.plot(interpolation_time, cubic_results, label='cubic interp')
plt.legend()
plt.show()

Download Python source code: plot_interpolation.py
Download Jupyter notebook: plot_interpolation.ipynb

5.12.11 Demo mathematical morphology
A basic demo of binary opening and closing.
# Generate some binary data

import numpy as np
np.random.seed(0)
a = np.zeros((50, 50))
a[10:-10, 10:-10] = 1
a += 0.25 * np.random.standard_normal(a.shape)
mask = a>=0.5
# Apply mathematical morphology

from scipy import ndimage
opened_mask = ndimage.binary_opening(mask)
closed_mask = ndimage.binary_closing(opened_mask)
# Plot
plt.figure(figsize=(12, 3.5))
plt.subplot(141)
plt.imshow(a, cmap=plt.cm.gray)
plt.axis('off')
plt.title('a')
plt.subplot(142)
plt.imshow(mask, cmap=plt.cm.gray)
plt.axis('off')
plt.title('mask')
plt.subplot(143)
plt.imshow(opened_mask, cmap=plt.cm.gray)
plt.axis('off')
plt.title('opened_mask')
plt.subplot(144)
plt.imshow(closed_mask, cmap=plt.cm.gray)
plt.title('closed_mask')
plt.axis('off')
plt.subplots_adjust(wspace=.05, left=.01, bottom=.01, right=.99, top=.99)
plt.show()

Download Python source code: plot_mathematical_morpho.py
Download Jupyter notebook: plot_mathematical_morpho.ipynb

5.12.12 Plot geometrical transformations on images
Demo geometrical transformations of images.
# Load some data

from scipy import misc
face = misc.face(gray=True)
# Apply a variety of transformations

shifted_face = ndimage.shift(face, (50, 50))
shifted_face2 = ndimage.shift(face, (50, 50), mode='nearest')
rotated_face = ndimage.rotate(face, 30)
cropped_face = face[50:-50, 50:-50]
zoomed_face = ndimage.zoom(face, 2)
zoomed_face.shape
plt.subplot(151)
plt.imshow(shifted_face, cmap=plt.cm.gray)
plt.axis('off')
plt.subplot(152)
plt.imshow(shifted_face2, cmap=plt.cm.gray)
plt.axis('off')
plt.subplot(153)
plt.imshow(rotated_face, cmap=plt.cm.gray)
plt.axis('off')
plt.subplot(154)
plt.imshow(cropped_face, cmap=plt.cm.gray)
plt.axis('off')
plt.subplot(155)
plt.imshow(zoomed_face, cmap=plt.cm.gray)
plt.axis('off')
plt.show()

Download Python source code: plot_image_transform.py
Download Jupyter notebook: plot_image_transform.ipynb

5.12.13 Demo connected components
Extracting and labeling connected components in a 2D array
import numpy as np
Generate some binary data
np.random.seed(0)
x, y = np.indices((100, 100))
sig = np.sin(2*np.pi*x/50.) * np.sin(2*np.pi*y/50.) * (1+x*y/50.**2)**2
mask = sig > 1
plt.imshow(sig)
plt.axis('off')
plt.title('sig')
plt.axis('off')
plt.title('mask')
Label connected components

labels, nb = ndimage.label(mask)
plt.figure(figsize=(3.5, 3.5))
plt.imshow(labels)
plt.title('label')
plt.axis('off')

Extract the 4th connected component, and crop the array around it
sl = ndimage.find_objects(labels==4)
plt.imshow(sig[sl[0]])
plt.title('Cropped connected component')
plt.axis('off')
plt.show()

Download Python source code: plot_connect_measurements.py
Download Jupyter notebook: plot_connect_measurements.ipynb

5.12.14 Minima and roots of a function
Demos finding minima and roots of a function.
Define the function
import numpy as np
x = np.arange(-10, 10, 0.1)

def f(x):
return x**2 + 10*np.sin(x)
Find minima
# Global optimization
grid = (-10, 10, 0.1)
xmin_global = optimize.brute(f, (grid, ))
print("Global minima found %s " % xmin_global)
# Constrain optimization
xmin_local = optimize.fminbound(f, 0, 10)
print("Local minimum found %s " % xmin_local)
Out:
Global minima found [-1.30641113]

Local minimum found 3.8374671195
Root finding
root = optimize.root(f, 1) # our initial guess is 1

print("First root found %s " % root.x)
root2 = optimize.root(f, -2.5)
print("Second root found %s " % root2.x)
Out:
First root found [ 0.]

Second root found [-2.47948183]
Plot function, minima, and roots

ax = fig.add_subplot(111)
# Plot the function

ax.plot(x, f(x), 'b-', label="f(x)")
# Plot the minima

xmins = np.array([xmin_global[0], xmin_local])

ax.plot(xmins, f(xmins), 'go', label="Minima")
# Plot the roots

roots = np.array([root.x, root2.x])
ax.plot(roots, f(roots), 'kv', label="Roots")
# Decorate the figure

ax.legend(loc='best')
ax.set_xlabel('x')
ax.set_ylabel('f(x)')
ax.axhline(0, color='gray')
plt.show()

Download Python source code: plot_optimize_example2.py
Download Jupyter notebook: plot_optimize_example2.ipynb
5.12.15 Optimization of a two-parameter function
import numpy as np
# Define the function that we are interested in

def sixhump(x):
return ((4 - 2.1*x[0]**2 + x[0]**4 / 3.) * x[0]**2 + x[0] * x[1]
+ (-4 + 4*x[1]**2) * x[1] **2)
# Make a grid to evaluate the function (for plotting)

x = np.linspace(-2, 2)
y = np.linspace(-1, 1)
xg, yg = np.meshgrid(x, y)

A 2D image plot of the function
Simple visualization in 2D

plt.figure()
plt.imshow(sixhump([xg, yg]), extent=[-2, 2, -1, 1])
plt.colorbar()
A 3D surface plot of the function

fig = plt.figure()
ax = fig.add_subplot(111, projection='3d')
surf = ax.plot_surface(xg, yg, sixhump([xg, yg]), rstride=1, cstride=1,
cmap=plt.cm.jet, linewidth=0, antialiased=False)
ax.set_xlabel('x')
ax.set_ylabel('y')
ax.set_zlabel('f(x, y)')
ax.set_title('Six-hump Camelback function')

Find the minima
x_min = optimize.minimize(sixhump, x0=[0, 0])
plt.figure()
# Show the function in 2D
plt.imshow(sixhump([xg, yg]), extent=[-2, 2, -1, 1])
plt.colorbar()
# And the minimum that we've found:
plt.scatter(x_min.x[0], x_min.x[1])
plt.show()


Download Python source code: plot_2d_minimization.py
Download Jupyter notebook: plot_2d_minimization.ipynb
5.12.16 Plot filtering on images
Demo filtering for denoising of images.
# Load some data

face = misc.face(gray=True)
face = face[:512, -512:] # crop out square on right
# Apply a variety of filters



import numpy as np
noisy_face = np.copy(face).astype(np.float)
noisy_face += face.std() * 0.5 * np.random.standard_normal(face.shape)
blurred_face = ndimage.gaussian_filter(noisy_face, sigma=3)
median_face = ndimage.median_filter(noisy_face, size=5)
wiener_face = signal.wiener(noisy_face, (5, 5))
plt.subplot(141)
plt.imshow(noisy_face, cmap=plt.cm.gray)
plt.axis('off')
plt.title('noisy')
plt.subplot(142)
plt.imshow(blurred_face, cmap=plt.cm.gray)
plt.axis('off')
plt.title('Gaussian filter')
plt.subplot(143)
plt.imshow(median_face, cmap=plt.cm.gray)
plt.axis('off')
plt.title('median filter')
plt.subplot(144)
plt.imshow(wiener_face, cmap=plt.cm.gray)
plt.title('Wiener filter')
plt.axis('off')
plt.show()

Download Python source code: plot_image_filters.py
Download Jupyter notebook: plot_image_filters.ipynb
5.12.17 Plotting and manipulating FFTs for filtering
Plot the power of the FFT of a signal and inverse FFT back to reconstruct a signal.
This example demonstrate scipy.fftpack.fft(), scipy.fftpack.fftfreq() and scipy.fftpack.
ifft(). It implements a basic filter that is very suboptimal, and should not be used.
import numpy as np
from scipy import fftpack
Generate the signal
# Seed the random number generator

np.random.seed(1234)
time_step = 0.02
period = 5.

time_vec = np.arange(0, 20, time_step)

sig = (np.sin(2 * np.pi / period * time_vec)
+ 0.5 * np.random.randn(time_vec.size))
plt.plot(time_vec, sig, label='Original signal')
Compute and plot the power
# The FFT of the signal

sig_fft = fftpack.fft(sig)
# And the power (sig_fft is of complex dtype)

power = np.abs(sig_fft)
# The corresponding frequencies

sample_freq = fftpack.fftfreq(sig.size, d=time_step)
# Plot the FFT power

plt.plot(sample_freq, power)
plt.xlabel('Frequency [Hz]')
plt.ylabel('plower')
# Find the peak frequency: we can focus on only the positive frequencies
pos_mask = np.where(sample_freq > 0)

freqs = sample_freq[pos_mask]
peak_freq = freqs[power[pos_mask].argmax()]
# Check that it does indeed correspond to the frequency that we generate

# the signal with
np.allclose(peak_freq, 1./period)
# An inner plot to show the peak frequency

axes = plt.axes([0.55, 0.3, 0.3, 0.5])
plt.title('Peak frequency')
plt.plot(freqs[:8], power[:8])
plt.setp(axes, yticks=[])
# scipy.signal.find_peaks_cwt can also be used for more advanced

# peak detection
Remove all the high frequencies
We now remove all the high frequencies and transform back from frequencies to signal.
high_freq_fft = sig_fft.copy()
high_freq_fft[np.abs(sample_freq) > peak_freq] = 0
filtered_sig = fftpack.ifft(high_freq_fft)
plt.plot(time_vec, sig, label='Original signal')
plt.plot(time_vec, filtered_sig, linewidth=3, label='Filtered signal')
plt.xlabel('Time [s]')

plt.ylabel('Amplitude')
Note This is actually a bad way of creating a filter: such brutal cut-off in frequency space does not control
distorsion on the signal.
Filters should be created using the scipy filter design code
plt.show()

Download Python source code: plot_fftpack.py
Download Jupyter notebook: plot_fftpack.ipynb
5.12.18 Solutions of the exercises for scipy
Crude periodicity finding
Discover the periods in evolution of animal populations (../../data/populations.txt)

Load the data
import numpy as np
data = np.loadtxt('../../../../data/populations.txt')
years = data[:, 0]
populations = data[:, 1:]
Plot the data

plt.figure()
plt.plot(years, populations * 1e-3)
plt.xlabel('Year')
plt.ylabel('Population number ($\cdot10^3$)')
plt.legend(['hare', 'lynx', 'carrot'], loc=1)
Plot its periods
ft_populations = fftpack.fft(populations, axis=0)

frequencies = fftpack.fftfreq(populations.shape[0], years[1] - years[0])
periods = 1 / frequencies
plt.figure()
plt.plot(periods, abs(ft_populations) * 1e-3, 'o')

plt.xlim(0, 22)
plt.xlabel('Period')
plt.ylabel('Power ($\cdot10^3$)')
plt.show()
Theres probably a period of around 10 years (obvious from the plot), but for this crude a method, theres not
enough data to say much more.
Download Python source code: plot_periodicity_finder.py
Download Jupyter notebook: plot_periodicity_finder.ipynb
Curve fitting: temperature as a function of month of the year
We have the min and max temperatures in Alaska for each months of the year. We would like to find a function
to describe this yearly evolution.
For this, we will fit a periodic function.
The data
import numpy as np
temp_max = np.array([17, 19, 21, 28, 33, 38, 37, 37, 31, 23, 19, 18])
temp_min = np.array([-62, -59, -56, -46, -32, -18, -9, -13, -25, -46, -52, -58])


months = np.arange(12)
plt.plot(months, temp_max, 'ro')
plt.plot(months, temp_min, 'bo')
plt.xlabel('Month')
plt.ylabel('Min and max temperature')
Fitting it to a periodic function

def yearly_temps(times, avg, ampl, time_offset):
return (avg
+ ampl * np.cos((times + time_offset) * 2 * np.pi / times.max()))
res_max, cov_max = optimize.curve_fit(yearly_temps, months,

temp_max, [20, 10, 0])
res_min, cov_min = optimize.curve_fit(yearly_temps, months,
temp_min, [-40, 20, 0])
Plotting the fit
days = np.linspace(0, 12, num=365)
plt.figure()
plt.plot(months, temp_max, 'ro')
plt.plot(days, yearly_temps(days, *res_max), 'r-')

plt.plot(months, temp_min, 'bo')

plt.plot(days, yearly_temps(days, *res_min), 'b-')
plt.xlabel('Month')
plt.ylabel('Temperature ($^\circ$C)')
plt.show()

Download Python source code: plot_curvefit_temperature_data.py
Download Jupyter notebook: plot_curvefit_temperature_data.ipynb
Simple image blur by convolution with a Gaussian kernel
Blur an an image (../../../../data/elephant.png) using a Gaussian kernel.

Convolution is easy to perform with FFT: convolving two signals boils down to multiplying their FFTs (and
performing an inverse FFT).
import numpy as np

The original image
# read image
img = plt.imread('../../../../data/elephant.png')
plt.figure()
plt.imshow(img)
Prepare an Gaussian convolution kernel
# First a 1-D Gaussian

t = np.linspace(-10, 10, 30)
bump = np.exp(-0.1*t**2)
bump /= np.trapz(bump) # normalize the integral to 1
# make a 2-D kernel out of it

kernel = bump[:, np.newaxis] * bump[np.newaxis, :]
Implement convolution via FFT
# Padded fourier transform, with the same shape as the image

# We use :func:`scipy.signal.fftpack.fft2` to have a 2D FFT
kernel_ft = fftpack.fft2(kernel, shape=img.shape[:2], axes=(0, 1))
# convolve
img_ft = fftpack.fft2(img, axes=(0, 1))
# the 'newaxis' is to match to color direction

img2_ft = kernel_ft[:, :, np.newaxis] * img_ft

img2 = fftpack.ifft2(img2_ft, axes=(0, 1)).real
# clip values to range

img2 = np.clip(img2, 0, 1)
# plot output
plt.figure()
plt.imshow(img2)
Further exercise (only if you are familiar with this stuff):

A wrapped border appears in the upper left and top edges of the image. This is because the padding is not
done correctly, and does not take the kernel size into account (so the convolution flows out of bounds of the
image). Try to remove this artifact.
A function to do it: scipy.signal.fftconvolve()
The above exercise was only for didactic reasons: there exists a function in scipy that will do this
for us, and probably do a better job: scipy.signal.fftconvolve()

# mode='same' is there to enforce the same output shape as input arrays
# (ie avoid border effects)
img3 = signal.fftconvolve(img, kernel[:, :, np.newaxis], mode='same')
plt.figure()
plt.imshow(img3)

Note that we still have a decay to zero at the border of the image. Using scipy.ndimage.gaussian_filter()
would get rid of this artifact
plt.show()

Download Python source code: plot_image_blur.py
Download Jupyter notebook: plot_image_blur.ipynb
Image denoising by FFT
Denoise an image (../../../../data/moonlanding.png) by implementing a blur with an FFT.

Implements, via FFT, the following convolution:
Z
f 1 (t ) = d t 0 K (t t 0 ) f 0 (t 0 )
f1 () = K () f0 ()
Read and plot the image
import numpy as np
im = plt.imread('../../../../data/moonlanding.png').astype(float)

plt.figure()
plt.imshow(im, plt.cm.gray)
plt.title('Original image')
Compute the 2d FFT of the input image

im_fft = fftpack.fft2(im)
# Show the results
def plot_spectrum(im_fft):
from matplotlib.colors import LogNorm
# A logarithmic colormap
plt.imshow(np.abs(im_fft), norm=LogNorm(vmin=5))
plt.colorbar()
plt.figure()
plot_spectrum(im_fft)
plt.title('Fourier transform')

Filter in FFT
# In the lines following, we'll make a copy of the original spectrum and
# truncate coefficients.
# Define the fraction of coefficients (in each direction) we keep

keep_fraction = 0.1
# Call ff a copy of the original transform. Numpy arrays have a copy

# method for this purpose.
im_fft2 = im_fft.copy()
# Set r and c to be the number of rows and columns of the array.

r, c = im_fft2.shape
# Set to zero all rows with indices between r*keep_fraction and

# r*(1-keep_fraction):
im_fft2[int(r*keep_fraction):int(r*(1-keep_fraction))] = 0
# Similarly with the columns:

im_fft2[:, int(c*keep_fraction):int(c*(1-keep_fraction))] = 0
plt.figure()
plot_spectrum(im_fft2)
plt.title('Filtered Spectrum')

Reconstruct the final image
# Reconstruct the denoised image from the filtered spectrum, keep only the
# real part for display.
im_new = fftpack.ifft2(im_fft2).real
plt.figure()
plt.imshow(im_new, plt.cm.gray)
plt.title('Reconstructed Image')

Easier and better: scipy.ndimage.gaussian_filter()
Implementing filtering directly with FFTs is tricky and time consuming. We can use the Gaussian
filter from scipy.ndimage

im_blur = ndimage.gaussian_filter(im, 4)
plt.figure()
plt.imshow(im_blur, plt.cm.gray)
plt.title('Blurred image')
plt.show()


Download Python source code: plot_fft_image_denoise.py
Download Jupyter notebook: plot_fft_image_denoise.ipynb
See also:
References to go further
Some chapters of the advanced and the packages and applications parts of the scipy lectures
The scipy cookbook

CHAPTER 6
Getting help and finding documentation
Author: Emmanuelle Gouillart

Rather than knowing all functions in Numpy and Scipy, it is important to find rapidly information throughout
the documentation and the available help. Here are some ways to get information:
In Ipython, help function opens the docstring of the function. Only type the beginning of the func-
tions name and use tab completion to display the matching functions.
In [204]: help np.v

np.vander np.vdot np.version np.void0 np.vstack
np.var np.vectorize np.void np.vsplit
In [204]: help np.vander
In Ipython it is not possible to open a separated window for help and documentation; how-
ever one can always open a second Ipython shell just to display help and docstrings. . .
Numpys and Scipys documentations can be browsed online on http://docs.scipy.org/doc. The search
button is quite useful inside the reference documentation of the two packages (http://docs.scipy.org/
doc/numpy/reference/ and http://docs.scipy.org/doc/scipy/reference/).
259
Tutorials on various topics as well as the complete API with all docstrings are found on this website.
Numpys and Scipys documentation is enriched and updated on a regular basis by users on a wiki http:
//docs.scipy.org/doc/numpy/. As a result, some docstrings are clearer or more detailed on the wiki,
and you may want to read directly the documentation on the wiki instead of the official documentation
website. Note that anyone can create an account on the wiki and write better documentation; this is an
easy way to contribute to an open-source project and improve the tools you are using!
Scipy central http://central.scipy.org/ gives recipes on many common problems frequently encoun-
tered, such as fitting data points, solving ODE, etc.

Matplotlibs website http://matplotlib.org/ features a very nice gallery with a large number of plots, each
of them shows both the source code and the resulting plot. This is very useful for learning by example.
More standard documentation is also available.
Finally, two more technical possibilities are useful as well:
In Ipython, the magical function %psearch search for objects matching patterns. This is useful if, for
example, one does not know the exact name of a function.

In [4]: %psearch np.diag*
np.diag
np.diagflat
np.diagonal
numpy.lookfor looks for keywords inside the docstrings of specified modules.
In [45]: numpy.lookfor('convolution')
Search results for 'convolution'
--------------------------------
numpy.convolve
Returns the discrete, linear convolution of two one-dimensional
sequences.
numpy.bartlett
Return the Bartlett window.
numpy.correlate
Discrete, linear correlation of two 1-dimensional sequences.
In [46]: numpy.lookfor('remove', module='os')
Search results for 'remove'
---------------------------
os.remove
260
remove(path)
os.removedirs
removedirs(path)
os.rmdir
rmdir(path)
os.unlink
unlink(path)
os.walk
Directory tree generator.
If everything listed above fails (and Google doesnt have the answer). . . dont despair! Write to the
mailing-list suited to your problem: you should have a quick answer if you describe your problem well.
Experts on scientific python often give very enlightening explanations on the mailing-list.
Numpy discussion (numpy-discussion@scipy.org): all about numpy arrays, manipulating them,
indexation questions, etc.
SciPy Users List (scipy-user@scipy.org): scientific computing with Python, high-level data process-
ing, in particular with the scipy package.
matplotlib-users@lists.sourceforge.net for plotting with matplotlib.
261
Part II
Advanced topics
262
This part of the Scipy lecture notes is dedicated to advanced usage. It strives to educate the proficient Python
coder to be an expert and tackles various specific topics.
263
CHAPTER 7
Advanced Python Constructs
Author Zbigniew Jedrzejewski-Szmek

This section covers some features of the Python language which can be considered advanced in the sense
that not every language has them, and also in the sense that they are more useful in more complicated pro-
grams or libraries, but not in the sense of being particularly specialized, or particularly complicated.
It is important to underline that this chapter is purely about the language itself about features supported
through special syntax complemented by functionality of the Python stdlib, which could not be implemented
through clever external modules.
The process of developing the Python programming language, its syntax, is very transparent; proposed
changes are evaluated from various angles and discussed via Python Enhancement Proposals PEPs. As a
result, features described in this chapter were added after it was shown that they indeed solve real problems
and that their use is as simple as possible.
Chapter contents
Iterators, generator expressions and generators

Iterators
Generator expressions
Generators
Bidirectional communication
Chaining generators
Decorators
Replacing or tweaking the original object
264
Decorators implemented as classes and as functions

Copying the docstring and other attributes of the original function
Examples in the standard library
Deprecation of functions
A while-loop removing decorator
A plugin registration system
Context managers
Catching exceptions
Using generators to define context managers
7.1 Iterators, generator expressions and generators
7.1.1 Iterators
Simplicity
Duplication of effort is wasteful, and replacing the various home-grown approaches with a standard feature
usually ends up making things more readable, and interoperable as well.
Guido van Rossum Adding Optional Static Typing to Python
An iterator is an object adhering to the iterator protocol basically this means that it has a next method,
which, when called, returns the next item in the sequence, and when theres nothing to return, raises the
StopIteration exception.
An iterator object allows to loop just once. It holds the state (position) of a single iteration, or from the other
side, each loop over a sequence requires a single iterator object. This means that we can iterate over the same
sequence more than once concurrently. Separating the iteration logic from the sequence allows us to have
more than one way of iteration.
Calling the __iter__ method on a container to create an iterator object is the most straightforward way to get
hold of an iterator. The iter function does that for us, saving a few keystrokes.
>>> nums = [1, 2, 3] # note that ... varies: these are different objects
>>> iter(nums)
<...iterator object at ...>
>>> nums.__iter__()
<...iterator object at ...>
>>> nums.__reversed__()
<...reverseiterator object at ...>
>>> it = iter(nums)
>>> next(it)
1
>>> next(it)
2
>>> next(it)
3
>>> next(it)
StopIteration
7.1. Iterators, generator expressions and generators 265

When used in a loop, StopIteration is swallowed and causes the loop to finish. But with explicit invocation,
we can see that once the iterator is exhausted, accessing it raises an exception.
Using the for..in loop also uses the __iter__ method. This allows us to transparently start the iteration over a
sequence. But if we already have the iterator, we want to be able to use it in an for loop in the same way. In
order to achieve this, iterators in addition to next are also required to have a method called __iter__ which
returns the iterator (self).
Support for iteration is pervasive in Python: all sequences and unordered containers in the standard library
allow this. The concept is also stretched to other things: e.g. file objects support iteration over lines.
>>> f = open('/etc/fstab')
>>> f is f.__iter__()
True
The file is an iterator itself and its __iter__ method doesnt create a separate object: only a single thread of
sequential access is allowed.
7.1.2 Generator expressions
A second way in which iterator objects are created is through generator expressions, the basis for list compre-
hensions. To increase clarity, a generator expression must always be enclosed in parentheses or an expression.
If round parentheses are used, then a generator iterator is created. If rectangular parentheses are used, the pro-
cess is short-circuited and we get a list.
>>> (i for i in nums)

<generator object <genexpr> at 0x...>
>>> [i for i in nums]
[1, 2, 3]
>>> list(i for i in nums)
[1, 2, 3]
The list comprehension syntax also extends to dictionary and set comprehensions. A set is created when the
generator expression is enclosed in curly braces. A dict is created when the generator expression contains
pairs of the form key:value:
>>> {i for i in range(3)}

set([0, 1, 2])
>>> {i:i**2 for i in range(3)}
{0: 0, 1: 1, 2: 4}
One gotcha should be mentioned: in old Pythons the index variable (i) would leak, and in versions >= 3 this is
fixed.
7.1.3 Generators
Generators
A generator is a function that produces a sequence of results instead of a single value.

David Beazley A Curious Course on Coroutines and Concurrency
A third way to create iterator objects is to call a generator function. A generator is a function containing the
keyword yield. It must be noted that the mere presence of this keyword completely changes the nature of the
function: this yield statement doesnt have to be invoked, or even reachable, but causes the function to be
marked as a generator. When a normal function is called, the instructions contained in the body start to be
executed. When a generator is called, the execution stops before the first instruction in the body. An invocation

of a generator function creates a generator object, adhering to the iterator protocol. As with normal function
invocations, concurrent and recursive invocations are allowed.
When next is called, the function is executed until the first yield. Each encountered yield statement gives a
value becomes the return value of next. After executing the yield statement, the execution of this function is
suspended.
>>> def f():

... yield 1
... yield 2
>>> f()
<generator object f at 0x...>
>>> gen = f()
>>> next(gen)
1
>>> next(gen)
2
>>> next(gen)
StopIteration
Lets go over the life of the single invocation of the generator function.
>>> def f():

... print("-- start --")
... yield 3
... print("-- middle --")
... yield 4
... print("-- finished --")
>>> gen = f()
>>> next(gen)
-- start --
3
>>> next(gen)
-- middle --
4
>>> next(gen)
-- finished --
...
StopIteration
Contrary to a normal function, where executing f() would immediately cause the first print to be executed,
gen is assigned without executing any statements in the function body. Only when gen.next() is invoked by
next, the statements up to the first yield are executed. The second next prints -- middle -- and execution
halts on the second yield. The third next prints -- finished -- and falls of the end of the function. Since
no yield was reached, an exception is raised.
What happens with the function after a yield, when the control passes to the caller? The state of each generator
is stored in the generator object. From the point of view of the generator function, is looks almost as if it was
running in a separate thread, but this is just an illusion: execution is strictly single-threaded, but the interpreter
keeps and restores the state in between the requests for the next value.
Why are generators useful? As noted in the parts about iterators, a generator function is just a different way to
create an iterator object. Everything that can be done with yield statements, could also be done with next
methods. Nevertheless, using a function and having the interpreter perform its magic to create an iterator
has advantages. A function can be much shorter than the definition of a class with the required next and
__iter__ methods. What is more important, it is easier for the author of the generator to understand the state
which is kept in local variables, as opposed to instance attributes, which have to be used to pass data between
consecutive invocations of next on an iterator object.
A broader question is why are iterators useful? When an iterator is used to power a loop, the loop becomes very

simple. The code to initialise the state, to decide if the loop is finished, and to find the next value is extracted
into a separate place. This highlights the body of the loop the interesting part. In addition, it is possible to
reuse the iterator code in other places.
7.1.4 Bidirectional communication
Each yield statement causes a value to be passed to the caller. This is the reason for the introduction of
generators by PEP 255 (implemented in Python 2.2). But communication in the reverse direction is also useful.
One obvious way would be some external state, either a global variable or a shared mutable object. Direct
communication is possible thanks to PEP 342 (implemented in 2.5). It is achieved by turning the previously
boring yield statement into an expression. When the generator resumes execution after a yield statement,
the caller can call a method on the generator object to either pass a value into the generator, which then is
returned by the yield statement, or a different method to inject an exception into the generator.
The first of the new methods is send(value), which is similar to next(), but passes value into the generator
to be used for the value of the yield expression. In fact, g.next() and g.send(None) are equivalent.
The second of the new methods is throw(type, value=None, traceback=None) which is equivalent to:
raise type, value, traceback
at the point of the yield statement.

Unlike raise (which immediately raises an exception from the current execution point), throw() first resumes
the generator, and only then raises the exception. The word throw was picked because it is suggestive of putting
the exception in another location, and is associated with exceptions in other languages.
What happens when an exception is raised inside the generator? It can be either raised explicitly or when
executing some statements or it can be injected at the point of a yield statement by means of the throw()
method. In either case, such an exception propagates in the standard manner: it can be intercepted by an
except or finally clause, or otherwise it causes the execution of the generator function to be aborted and
propagates in the caller.
For completeness sake, its worth mentioning that generator iterators also have a close() method, which can
be used to force a generator that would otherwise be able to provide more values to finish immediately. It
allows the generator __del__ method to destroy objects holding the state of generator.
Lets define a generator which just prints what is passed in through send and throw.
>>> import itertools

>>> def g():
... print('--start--')
... for i in itertools.count():
... print('--yielding %i --' % i)
... try:
... ans = yield i
... except GeneratorExit:
... print('--closing--')
... raise
... except Exception as e:
... print('--yield raised %r --' % e)
... else:
... print('--yield returned %s --' % ans)
>>> it = g()
>>> next(it)
--start--
--yielding 0--
0
>>> it.send(11)
--yield returned 11--
--yielding 1--

1
>>> it.throw(IndexError)
--yield raised IndexError()--
--yielding 2--
2
>>> it.close()
--closing--
next or __next__?
In Python 2.x, the iterator method to retrieve the next value is called next. It is invoked implicitly through
the global function next, which means that it should be called __next__. Just like the global function iter
calls __iter__. This inconsistency is corrected in Python 3.x, where it.next becomes it.__next__. For
other generator methods send and throw the situation is more complicated, because they are not
called implicitly by the interpreter. Nevertheless, theres a proposed syntax extension to allow continue
to take an argument which will be passed to send of the loops iterator. If this extension is accepted, its
likely that gen.send will become gen.__send__. The last of generator methods, close, is pretty obviously
named incorrectly, because it is already invoked implicitly.
7.1.5 Chaining generators
Note: This is a preview of PEP 380 (not yet implemented, but accepted for Python 3.3).
Lets say we are writing a generator and we want to yield a number of values generated by a second generator,
a subgenerator. If yielding of values is the only concern, this can be performed without much difficulty using
a loop such as
subgen = some_other_generator()
for v in subgen:
yield v
However, if the subgenerator is to interact properly with the caller in the case of calls to send(), throw()
and close(), things become considerably more difficult. The yield statement has to be guarded by a
try..except..finally structure similar to the one defined in the previous section to debug the generator func-
tion. Such code is provided in PEP 380#id13, here it suffices to say that new syntax to properly yield from a
subgenerator is being introduced in Python 3.3:
yield from some_other_generator()
This behaves like the explicit loop above, repeatedly yielding values from some_other_generator until it is
exhausted, but also forwards send, throw and close to the subgenerator.
7.2 Decorators
Summary
This amazing feature appeared in the language almost apologetically and with concern that it might not be
that useful.
Bruce Eckel An Introduction to Python Decorators
7.2. Decorators 269

Since functions and classes are objects, they can be passed around. Since they are mutable objects, they can
be modified. The act of altering a function or class object after it has been constructed but before is is bound
to its name is called decorating.
There are two things hiding behind the name decorator one is the function which does the work of deco-
rating, i.e. performs the real work, and the other one is the expression adhering to the decorator syntax, i.e. an
at-symbol and the name of the decorating function.
Function can be decorated by using the decorator syntax for functions:
@decorator #
def function(): #
pass
A function is defined in the standard way.

An expression starting with @ placed before the function definition is the decorator . The part after @
must be a simple expression, usually this is just the name of a function or class. This part is evaluated
first, and after the function defined below is ready, the decorator is called with the newly defined function
object as the single argument. The value returned by the decorator is attached to the original name of
the function.
Decorators can be applied to functions and to classes. For classes the semantics are identical the original
class definition is used as an argument to call the decorator and whatever is returned is assigned under the
original name.
Before the decorator syntax was implemented (PEP 318), it was possible to achieve the same effect by assign-
ing the function or class object to a temporary variable and then invoking the decorator explicitly and then
assigning the return value to the name of the function. This sounds like more typing, and it is, and also the
name of the decorated function doubling as a temporary variable must be used at least three times, which is
prone to errors. Nevertheless, the example above is equivalent to:
def function(): #
pass
function = decorator(function) #
Decorators can be stacked the order of application is bottom-to-top, or inside-out. The semantics are such
that the originally defined function is used as an argument for the first decorator, whatever is returned by
the first decorator is used as an argument for the second decorator, . . . , and whatever is returned by the last
decorator is attached under the name of the original function.
The decorator syntax was chosen for its readability. Since the decorator is specified before the header of the
function, it is obvious that its is not a part of the function body and its clear that it can only operate on the
whole function. Because the expression is prefixed with @ is stands out and is hard to miss (in your face,
according to the PEP :) ). When more than one decorator is applied, each one is placed on a separate line in an
easy to read way.
7.2.1 Replacing or tweaking the original object
Decorators can either return the same function or class object or they can return a completely different ob-
ject. In the first case, the decorator can exploit the fact that function and class objects are mutable and add
attributes, e.g. add a docstring to a class. A decorator might do something useful even without modifying
the object, for example register the decorated class in a global registry. In the second case, virtually anything is
possible: when something different is substituted for the original function or class, the new object can be com-
pletely different. Nevertheless, such behaviour is not the purpose of decorators: they are intended to tweak
the decorated object, not do something unpredictable. Therefore, when a function is decorated by replacing
it with a different function, the new function usually calls the original function, after doing some preparatory
work. Likewise, when a class is decorated by replacing if with a new class, the new class is usually derived
from the original class. When the purpose of the decorator is to do something every time, like to log every call
to a decorated function, only the second type of decorators can be used. On the other hand, if the first type is
sufficient, it is better to use it, because it is simpler.
7.2. Decorators 270

7.2.2 Decorators implemented as classes and as functions
The only requirement on decorators is that they can be called with a single argument. This means that deco-
rators can be implemented as normal functions, or as classes with a __call__ method, or in theory, even as
lambda functions.
Lets compare the function and class approaches. The decorator expression (the part after @) can be either just
a name, or a call. The bare-name approach is nice (less to type, looks cleaner, etc.), but is only possible when
no arguments are needed to customise the decorator. Decorators written as functions can be used in those
two cases:
>>> def simple_decorator(function):

... print("doing decoration")
... return function
>>> @simple_decorator
... def function():
... print("inside function")
doing decoration
>>> function()
inside function
>>> def decorator_with_arguments(arg):

... print("defining the decorator")
... def _decorator(function):
... # in this inner function, arg is available too
... print("doing decoration, %r " % arg)
... return function
... return _decorator
>>> @decorator_with_arguments("abc")
... def function():
defining the decorator
doing decoration, 'abc'
>>> function()
inside function
The two trivial decorators above fall into the category of decorators which return the original function. If they
were to return a new function, an extra level of nestedness would be required. In the worst case, three levels of
nested functions.
>>> def replacing_decorator_with_args(arg):

... # in this inner function, arg is available too
... def _wrapper(*args, **kwargs):
... print("inside wrapper, %r %r " % (args, kwargs))
... return function(*args, **kwargs)
... return _wrapper
>>> @replacing_decorator_with_args("abc")
... def function(*args, **kwargs):
... print("inside function, %r %r " % (args, kwargs))
... return 14
>>> function(11, 12)
inside wrapper, (11, 12) {}
inside function, (11, 12) {}
14
The _wrapper function is defined to accept all positional and keyword arguments. In general we cannot know
what arguments the decorated function is supposed to accept, so the wrapper function just passes everything
7.2. Decorators 271

to the wrapped function. One unfortunate consequence is that the apparent argument list is misleading.
Compared to decorators defined as functions, complex decorators defined as classes are simpler. When an
object is created, the __init__ method is only allowed to return None, and the type of the created object
cannot be changed. This means that when a decorator is defined as a class, it doesnt make much sense to use
the argument-less form: the final decorated object would just be an instance of the decorating class, returned
by the constructor call, which is not very useful. Therefore its enough to discuss class-based decorators where
arguments are given in the decorator expression and the decorator __init__ method is used for decorator
construction.
>>> class decorator_class(object):

... def __init__(self, arg):
... # this method is called in the decorator expression
... print("in decorator init, %s " % arg)
... self.arg = arg
... def __call__(self, function):
... # this method is called to do the job
... print("in decorator call, %s " % self.arg)
... return function
>>> deco_instance = decorator_class('foo')
in decorator init, foo
>>> @deco_instance
... print("in function, %s %s " % (args, kwargs))
in decorator call, foo
>>> function()
in function, () {}
Contrary to normal rules (PEP 8) decorators written as classes behave more like functions and therefore their
name often starts with a lowercase letter.
In reality, it doesnt make much sense to create a new class just to have a decorator which returns the original
function. Objects are supposed to hold state, and such decorators are more useful when the decorator returns
a new object.
>>> class replacing_decorator_class(object):

... def __init__(self, arg):
... # this method is called in the decorator expression
... print("in decorator init, %s " % arg)
... self.arg = arg
... def __call__(self, function):
... # this method is called to do the job
... print("in decorator call, %s " % self.arg)
... self.function = function
... return self._wrapper
... def _wrapper(self, *args, **kwargs):
... print("in the wrapper, %s %s " % (args, kwargs))
... return self.function(*args, **kwargs)
>>> deco_instance = replacing_decorator_class('foo')
in decorator init, foo
>>> @deco_instance
... print("in function, %s %s " % (args, kwargs))
in decorator call, foo
>>> function(11, 12)
in the wrapper, (11, 12) {}
in function, (11, 12) {}
A decorator like this can do pretty much anything, since it can modify the original function object and mangle
the arguments, call the original function or not, and afterwards mangle the return value.
7.2. Decorators 272

7.2.3 Copying the docstring and other attributes of the original function
When a new function is returned by the decorator to replace the original function, an unfortunate conse-
quence is that the original function name, the original docstring, the original argument list are lost. Those
attributes of the original function can partially be transplanted to the new function by setting __doc__ (the
docstring), __module__ and __name__ (the full name of the function), and __annotations__ (extra informa-
tion about arguments and the return value of the function available in Python 3). This can be done automati-
cally by using functools.update_wrapper.
functools.update_wrapper(wrapper, wrapped)
Update a wrapper function to look like the wrapped function.

>>> import functools
>>> def replacing_decorator_with_args(arg):
... def _wrapper(*args, **kwargs):
... print("inside wrapper, %r %r " % (args, kwargs))
... return function(*args, **kwargs)
... return functools.update_wrapper(_wrapper, function)
>>> @replacing_decorator_with_args("abc")
... def function():
... "extensive documentation"
... return 14
>>> function
<function function at 0x...>
>>> print(function.__doc__)
extensive documentation
One important thing is missing from the list of attributes which can be copied to the replacement func-
tion: the argument list. The default values for arguments can be modified through the __defaults__,
__kwdefaults__ attributes, but unfortunately the argument list itself cannot be set as an attribute. This
means that help(function) will display a useless argument list which will be confusing for the user of the
function. An effective but ugly way around this problem is to create the wrapper dynamically, using eval. This
can be automated by using the external decorator module. It provides support for the decorator decorator,
which takes a wrapper and turns it into a decorator which preserves the function signature.
To sum things up, decorators should always use functools.update_wrapper or some other means of copy-
ing function attributes.
7.2.4 Examples in the standard library
First, it should be mentioned that theres a number of useful decorators available in the standard library. There
are three decorators which really form a part of the language:
classmethod causes a method to become a class method, which means that it can be invoked without
creating an instance of the class. When a normal method is invoked, the interpreter inserts the instance
object as the first positional parameter, self. When a class method is invoked, the class itself is given as
the first parameter, often called cls.
Class methods are still accessible through the class namespace, so they dont pollute the modules
namespace. Class methods can be used to provide alternative constructors:
7.2. Decorators 273

class Array(object):
def __init__(self, data):
self.data = data
@classmethod
def fromfile(cls, file):
data = numpy.load(file)
return cls(data)
This is cleaner then using a multitude of flags to __init__.

staticmethod is applied to methods to make them static, i.e. basically a normal function, but acces-
sible through the class namespace. This can be useful when the function is only needed inside this class
(its name would then be prefixed with _), or when we want the user to think of the method as connected
to the class, despite an implementation which doesnt require this.
property is the pythonic answer to the problem of getters and setters. A method decorated with
property becomes a getter which is automatically called on attribute access.
>>> class A(object):

... @property
... def a(self):
... "an important attribute"
... return "a value"
>>> A.a
<property object at 0x...>
>>> A().a
'a value'
In this example, A.a is an read-only attribute. It is also documented: help(A) includes the docstring for
attribute a taken from the getter method. Defining a as a property allows it to be a calculated on the fly,
and has the side effect of making it read-only, because no setter is defined.
To have a setter and a getter, two methods are required, obviously. Since Python 2.6 the following syntax
is preferred:
class Rectangle(object):
def __init__(self, edge):
self.edge = edge
@property
def area(self):
"""Computed area.
Setting this updates the edge length to the proper value.

"""
return self.edge**2
@area.setter
def area(self, area):
self.edge = area ** 0.5
The way that this works, is that the property decorator replaces the getter method with a property
object. This object in turn has three methods, getter, setter, and deleter, which can be used as
decorators. Their job is to set the getter, setter and deleter of the property object (stored as attributes
fget, fset, and fdel). The getter can be set like in the example above, when creating the object. When
defining the setter, we already have the property object under area, and we add the setter to it by using
the setter method. All this happens when we are creating the class.
Afterwards, when an instance of the class has been created, the property object is special. When the
interpreter executes attribute access, assignment, or deletion, the job is delegated to the methods of the
property object.
7.2. Decorators 274

To make everything crystal clear, lets define a debug example:
>>> class D(object):

... @property
... def a(self):
... print("getting 1")
... return 1
... @a.setter
... def a(self, value):
... print("setting %r " % value)
... @a.deleter
... def a(self):
... print("deleting")
>>> D.a
<property object at 0x...>
>>> D.a.fget
<function ...>
>>> D.a.fset
<function ...>
>>> D.a.fdel
<function ...>
>>> d = D() # ... varies, this is not the same à` function
>>> d.a
getting 1
1
>>> d.a = 2
setting 2
>>> del d.a
deleting
>>> d.a
getting 1
1
Properties are a bit of a stretch for the decorator syntax. One of the premises of the decorator syntax
that the name is not duplicated is violated, but nothing better has been invented so far. It is just good
style to use the same name for the getter, setter, and deleter methods.
Some newer examples include:
functools.lru_cache memoizes an arbitrary function maintaining a limited cache of argu-
ments:answer pairs (Python 3.2)
functools.total_ordering is a class decorator which fills in missing ordering methods (__lt__,
__gt__, __le__, . . . ) based on a single available one (Python 2.7).
7.2.5 Deprecation of functions
Lets say we want to print a deprecation warning on stderr on the first invocation of a function we dont like
anymore. If we dont want to modify the function, we can use a decorator:
class deprecated(object):
"""Print a deprecation warning once on first use of the function.
>>> @deprecated() # doctest: +SKIP

... def f():
... pass
>>> f() # doctest: +SKIP
f is deprecated
"""
def __call__(self, func):
self.func = func
self.count = 0
7.2. Decorators 275

return self._wrapper
def _wrapper(self, *args, **kwargs):
self.count += 1
if self.count == 1:
print self.func.__name__, 'is deprecated'
return self.func(*args, **kwargs)
It can also be implemented as a function:
def deprecated(func):
"""Print a deprecation warning once on first use of the function.
>>> @deprecated # doctest: +SKIP

... def f():
... pass
>>> f() # doctest: +SKIP
f is deprecated
"""
count = [0]
def wrapper(*args, **kwargs):
count[0] += 1
if count[0] == 1:
print func.__name__, 'is deprecated'
return func(*args, **kwargs)
return wrapper
7.2.6 A while-loop removing decorator
Lets say we have function which returns a lists of things, and this list created by running a loop. If we dont
know how many objects will be needed, the standard way to do this is something like:
def find_answers():
answers = []
while True:
ans = look_for_next_answer()
if ans is None:
break
answers.append(ans)
return answers
This is fine, as long as the body of the loop is fairly compact. Once it becomes more complicated, as often
happens in real code, this becomes pretty unreadable. We could simplify this by using yield statements, but
then the user would have to explicitly call list(find_answers()).
We can define a decorator which constructs the list for us:
def vectorized(generator_func):
def wrapper(*args, **kwargs):
return list(generator_func(*args, **kwargs))
return functools.update_wrapper(wrapper, generator_func)
Our function then becomes:
@vectorized
def find_answers():
while True:
ans = look_for_next_answer()
if ans is None:
break
yield ans
7.2. Decorators 276

7.2.7 A plugin registration system
This is a class decorator which doesnt modify the class, but just puts it in a global registry. It falls into the
category of decorators returning the original object:
class WordProcessor(object):
PLUGINS = []
def process(self, text):
for plugin in self.PLUGINS:
text = plugin().cleanup(text)
return text
@classmethod
def plugin(cls, plugin):
cls.PLUGINS.append(plugin)
@WordProcessor.plugin
class CleanMdashesExtension(object):
def cleanup(self, text):
return text.replace('—', u'\N{em dash}')
Here we use a decorator to decentralise the registration of plugins. We call our decorator with a noun, instead
of a verb, because we use it to declare that our class is a plugin for WordProcessor. Method plugin simply
appends the class to the list of plugins.
A word about the plugin itself: it replaces HTML entity for em-dash with a real Unicode em-dash character.
It exploits the unicode literal notation to insert a character by using its name in the unicode database (EM
DASH). If the Unicode character was inserted directly, it would be impossible to distinguish it from an en-
dash in the source of a program.
See also:
More examples and reading
PEP 318 (function and method decorator syntax)
PEP 3129 (class decorator syntax)
http://wiki.python.org/moin/PythonDecoratorLibrary
https://docs.python.org/dev/library/functools.html
http://pypi.python.org/pypi/decorator
Bruce Eckel
Decorators I: Introduction to Python Decorators
Python Decorators II: Decorator Arguments
Python Decorators III: A Decorator-Based Build System
7.3 Context managers
A context manager is an object with __enter__ and __exit__ methods which can be used in the with state-
ment:
with manager as var:

do_something(var)
is in the simplest case equivalent to
7.3. Context managers 277

var = manager.__enter__()
try:
do_something(var)
finally:
manager.__exit__()
In other words, the context manager protocol defined in PEP 343 permits the extraction of the boring part of a
try..except..finally structure into a separate class leaving only the interesting do_something block.
1. The __enter__ method is called first. It can return a value which will be assigned to var. The as-part is
optional: if it isnt present, the value returned by __enter__ is simply ignored.
2. The block of code underneath with is executed. Just like with try clauses, it can either execute success-
fully to the end, or it can break, continue or return, or it can throw an exception. Either way, after the
block is finished, the __exit__ method is called. If an exception was thrown, the information about the
exception is passed to __exit__, which is described below in the next subsection. In the normal case,
exceptions can be ignored, just like in a finally clause, and will be rethrown after __exit__ is finished.
Lets say we want to make sure that a file is closed immediately after we are done writing to it:
>>> class closing(object):

... def __init__(self, obj):
... self.obj = obj
... def __enter__(self):
... return self.obj
... def __exit__(self, *args):
... self.obj.close()
>>> with closing(open('/tmp/file', 'w')) as f:
... f.write('the contents\n')
Here we have made sure that the f.close() is called when the with block is exited. Since closing files is such
a common operation, the support for this is already present in the file class. It has an __exit__ method
which calls close and can be used as a context manager itself:
>>> with open('/tmp/file', 'a') as f:

... f.write('more contents\n')
The common use for try..finally is releasing resources. Various different cases are implemented similarly:
in the __enter__ phase the resource is acquired, in the __exit__ phase it is released, and the exception, if
thrown, is propagated. As with files, theres often a natural operation to perform after the object has been used
and it is most convenient to have the support built in. With each release, Python provides support in more
places:
all file-like objects:
file automatically closed
fileinput, tempfile (py >= 3.2)
bz2.BZ2File, gzip.GzipFile, tarfile.TarFile, zipfile.ZipFile
ftplib, nntplib close connection (py >= 3.2 or 3.3)
locks
multiprocessing.RLock lock and unlock
multiprocessing.Semaphore
memoryview automatically release (py >= 3.2 and 2.7)
decimal.localcontext modify precision of computations temporarily
_winreg.PyHKEY open and close hive key
warnings.catch_warnings kill warnings temporarily

contextlib.closing the same as the example above, call close

parallel programming
concurrent.futures.ThreadPoolExecutor invoke in parallel then kill thread pool (py >= 3.2)
concurrent.futures.ProcessPoolExecutor invoke in parallel then kill process pool (py >=
3.2)
nogil solve the GIL problem temporarily (cython only :( )
7.3.1 Catching exceptions
When an exception is thrown in the with-block, it is passed as arguments to __exit__. Three arguments are
used, the same as returned by sys.exc_info(): type, value, traceback. When no exception is thrown, None is
used for all three arguments. The context manager can swallow the exception by returning a true value from
__exit__. Exceptions can be easily ignored, because if __exit__ doesnt use return and just falls of the end,
None is returned, a false value, and therefore the exception is rethrown after __exit__ is finished.
The ability to catch exceptions opens interesting possibilities. A classic example comes from unit-tests we
want to make sure that some code throws the right kind of exception:
class assert_raises(object):
# based on pytest and unittest.TestCase
def __init__(self, type):
self.type = type
def __enter__(self):
pass
def __exit__(self, type, value, traceback):
if type is None:
raise AssertionError('exception expected')
if issubclass(type, self.type):
return True # swallow the expected exception
raise AssertionError('wrong exception type')
with assert_raises(KeyError):
{}['foo']
7.3.2 Using generators to define context managers
When discussing generators, it was said that we prefer generators to iterators implemented as classes because
they are shorter, sweeter, and the state is stored as local, not instance, variables. On the other hand, as de-
scribed in Bidirectional communication, the flow of data between the generator and its caller can be bidirec-
tional. This includes exceptions, which can be thrown into the generator. We would like to implement context
managers as special generator functions. In fact, the generator protocol was designed to support this use case.
@contextlib.contextmanager
def some_generator(<arguments>):
<setup>
try:
yield <value>
finally:
<cleanup>
The contextlib.contextmanager helper takes a generator and turns it into a context manager. The gener-
ator has to obey some rules which are enforced by the wrapper function most importantly it must yield
exactly once. The part before the yield is executed from __enter__, the block of code protected by the con-
text manager is executed when the generator is suspended in yield, and the rest is executed in __exit__. If
an exception is thrown, the interpreter hands it to the wrapper through __exit__ arguments, and the wrap-
per function then throws it at the point of the yield statement. Through the use of generators, the context
manager is shorter and simpler.

Lets rewrite the closing example as a generator:
def closing(obj):
try:
yield obj
finally:
obj.close()
Lets rewrite the assert_raises example as a generator:
def assert_raises(type):
try:
yield
except type:
return
except Exception as value:
raise AssertionError('wrong exception type')
else:
raise AssertionError('exception expected')
Here we use a decorator to turn generator functions into context managers!

CHAPTER 8
Advanced NumPy
Author: Pauli Virtanen

NumPy is at the base of Pythons scientific stack of tools. Its purpose to implement efficient operations on
many items in a block of memory. Understanding how it works in detail helps in making efficient use of its
flexibility, taking useful shortcuts.
This section covers:
Anatomy of NumPy arrays, and its consequences. Tips and tricks.
Universal functions: what, why, and what to do if you want a new one.
Integration with other tools: NumPy offers several ways to wrap any data in an ndarray, without unnec-
essary copies.
Recently added features, and whats in them: PEP 3118 buffers, generalized ufuncs, . . .
Prerequisites
NumPy
Cython
Pillow (Python imaging library, used in a couple of examples)
Chapter contents
Life of ndarray
Its. . .
281
Block of memory
Data types
Indexing scheme: strides
Findings in dissection
Universal functions
What they are?
Exercise: building an ufunc from scratch
Solution: building an ufunc from scratch
Generalized ufuncs
Interoperability features
Sharing multidimensional, typed data
The old buffer protocol
The old buffer protocol
Array interface protocol
Array siblings: chararray, maskedarray, matrix
chararray: vectorized string operations
masked_array missing data
recarray: purely convenience
matrix: convenience?
Summary
Contributing to NumPy/Scipy
Why
Reporting bugs
Contributing to documentation
Contributing features
How to help, in general
Tip: In this section, numpy will be imported as follows:
8.1 Life of ndarray
8.1.1 Its. . .
ndarray =
block of memory + indexing scheme + data type descriptor
raw data
how to locate an element
8.1. Life of ndarray 282

how to interpret an element
typedef struct PyArrayObject {

PyObject_HEAD
/* Block of memory */
char *data;
/* Data type descriptor */

PyArray_Descr *descr;
/* Indexing scheme */
int nd;
npy_intp *dimensions;
npy_intp *strides;
/* Other stuff */
PyObject *base;
int flags;
PyObject *weakreflist;
} PyArrayObject;
8.1.2 Block of memory
>>> x = np.array([1, 2, 3], dtype=np.int32)

>>> x.data
<... at ...>
>>> str(x.data)
'\x01\x00\x00\x00\x02\x00\x00\x00\x03\x00\x00\x00'
Memory address of the data:
>>> x.__array_interface__['data'][0]
64803824
The whole __array_interface__:
>>> x.__array_interface__
{'data': (35828928, False),
'descr': [('', '<i4')],
'shape': (4,),
'strides': None,
'typestr': '<i4',
'version': 3}
Reminder: two ndarrays may share the same memory:

>>> x = np.array([1, 2, 3, 4])

>>> y = x[:-1]
>>> x[0] = 9
>>> y
array([9, 2, 3])
Memory does not need to be owned by an ndarray:
>>> x = b'1234' # The 'b' is for "bytes", necessary in Python 3
x is a string (in Python 3 a bytes), we can represent its data as an array of ints:
>>> y = np.frombuffer(x, dtype=np.int8)

>>> y.data
<... at ...>
>>> y.base is x
True
>>> y.flags
C_CONTIGUOUS : True
F_CONTIGUOUS : True
OWNDATA : False
WRITEABLE : False
ALIGNED : True
UPDATEIFCOPY : False
The owndata and writeable flags indicate status of the memory block.
8.1.3 Data types
The descriptor
dtype describes a single item in the array:
type scalar type of the data, one of:

int8, int16, float64, et al. (fixed size)
str, unicode, void (flexible size)
itemsize size of the data block
byte- byte order: big-endian > / little-endian < / not applicable |
order
fields sub-dtypes, if its a structured data type
shape shape of the array, if its a sub-array
>>> np.dtype(int).type
<type 'numpy.int64'>
>>> np.dtype(int).itemsize
8
>>> np.dtype(int).byteorder
'='
Example: reading .wav files
The .wav file header:

chunk_id "RIFF"
chunk_size 4-byte unsigned little-endian integer
format "WAVE"
fmt_id "fmt "
fmt_size 4-byte unsigned little-endian integer
audio_fmt 2-byte unsigned little-endian integer
num_channels 2-byte unsigned little-endian integer
sample_rate 4-byte unsigned little-endian integer
byte_rate 4-byte unsigned little-endian integer
block_align 2-byte unsigned little-endian integer
bits_per_sample 2-byte unsigned little-endian integer
data_id "data"
data_size 4-byte unsigned little-endian integer
44-byte block of raw data (in the beginning of the file)

. . . followed by data_size bytes of actual sound data.
The .wav file header as a NumPy structured data type:
>>> wav_header_dtype = np.dtype([

... ("chunk_id", (bytes, 4)), # flexible-sized scalar type, item size 4
... ("chunk_size", "<u4"), # little-endian unsigned 32-bit integer
... ("format", "S4"), # 4-byte string
... ("fmt_id", "S4"),
... ("fmt_size", "<u4"),
... ("audio_fmt", "<u2"), #
... ("num_channels", "<u2"), # .. more of the same ...
... ("sample_rate", "<u4"), #
... ("byte_rate", "<u4"),
... ("block_align", "<u2"),
... ("bits_per_sample", "<u2"),
... ("data_id", ("S1", (2, 2))), # sub-array, just for fun!
... ("data_size", "u4"),
... #
... # the sound data itself cannot be represented here:
... # it does not have a fixed size
... ])
See also:
wavreader.py
>>> wav_header_dtype['format']
dtype('S4')
>>> wav_header_dtype.fields
dict_proxy({'block_align': (dtype('uint16'), 32), 'format': (dtype('S4'), 8), 'data_id':
,(dtype(('S1', (2, 2))), 36), 'fmt_id': (dtype('S4'), 12), 'byte_rate': (dtype('uint32'), 28),
, 'chunk_id': (dtype('S4'), 0), 'num_channels': (dtype('uint16'), 22), 'sample_rate': (dtype(
,'uint32'), 24), 'bits_per_sample': (dtype('uint16'), 34), 'chunk_size': (dtype('uint32'), 4),
, 'fmt_size': (dtype('uint32'), 16), 'data_size': (dtype('uint32'), 40), 'audio_fmt': (dtype(
,'uint16'), 20)})
>>> wav_header_dtype.fields['format']
(dtype('S4'), 8)
The first element is the sub-dtype in the structured data, corresponding to the name format
The second one is its offset (in bytes) from the beginning of the item

Exercise
Mini-exercise, make a sparse dtype by using offsets, and only some of the fields:
>>> wav_header_dtype = np.dtype(dict(
... names=['format', 'sample_rate', 'data_id'],
... offsets=[offset_1, offset_2, offset_3], # counted from start of structure in bytes
... formats=list of dtypes for each of the fields,
... ))
and use that to read the sample rate, and data_id (as sub-array).
>>> f = open('data/test.wav', 'r')

>>> wav_header = np.fromfile(f, dtype=wav_header_dtype, count=1)
>>> f.close()
>>> print(wav_header)
[ ('RIFF', 17402L, 'WAVE', 'fmt ', 16L, 1, 1, 16000L, 32000L, 2, 16, [['d', 'a'], ['t', 'a']],
,17366L)]
>>> wav_header['sample_rate']
array([16000], dtype=uint32)
Lets try accessing the sub-array:
>>> wav_header['data_id']
array([[['d', 'a'],
['t', 'a']]],
dtype='|S1')
>>> wav_header.shape
(1,)
>>> wav_header['data_id'].shape
(1, 2, 2)
When accessing sub-arrays, the dimensions get added to the end!
Note: There are existing modules such as wavfile, audiolab, etc. for loading sound data. . .
Casting and re-interpretation/views
casting
on assignment
on array construction
on arithmetic
etc.
and manually: .astype(dtype)
data re-interpretation
manually: .view(dtype)
Casting
Casting in arithmetic, in nutshell:

only type (not value!) of operands matters
largest safe type able to represent both is picked

scalars can lose to arrays in some situations

Casting in general copies data:
>>> x = np.array([1, 2, 3, 4], dtype=np.float)

>>> x
array([ 1., 2., 3., 4.])
>>> y = x.astype(np.int8)
>>> y
array([1, 2, 3, 4], dtype=int8)
>>> y + 1
>>> y + 256
array([257, 258, 259, 260], dtype=int16)
>>> y + 256.0
array([ 257., 258., 259., 260.])
>>> y + np.array([256], dtype=np.int32)
array([257, 258, 259, 260], dtype=int32)
Casting on setitem: dtype of the array is not changed on item assignment:
>>> y[:] = y + 1.5

>>> y
Note: Exact rules: see numpy documentation
Re-interpretation / viewing
Data block in memory (4 bytes)
0x01 || 0x02 || 0x03 || 0x04
4 of uint8, OR,
4 of int8, OR,
2 of int16, OR,
1 of int32, OR,
1 of float32, OR,
...
How to switch from one to another?
1. Switch the dtype:
>>> x = np.array([1, 2, 3, 4], dtype=np.uint8)

>>> x.dtype = "<i2"
>>> x
array([ 513, 1027], dtype=int16)
>>> 0x0201, 0x0403
(513, 1027)
0x01 0x02 || 0x03 0x04
Note: little-endian: least significant byte is on the left in memory

2. Create a new view:
>>> y = x.view("<i4")
>>> y
array([67305985], dtype=int32)
>>> 0x04030201
67305985
0x01 0x02 0x03 0x04
Note:
.view() makes views, does not copy (or alter) the memory block
only changes the dtype (and adjusts array shape):
>>> x[1] = 5
>>> y
array([328193], dtype=int32)
>>> y.base is x
True
Mini-exercise: data re-interpretation
See also:
view-colors.py
You have RGBA data in an array:
>>> x = np.zeros((10, 10, 4), dtype=np.int8)

>>> x[:, :, 0] = 1
>>> x[:, :, 1] = 2
>>> x[:, :, 2] = 3
>>> x[:, :, 3] = 4
where the last three dimensions are the R, B, and G, and alpha channels.
How to make a (10, 10) structured array with field names r, g, b, a without copying data?
>>> y = ...
>>> assert (y['r'] == 1).all()

>>> assert (y['g'] == 2).all()
>>> assert (y['b'] == 3).all()
>>> assert (y['a'] == 4).all()
Solution
>>> y = x.view([('r', 'i1'),

... ('g', 'i1'),
... ('b', 'i1'),
... ('a', 'i1')]
... )[:, :, 0]
Warning: Another array taking exactly 4 bytes of memory:

>>> y = np.array([[1, 3], [2, 4]], dtype=np.uint8).transpose()

>>> x = y.copy()
>>> x
array([[1, 2],
[3, 4]], dtype=uint8)
>>> y
array([[1, 2],
[3, 4]], dtype=uint8)
>>> x.view(np.int16)
array([[ 513],
[1027]], dtype=int16)
>>> 0x0201, 0x0403
(513, 1027)
>>> y.view(np.int16)
array([[ 769, 1026]], dtype=int16)
What happened?
. . . we need to look into what x[0,1] actually means
>>> 0x0301, 0x0402
(769, 1026)
8.1.4 Indexing scheme: strides
Main point
The question:
>>> x = np.array([[1, 2, 3],

... [4, 5, 6],
... [7, 8, 9]], dtype=np.int8)
>>> str(x.data)
'\x01\x02\x03\x04\x05\x06\x07\x08\t'
At which byte in ``x.data`` does the item ``x[1, 2]`` begin?
The answer (in NumPy)

strides: the number of bytes to jump to find the next element
1 stride per dimension
>>> x.strides
(3, 1)
>>> byte_offset = 3*1 + 1*2 # to find x[1, 2]
>>> x.flat[byte_offset]
6
>>> x[1, 2]
6
- simple, **flexible**
C and Fortran order
>>> x = np.array([[1, 2, 3],

... [4, 5, 6]], dtype=np.int16, order='C')
>>> x.strides
(6, 2)

>>> str(x.data)
'\x01\x00\x02\x00\x03\x00\x04\x00\x05\x00\x06\x00'
Need to jump 6 bytes to find the next row

Need to jump 2 bytes to find the next column
>>> y = np.array(x, order='F')

>>> y.strides
(2, 4)
>>> str(y.data)
'\x01\x00\x04\x00\x02\x00\x05\x00\x03\x00\x06\x00'
Need to jump 2 bytes to find the next row

Need to jump 4 bytes to find the next column
Similarly to higher dimensions:
C: last dimensions vary fastest (= smaller strides)
F: first dimensions vary fastest
shape = (d 1 , d 2 , ..., d n )
strides = (s 1 , s 2 , ..., s n )
s Cj = d j +1 d j +2 ...d n itemsize
s Fj = d 1 d 2 ...d j 1 itemsize
Note: Now we can understand the behavior of .view():
>>> y = np.array([[1, 3], [2, 4]], dtype=np.uint8).transpose()

>>> x = y.copy()
Transposition does not affect the memory layout of the data, only strides
>>> x.strides
(2, 1)
>>> y.strides
(1, 2)
>>> str(x.data)
'\x01\x02\x03\x04'
>>> str(y.data)
'\x01\x03\x02\x04'
the results are different when interpreted as 2 of int16

.copy() creates new arrays in the C order (by default)
Slicing with integers
Everything can be represented by changing only shape, strides, and possibly adjusting the data
pointer!
Never makes copies of the data
>>> x = np.array([1, 2, 3, 4, 5, 6], dtype=np.int32)

>>> y = x[::-1]
>>> y
array([6, 5, 4, 3, 2, 1], dtype=int32)

>>> y.strides
(-4,)
>>> y = x[2:]
>>> y.__array_interface__['data'][0] - x.__array_interface__['data'][0]
8
>>> x = np.zeros((10, 10, 10), dtype=np.float)

>>> x.strides
(800, 80, 8)
>>> x[::2,::3,::4].strides
(1600, 240, 32)
Similarly, transposes never make copies (it just swaps strides):
>>> x = np.zeros((10, 10, 10), dtype=np.float)

>>> x.strides
(800, 80, 8)
>>> x.T.strides
(8, 80, 800)
But: not all reshaping operations can be represented by playing with strides:
>>> a = np.arange(6, dtype=np.int8).reshape(3, 2)

>>> b = a.T
>>> b.strides
(1, 2)
So far, so good. However:
>>> str(a.data)
'\x00\x01\x02\x03\x04\x05'
>>> b
array([[0, 2, 4],
[1, 3, 5]], dtype=int8)
>>> c = b.reshape(3*2)
>>> c
array([0, 2, 4, 1, 3, 5], dtype=int8)
Here, there is no way to represent the array c given one stride and the block of memory for a. Therefore, the
reshape operation needs to make a copy here.
Example: fake dimensions with strides
Stride manipulation
>>> from numpy.lib.stride_tricks import as_strided

>>> help(as_strided)
as_strided(x, shape=None, strides=None)
Make an ndarray from the given array with the given shape and strides
Warning: as_strided does not check that you stay inside the memory block bounds. . .
>>> x = np.array([1, 2, 3, 4], dtype=np.int16)

>>> as_strided(x, strides=(2*2, ), shape=(2, ))
array([1, 3], dtype=int16)
>>> x[::2]

See also:
stride-fakedims.py
Exercise
array([1, 2, 3, 4], dtype=np.int8)
-> array([[1, 2, 3, 4],

[1, 2, 3, 4],
[1, 2, 3, 4]], dtype=np.int8)
using only as_strided.:
Hint: byte_offset = stride[0]*index[0] + stride[1]*index[1] + ...
Spoiler
Stride can also be 0:

>>> y = as_strided(x, strides=(0, 1), shape=(3, 4))
>>> y
array([[1, 2, 3, 4],
[1, 2, 3, 4],
[1, 2, 3, 4]], dtype=int8)
>>> y.base.base is x
True
Broadcasting
Doing something useful with it: outer product of [1, 2, 3, 4] and [5, 6, 7]

>>> x2 = as_strided(x, strides=(0, 1*2), shape=(3, 4))
>>> x2
array([[1, 2, 3, 4],
[1, 2, 3, 4],
[1, 2, 3, 4]], dtype=int16)
>>> y = np.array([5, 6, 7], dtype=np.int16)

>>> y2 = as_strided(y, strides=(1*2, 0), shape=(3, 4))
>>> y2
array([[5, 5, 5, 5],
[6, 6, 6, 6],
[7, 7, 7, 7]], dtype=int16)
>>> x2 * y2
array([[ 5, 10, 15, 20],
[ 6, 12, 18, 24],
[ 7, 14, 21, 28]], dtype=int16)
. . . seems somehow familiar . . .

>>> y = np.array([5, 6, 7], dtype=np.int16)
>>> x[np.newaxis,:] * y[:,np.newaxis]
array([[ 5, 10, 15, 20],
[ 6, 12, 18, 24],
[ 7, 14, 21, 28]], dtype=int16)

Internally, array broadcasting is indeed implemented using 0-strides.
More tricks: diagonals
See also:
stride-diagonals.py
Challenge
Pick diagonal entries of the matrix: (assume C memory order):
>>> x = np.array([[1, 2, 3],

... [4, 5, 6],
... [7, 8, 9]], dtype=np.int32)
>>> x_diag = as_strided(x, shape=(3,), strides=(???,))
Pick the first super-diagonal entries [2, 6].

And the sub-diagonals?
(Hint to the last two: slicing first moves the point where striding starts from.)
Solution
Pick diagonals:
>>> x_diag = as_strided(x, shape=(3, ), strides=((3+1)*x.itemsize, ))

>>> x_diag
array([1, 5, 9], dtype=int32)
Slice first, to adjust the data pointer:
>>> as_strided(x[0, 1:], shape=(2, ), strides=((3+1)*x.itemsize, ))

>>> as_strided(x[1:, 0], shape=(2, ), strides=((3+1)*x.itemsize, ))

Note: Using np.diag
>>> y = np.diag(x, k=1)

>>> y
However,
>>> y.flags.owndata
False
Note This behavior has changed: before numpy 1.9, np.diag would make a copy.
See also:
stride-diagonals.py
Challenge
Compute the tensor trace:

>>> x = np.arange(5*5*5*5).reshape(5, 5, 5, 5)
>>> s = 0
... for j in range(5):
... s += x[j, i, j, i]
by striding, and using sum() on the result.
>>> y = as_strided(x, shape=(5, 5), strides=(TODO, TODO))

>>> s2 = ...
>>> assert s == s2
Solution
>>> y = as_strided(x, shape=(5, 5), strides=((5*5*5 + 5)*x.itemsize,

... (5*5 + 1)*x.itemsize))
>>> s2 = y.sum()
CPU cache effects
Memory layout can affect performance:
In [1]: x = np.zeros((20000,))
In [2]: y = np.zeros((20000*67,))[::67]
In [3]: x.shape, y.shape

((20000,), (20000,))
In [4]: %timeit x.sum()

In [5]: %timeit y.sum()

In [6]: x.strides, y.strides

((8,), (536,))
Smaller strides are faster?
CPU pulls data from main memory to its cache in blocks

If many array items consecutively operated on fit in a single block (small stride):
fewer transfers needed
faster
See also:
numexpr is designed to mitigate cache effects in array computing.

Example: inplace operations (caveat emptor)
Sometimes,
>>> a -= b
is not the same as
>>> a -= b.copy()
>>> x = np.array([[1, 2], [3, 4]])

>>> x -= x.transpose()
>>> x
array([[ 0, -1],
[ 4, 0]])
>>> y = np.array([[1, 2], [3, 4]])

>>> y -= y.T.copy()
>>> y
array([[ 0, -1],
[ 1, 0]])
x and x.transpose() share data

x -= x.transpose() modifies the data element-by-element. . .
because x and x.transpose() have different striding, modified data re-appears on the RHS
8.1.5 Findings in dissection
memory block: may be shared, .base, .data

data type descriptor: structured data, sub-arrays, byte order, casting, viewing, .astype(), .view()
strided indexing: strides, C/F-order, slicing w/ integers, as_strided, broadcasting, stride tricks, diag,
CPU cache coherence
8.2 Universal functions
8.2.1 What they are?
Ufunc performs and elementwise operation on all elements of an array.

Examples:
8.2. Universal functions 295

np.add, np.subtract, scipy.special.*, ...
Automatically support: broadcasting, casting, . . .

The author of an ufunc only has to supply the elementwise operation, NumPy takes care of the rest.
The elementwise operation needs to be implemented in C (or, e.g., Cython)
Parts of an Ufunc
1. Provided by user
void ufunc_loop(void **args, int *dimensions, int *steps, void *data)

{
/*
* int8 output = elementwise_function(int8 input_1, int8 input_2)
*
* This function must compute the ufunc for many values at once,
* in the way shown below.
*/
char *input_1 = (char*)args[0];
char *output = (char*)args[2];
int i;
for (i = 0; i < dimensions[0]; ++i) {

*output = elementwise_function(*input_1, *input_2);
input_1 += steps[0];
output += steps[2];
}
}
2. The NumPy part, built by
char types[3]
types[0] = NPY_BYTE /* type of first input arg */

types[1] = NPY_BYTE /* type of second input arg */
types[2] = NPY_BYTE /* type of third input arg */
PyObject *python_ufunc = PyUFunc_FromFuncAndData(

ufunc_loop,
NULL,
types,
1, /* ntypes */
2, /* num_inputs */
1, /* num_outputs */
identity_element,
name,
docstring,
unused)
A ufunc can also support multiple different input-output type combinations.
Making it easier
3. ufunc_loop is of very generic form, and NumPy provides pre-made ones

PyUfunc_f_f float elementwise_func(float input_1)

PyUfunc_ff_f float elementwise_func(float input_1, float input_2)
PyUfunc_d_d double elementwise_func(double input_1)
PyUfunc_dd_d double elementwise_func(double input_1, double input_2)
PyUfunc_D_D elementwise_func(npy_cdouble *input, npy_cdouble* output)
PyUfunc_DD_D elementwise_func(npy_cdouble *in1, npy_cdouble *in2, npy_cdouble*
out)
Only elementwise_func needs to be supplied

. . . except when your elementwise function is not in one of the above forms
8.2.2 Exercise: building an ufunc from scratch
The Mandelbrot fractal is defined by the iteration
z z2 + c
where c = x + i y is a complex number. This iteration is repeated if z stays finite no matter how long the
iteration runs, c belongs to the Mandelbrot set.
Make ufunc called mandel(z0, c) that computes:
z = z0
for k in range(iterations):
z = z*z + c
say, 100 iterations or until z.real**2 + z.imag**2 > 1000. Use it to determine which c are in the
Mandelbrot set.
Our function is a simple one, so make use of the PyUFunc_* helpers.
Write it in Cython
See also:
mandel.pyx, mandelplot.py
#
# Fix the parts marked by TODO
#
#
# Compile this file by (Cython >= 0.12 required because of the complex vars)
#
# cython mandel.pyx
# python setup.py build_ext -i
#
# and try it out with, in this directory,
#
# >>> import mandel
# >>> mandel.mandel(0, 1 + 2j)
#
#
# The elementwise function

# ------------------------
cdef void mandel_single_point(double complex *z_in,

double complex *c_in,
double complex *z_out) nogil:
#

# The Mandelbrot iteration

#
#
# Some points of note:
#
# - It's *NOT* allowed to call any Python functions here.
#
# The Ufunc loop runs with the Python Global Interpreter Lock released.
# Hence, the ``nogil``.
#
# - And so all local variables must be declared with ``cdef``
#
# - Note also that this function receives *pointers* to the data
#
cdef double complex z = z_in[0]

cdef double complex c = c_in[0]
cdef int k # the integer we use in the for loop
#
# TODO: write the Mandelbrot iteration for one point here,
# as you would write it in Python.
#
# Say, use 100 as the maximum number of iterations, and 1000
# as the cutoff for z.real**2 + z.imag**2.
#
TODO: mandelbrot iteration should go here
# Return the answer for this point

z_out[0] = z
# Boilerplate Cython definitions

#
# Pulls definitions from the Numpy C headers.
# -------------------------------------------
from numpy cimport import_array, import_ufunc

from numpy cimport (PyUFunc_FromFuncAndData,
PyUFuncGenericFunction)
from numpy cimport NPY_CDOUBLE, NP_DOUBLE, NPY_LONG
# Import all pre-defined loop functions

# you won't need all of them - keep the relevant ones
from numpy cimport (

PyUFunc_f_f_As_d_d,
PyUFunc_d_d,
PyUFunc_f_f,
PyUFunc_g_g,
PyUFunc_F_F_As_D_D,
PyUFunc_F_F,
PyUFunc_D_D,
PyUFunc_G_G,
PyUFunc_ff_f_As_dd_d,
PyUFunc_ff_f,
PyUFunc_dd_d,
PyUFunc_gg_g,
PyUFunc_FF_F_As_DD_D,
PyUFunc_DD_D,
PyUFunc_FF_F,

PyUFunc_GG_G)
# Required module initialization

# ------------------------------
import_array()
import_ufunc()
# The actual ufunc declaration

# ----------------------------
cdef PyUFuncGenericFunction loop_func[1]

cdef char input_output_types[3]
cdef void *elementwise_funcs[1]
#
# Reminder: some pre-made Ufunc loops:
#
# ================ =======================================================
# ``PyUfunc_f_f`` ``float elementwise_func(float input_1)``
# ``PyUfunc_ff_f`` ``float elementwise_func(float input_1, float input_2)``
# ``PyUfunc_d_d`` ``double elementwise_func(double input_1)``
# ``PyUfunc_dd_d`` ``double elementwise_func(double input_1, double input_2)``
# ``PyUfunc_D_D`` `èlementwise_func(complex_double *input, complex_double* complex_double)``
# ``PyUfunc_DD_D`` `èlementwise_func(complex_double *in1, complex_double *in2, complex_
,double* out)``
# ================ =======================================================
#
# The full list is above.
#
#
# Type codes:
#
# NPY_BOOL, NPY_BYTE, NPY_UBYTE, NPY_SHORT, NPY_USHORT, NPY_INT, NPY_UINT,
# NPY_LONG, NPY_ULONG, NPY_LONGLONG, NPY_ULONGLONG, NPY_FLOAT, NPY_DOUBLE,
# NPY_LONGDOUBLE, NPY_CFLOAT, NPY_CDOUBLE, NPY_CLONGDOUBLE, NPY_DATETIME,
# NPY_TIMEDELTA, NPY_OBJECT, NPY_STRING, NPY_UNICODE, NPY_VOID
#
loop_func[0] = ... TODO: suitable PyUFunc_* ...

input_output_types[0] = ... TODO ...
... TODO: fill in rest of input_output_types ...
# This thing is passed as the ``data`` parameter for the generic

# PyUFunc_* loop, to let it know which function it should call.
elementwise_funcs[0] = <void*>mandel_single_point
# Construct the ufunc:
mandel = PyUFunc_FromFuncAndData(
loop_func,
elementwise_funcs,
input_output_types,
1, # number of supported input types
TODO, # number of input args
TODO, # number of output args
0, # ìdentity` element, never mind this
"mandel", # function name
"mandel(z, c) -> computes z*z + c", # docstring
0 # unused
)

Reminder: some pre-made Ufunc loops:
PyUfunc_f_f float elementwise_func(float input_1)

PyUfunc_ff_f float elementwise_func(float input_1, float input_2)
PyUfunc_d_d double elementwise_func(double input_1)
PyUfunc_dd_d double elementwise_func(double input_1, double input_2)
PyUfunc_D_D elementwise_func(complex_double *input, complex_double* output)
PyUfunc_DD_D elementwise_func(complex_double *in1, complex_double *in2,
complex_double* out)
Type codes:
NPY_BOOL, NPY_BYTE, NPY_UBYTE, NPY_SHORT, NPY_USHORT, NPY_INT, NPY_UINT,

NPY_LONG, NPY_ULONG, NPY_LONGLONG, NPY_ULONGLONG, NPY_FLOAT, NPY_DOUBLE,
NPY_LONGDOUBLE, NPY_CFLOAT, NPY_CDOUBLE, NPY_CLONGDOUBLE, NPY_DATETIME,
NPY_TIMEDELTA, NPY_OBJECT, NPY_STRING, NPY_UNICODE, NPY_VOID
8.2.3 Solution: building an ufunc from scratch
# The elementwise function

# ------------------------

#
# The Mandelbrot iteration
#
#
# Some points of note:
#
# - It's *NOT* allowed to call any Python functions here.
#
# The Ufunc loop runs with the Python Global Interpreter Lock released.
# Hence, the ``nogil``.
#
# - And so all local variables must be declared with ``cdef``
#
# - Note also that this function receives *pointers* to the data;
# the "traditional" solution to passing complex variables around
#
cdef double complex z = z_in[0]

cdef double complex c = c_in[0]
cdef int k # the integer we use in the for loop
# Straightforward iteration
for k in range(100):
z = z*z + c
if z.real**2 + z.imag**2 > 1000:
break
# Return the answer for this point

z_out[0] = z

# Boilerplate Cython definitions

#
# Pulls definitions from the Numpy C headers.
# -------------------------------------------
from numpy cimport import_array, import_ufunc

from numpy cimport (PyUFunc_FromFuncAndData,
PyUFuncGenericFunction)
from numpy cimport NPY_CDOUBLE
from numpy cimport PyUFunc_DD_D
# Required module initialization

# ------------------------------
import_array()
import_ufunc()
# The actual ufunc declaration

# ----------------------------
cdef PyUFuncGenericFunction loop_func[1]

cdef char input_output_types[3]
cdef void *elementwise_funcs[1]
loop_func[0] = PyUFunc_DD_D
input_output_types[0] = NPY_CDOUBLE
loop_func,
elementwise_funcs,
input_output_types,
1, # number of supported input types
2, # number of input args
1, # number of output args
"mandel(z, c) -> computes iterated z*z + c", # docstring
0 # unused
)
"""
Plot Mandelbrot
================
Plot the Mandelbrot ensemble.
"""
import numpy as np
import mandel
x = np.linspace(-1.7, 0.6, 1000)
y = np.linspace(-1.4, 1.4, 1000)
c = x[None,:] + 1j*y[:,None]
z = mandel.mandel(c, c)

plt.imshow(abs(z)**2 < 1000, extent=[-1.7, 0.6, -1.4, 1.4])

plt.gray()
plt.show()
Note: Most of the boilerplate could be automated by these Cython modules:

http://wiki.cython.org/MarkLodato/CreatingUfuncs
Several accepted input types
E.g. supporting both single- and double-precision versions

...
cdef void mandel_single_point_singleprec(float complex *z_in,

float complex *c_in,
float complex *z_out) nogil:
...
cdef PyUFuncGenericFunction loop_funcs[2]

cdef char input_output_types[3*2]
cdef void *elementwise_funcs[1*2]
loop_funcs[0] = PyUFunc_DD_D
loop_funcs[1] = PyUFunc_FF_F
input_output_types[3] = NPY_CFLOAT
elementwise_funcs[1] = <void*>mandel_single_point_singleprec
loop_func,
elementwise_funcs,
input_output_types,
2, # number of supported input types <----------------

2, # number of input args

1, # number of output args
"mandel(z, c) -> computes iterated z*z + c", # docstring
0 # unused
)
8.2.4 Generalized ufuncs
ufunc
output = elementwise_function(input)
Both output and input can be a single array element only.
generalized ufunc
output and input can be arrays with a fixed number of dimensions
For example, matrix trace (sum of diag elements):
input shape = (n, n)

output shape = () i.e. scalar
(n, n) -> ()
Matrix product:
input_1 shape = (m, n)

input_2 shape = (n, p)
output shape = (m, p)
(m, n), (n, p) -> (m, p)
This is called the signature of the generalized ufunc

The dimensions on which the g-ufunc acts, are core dimensions
Status in NumPy
g-ufuncs are in NumPy already . . .

new ones can be created with PyUFunc_FromFuncAndDataAndSignature
most linear-algebra functions are implemented as g-ufuncs to enable working with stacked arrays:

>>> np.linalg.det(np.random.rand(3, 5, 5))
array([ 0.00965823, -0.13344729, 0.04583961])
>>> np.linalg._umath_linalg.det.signature
'(m,m)->()'
we also ship with a few g-ufuncs for testing, ATM:
>>> import numpy.core.umath_tests as ut

>>> ut.matrix_multiply.signature
'(m,n),(n,p)->(m,p)'
>>> x = np.ones((10, 2, 4))

>>> y = np.ones((10, 4, 5))

>>> ut.matrix_multiply(x, y).shape

(10, 2, 5)
in both examples the last two dimensions became core dimensions, and are modified as per the signature
otherwise, the g-ufunc operates elementwise
matrix multiplication this way could be useful for operating on many small matrices at once
Generalized ufunc loop
Matrix multiplication (m,n),(n,p) -> (m,p)
void gufunc_loop(void **args, int *dimensions, int *steps, void *data)

{
char *input_1 = (char*)args[0]; /* these are as previously */
char *output = (char*)args[2];
int input_1_stride_m = steps[3]; /* strides for the core dimensions */

int input_1_stride_n = steps[4]; /* are added after the non-core */
int input_2_strides_n = steps[5]; /* steps */
int input_2_strides_p = steps[6];
int output_strides_n = steps[7];
int output_strides_p = steps[8];
int m = dimension[1]; /* core dimensions are added after */

int n = dimension[2]; /* the main dimension; order as in */
int p = dimension[3]; /* signature */
int i;
for (i = 0; i < dimensions[0]; ++i) {

matmul_for_strided_matrices(input_1, input_2, output,
strides for each array...);
output += steps[2];
}
}
8.3 Interoperability features
8.3.1 Sharing multidimensional, typed data
Suppose you
1. Write a library than handles (multidimensional) binary data,
2. Want to make it easy to manipulate the data with NumPy, or whatever other library,
3. . . . but would not like to have NumPy as a dependency.
Currently, 3 solutions:
1. the old buffer interface
2. the array interface
3. the new buffer interface (PEP 3118)
8.3. Interoperability features 304

8.3.2 The old buffer protocol
Only 1-D buffers

No data type information
C-level interface; PyBufferProcs tp_as_buffer in the type object
But its integrated into Python (e.g. strings support it)
Mini-exercise using Pillow (Python Imaging Library):
See also:
pilbuffer.py
>>> from PIL import Image

>>> data = np.zeros((200, 200, 4), dtype=np.int8)
>>> data[:, :] = [255, 0, 0, 255] # Red
>>> # In PIL, RGBA images consist of 32-bit integers whose bytes are [RR,GG,BB,AA]
>>> data = data.view(np.int32).squeeze()
>>> img = Image.frombuffer("RGBA", (200, 200), data, "raw", "RGBA", 0, 1)
>>> img.save('test.png')
Q:
Check what happens if data is now modified, and img saved again.
8.3.3 The old buffer protocol
"""
From buffer
============
Show how to exchange data between numpy and a library that only knows
the buffer interface.
"""
import numpy as np
import Image
# Let's make a sample image, RGBA format
x = np.zeros((200, 200, 4), dtype=np.int8)
x[:,:,0] = 254 # red

x[:,:,3] = 255 # opaque
data = x.view(np.int32) # Check that you understand why this is OK!
img = Image.frombuffer("RGBA", (200, 200), data)

img.save('test.png')
#
# Modify the original data, and save again.
#
# It turns out that PIL, which knows next to nothing about Numpy,
# happily shares the same data.
#
x[:,:,1] = 254
img.save('test2.png')

8.3.4 Array interface protocol
Multidimensional buffers
Data type information present
NumPy-specific approach; slowly deprecated (but not going away)
Not integrated in Python otherwise
See also:
Documentation: http://docs.scipy.org/doc/numpy/reference/arrays.interface.html
>>> x = np.array([[1, 2], [3, 4]])

>>> x.__array_interface__
{'data': (171694552, False), # memory address of data, is readonly?
'descr': [('', '<i4')], # data type descriptor
'typestr': '<i4', # same, in another form
'strides': None, # strides; or None if in C-order
'shape': (2, 2),
'version': 3,
}
::

>>> from PIL import Image

>>> img = Image.open('data/test.png')
>>> img.__array_interface__
{'data': ...,
'shape': (200, 200, 4),
'typestr': '|u1'}
>>> x = np.asarray(img)
>>> x.shape
(200, 200, 4)
Note: A more C-friendly variant of the array interface is also defined.
8.4 Array siblings: chararray, maskedarray, matrix
8.4.1 chararray: vectorized string operations
>>> x = np.array(['a', ' bbb', ' ccc']).view(np.chararray)

>>> x.lstrip(' ')
chararray(['a', 'bbb', 'ccc'],
dtype='...')
>>> x.upper()
chararray(['A', ' BBB', ' CCC'],
dtype='...')
Note: .view() has a second meaning: it can make an ndarray an instance of a specialized ndarray subclass
8.4.2 masked_array missing data
Masked arrays are arrays that may have missing or invalid entries.
For example, suppose we have an array where the fourth entry is invalid:
>>> x = np.array([1, 2, 3, -99, 5])
One way to describe this is to create a masked array:
>>> mx = np.ma.masked_array(x, mask=[0, 0, 0, 1, 0])

>>> mx
masked_array(data = [1 2 3 -- 5],
mask = [False False False True False],
Masked mean ignores masked data:
>>> mx.mean()
2.75
>>> np.mean(mx)
2.75
Warning: Not all NumPy functions respect masks, for instance np.dot, so check the return types.
The masked_array returns a view to the original array:
8.4. Array siblings: chararray, maskedarray, matrix 307

>>> mx[1] = 9
>>> x
array([ 1, 9, 3, -99, 5])
The mask
You can modify the mask by assigning:
>>> mx[1] = np.ma.masked

>>> mx
masked_array(data = [1 -- 3 -- 5],
mask = [False True False True False],
The mask is cleared on assignment:
>>> mx[1] = 9
>>> mx
masked_array(data = [1 9 3 -- 5],
mask = [False False False True False],
The mask is also available directly:
>>> mx.mask
array([False, False, False, True, False], dtype=bool)
The masked entries can be filled with a given value to get an usual array back:
>>> x2 = mx.filled(-1)
>>> x2
array([ 1, 9, 3, -1, 5])
The mask can also be cleared:
>>> mx.mask = np.ma.nomask

>>> mx
masked_array(data = [1 9 3 -99 5],
mask = [False False False False False],
Domain-aware functions
The masked array package also contains domain-aware functions:
>>> np.ma.log(np.array([1, 2, -1, -2, 3, -5]))

masked_array(data = [0.0 0.6931471805599453 -- -- 1.0986122886681098 --],
mask = [False False True True False True],
fill_value = 1e+20)
Note: Streamlined and more seamless support for dealing with missing data in arrays is making its way into
NumPy 1.7. Stay tuned!

Example: Masked statistics
Canadian rangers were distracted when counting hares and lynxes in 1903-1910 and 1917-1918, and got
the numbers are wrong. (Carrot farmers stayed alert, though.) Compute the mean populations over time,
ignoring the invalid numbers.
>>> populations = np.ma.masked_array(data[:,1:])
>>> year = data[:, 0]
>>> bad_years = (((year >= 1903) & (year <= 1910))

... | ((year >= 1917) & (year <= 1918)))
>>> # '&' means 'and' and '|' means 'or'
>>> populations[bad_years, 0] = np.ma.masked
>>> populations[bad_years, 1] = np.ma.masked
>>> populations.mean(axis=0)
masked_array(data = [40472.72727272727 18627.272727272728 42400.0],
mask = [False False False],
fill_value = 1e+20)
>>> populations.std(axis=0)
masked_array(data = [21087.656489006717 15625.799814240254 3322.5062255844787],
mask = [False False False],
fill_value = 1e+20)
Note that Matplotlib knows about masked arrays:

>>> plt.plot(year, populations, 'o-')
[<matplotlib.lines.Line2D object at ...>, ...]
8.4.3 recarray: purely convenience
>>> arr = np.array([('a', 1), ('b', 2)], dtype=[('x', 'S1'), ('y', int)])
>>> arr2 = arr.view(np.recarray)
>>> arr2.x
chararray(['a', 'b'],
dtype='|S1')
>>> arr2.y
array([1, 2])

8.4.4 matrix: convenience?
always 2-D
* is the matrix product, not the elementwise one
>>> np.matrix([[1, 0], [0, 1]]) * np.matrix([[1, 2], [3, 4]])

matrix([[1, 2],
[3, 4]])
8.5 Summary
Anatomy of the ndarray: data, dtype, strides.

Universal functions: elementwise operations, how to make new ones
Ndarray subclasses
Various buffer interfaces for integration with other tools
Recent additions: PEP 3118, generalized ufuncs
8.6 Contributing to NumPy/Scipy
Get this tutorial: http://www.euroscipy.org/talk/882
8.6.1 Why
Theres a bug?
I dont understand what this is supposed to do?
I have this fancy code. Would you like to have it?
Id like to help! What can I do?
8.6.2 Reporting bugs
Bug tracker (prefer this)

http://projects.scipy.org/numpy
http://projects.scipy.org/scipy
Click the Register link to get an account
Mailing lists ( scipy.org/Mailing_Lists )
If youre unsure
No replies in a week or so? Just file a bug ticket.
Good bug report
Title: numpy.random.permutations fails for non-integer arguments
I'm trying to generate random permutations, using numpy.random.permutations
When calling numpy.random.permutation with non-integer arguments
8.5. Summary 310

it fails with a cryptic error message::
>>> np.random.permutation(12)
array([11, 5, 8, 4, 6, 1, 9, 3, 7, 2, 10, 0])
>>> np.random.permutation(12.)
File "mtrand.pyx", line 3311, in mtrand.RandomState.permutation
File "mtrand.pyx", line 3254, in mtrand.RandomState.shuffle
TypeError: len() of unsized object
This also happens with long arguments, and so

np.random.permutation(X.shape[0]) where X is an array fails on 64
bit windows (where shape is a tuple of longs).
It would be great if it could cast to integer or at least raise a

proper error for non-integer types.
I'm using NumPy 1.4.1, built from the official tarball, on Windows
64 with Visual studio 2008, on Python.org 64-bit Python.
0. What are you trying to do?

1. Small code snippet reproducing the bug (if possible)
What actually happens
What youd expect
2. Platform (Windows / Linux / OSX, 32/64 bits, x86/PPC, . . . )
3. Version of NumPy/Scipy
>>> print(np.__version__)
1...
Check that the following is what you expect
>>> print(np.__file__)
/...
In case you have old/broken NumPy installations lying around.

If unsure, try to remove existing NumPy installations, and reinstall. . .
8.6.3 Contributing to documentation
1. Documentation editor
http://docs.scipy.org/doc/numpy
Registration
Register an account
Subscribe to scipy-dev mailing list (subscribers-only)
Problem with mailing lists: you get mail
* But: you can turn mail delivery off

* change your subscription options, at the bottom of
http://mail.scipy.org/mailman/listinfo/scipy-dev
Send a mail @ scipy-dev mailing list; ask for activation:
8.6. Contributing to NumPy/Scipy 311

To: scipy-dev@scipy.org
Hi,
I'd like to edit NumPy/Scipy docstrings. My account is XXXXX
Cheers,
N. N.
Check the style guide:

http://docs.scipy.org/doc/numpy/
Dont be intimidated; to fix a small thing, just fix it
Edit
2. Edit sources and send patches (as for bugs)
3. Complain on the mailing list
8.6.4 Contributing features
0. Ask on mailing list, if unsure where it should go

1. Write a patch, add an enhancement ticket on the bug tracket
2. OR, create a Git branch implementing the feature + add enhancement ticket.
Especially for big/invasive additions
http://projects.scipy.org/numpy/wiki/GitMirror
http://www.spheredev.org/wiki/Git_for_the_lazy
# Clone numpy repository

git clone --origin svn http://projects.scipy.org/git/numpy.git numpy
cd numpy
# Create a feature branch

git checkout -b name-of-my-feature-branch svn/trunk
<edit stuff>
git commit -a
Create account on https://github.com (or anywhere)

Create a new repository @ Github
Push your work to github
git remote add github git@github:USERNAME/REPOSITORYNAME.git

git push github name-of-my-feature-branch
8.6.5 How to help, in general
Bug fixes always welcome!

What irks you most
Browse the tracker
Documentation work

API docs: improvements to docstrings
* Know some Scipy module well?

User guide
* Needs to be done eventually.

* Want to think? Come up with a Table of Contents
http://scipy.org/Developer_Zone/UG_Toc
Ask on communication channels:
numpy-discussion list
scipy-dev list

CHAPTER 9
Debugging code
Author: Gal Varoquaux

This section explores tools to understand better your code base: debugging, to find and fix bugs.
It is not specific to the scientific Python community, but the strategies that we will employ are tailored to its
needs.
Prerequisites
Numpy
IPython
nosetests
pyflakes
gdb for the C-debugging part.
Chapter contents
Avoiding bugs
Coding best practices to avoid getting in trouble
pyflakes: fast static analysis
Debugging workflow
Using the Python debugger
Invoking the debugger
314
Debugger commands and interaction

Debugging segmentation faults using gdb
9.1 Avoiding bugs
9.1.1 Coding best practices to avoid getting in trouble
Brian Kernighan
Everyone knows that debugging is twice as hard as writing a program in the first place. So if youre as clever
as you can be when you write it, how will you ever debug it?
We all write buggy code. Accept it. Deal with it.

Write your code with testing and debugging in mind.
Keep It Simple, Stupid (KISS).
What is the simplest thing that could possibly work?
Dont Repeat Yourself (DRY).
Every piece of knowledge must have a single, unambiguous, authoritative representation within a
system.
Constants, algorithms, etc. . .
Try to limit interdependencies of your code. (Loose Coupling)
Give your variables, functions and modules meaningful names (not mathematics names)
9.1.2 pyflakes: fast static analysis
They are several static analysis tools in Python; to name a few:

pylint
pychecker
pyflakes
flake8
Here we focus on pyflakes, which is the simplest tool.
Fast, simple
Detects syntax errors, missing imports, typos on names.
Another good recommendation is the flake8 tool which is a combination of pyflakes and pep8. Thus, in addi-
tion to the types of errors that pyflakes catches, flake8 detects violations of the recommendation in PEP8 style
guide.
Integrating pyflakes (or flake8) in your editor or IDE is highly recommended, it does yield productivity gains.
Running pyflakes on the current edited file
You can bind a key to run pyflakes in the current buffer.

In kate Menu: settings -> configure kate
9.1. Avoiding bugs 315

In plugins enable external tools

In external Tools, add pyflakes:
kdialog --title "pyflakes %f ilename" --msgbox "$(pyflakes %f ilename)"
In TextMate
Menu: TextMate -> Preferences -> Advanced -> Shell variables, add a shell variable:
TM_PYCHECKER = /Library/Frameworks/Python.framework/Versions/Current/bin/pyflakes
Then Ctrl-Shift-V is binded to a pyflakes report

In vim In your .vimrc (binds F5 to pyflakes):
autocmd FileType python let &mp = 'echo "*** running % ***" ; pyflakes %'
autocmd FileType tex,mp,rst,python imap <Esc>[15~ <C-O>:make!^M
autocmd FileType tex,mp,rst,python map <Esc>[15~ :make!^M
autocmd FileType tex,mp,rst,python set autowrite
In emacs In your .emacs (binds F5 to pyflakes):
(defun pyflakes-thisfile () (interactive)

(compile (format "pyflakes %s " (buffer-file-name)))
)
(define-minor-mode pyflakes-mode
"Toggle pyflakes mode.
With no argument, this command toggles the mode.
Non-null prefix argument turns on the mode.
Null prefix argument turns off the mode."
;; The initial value.
nil
;; The indicator for the mode line.
" Pyflakes"
;; The minor mode bindings.
'( ([f5] . pyflakes-thisfile) )
)
(add-hook 'python-mode-hook (lambda () (pyflakes-mode t)))
A type-as-go spell-checker like integration
In vim
Use the pyflakes.vim plugin:
1. download the zip file from http://www.vim.org/scripts/script.php?script_id=2441
2. extract the files in ~/.vim/ftplugin/python
3. make sure your vimrc has filetype plugin indent on
Alternatively: use the syntastic plugin. This can be configured to use flake8 too and also handles
on-the-fly checking for many other languages.
9.1. Avoiding bugs 316

In emacs Use the flymake mode with pyflakes, documented on http://www.plope.com/Members/

chrism/flymake-mode : add the following to your .emacs file:
(when (load "flymake" t)

(defun flymake-pyflakes-init ()
(let* ((temp-file (flymake-init-create-temp-buffer-copy
'flymake-create-temp-inplace))
(local-file (file-relative-name
temp-file
(file-name-directory buffer-file-name))))
(list "pyflakes" (list local-file))))
(add-to-list 'flymake-allowed-file-name-masks
'("\\.py\\'" flymake-pyflakes-init)))
(add-hook 'find-file-hook 'flymake-find-file-hook)
9.2 Debugging workflow
If you do have a non trivial bug, this is when debugging strategies kick in. There is no silver bullet. Yet, strategies
help:
For debugging a given problem, the favorable situation is when the problem is isolated in a
small number of lines of code, outside framework or application code, with short modify-run-
fail cycles
1. Make it fail reliably. Find a test case that makes the code fail every time.
2. Divide and Conquer. Once you have a failing test case, isolate the failing code.
Which module.
Which function.
Which line of code.
=> isolate a small reproducible failure: a test case
3. Change one thing at a time and re-run the failing test case.
4. Use the debugger to understand what is going wrong.
5. Take notes and be patient. It may take a while.
Note: Once you have gone through this process: isolated a tight piece of code reproducing the bug and fix the
bug using this piece of code, add the corresponding code to your test suite.
9.2. Debugging workflow 317

9.3 Using the Python debugger
The python debugger, pdb: https://docs.python.org/library/pdb.html, allows you to inspect your code inter-
actively.
Specifically it allows you to:
View the source code.
Walk up and down the call stack.
Inspect values of variables.
Modify values of variables.
Set breakpoints.
print
Yes, print statements do work as a debugging tool. However to inspect runtime, it is often more efficient to
use the debugger.
9.3.1 Invoking the debugger
Ways to launch the debugger:

1. Postmortem, launch debugger after module errors.
2. Launch the module with the debugger.
3. Call the debugger inside the module
Postmortem
Situation: Youre working in IPython and you get a traceback.

Here we debug the file index_error.py. When running it, an IndexError is raised. Type %debug and drop
into the debugger.
In [1]: %run index_error.py

---------------------------------------------------------------------------
IndexError Traceback (most recent call last)
/home/varoquau/dev/scipy-lecture-notes/advanced/debugging/index_error.py in <module>()
6
7 if __name__ == '__main__':
----> 8 index_error()
9
/home/varoquau/dev/scipy-lecture-notes/advanced/debugging/index_error.py in index_error()
3 def index_error():
4 lst = list('foobar')
----> 5 print lst[len(lst)]
6
7 if __name__ == '__main__':
In [2]: %debug
> /home/varoquau/dev/scipy-lecture-notes/advanced/debugging/index_error.py(5)index_error()
9.3. Using the Python debugger 318

ipdb> list
1 """Small snippet to raise an IndexError."""
2
3 def index_error():
6
7 if __name__ == '__main__':
8 index_error()
9
ipdb> len(lst)
6
ipdb> print lst[len(lst)-1]
r
ipdb> quit
In [3]:
Post-mortem debugging without IPython
In some situations you cannot use IPython, for instance to debug a script that wants to be called from the
command line. In this case, you can call the script with python -m pdb script.py:
$ python -m pdb index_error.py
> /home/varoquau/dev/scipy-lecture-notes/advanced/optimizing/index_error.py(1)<module>()
-> """Small snippet to raise an IndexError."""
(Pdb) continue
File "/usr/lib/python2.6/pdb.py", line 1296, in main
pdb._runscript(mainpyfile)
File "/usr/lib/python2.6/pdb.py", line 1215, in _runscript
self.run(statement)
File "/usr/lib/python2.6/bdb.py", line 372, in run
exec cmd in globals, locals
File "<string>", line 1, in <module>
File "index_error.py", line 8, in <module>
index_error()
File "index_error.py", line 5, in index_error
print lst[len(lst)]
Uncaught exception. Entering post mortem debugging
Running 'cont' or 'step' will restart the program
> /home/varoquau/dev/scipy-lecture-notes/advanced/optimizing/index_error.py(5)index_error()
-> print lst[len(lst)]
(Pdb)
Step-by-step execution
Situation: You believe a bug exists in a module but are not sure where.
For instance we are trying to debug wiener_filtering.py. Indeed the code runs, but the filtering does not
work well.
Run the script in IPython with the debugger using %run -d wiener_filtering.p :
In [1]: %run -d wiener_filtering.py

*** Blank or comment


Breakpoint 1 at /home/varoquau/dev/scipy-lecture-notes/advanced/optimizing/wiener_
,filtering.py:4
NOTE: Enter 'c' at the ipdb> prompt to start your script.
> <string>(1)<module>()
Set a break point at line 34 using b 34:
ipdb> n
> /home/varoquau/dev/scipy-lecture-notes/advanced/optimizing/wiener_filtering.py(4)
,<module>()
3
1---> 4 import numpy as np
5 import scipy as sp
ipdb> b 34
Breakpoint 2 at /home/varoquau/dev/scipy-lecture-notes/advanced/optimizing/wiener_
,filtering.py:34
Continue execution to next breakpoint with c(ont(inue)):
ipdb> c
> /home/varoquau/dev/scipy-lecture-notes/advanced/optimizing/wiener_filtering.
,py(34)iterated_wiener()
33 """
2--> 34 noisy_img = noisy_img
35 denoised_img = local_mean(noisy_img, size=size)
Step into code with n(ext) and s(tep): next jumps to the next statement in the current execution
context, while step will go across execution contexts, i.e. enable exploring inside function calls:
ipdb> s
2 34 noisy_img = noisy_img
---> 35 denoised_img = local_mean(noisy_img, size=size)
36 l_var = local_var(noisy_img, size=size)
ipdb> n
35 denoised_img = local_mean(noisy_img, size=size)
---> 36 l_var = local_var(noisy_img, size=size)
37 for i in range(3):
Step a few lines and explore the local variables:
ipdb> n
36 l_var = local_var(noisy_img, size=size)
---> 37 for i in range(3):
38 res = noisy_img - denoised_img
ipdb> print l_var
[[5868 5379 5316 ..., 5071 4799 5149]
[5013 363 437 ..., 346 262 4355]
[5379 410 344 ..., 392 604 3377]
...,
[ 435 362 308 ..., 275 198 1632]
[ 548 392 290 ..., 248 263 1653]
[ 466 789 736 ..., 1835 1725 1940]]

ipdb> print l_var.min()

0
Oh dear, nothing but integers, and 0 variation. Here is our bug, we are doing integer arithmetic.
Raising exception on numerical errors
When we run the wiener_filtering.py file, the following warnings are raised:
In [2]: %run wiener_filtering.py
wiener_filtering.py:40: RuntimeWarning: divide by zero encountered in divide
noise_level = (1 - noise/l_var )
We can turn these warnings in exception, which enables us to do post-mortem debugging on them, and find
our problem more quickly:
In [3]: np.seterr(all='raise')
Out[3]: {'divide': 'print', 'invalid': 'print', 'over': 'print', 'under': 'ignore'}
In [4]: %run wiener_filtering.py
---------------------------------------------------------------------------
FloatingPointError Traceback (most recent call last)
/home/esc/anaconda/lib/python2.7/site-packages/IPython/utils/py3compat.pyc in execfile(fname,
, *where)
176 else:
177 filename = fname
--> 178 __builtin__.execfile(filename, *where)
/home/esc/physique-cuso-python-2013/scipy-lecture-notes/advanced/debugging/wiener_filtering.
,py in <module>()
55 pl.matshow(noisy_face[cut], cmap=pl.cm.gray)
56
---> 57 denoised_face = iterated_wiener(noisy_face)
58 pl.matshow(denoised_face[cut], cmap=pl.cm.gray)
59
/home/esc/physique-cuso-python-2013/scipy-lecture-notes/advanced/debugging/wiener_filtering.
,py in iterated_wiener(noisy_img, size)
38 res = noisy_img - denoised_img
39 noise = (res**2).sum()/res.size
---> 40 noise_level = (1 - noise/l_var )
41 noise_level[noise_level<0] = 0
42 denoised_img += noise_level*res
FloatingPointError: divide by zero encountered in divide
Other ways of starting a debugger
Raising an exception as a poor man break point

If you find it tedious to note the line number to set a break point, you can simply raise an exception at
the point that you want to inspect and use IPythons %debug. Note that in this case you cannot step or
continue the execution.
Debugging test failures using nosetests
You can run nosetests --pdb to drop in post-mortem debugging on exceptions, and nosetests
--pdb-failure to inspect test failures using the debugger.
In addition, you can use the IPython interface for the debugger in nose by installing the nose plugin
ipdbplugin. You can than pass --ipdb and --ipdb-failure options to nosetests.

Calling the debugger explicitly

Insert the following line where you want to drop in the debugger:
import pdb; pdb.set_trace()
Warning: When running nosetests, the output is captured, and thus it seems that the debugger does not
work. Simply run the nosetests with the -s flag.
Graphical debuggers and alternatives
For stepping through code and inspecting variables, you might find it more convenient to use a graph-
ical debugger such as winpdb.
Alternatively, pudb is a good semi-graphical debugger with a text user interface in the console.
Also, the pydbgr project is probably worth looking at.
9.3.2 Debugger commands and interaction
l(list) Lists the code at the current position

u(p) Walk up the call stack
d(own) Walk down the call stack
n(ext) Execute the next line (does not go down in new functions)
s(tep) Execute the next statement (goes down in new functions)
bt Print the call stack
a Print the local variables
!command Execute the given Python command (by opposition to pdb commands
Warning: Debugger commands are not Python code

You cannot name the variables the way you want. For instance, if in you cannot override the variables in
the current frame with the same name: use different names than your local variable when typing code in
the debugger.
Getting help when in the debugger
Type h or help to access the interactive help:
ipdb> help
Documented commands (type help <topic>):

========================================
EOF bt cont enable jump pdef r tbreak w
a c continue exit l pdoc restart u whatis
alias cl d h list pinfo return unalias where
args clear debug help n pp run unt
b commands disable ignore next q s until
break condition down j p quit step up
Miscellaneous help topics:

==========================
exec pdb

Undocumented commands:
======================
retval rv
9.4 Debugging segmentation faults using gdb
If you have a segmentation fault, you cannot debug it with pdb, as it crashes the Python interpreter before it
can drop in the debugger. Similarly, if you have a bug in C code embedded in Python, pdb is useless. For this
we turn to the gnu debugger, gdb, available on Linux.
Before we start with gdb, let us add a few Python-specific tools to it. For this we add a few macros to our ~/.
gdbinit. The optimal choice of macro depends on your Python version and your gdb version. I have added a
simplified version in gdbinit, but feel free to read DebuggingWithGdb.
To debug with gdb the Python script segfault.py, we can run the script in gdb as follows
$ gdb python
...
(gdb) run segfault.py
Starting program: /usr/bin/python segfault.py
[Thread debugging using libthread_db enabled]
Program received signal SIGSEGV, Segmentation fault.

_strided_byte_copy (dst=0x8537478 "\360\343G", outstrides=4, src=
0x86c0690 <Address 0x86c0690 out of bounds>, instrides=32, N=3,
elsize=4)
at numpy/core/src/multiarray/ctors.c:365
365 _FAST_MOVE(Int32);
(gdb)
We get a segfault, and gdb captures it for post-mortem debugging in the C level stack (not the Python call
stack). We can debug the C call stack using gdbs commands:
(gdb) up
#1 0x004af4f5 in _copy_from_same_shape (dest=<value optimized out>,
src=<value optimized out>, myfunc=0x496780 <_strided_byte_copy>,
swap=0)
at numpy/core/src/multiarray/ctors.c:748
748 myfunc(dit->dataptr, dest->strides[maxaxis],
As you can see, right now, we are in the C code of numpy. We would like to know what is the Python code that
triggers this segfault, so we go up the stack until we hit the Python execution loop:
(gdb) up
#8 0x080ddd23 in call_function (f=
Frame 0x85371ec, for file /home/varoquau/usr/lib/python2.6/site-packages/numpy/core/
,arrayprint.py, line 156, in _leading_trailing (a=<numpy.ndarray at remote 0x85371b0>, _nc=
,<module at remote 0xb7f93a64>), throwflag=0)
at ../Python/ceval.c:3750
3750 ../Python/ceval.c: No such file or directory.
in ../Python/ceval.c
(gdb) up
#9 PyEval_EvalFrameEx (f=
Frame 0x85371ec, for file /home/varoquau/usr/lib/python2.6/site-packages/numpy/core/
,arrayprint.py, line 156, in _leading_trailing (a=<numpy.ndarray at remote 0x85371b0>, _nc=
,<module at remote 0xb7f93a64>), throwflag=0)
at ../Python/ceval.c:2412
9.4. Debugging segmentation faults using gdb 323

2412 in ../Python/ceval.c
(gdb)
Once we are in the Python execution loop, we can use our special Python helper function. For instance we can
find the corresponding Python code:
(gdb) pyframe
/home/varoquau/usr/lib/python2.6/site-packages/numpy/core/arrayprint.py (158): _leading_
,trailing
(gdb)
This is numpy code, we need to go up until we find code that we have written:
(gdb) up
...
(gdb) up
#34 0x080dc97a in PyEval_EvalFrameEx (f=
Frame 0x82f064c, for file segfault.py, line 11, in print_big_array (small_array=<numpy.
,ndarray at remote 0x853ecf0>, big_array=<numpy.ndarray at remote 0x853ed20>), throwflag=0)
,at ../Python/ceval.c:1630
1630 ../Python/ceval.c: No such file or directory.
in ../Python/ceval.c
(gdb) pyframe
segfault.py (12): print_big_array
The corresponding code is:
def make_big_array(small_array):
big_array = stride_tricks.as_strided(small_array,
shape=(2e6, 2e6), strides=(32, 32))
return big_array
def print_big_array(small_array):
big_array = make_big_array(small_array)
Thus the segfault happens when printing big_array[-10:]. The reason is simply that big_array has been
allocated with its end outside the program memory.
Note: For a list of Python-specific commands defined in the gdbinit, read the source of this file.
Wrap up exercise
The following script is well documented and hopefully legible. It seeks to answer a problem of actual interest
for numerical computing, but it does not work. . . Can you debug it?
Python source code: to_debug.py
9.4. Debugging segmentation faults using gdb 324

CHAPTER 10
Optimizing code
Donald Knuth
Premature optimization is the root of all evil

This chapter deals with strategies to make Python code go faster.
Prerequisites
line_profiler
Chapters contents
Optimization workflow
Profiling Python code
Timeit
Profiler
Line-profiler
Making code go faster
Algorithmic optimization
* Example of the SVD
325
Writing faster numerical code

Additional Links
10.1 Optimization workflow
1. Make it work: write the code in a simple legible ways.

2. Make it work reliably: write automated test cases, make really sure that your algorithm is right and that
if you break it, the tests will capture the breakage.
3. Optimize the code by profiling simple use-cases to find the bottlenecks and speeding up these bottle-
neck, finding a better algorithm or implementation. Keep in mind that a trade off should be found be-
tween profiling on a realistic example and the simplicity and speed of execution of the code. For efficient
work, it is best to work with profiling runs lasting around 10s.
10.2 Profiling Python code
No optimization without measuring!
Measure: profiling, timing

Youll have surprises: the fastest code is not always what you think
10.2.1 Timeit
In IPython, use timeit (https://docs.python.org/library/timeit.html) to time elementary operations:
In [2]: a = np.arange(1000)
In [3]: %timeit a ** 2
In [4]: %timeit a ** 2.1

In [5]: %timeit a * a
Use this to guide your choice between strategies.
Note: For long running calls, using %time instead of %timeit; it is less precise but faster
10.2.2 Profiler
Useful when you have a large program to profile, for example the following file:
# For this example to run, you also need the 'ica.py' file
import numpy as np
10.1. Optimization workflow 326

from scipy import linalg
from ica import fastica
def test():
data = np.random.random((5000, 100))
u, s, v = linalg.svd(data)
pca = np.dot(u[:, :10].T, data)
results = fastica(pca.T, whiten=False)
if __name__ == '__main__':
test()
Note: This is a combination of two unsupervised learning techniques, principal component analysis (PCA)
and independent component analysis (ICA). PCA is a technique for dimensionality reduction, i.e. an algorithm
to explain the observed variance in your data using less dimensions. ICA is a source seperation technique,
for example to unmix multiple signals that have been recorded through multiple sensors. Doing a PCA first
and then an ICA can be useful if you have more sensors than signals. For more information see: the FastICA
example from scikits-learn.
To run it, you also need to download the ica module. In IPython we can time the script:
In [1]: %run -t demo.py
IPython CPU timings (estimated):

User : 14.3929 s.
System: 0.256016 s.
and profile it:
In [2]: %run -p demo.py
916 function calls in 14.551 CPU seconds
Ordered by: internal time
ncalls tottime percall cumtime percall filename:lineno (function)

1 14.457 14.457 14.479 14.479 decomp.py:849 (svd)
1 0.054 0.054 0.054 0.054 {method 'random_sample' of 'mtrand.RandomState'
,objects}
1 0.017 0.017 0.021 0.021 function_base.py:645 (asarray_chkfinite)
54 0.011 0.000 0.011 0.000 {numpy.core._dotblas.dot}
2 0.005 0.002 0.005 0.002 {method 'any' of 'numpy.ndarray' objects}
6 0.001 0.000 0.001 0.000 ica.py:195 (gprime)
6 0.001 0.000 0.001 0.000 ica.py:192 (g)
14 0.001 0.000 0.001 0.000 {numpy.linalg.lapack_lite.dsyevd}
19 0.001 0.000 0.001 0.000 twodim_base.py:204 (diag)
1 0.001 0.001 0.008 0.008 ica.py:69 (_ica_par)
1 0.001 0.001 14.551 14.551 {execfile}
107 0.000 0.000 0.001 0.000 defmatrix.py:239 (__array_finalize__)
7 0.000 0.000 0.004 0.001 ica.py:58 (_sym_decorrelation)
7 0.000 0.000 0.002 0.000 linalg.py:841 (eigh)
172 0.000 0.000 0.000 0.000 {isinstance}
1 0.000 0.000 14.551 14.551 demo.py:1 (<module>)
29 0.000 0.000 0.000 0.000 numeric.py:180 (asarray)
35 0.000 0.000 0.000 0.000 defmatrix.py:193 (__new__)
35 0.000 0.000 0.001 0.000 defmatrix.py:43 (asmatrix)
21 0.000 0.000 0.001 0.000 defmatrix.py:287 (__mul__)
41 0.000 0.000 0.000 0.000 {numpy.core.multiarray.zeros}
10.2. Profiling Python code 327

28 0.000 0.000 0.000 0.000 {method 'transpose' of 'numpy.ndarray' objects}

1 0.000 0.000 0.008 0.008 ica.py:97 (fastica)
...
Clearly the svd (in decomp.py) is what takes most of our time, a.k.a. the bottleneck. We have to find a way to
make this step go faster, or to avoid this step (algorithmic optimization). Spending time on the rest of the code
is useless.
Profiling outside of IPython, running cProfile
Similar profiling can be done outside of IPython, simply calling the built-in Python profilers cProfile and
profile.
$ python -m cProfile -o demo.prof demo.py
Using the -o switch will output the profiler results to the file demo.prof to view with an external tool. This
can be useful if you wish to process the profiler output with a visualization tool.
10.2.3 Line-profiler
The profiler tells us which function takes most of the time, but not where it is called.
For this, we use the line_profiler: in the source file, we decorate a few functions that we want to inspect with
@profile (no need to import it)
@profile
def test():
u, s, v = linalg.svd(data)
pca = np.dot(u[:, :10], data)
Then we run the script using the kernprof.py program, with switches -l, --line-by-line and -v, --view
to use the line-by-line profiler and view the results in addition to saving them:
$ kernprof.py -l -v demo.py
Wrote profile results to demo.py.lprof

Timer unit: 1e-06 s
File: demo.py
Function: test at line 5
Total time: 14.2793 s
Line # Hits Time Per Hit % Time Line Contents

=========== ============ ===== ========= ======= ==== ========
5 @profile
6 def test():
7 1 19015 19015.0 0.1 data = np.random.random((5000, 100))
8 1 14242163 14242163.0 99.7 u, s, v = linalg.svd(data)
9 1 10282 10282.0 0.1 pca = np.dot(u[:10, :], data)
10 1 7799 7799.0 0.1 results = fastica(pca.T, whiten=False)
The SVD is taking all the time. We need to optimise this line.
10.3 Making code go faster
Once we have identified the bottlenecks, we need to make the corresponding code go faster.
10.3. Making code go faster 328

10.3.1 Algorithmic optimization
The first thing to look for is algorithmic optimization: are there ways to compute less, or better?
For a high-level view of the problem, a good understanding of the maths behind the algorithm helps. However,
it is not uncommon to find simple changes, like moving computation or memory allocation outside a for
loop, that bring in big gains.
Example of the SVD
In both examples above, the SVD - Singular Value Decomposition - is what takes most of the time. Indeed, the
computational cost of this algorithm is roughly n 3 in the size of the input matrix.
However, in both of these example, we are not using all the output of the SVD, but only the first few rows of
its first return argument. If we use the svd implementation of scipy, we can ask for an incomplete version of
the SVD. Note that implementations of linear algebra in scipy are richer then those in numpy and should be
preferred.
In [3]: %timeit np.linalg.svd(data)

1 loops, best of 3: 14.5 s per loop
In [4]: from scipy import linalg
In [5]: %timeit linalg.svd(data)

In [6]: %timeit linalg.svd(data, full_matrices=False)

1 loops, best of 3: 295 ms per loop
In [7]: %timeit np.linalg.svd(data, full_matrices=False)

We can then use this insight to optimize the previous code:
def test():
u, s, v = linalg.svd(data, full_matrices=False)
pca = np.dot(u[:, :10].T, data)
In [1]: import demo
In [2]: %timeit demo.

demo.fastica demo.np demo.prof.pdf demo.py demo.pyc
demo.linalg demo.prof demo.prof.png demo.py.lprof demo.test
In [2]: %timeit demo.test()

ica.py:65: RuntimeWarning: invalid value encountered in sqrt
W = (u * np.diag(1.0/np.sqrt(s)) * u.T) * W # W = (W * W.T) ^{-1/2} * W
In [3]: import demo_opt
In [4]: %timeit demo_opt.test()

Real incomplete SVDs, e.g. computing only the first 10 eigenvectors, can be computed with arpack, available
in scipy.sparse.linalg.eigsh.
10.3. Making code go faster 329

Computational linear algebra
For certain algorithms, many of the bottlenecks will be linear algebra computations. In this case, using
the right function to solve the right problem is key. For instance, an eigenvalue problem with a symmetric
matrix is easier to solve than with a general matrix. Also, most often, you can avoid inverting a matrix and
use a less costly (and more numerically stable) operation.
Know your computational linear algebra. When in doubt, explore scipy.linalg, and use %timeit to try
out different alternatives on your data.
10.4 Writing faster numerical code
A complete discussion on advanced use of numpy is found in chapter Advanced NumPy, or in the article The
NumPy array: a structure for efficient numerical computation by van der Walt et al. Here we discuss only some
commonly encountered tricks to make code faster.
Vectorizing for loops
Find tricks to avoid for loops using numpy arrays. For this, masks and indices arrays can be useful.
Broadcasting
Use broadcasting to do operations on arrays as small as possible before combining them.
In place operations
In [1]: a = np.zeros(1e7)
In [2]: %timeit global a ; a = 0*a

In [3]: %timeit global a ; a *= 0

note: we need global a in the timeit so that it work, as it is assigning to a, and thus considers it as a local
variable.
Be easy on the memory: use views, and not copies
Copying big arrays is as costly as making simple numerical operations on them:
In [1]: a = np.zeros(1e7)
In [2]: %timeit a.copy()

In [3]: %timeit a + 1
Beware of cache effects

Memory access is cheaper when it is grouped: accessing a big array in a continuous way is much faster
than random access. This implies amongst other things that smaller strides are faster (see CPU cache
effects):
In [1]: c = np.zeros((1e4, 1e4), order='C')
In [2]: %timeit c.sum(axis=0)

In [3]: %timeit c.sum(axis=1)

10.4. Writing faster numerical code 330

In [4]: c.strides
Out[4]: (80000, 8)
This is the reason why Fortran ordering or C ordering may make a big difference on operations:
In [5]: a = np.random.rand(20, 2**18)
In [6]: b = np.random.rand(20, 2**18)
In [7]: %timeit np.dot(b, a.T)

In [8]: c = np.ascontiguousarray(a.T)
In [9]: %timeit np.dot(b, c)

Note that copying the data to work around this effect may not be worth it:
In [10]: %timeit c = np.ascontiguousarray(a.T)

Using numexpr can be useful to automatically optimize code for such effects.
Use compiled code
The last resort, once you are sure that all the high-level optimizations have been explored, is to transfer
the hot spots, i.e. the few lines or functions in which most of the time is spent, to compiled code. For
compiled code, the preferred option is to use Cython: it is easy to transform exiting Python code in
compiled code, and with a good use of the numpy support yields efficient code on numpy arrays, for
instance by unrolling loops.
Warning: For all the above: profile and time your choices. Dont base your optimization on theoretical
considerations.
10.4.1 Additional Links
If you need to profile memory usage, you could try the memory_profiler
If you need to profile down into C extensions, you could try using gperftools from Python with yep.
If you would like to track performace of your code across time, i.e. as you make new commits to your
repository, you could try: vbench
If you need some interactive visualization why not try RunSnakeRun
10.4. Writing faster numerical code 331

CHAPTER 11
Sparse Matrices in SciPy
Author: Robert Cimrman
11.1 Introduction
(dense) matrix is:

mathematical object
data structure for storing a 2D array of values
important features:
memory allocated once for all items
usually a contiguous chunk, think NumPy ndarray
fast access to individual items (*)
11.1.1 Why Sparse Matrices?
the memory, that grows like n**2

small example (double precision matrix):

>>> x = np.linspace(0, 1e6, 10)
332
>>> plt.plot(x, 8.0 * (x**2) / 1e6, lw=5)

>>> plt.xlabel('size n')
>>> plt.ylabel('memory [MB]')
11.1.2 Sparse Matrices vs. Sparse Matrix Storage Schemes
sparse matrix is a matrix, which is almost empty

storing all the zeros is wasteful -> store only nonzero items
think compression
pros: huge memory savings
cons: depends on actual storage scheme, (*) usually does not hold
11.1.3 Typical Applications
solution of partial differential equations (PDEs)

the finite element method
mechanical engineering, electrotechnics, physics, . . .
graph theory
nonzero at (i, j) means that node i is connected to node j
...
11.1.4 Prerequisites
numpy
scipy
matplotlib (optional)
ipython (the enhancements come handy)
11.1.5 Sparsity Structure Visualization
spy() from matplotlib

example plots:
11.1. Introduction 333

11.2 Storage Schemes

seven sparse matrix types in scipy.sparse:
1. csc_matrix: Compressed Sparse Column format
2. csr_matrix: Compressed Sparse Row format
3. bsr_matrix: Block Sparse Row format
4. lil_matrix: List of Lists format
5. dok_matrix: Dictionary of Keys format
6. coo_matrix: COOrdinate format (aka IJV, triplet format)
7. dia_matrix: DIAgonal format
11.2. Storage Schemes 334

each suitable for some tasks

many employ sparsetools C++ module by Nathan Bell
assume the following is imported:

>>> import scipy.sparse as sps
warning for NumPy users:

the multiplication with * is the matrix multiplication (dot product)
not part of NumPy!
* passing a sparse matrix object to NumPy functions expecting ndarray/matrix does not
work
11.2.1 Common Methods
all scipy.sparse classes are subclasses of spmatrix

default implementation of arithmetic operations
* always converts to CSR

* subclasses override for efficiency
shape, data type set/get
nonzero indices
format conversion, interaction with NumPy (toarray(), todense())
...
attributes:
mtx.A - same as mtx.toarray()
mtx.T - transpose (same as mtx.transpose())
mtx.H - Hermitian (conjugate) transpose
mtx.real - real part of complex matrix
mtx.imag - imaginary part of complex matrix
mtx.size - the number of nonzeros (same as self.getnnz())
mtx.shape - the number of rows and columns (tuple)
data usually stored in NumPy arrays
11.2.2 Sparse Matrix Classes
Diagonal Format (DIA)
very simple scheme

diagonals in dense NumPy array of shape (n_diag, length)
fixed length -> waste space a bit when far from main diagonal
subclass of _data_matrix (sparse matrix classes with .data attribute)
offset for each diagonal
0 is the main diagonal

negative offset = below

positive offset = above
fast matrix * vector (sparsetools)
fast and easy item-wise operations
manipulate data array directly (fast NumPy machinery)
constructor accepts:
dense matrix (array)
sparse matrix
shape tuple (create empty matrix)
(data, offsets) tuple
no slicing, no individual item access
use:
rather specialized
solving PDEs by finite differences
with an iterative solver
Examples
create some DIA matrices:
>>> data = np.array([[1, 2, 3, 4]]).repeat(3, axis=0)

>>> data
array([[1, 2, 3, 4],
[1, 2, 3, 4],
[1, 2, 3, 4]])
>>> offsets = np.array([0, -1, 2])
>>> mtx = sparse.dia_matrix((data, offsets), shape=(4, 4))
>>> mtx
<4x4 sparse matrix of type '<... 'numpy.int64'>'
with 9 stored elements (3 diagonals) in DIAgonal format>
>>> mtx.todense()
matrix([[1, 0, 3, 0],
[1, 2, 0, 4],
[0, 2, 3, 0],
[0, 0, 3, 4]])
>>> data = np.arange(12).reshape((3, 4)) + 1

>>> data
array([[ 1, 2, 3, 4],
[ 5, 6, 7, 8],
[ 9, 10, 11, 12]])
>>> mtx = sparse.dia_matrix((data, offsets), shape=(4, 4))
>>> mtx.data
array([[ 1, 2, 3, 4],
[ 5, 6, 7, 8],
[ 9, 10, 11, 12]]...)
>>> mtx.offsets
array([ 0, -1, 2], dtype=int32)
>>> print(mtx)
(0, 0) 1
(1, 1) 2
(2, 2) 3
(3, 3) 4

(1, 0) 5
(2, 1) 6
(3, 2) 7
(0, 2) 11
(1, 3) 12
>>> mtx.todense()
matrix([[ 1, 0, 11, 0],
[ 5, 2, 0, 12],
[ 0, 6, 3, 0],
[ 0, 0, 7, 4]])
explanation with a scheme:
offset: row
2: 9
1: --10------
0: 1 . 11 .
-1: 5 2 . 12
-2: . 6 3 .
-3: . . 7 4
---------8
matrix-vector multiplication
>>> vec = np.ones((4, ))

>>> vec
array([ 1., 1., 1., 1.])
>>> mtx * vec
array([ 12., 19., 9., 11.])
>>> mtx.toarray() * vec
array([[ 1., 0., 11., 0.],
[ 5., 2., 0., 12.],
[ 0., 6., 3., 0.],
[ 0., 0., 7., 4.]])
List of Lists Format (LIL)
row-based linked list

each row is a Python list (sorted) of column indices of non-zero elements
rows stored in a NumPy array (dtype=np.object)
non-zero values data stored analogously
efficient for constructing sparse matrices incrementally
sparse matrix
flexible slicing, changing sparsity structure is efficient
slow arithmetics, slow column slicing due to being row-based
use:
when sparsity pattern is not known apriori or changes
example: reading a sparse matrix from a text file

Examples
create an empty LIL matrix:
>>> mtx = sparse.lil_matrix((4, 5))
prepare random data:
>>> from numpy.random import rand

>>> data = np.round(rand(2, 3))
>>> data
array([[ 1., 1., 1.],
[ 1., 0., 1.]])
assign the data using fancy indexing:
>>> mtx[:2, [1, 2, 3]] = data

>>> mtx
<4x5 sparse matrix of type '<... 'numpy.float64'>'
with 5 stored elements in LInked List format>
>>> print(mtx)
(0, 1) 1.0
(0, 2) 1.0
(0, 3) 1.0
(1, 1) 1.0
(1, 3) 1.0
>>> mtx.todense()
matrix([[ 0., 1., 1., 1., 0.],
[ 0., 1., 0., 1., 0.],
[ 0., 0., 0., 0., 0.],
[ 0., 0., 0., 0., 0.]])
>>> mtx.toarray()
array([[ 0., 1., 1., 1., 0.],
[ 0., 1., 0., 1., 0.],
[ 0., 0., 0., 0., 0.],
[ 0., 0., 0., 0., 0.]])
more slicing and indexing:
>>> mtx = sparse.lil_matrix([[0, 1, 2, 0], [3, 0, 1, 0], [1, 0, 0, 1]])

>>> mtx.todense()
matrix([[0, 1, 2, 0],
[3, 0, 1, 0],
[1, 0, 0, 1]]...)
>>> print(mtx)
(0, 1) 1
(0, 2) 2
(1, 0) 3
(1, 2) 1
(2, 0) 1
(2, 3) 1
>>> mtx[:2, :]
with 4 stored elements in LInked List format>
>>> mtx[:2, :].todense()
matrix([[0, 1, 2, 0],
[3, 0, 1, 0]]...)
>>> mtx[1:2, [0,2]].todense()
matrix([[3, 1]]...)
>>> mtx.todense()
matrix([[0, 1, 2, 0],
[3, 0, 1, 0],
[1, 0, 0, 1]]...)

Dictionary of Keys Format (DOK)
subclass of Python dict

keys are (row, column) index tuples (no duplicate entries allowed)
values are corresponding non-zero values
efficient for constructing sparse matrices incrementally
sparse matrix
efficient O(1) access to individual elements
flexible slicing, changing sparsity structure is efficient
can be efficiently converted to a coo_matrix once constructed
slow arithmetics (for loops with dict.iteritems())
use:
when sparsity pattern is not known apriori or changes
Examples
create a DOK matrix element by element:
>>> mtx = sparse.dok_matrix((5, 5), dtype=np.float64)

>>> mtx
with 0 stored elements in Dictionary Of Keys format>
>>> for ir in range(5):
... for ic in range(5):
... mtx[ir, ic] = 1.0 * (ir != ic)
>>> mtx
>>> mtx.todense()
matrix([[ 0., 1., 1., 1., 1.],
[ 1., 0., 1., 1., 1.],
[ 1., 1., 0., 1., 1.],
[ 1., 1., 1., 0., 1.],
[ 1., 1., 1., 1., 0.]])
slicing and indexing:
>>> mtx[1, 1]
0.0
>>> mtx[1, 1:3]
>>> mtx[1, 1:3].todense()
matrix([[ 0., 1.]])
>>> mtx[[2,1], 1:3].todense()
matrix([[ 1., 0.],
[ 0., 1.]])

Coordinate Format (COO)
also known as the ijv or triplet format

three NumPy arrays: row, col, data
data[i] is value at (row[i], col[i]) position
permits duplicate entries
subclass of _data_matrix (sparse matrix classes with .data attribute)
fast format for constructing sparse matrices
sparse matrix
(data, ij) tuple
very fast conversion to and from CSR/CSC formats
fast matrix * vector (sparsetools)
fast and easy item-wise operations
manipulate data array directly (fast NumPy machinery)
no slicing, no arithmetics (directly)
use:
facilitates fast conversion among sparse formats
when converting to other format (usually CSR or CSC), duplicate entries are summed together
* facilitates efficient construction of finite element matrices
Examples
create empty COO matrix:
>>> mtx = sparse.coo_matrix((3, 4), dtype=np.int8)

>>> mtx.todense()
matrix([[0, 0, 0, 0],
[0, 0, 0, 0],
[0, 0, 0, 0]], dtype=int8)
create using (data, ij) tuple:
>>> row = np.array([0, 3, 1, 0])

>>> col = np.array([0, 3, 1, 2])
>>> data = np.array([4, 5, 7, 9])
>>> mtx = sparse.coo_matrix((data, (row, col)), shape=(4, 4))
>>> mtx
with 4 stored elements in COOrdinate format>
>>> mtx.todense()
matrix([[4, 0, 9, 0],
[0, 7, 0, 0],
[0, 0, 0, 0],
[0, 0, 0, 5]])
duplicates entries are summed together:

>>> row = np.array([0, 0, 1, 3, 1, 0, 0])

>>> col = np.array([0, 2, 1, 3, 1, 0, 0])
>>> data = np.array([1, 1, 1, 1, 1, 1, 1])
>>> mtx = sparse.coo_matrix((data, (row, col)), shape=(4, 4))
>>> mtx.todense()
matrix([[3, 0, 1, 0],
[0, 2, 0, 0],
[0, 0, 0, 0],
[0, 0, 0, 1]])
no slicing. . . :
>>> mtx[2, 3]
...
TypeError: 'coo_matrix' object ...
Compressed Sparse Row Format (CSR)
row oriented
three NumPy arrays: indices, indptr, data
* indices is array of column indices

* data is array of corresponding nonzero values
* indptr points to row starts in indices and data
* length is n_row + 1, last item = number of values = length of both indices and data
* nonzero values of the i-th row are data[indptr[i]:indptr[i+1]] with column indices in-
dices[indptr[i]:indptr[i+1]]
* item (i, j) can be accessed as data[indptr[i]+k], where k is position of j in in-

subclass of _cs_matrix (common CSR/CSC functionality)
* subclass of _data_matrix (sparse matrix classes with .data attribute)

fast matrix vector products and other arithmetics (sparsetools)
sparse matrix
(data, ij) tuple
(data, indices, indptr) tuple
efficient row slicing, row-oriented operations
slow column slicing, expensive changes to the sparsity structure
use:
actual computations (most linear solvers support this format)

Examples
create empty CSR matrix:
>>> mtx = sparse.csr_matrix((3, 4), dtype=np.int8)

>>> mtx.todense()
matrix([[0, 0, 0, 0],
[0, 0, 0, 0],
[0, 0, 0, 0]], dtype=int8)
>>> row = np.array([0, 0, 1, 2, 2, 2])

>>> col = np.array([0, 2, 2, 0, 1, 2])
>>> data = np.array([1, 2, 3, 4, 5, 6])
>>> mtx = sparse.csr_matrix((data, (row, col)), shape=(3, 3))
>>> mtx
with 6 stored elements in Compressed Sparse Row format>
>>> mtx.todense()
matrix([[1, 0, 2],
[0, 0, 3],
[4, 5, 6]]...)
>>> mtx.data
array([1, 2, 3, 4, 5, 6]...)
>>> mtx.indices
array([0, 2, 2, 0, 1, 2], dtype=int32)
>>> mtx.indptr
create using (data, indices, indptr) tuple:
>>> data = np.array([1, 2, 3, 4, 5, 6])

>>> indices = np.array([0, 2, 2, 0, 1, 2])
>>> indptr = np.array([0, 2, 3, 6])
>>> mtx = sparse.csr_matrix((data, indices, indptr), shape=(3, 3))
>>> mtx.todense()
matrix([[1, 0, 2],
[0, 0, 3],
[4, 5, 6]])
Compressed Sparse Column Format (CSC)
column oriented
* indices is array of row indices

* data is array of corresponding nonzero values
* indptr points to column starts in indices and data
* length is n_col + 1, last item = number of values = length of both indices and data
* nonzero values of the i-th column are data[indptr[i]:indptr[i+1]] with row indices in-
* item (i, j) can be accessed as data[indptr[j]+k], where k is position of i in in-

dices[indptr[j]:indptr[j+1]]


sparse matrix
(data, ij) tuple
efficient column slicing, column-oriented operations
slow row slicing, expensive changes to the sparsity structure
use:
actual computations (most linear solvers support this format)
Examples
create empty CSC matrix:
>>> mtx = sparse.csc_matrix((3, 4), dtype=np.int8)

>>> mtx.todense()
matrix([[0, 0, 0, 0],
[0, 0, 0, 0],
[0, 0, 0, 0]], dtype=int8)
>>> row = np.array([0, 0, 1, 2, 2, 2])

>>> col = np.array([0, 2, 2, 0, 1, 2])
>>> data = np.array([1, 2, 3, 4, 5, 6])
>>> mtx = sparse.csc_matrix((data, (row, col)), shape=(3, 3))
>>> mtx
with 6 stored elements in Compressed Sparse Column format>
>>> mtx.todense()
matrix([[1, 0, 2],
[0, 0, 3],
[4, 5, 6]]...)
>>> mtx.data
array([1, 4, 5, 2, 3, 6]...)
>>> mtx.indices
array([0, 2, 2, 0, 1, 2], dtype=int32)
>>> mtx.indptr
create using (data, indices, indptr) tuple:
>>> data = np.array([1, 4, 5, 2, 3, 6])

>>> indices = np.array([0, 2, 2, 0, 1, 2])
>>> indptr = np.array([0, 2, 3, 6])
>>> mtx = sparse.csc_matrix((data, indices, indptr), shape=(3, 3))
>>> mtx.todense()
matrix([[1, 0, 2],
[0, 0, 3],
[4, 5, 6]])

Block Compressed Row Format (BSR)
basically a CSR with dense sub-matrices of fixed shape instead of scalar items
block size (R, C) must evenly divide the shape of the matrix (M, N)
* indices is array of column indices for each block

* data is array of corresponding nonzero values of shape (nnz, R, C)
* ...

sparse matrix
(data, ij) tuple
many arithmetic operations considerably more efficient than CSR for sparse matrices with dense sub-
matrices
use:
like CSR
vector-valued finite element discretizations
Examples
create empty BSR matrix with (1, 1) block size (like CSR. . . ):
>>> mtx = sparse.bsr_matrix((3, 4), dtype=np.int8)

>>> mtx
with 0 stored elements (blocksize = 1x1) in Block Sparse Row format>
>>> mtx.todense()
matrix([[0, 0, 0, 0],
[0, 0, 0, 0],
[0, 0, 0, 0]], dtype=int8)
create empty BSR matrix with (3, 2) block size:
>>> mtx = sparse.bsr_matrix((3, 4), blocksize=(3, 2), dtype=np.int8)

>>> mtx
>>> mtx.todense()
matrix([[0, 0, 0, 0],
[0, 0, 0, 0],
[0, 0, 0, 0]], dtype=int8)
a bug?
create using (data, ij) tuple with (1, 1) block size (like CSR. . . ):

>>> row = np.array([0, 0, 1, 2, 2, 2])

>>> col = np.array([0, 2, 2, 0, 1, 2])
>>> data = np.array([1, 2, 3, 4, 5, 6])
>>> mtx = sparse.bsr_matrix((data, (row, col)), shape=(3, 3))
>>> mtx
>>> mtx.todense()
matrix([[1, 0, 2],
[0, 0, 3],
[4, 5, 6]]...)
>>> mtx.data
array([[[1]],
[[2]],
[[3]],
[[4]],
[[5]],
[[6]]]...)
>>> mtx.indices
array([0, 2, 2, 0, 1, 2], dtype=int32)
>>> mtx.indptr
create using (data, indices, indptr) tuple with (2, 2) block size:
>>> indptr = np.array([0, 2, 3, 6])

>>> indices = np.array([0, 2, 2, 0, 1, 2])
>>> data = np.array([1, 2, 3, 4, 5, 6]).repeat(4).reshape(6, 2, 2)
>>> mtx = sparse.bsr_matrix((data, indices, indptr), shape=(6, 6))
>>> mtx.todense()
matrix([[1, 1, 0, 0, 2, 2],
[1, 1, 0, 0, 2, 2],
[0, 0, 0, 0, 3, 3],
[0, 0, 0, 0, 3, 3],
[4, 4, 5, 5, 6, 6],
[4, 4, 5, 5, 6, 6]])
>>> data
array([[[1, 1],
[1, 1]],
[[2, 2],
[2, 2]],
[[3, 3],
[3, 3]],
[[4, 4],
[4, 4]],
[[5, 5],
[5, 5]],
[[6, 6],
[6, 6]]])

11.2.3 Summary
Table 11.1: Summary of storage schemes.

format matrix get fancy set item fancy solvers note
* vector item get set
DIA sparsetools. . . . iterative has data array, specialized
LIL via CSR yes yes yes yes iterative arithmetics via CSR, incre-
mental construction
DOK python yes one yes yes iterative O(1) item access, incremental
axis construction
only
COO sparsetools. . . . iterative has data array, facilitates fast
conversion
CSR sparsetoolsyes yes slow . any has data array, fast row-wise
ops
CSC sparsetoolsyes yes slow . any has data array, fast column-
wise ops
BSR sparsetools. . . . specializedhas data array, specialized
11.3 Linear System Solvers
sparse matrix/eigenvalue problem solvers live in scipy.sparse.linalg

the submodules:
dsolve: direct factorization methods for solving linear systems
isolve: iterative methods for solving linear systems
eigen: sparse eigenvalue problem solvers
all solvers are accessible from:
>>> import scipy.sparse.linalg as spla

>>> spla.__all__
['LinearOperator', 'Tester', 'arpack', 'aslinearoperator', 'bicg',
'bicgstab', 'cg', 'cgs', 'csc_matrix', 'csr_matrix', 'dsolve',
'eigen', 'eigen_symmetric', 'factorized', 'gmres', 'interface',
'isolve', 'iterative', 'lgmres', 'linsolve', 'lobpcg', 'lsqr',
'minres', 'np', 'qmr', 'speigs', 'spilu', 'splu', 'spsolve', 'svd',
'test', 'umfpack', 'use_solver', 'utils', 'warnings']
11.3.1 Sparse Direct Solvers
default solver: SuperLU 4.0

included in SciPy
real and complex systems
both single and double precision
optional: umfpack
real and complex systems
double precision only
recommended for performance
wrappers now live in scikits.umfpack
11.3. Linear System Solvers 346

check-out the new scikits.suitesparse by Nathaniel Smith
Examples
import the whole module, and see its docstring:
>>> from scipy.sparse.linalg import dsolve

>>> help(dsolve)
both superlu and umfpack can be used (if the latter is installed) as follows:
prepare a linear system:

>>> from scipy import sparse
>>> mtx = sparse.spdiags([[1, 2, 3, 4, 5], [6, 5, 8, 9, 10]], [0, 1], 5, 5)
>>> mtx.todense()
matrix([[ 1, 5, 0, 0, 0],
[ 0, 2, 8, 0, 0],
[ 0, 0, 3, 9, 0],
[ 0, 0, 0, 4, 10],
[ 0, 0, 0, 0, 5]])
>>> rhs = np.array([1, 2, 3, 4, 5], dtype=np.float32)
solve as single precision real:
>>> mtx1 = mtx.astype(np.float32)

>>> x = dsolve.spsolve(mtx1, rhs, use_umfpack=False)
>>> print(x)
[ 106. -21. 5.5 -1.5 1. ]
>>> print("Error: %s " % (mtx1 * x - rhs))
Error: [ 0. 0. 0. 0. 0.]
solve as double precision real:
>>> mtx2 = mtx.astype(np.float64)

>>> x = dsolve.spsolve(mtx2, rhs, use_umfpack=True)
>>> print(x)
[ 106. -21. 5.5 -1.5 1. ]
Error: [ 0. 0. 0. 0. 0.]
solve as single precision complex:
>>> mtx1 = mtx.astype(np.complex64)

>>> x = dsolve.spsolve(mtx1, rhs, use_umfpack=False)
>>> print(x)
[ 106.0+0.j -21.0+0.j 5.5+0.j -1.5+0.j 1.0+0.j]
Error: [ 0.+0.j 0.+0.j 0.+0.j 0.+0.j 0.+0.j]
solve as double precision complex:
>>> mtx2 = mtx.astype(np.complex128)

>>> x = dsolve.spsolve(mtx2, rhs, use_umfpack=True)
>>> print(x)
[ 106.0+0.j -21.0+0.j 5.5+0.j -1.5+0.j 1.0+0.j]
Error: [ 0.+0.j 0.+0.j 0.+0.j 0.+0.j 0.+0.j]
"""
Solve a linear system

=======================
Construct a 1000x1000 lil_matrix and add some values to it, convert it

to CSR format and solve A x = b for x:and solve a linear system with a
direct solver.
"""
import numpy as np
import scipy.sparse as sps
from scipy.sparse.linalg.dsolve import linsolve
rand = np.random.rand
mtx = sps.lil_matrix((1000, 1000), dtype=np.float64)

mtx[0, :100] = rand(100)
mtx[1, 100:200] = mtx[0, :100]
mtx.setdiag(rand(1000))
plt.clf()
plt.spy(mtx, marker='.', markersize=2)
plt.show()
mtx = mtx.tocsr()
rhs = rand(1000)
x = linsolve.spsolve(mtx, rhs)
print('rezidual: %r ' % np.linalg.norm(mtx * x - rhs))
examples/direct_solve.py
11.3.2 Iterative Solvers
the isolve module contains the following solvers:

bicg (BIConjugate Gradient)
bicgstab (BIConjugate Gradient STABilized)
cg (Conjugate Gradient) - symmetric positive definite matrices only
cgs (Conjugate Gradient Squared)
gmres (Generalized Minimal RESidual)
minres (MINimum RESidual)
qmr (Quasi-Minimal Residual)
Common Parameters
mandatory:
A [{sparse matrix, dense matrix, LinearOperator}] The N-by-N matrix of the linear system.
b [{array, matrix}] Right hand side of the linear system. Has shape (N,) or (N,1).
optional:
x0 [{array, matrix}] Starting guess for the solution.
tol [float] Relative tolerance to achieve before terminating.
maxiter [integer] Maximum number of iterations. Iteration will stop after maxiter steps even if the spec-
ified tolerance has not been achieved.

M [{sparse matrix, dense matrix, LinearOperator}] Preconditioner for A. The preconditioner should ap-
proximate the inverse of A. Effective preconditioning dramatically improves the rate of conver-
gence, which implies that fewer iterations are needed to reach a given error tolerance.
callback [function] User-supplied function to call after each iteration. It is called as callback(xk), where
xk is the current solution vector.
LinearOperator Class
from scipy.sparse.linalg.interface import LinearOperator
common interface for performing matrix vector products

useful abstraction that enables using dense and sparse matrices within the solvers, as well as matrix-free
solutions
has shape and matvec() (+ some optional parameters)
example:

>>> from scipy.sparse.linalg import LinearOperator
>>> def mv(v):
... return np.array([2*v[0], 3*v[1]])
...
>>> A = LinearOperator((2, 2), matvec=mv)
>>> A
<2x2 _CustomLinearOperator with dtype=float64>
>>> A.matvec(np.ones(2))
array([ 2., 3.])
>>> A * np.ones(2)
array([ 2., 3.])
A Few Notes on Preconditioning
problem specific
often hard to develop
if not sure, try ILU
available in dsolve as spilu()
11.3.3 Eigenvalue Problem Solvers
The eigen module
arpack * a collection of Fortran77 subroutines designed to solve large scale eigenvalue problems
lobpcg (Locally Optimal Block Preconditioned Conjugate Gradient Method) * works very well in combi-
nation with PyAMG * example by Nathan Bell:
"""
Compute eigenvectors and eigenvalues using a preconditioned eigensolver
========================================================================
In this example Smoothed Aggregation (SA) is used to precondition

the LOBPCG eigensolver on a two-dimensional Poisson problem with
Dirichlet boundary conditions.
"""

import scipy
from scipy.sparse.linalg import lobpcg
from pyamg import smoothed_aggregation_solver

from pyamg.gallery import poisson
N = 100
K = 9
A = poisson((N,N), format='csr')
# create the AMG hierarchy

ml = smoothed_aggregation_solver(A)
# initial approximation to the K eigenvectors

X = scipy.rand(A.shape[0], K)
# preconditioner based on ml
M = ml.aspreconditioner()
# compute eigenvalues and eigenvectors with LOBPCG

W,V = lobpcg(A, X, M=M, tol=1e-8, largest=False)
#plot the eigenvectors

import pylab
pylab.figure(figsize=(9,9))
for i in range(K):
pylab.subplot(3, 3, i+1)
pylab.title('Eigenvector %d ' % i)
pylab.pcolor(V[:,i].reshape(N,N))
pylab.axis('equal')
pylab.axis('off')
pylab.show()
examples/pyamg_with_lobpcg.py
example by Nils Wagner:
examples/lobpcg_sakurai.py
output:
$ python examples/lobpcg_sakurai.py
Results by LOBPCG for n=2500
[ 0.06250083 0.06250028 0.06250007]
Exact eigenvalues
[ 0.06250005 0.0625002 0.06250044]
Elapsed time 7.01

11.4 Other Interesting Packages

PyAMG
algebraic multigrid solvers
https://github.com/pyamg/pyamg
Pysparse
own sparse matrix classes
matrix and eigenvalue problem solvers
http://pysparse.sourceforge.net/
11.4. Other Interesting Packages 351

CHAPTER 12
Image manipulation and processing
using Numpy and Scipy
Authors: Emmanuelle Gouillart, Gal Varoquaux

This section addresses basic image manipulation and processing using the core scientific modules NumPy and
SciPy. Some of the operations covered by this tutorial may be useful for other kinds of multidimensional array
processing than image processing. In particular, the submodule scipy.ndimage provides functions operating
on n-dimensional NumPy arrays.
See also:
For more advanced image processing and image-specific routines, see the tutorial Scikit-image: image pro-
cessing, dedicated to the skimage module.
Image = 2-D numerical array
(or 3-D: CT, MRI, 2D + time; 4-D, . . . )

Here, image == Numpy array np.array
Tools used in this tutorial:

numpy: basic array manipulation
scipy: scipy.ndimage submodule dedicated to image processing (n-dimensional images). See the
documentation:
>>> from scipy import ndimage
Common tasks in image processing:
352
Input/Output, displaying images

Basic manipulations: cropping, flipping, rotating, . . .
Image filtering: denoising, sharpening
Image segmentation: labeling pixels corresponding to different objects
Classification
Feature extraction
Registration
...
Chapters contents
Opening and writing to image files

Displaying images
Basic manipulations
Statistical information
Geometrical transformations
Image filtering
Blurring/smoothing
Sharpening
Denoising
Mathematical morphology
Feature extraction
Edge detection
Segmentation
Measuring objects properties: ndimage.measurements
Full code examples
Examples for the image processing chapter
12.1 Opening and writing to image files
Writing an array to a file:

f = misc.face()
misc.imsave('face.png', f) # uses the Image module (PIL)

plt.imshow(f)
plt.show()
12.1. Opening and writing to image files 353

Creating a numpy array from an image file:

>>> face = misc.face()
>>> misc.imsave('face.png', face) # First we need to create the PNG file
>>> face = misc.imread('face.png')

>>> type(face)
<... 'numpy.ndarray'>
>>> face.shape, face.dtype
((768, 1024, 3), dtype('uint8'))
dtype is uint8 for 8-bit images (0-255)

Opening raw files (camera, 3-D images)
>>> face.tofile('face.raw') # Create raw file

>>> face_from_raw = np.fromfile('face.raw', dtype=np.uint8)
>>> face_from_raw.shape
(2359296,)
>>> face_from_raw.shape = (768, 1024, 3)
Need to know the shape and dtype of the image (how to separate data bytes).
For large data, use np.memmap for memory mapping:
>>> face_memmap = np.memmap('face.raw', dtype=np.uint8, shape=(768, 1024, 3))
(data are read from the file, and not loaded into memory)
Working on a list of image files
12.1. Opening and writing to image files 354


... im = np.random.randint(0, 256, 10000).reshape((100, 100))
... misc.imsave('random_%02d .png' % i, im)
>>> from glob import glob
>>> filelist = glob('random*.png')
>>> filelist.sort()
12.2 Displaying images
Use matplotlib and imshow to display an image inside a matplotlib figure:
>>> f = misc.face(gray=True) # retrieve a grayscale image

>>> plt.imshow(f, cmap=plt.cm.gray)
Increase contrast by setting min and max values:
>>> plt.imshow(f, cmap=plt.cm.gray, vmin=30, vmax=200)

>>> # Remove axes and ticks
>>> plt.axis('off')
(-0.5, 1023.5, 767.5, -0.5)
Draw contour lines:
>>> plt.contour(f, [50, 200])

<matplotlib.contour.QuadContourSet ...>
For fine inspection of intensity variations, use interpolation='nearest':
>>> plt.imshow(f[320:340, 510:530], cmap=plt.cm.gray)

>>> plt.imshow(f[320:340, 510:530], cmap=plt.cm.gray, interpolation='nearest')
See also:
3-D visualization: Mayavi
See 3D plotting with Mayavi.
Image plane widgets
Isosurfaces
12.2. Displaying images 355

...
12.3 Basic manipulations
Images are arrays: use the whole numpy machinery.
12.3. Basic manipulations 356


>>> face[0, 40]
127
>>> # Slicing
>>> face[10:13, 20:23]
array([[141, 153, 145],
[133, 134, 125],
[ 96, 92, 94]], dtype=uint8)
>>> face[100:120] = 255
>>>
>>> lx, ly = face.shape
>>> X, Y = np.ogrid[0:lx, 0:ly]
>>> mask = (X - lx / 2) ** 2 + (Y - ly / 2) ** 2 > lx * ly / 4
>>> # Masks
>>> face[mask] = 0
>>> # Fancy indexing
>>> face[range(400), range(400)] = 255

12.3.1 Statistical information

>>> face.mean()
113.48026784261067
>>> face.max(), face.min()
(250, 0)
np.histogram
Exercise
Open as an array the scikit-image logo (http://scikit-image.org/_static/img/logo.png), or an image

that you have on your computer.
Crop a meaningful part of the image, for example the python circle in the logo.
Display the image array using matplotlib. Change the interpolation method and zoom to see the
difference.
Transform your image to greyscale
Increase the contrast of the image by changing its minimum and maximum values. Optional: use
scipy.stats.scoreatpercentile (read the docstring!) to saturate 5% of the darkest pixels and 5%
of the lightest pixels.
Save the array to two different file formats (png, jpg, tiff)
12.3.2 Geometrical transformations

>>> lx, ly = face.shape
>>> # Cropping
>>> crop_face = face[lx // 4: - lx // 4, ly // 4: - ly // 4]
>>> # up <-> down flip
>>> flip_ud_face = np.flipud(face)
>>> # rotation
>>> rotate_face = ndimage.rotate(face, 45)
>>> rotate_face_noreshape = ndimage.rotate(face, 45, reshape=False)

12.4 Image filtering
Local filters: replace the value of pixels by a function of the values of neighboring pixels.
Neighbourhood: square (choose size), disk, or more complicated structuring element.
12.4.1 Blurring/smoothing
Gaussian filter from scipy.ndimage:

>>> blurred_face = ndimage.gaussian_filter(face, sigma=3)
>>> very_blurred = ndimage.gaussian_filter(face, sigma=5)
Uniform filter
>>> local_mean = ndimage.uniform_filter(face, size=11)
12.4.2 Sharpening
Sharpen a blurred image:

>>> face = misc.face(gray=True).astype(float)
>>> blurred_f = ndimage.gaussian_filter(face, 3)
increase the weight of edges by adding an approximation of the Laplacian:
>>> filter_blurred_f = ndimage.gaussian_filter(blurred_f, 1)

>>> alpha = 30
>>> sharpened = blurred_f + alpha * (blurred_f - filter_blurred_f)
12.4.3 Denoising
Noisy face:
12.4. Image filtering 359


>>> f = misc.face(gray=True)
>>> f = f[230:290, 220:320]
>>> noisy = f + 0.4 * f.std() * np.random.random(f.shape)
A Gaussian filter smoothes the noise out. . . and the edges as well:
>>> gauss_denoised = ndimage.gaussian_filter(noisy, 2)
Most local linear isotropic filters blur the image (ndimage.uniform_filter)

A median filter preserves better the edges:
>>> med_denoised = ndimage.median_filter(noisy, 3)
Median filter: better result for straight boundaries (low curvature):
>>> im = np.zeros((20, 20))

>>> im[5:-5, 5:-5] = 1
>>> im = ndimage.distance_transform_bf(im)
>>> im_noise = im + 0.2 * np.random.randn(*im.shape)
>>> im_med = ndimage.median_filter(im_noise, 3)

Other rank filter: ndimage.maximum_filter, ndimage.percentile_filter

Other local non-linear filters: Wiener (scipy.signal.wiener), etc.
Non-local filters
Exercise: denoising
Create a binary image (of 0s and 1s) with several objects (circles, ellipses, squares, or random shapes).
Add some noise (e.g., 20% of noise)
Try two different denoising methods for denoising the image: gaussian filtering and median filtering.
Compare the histograms of the two different denoised images. Which one is the closest to the his-
togram of the original (noise-free) image?
See also:
More denoising filters are available in skimage.denoising, see the Scikit-image: image processing tutorial.
See wikipedia for a definition of mathematical morphology.

Probe an image with a simple shape (a structuring element), and modify this image according to how the
shape locally fits or misses the image.
Structuring element:
>>> el = ndimage.generate_binary_structure(2, 1)
>>> el
array([[False, True, False],
[ True, True, True],
[False, True, False]], dtype=bool)
>>> el.astype(np.int)
array([[0, 1, 0],
[1, 1, 1],
[0, 1, 0]])
Erosion = minimum filter. Replace the value of a pixel by the minimal value covered by the structuring ele-
ment.:
>>> a = np.zeros((7,7), dtype=np.int)

>>> a[1:6, 2:5] = 1
>>> a
array([[0, 0, 0, 0, 0, 0, 0],
[0, 0, 1, 1, 1, 0, 0],

[0, 0, 1, 1, 1, 0, 0],
[0, 0, 1, 1, 1, 0, 0],
[0, 0, 1, 1, 1, 0, 0],
[0, 0, 1, 1, 1, 0, 0],
[0, 0, 0, 0, 0, 0, 0]])
>>> ndimage.binary_erosion(a).astype(a.dtype)
array([[0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 1, 0, 0, 0],
[0, 0, 0, 1, 0, 0, 0],
[0, 0, 0, 1, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0]])
>>> ndimage.binary_erosion(a, structure=np.ones((5,5))).astype(a.dtype)
array([[0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0]])
Dilation: maximum filter:
>>> a = np.zeros((5, 5))

>>> a[2, 2] = 1
>>> a
array([[ 0., 0., 0., 0., 0.],
[ 0., 0., 0., 0., 0.],
[ 0., 0., 1., 0., 0.],
[ 0., 0., 0., 0., 0.],
[ 0., 0., 0., 0., 0.]])
>>> ndimage.binary_dilation(a).astype(a.dtype)
array([[ 0., 0., 0., 0., 0.],
[ 0., 0., 1., 0., 0.],
[ 0., 1., 1., 1., 0.],
[ 0., 0., 1., 0., 0.],
[ 0., 0., 0., 0., 0.]])
Also works for grey-valued images:
>>> im = np.zeros((64, 64))
>>> x, y = (63*np.random.random((2, 8))).astype(np.int)
>>> im[x, y] = np.arange(8)
>>> bigger_points = ndimage.grey_dilation(im, size=(5, 5), structure=np.ones((5, 5)))

>>> square = np.zeros((16, 16))

>>> square[4:-4, 4:-4] = 1
>>> dist = ndimage.distance_transform_bf(square)
>>> dilate_dist = ndimage.grey_dilation(dist, size=(3, 3), \
... structure=np.ones((3, 3)))
Opening: erosion + dilation:

>>> a[1:4, 1:4] = 1; a[4, 4] = 1
>>> a
array([[0, 0, 0, 0, 0],
[0, 1, 1, 1, 0],
[0, 1, 1, 1, 0],
[0, 1, 1, 1, 0],
[0, 0, 0, 0, 1]])
>>> # Opening removes small objects
>>> ndimage.binary_opening(a, structure=np.ones((3,3))).astype(np.int)
array([[0, 0, 0, 0, 0],
[0, 1, 1, 1, 0],
[0, 1, 1, 1, 0],
[0, 1, 1, 1, 0],
[0, 0, 0, 0, 0]])
>>> # Opening can also smooth corners
>>> ndimage.binary_opening(a).astype(np.int)
array([[0, 0, 0, 0, 0],
[0, 0, 1, 0, 0],
[0, 1, 1, 1, 0],
[0, 0, 1, 0, 0],
[0, 0, 0, 0, 0]])
Application: remove noise:
>>> square = np.zeros((32, 32))

>>> square[10:-10, 10:-10] = 1
>>> x, y = (32*np.random.random((2, 20))).astype(np.int)
>>> square[x, y] = 1
>>> open_square = ndimage.binary_opening(square)
>>> eroded_square = ndimage.binary_erosion(square)

>>> reconstruction = ndimage.binary_propagation(eroded_square, mask=square)

Closing: dilation + erosion

Many other mathematical morphology operations: hit and miss transform, tophat, etc.
12.5 Feature extraction
12.5.1 Edge detection
Synthetic data:
>>> im = np.zeros((256, 256))

>>> im[64:-64, 64:-64] = 1
>>>
>>> im = ndimage.rotate(im, 15, mode='constant')
>>> im = ndimage.gaussian_filter(im, 8)
Use a gradient operator (Sobel) to find high intensity variations:
>>> sx = ndimage.sobel(im, axis=0, mode='constant')

>>> sy = ndimage.sobel(im, axis=1, mode='constant')
>>> sob = np.hypot(sx, sy)
12.5.2 Segmentation
Histogram-based segmentation (no spatial information)
>>> n = 10
>>> l = 256
>>> im = np.zeros((l, l))
>>> points = l*np.random.random((2, n**2))
>>> im[(points[0]).astype(np.int), (points[1]).astype(np.int)] = 1
>>> im = ndimage.gaussian_filter(im, sigma=l/(4.*n))
>>> mask = (im > im.mean()).astype(np.float)

>>> mask += 0.1 * im
>>> img = mask + 0.2*np.random.randn(*mask.shape)
>>> hist, bin_edges = np.histogram(img, bins=60)

>>> bin_centers = 0.5*(bin_edges[:-1] + bin_edges[1:])
>>> binary_img = img > 0.5
Use mathematical morphology to clean up the result:
12.5. Feature extraction 364

>>> # Remove small white regions

>>> open_img = ndimage.binary_opening(binary_img)
>>> # Remove small black hole
>>> close_img = ndimage.binary_closing(open_img)
Exercise
Check that reconstruction operations (erosion + propagation) produce a better result than opening/closing:
>>> eroded_img = ndimage.binary_erosion(binary_img)
>>> reconstruct_img = ndimage.binary_propagation(eroded_img, mask=binary_img)
>>> tmp = np.logical_not(reconstruct_img)
>>> eroded_tmp = ndimage.binary_erosion(tmp)
>>> reconstruct_final = np.logical_not(ndimage.binary_propagation(eroded_tmp, mask=tmp))
>>> np.abs(mask - close_img).mean()
0.00727836...
>>> np.abs(mask - reconstruct_final).mean()
0.00059502...
Exercise
Check how a first denoising step (e.g. with a median filter) modifies the histogram, and check that the
resulting histogram-based segmentation is more accurate.
See also:
More advanced segmentation algorithms are found in the scikit-image: see Scikit-image: image processing.
See also:
Other Scientific Packages provide algorithms that can be useful for image processing. In this example, we use
12.5. Feature extraction 365

the spectral clustering function of the scikit-learn in order to segment glued objects.
>>> from sklearn.feature_extraction import image

>>> from sklearn.cluster import spectral_clustering
>>> l = 100
>>> x, y = np.indices((l, l))
>>> center1 = (28, 24)

>>> center2 = (40, 50)
>>> center3 = (67, 58)
>>> center4 = (24, 70)
>>> radius1, radius2, radius3, radius4 = 16, 14, 15, 14
>>> circle1 = (x - center1[0])**2 + (y - center1[1])**2 < radius1**2

>>> # 4 circles
>>> img = circle1 + circle2 + circle3 + circle4
>>> mask = img.astype(bool)
>>> img = img.astype(float)
>>> img += 1 + 0.2*np.random.randn(*img.shape)

>>> # Convert the image into a graph with the value of the gradient on
>>> # the edges.
>>> graph = image.img_to_graph(img, mask=mask)
>>> # Take a decreasing function of the gradient: we take it weakly

>>> # dependant from the gradient the segmentation is close to a voronoi
>>> graph.data = np.exp(-graph.data/graph.data.std())
>>> labels = spectral_clustering(graph, n_clusters=4, eigen_solver='arpack')

>>> label_im = -np.ones(mask.shape)
>>> label_im[mask] = labels
12.6 Measuring objects properties: ndimage.measurements
Synthetic data:
>>> n = 10
>>> l = 256
12.6. Measuring objects properties: ndimage.measurements 366


>>> mask = im > im.mean()
Analysis of connected components

Label connected components: ndimage.label:
>>> label_im, nb_labels = ndimage.label(mask)

>>> nb_labels # how many regions?
28
>>> plt.imshow(label_im)
Compute size, mean_value, etc. of each region:
>>> sizes = ndimage.sum(mask, label_im, range(nb_labels + 1))

>>> mean_vals = ndimage.sum(im, label_im, range(1, nb_labels + 1))
Clean up small connect components:
>>> mask_size = sizes < 1000

>>> remove_pixel = mask_size[label_im]
>>> remove_pixel.shape
(256, 256)
>>> label_im[remove_pixel] = 0
>>> plt.imshow(label_im)
Now reassign labels with np.searchsorted:
>>> labels = np.unique(label_im)

>>> label_im = np.searchsorted(labels, label_im)
Find region of interest enclosing object:
>>> slice_x, slice_y = ndimage.find_objects(label_im==4)[0]

>>> roi = im[slice_x, slice_y]
>>> plt.imshow(roi)
Other spatial measures: ndimage.center_of_mass, ndimage.maximum_position, etc.

Can be used outside the limited scope of segmentation applications.
Example: block mean:



>>> f = misc.face(gray=True)
>>> sx, sy = f.shape
>>> X, Y = np.ogrid[0:sx, 0:sy]
>>> regions = (sy//6) * (X//4) + (Y//6) # note that we use broadcasting
>>> block_mean = ndimage.mean(f, labels=regions, index=np.arange(1,
... regions.max() +1))
>>> block_mean.shape = (sx // 4, sy // 6)
When regions are regular blocks, it is more efficient to use stride tricks (Example: fake dimensions with strides).
Non-regularly-spaced blocks: radial mean:
>>> sx, sy = f.shape

>>> X, Y = np.ogrid[0:sx, 0:sy]
>>> r = np.hypot(X - sx/2, Y - sy/2)
>>> rbin = (20* r/r.max()).astype(np.int)
>>> radial_mean = ndimage.mean(f, labels=rbin, index=np.arange(1, rbin.max() +1))
Other measures
Correlation function, Fourier/wavelet spectrum, etc.
One example with mathematical morphology: granulometry
>>> def disk_structure(n):

... struct = np.zeros((2 * n + 1, 2 * n + 1))
... x, y = np.indices((2 * n + 1, 2 * n + 1))
... mask = (x - n)**2 + (y - n)**2 <= n**2
... struct[mask] = 1
... return struct.astype(np.bool)
...
>>>
>>> def granulometry(data, sizes=None):
... s = max(data.shape)
... if sizes is None:
... sizes = range(1, s/2, 2)
... granulo = [ndimage.binary_opening(data, \
... structure=disk_structure(n)).sum() for n in sizes]

... return granulo

...
>>>
>>> n = 10
>>> l = 256
>>>
>>> mask = im > im.mean()
>>>
>>> granulo = granulometry(mask, sizes=np.arange(2, 19, 4))

12.8 Examples for the image processing chapter
12.8.1 Displaying a Racoon Face
Small example to plot a racoon face.

f = misc.face()
misc.imsave('face.png', f) # uses the Image module (PIL)

plt.imshow(f)
plt.show()

Download Python source code: plot_face.py
Download Jupyter notebook: plot_face.ipynb
12.8.2 Image interpolation
The example demonstrates image interpolation on a Racoon face.

import scipy.misc
f = scipy.misc.face(gray=True)
plt.imshow(f[320:340, 510:530], cmap=plt.cm.gray)
plt.axis('off')
plt.imshow(f[320:340, 510:530], cmap=plt.cm.gray, interpolation='nearest')
plt.axis('off')
plt.subplots_adjust(wspace=0.02, hspace=0.02, top=1, bottom=0, left=0, right=1)

plt.show()

Download Python source code: plot_interpolation_face.py
Download Jupyter notebook: plot_interpolation_face.ipynb
12.8.3 Image manipulation and numpy arrays
This example shows how to do image manipulation using common numpy arrays tricks.
12.8. Examples for the image processing chapter 372

import numpy as np
import scipy
import scipy.misc
face = scipy.misc.face(gray=True)
face[10:13, 20:23]
face[100:120] = 255
lx, ly = face.shape
X, Y = np.ogrid[0:lx, 0:ly]
mask = (X - lx/2)**2 + (Y - ly/2)**2 > lx*ly/4
face[mask] = 0
face[range(400), range(400)] = 255
plt.axes([0, 0, 1, 1])
plt.imshow(face, cmap=plt.cm.gray)
plt.axis('off')
plt.show()

Download Python source code: plot_numpy_array.py
Download Jupyter notebook: plot_numpy_array.ipynb
12.8.4 Radial mean
This example shows how to do a radial mean with scikit-image.

import numpy as np
import scipy
sx, sy = f.shape
X, Y = np.ogrid[0:sx, 0:sy]
r = np.hypot(X - sx/2, Y - sy/2)
rbin = (20* r/r.max()).astype(np.int)

radial_mean = ndimage.mean(f, labels=rbin, index=np.arange(1, rbin.max() +1))
plt.axes([0, 0, 1, 1])
plt.imshow(rbin, cmap=plt.cm.spectral)
plt.axis('off')
plt.show()

Download Python source code: plot_radial_mean.py
Download Jupyter notebook: plot_radial_mean.ipynb

12.8.5 Plot the block mean of an image
An example showing how to use broad-casting to plot the mean of blocks of an image.
import numpy as np
import scipy.misc
sx, sy = f.shape
X, Y = np.ogrid[0:sx, 0:sy]
regions = sy//6 * (X//4) + Y//6

block_mean = ndimage.mean(f, labels=regions,
index=np.arange(1, regions.max() +1))
block_mean.shape = (sx//4, sy//6)
plt.imshow(block_mean, cmap=plt.cm.gray)
plt.axis('off')
plt.show()

Download Python source code: plot_block_mean.py
Download Jupyter notebook: plot_block_mean.ipynb

12.8.6 Display a Racoon Face
An example that displays a racoon face with matplotlib.
import scipy.misc
plt.subplot(131)
plt.imshow(f, cmap=plt.cm.gray)
plt.subplot(132)
plt.imshow(f, cmap=plt.cm.gray, vmin=30, vmax=200)
plt.axis('off')
plt.subplot(133)
plt.contour(f, [50, 200])
plt.axis('off')
plt.subplots_adjust(wspace=0, hspace=0., top=0.99, bottom=0.01, left=0.05,

right=0.99)
plt.show()

Download Python source code: plot_display_face.py
Download Jupyter notebook: plot_display_face.ipynb
12.8.7 Image sharpening
This example shows how to sharpen an image in noiseless situation by applying the filter inverse to the blur.

import scipy
f = scipy.misc.face(gray=True).astype(float)
blurred_f = ndimage.gaussian_filter(f, 3)
filter_blurred_f = ndimage.gaussian_filter(blurred_f, 1)
alpha = 30
sharpened = blurred_f + alpha * (blurred_f - filter_blurred_f)
plt.subplot(131)
plt.axis('off')
plt.subplot(132)
plt.imshow(blurred_f, cmap=plt.cm.gray)
plt.axis('off')
plt.subplot(133)
plt.imshow(sharpened, cmap=plt.cm.gray)
plt.axis('off')
plt.tight_layout()
plt.show()

Download Python source code: plot_sharpen.py
Download Jupyter notebook: plot_sharpen.ipynb
12.8.8 Blurring of images
An example showing various processes that blur an image.

import scipy.misc
blurred_face = ndimage.gaussian_filter(face, sigma=3)
very_blurred = ndimage.gaussian_filter(face, sigma=5)
local_mean = ndimage.uniform_filter(face, size=11)
plt.subplot(131)
plt.imshow(blurred_face, cmap=plt.cm.gray)
plt.axis('off')
plt.subplot(132)
plt.imshow(very_blurred, cmap=plt.cm.gray)
plt.axis('off')
plt.subplot(133)
plt.imshow(local_mean, cmap=plt.cm.gray)
plt.axis('off')
plt.subplots_adjust(wspace=0, hspace=0., top=0.99, bottom=0.01,

left=0.01, right=0.99)
plt.show()

Download Python source code: plot_blur.py
Download Jupyter notebook: plot_blur.ipynb
12.8.9 Synthetic data
The example generates and displays simple synthetic data.

import numpy as np
np.random.seed(1)
n = 10
l = 256
im = np.zeros((l, l))
points = l*np.random.random((2, n**2))
im[(points[0]).astype(np.int), (points[1]).astype(np.int)] = 1
im = ndimage.gaussian_filter(im, sigma=l/(4.*n))
mask = im > im.mean()
label_im, nb_labels = ndimage.label(mask)
plt.figure(figsize=(9,3))
plt.subplot(131)
plt.imshow(im)
plt.axis('off')
plt.subplot(132)
plt.axis('off')
plt.subplot(133)
plt.imshow(label_im, cmap=plt.cm.spectral)
plt.axis('off')

plt.show()

Download Python source code: plot_synthetic_data.py
Download Jupyter notebook: plot_synthetic_data.ipynb
12.8.10 Opening, erosion, and propagation
This example shows simple operations of mathematical morphology.

import numpy as np
square = np.zeros((32, 32))

square[10:-10, 10:-10] = 1
np.random.seed(2)
x, y = (32*np.random.random((2, 20))).astype(np.int)
square[x, y] = 1
open_square = ndimage.binary_opening(square)
eroded_square = ndimage.binary_erosion(square)
reconstruction = ndimage.binary_propagation(eroded_square, mask=square)
plt.figure(figsize=(9.5, 3))
plt.subplot(131)
plt.imshow(square, cmap=plt.cm.gray, interpolation='nearest')
plt.axis('off')
plt.subplot(132)
plt.imshow(open_square, cmap=plt.cm.gray, interpolation='nearest')
plt.axis('off')
plt.subplot(133)
plt.imshow(reconstruction, cmap=plt.cm.gray, interpolation='nearest')
plt.axis('off')
plt.subplots_adjust(wspace=0, hspace=0.02, top=0.99, bottom=0.01, left=0.01, right=0.99)

plt.show()

Download Python source code: plot_propagation.py
Download Jupyter notebook: plot_propagation.ipynb
12.8.11 Image denoising
This example demoes image denoising on a Racoon face.

import numpy as np
import scipy
import scipy.misc
f = f[230:290, 220:320]
noisy = f + 0.4*f.std()*np.random.random(f.shape)
gauss_denoised = ndimage.gaussian_filter(noisy, 2)
med_denoised = ndimage.median_filter(noisy, 3)
plt.figure(figsize=(12,2.8))
plt.subplot(131)
plt.imshow(noisy, cmap=plt.cm.gray, vmin=40, vmax=220)
plt.axis('off')
plt.title('noisy', fontsize=20)
plt.subplot(132)
plt.imshow(gauss_denoised, cmap=plt.cm.gray, vmin=40, vmax=220)
plt.axis('off')
plt.title('Gaussian filter', fontsize=20)
plt.subplot(133)
plt.imshow(med_denoised, cmap=plt.cm.gray, vmin=40, vmax=220)
plt.axis('off')
plt.title('Median filter', fontsize=20)
plt.subplots_adjust(wspace=0.02, hspace=0.02, top=0.9, bottom=0, left=0,

right=1)
plt.show()

Download Python source code: plot_face_denoise.py
Download Jupyter notebook: plot_face_denoise.ipynb
12.8.12 Find the bounding box of an object
This example shows how to extract the bounding box of the largest object

import numpy as np
np.random.seed(1)
n = 10
l = 256
# Find the largest connected component

sizes = ndimage.sum(mask, label_im, range(nb_labels + 1))
mask_size = sizes < 1000
remove_pixel = mask_size[label_im]
label_im[remove_pixel] = 0
labels = np.unique(label_im)
label_im = np.searchsorted(labels, label_im)
# Now that we have only one connected component, extract it's bounding box
slice_x, slice_y = ndimage.find_objects(label_im==4)[0]
roi = im[slice_x, slice_y]
plt.axes([0, 0, 1, 1])
plt.imshow(roi)
plt.axis('off')
plt.show()

Download Python source code: plot_find_object.py
Download Jupyter notebook: plot_find_object.ipynb
12.8.13 Measurements from images
This examples shows how to measure quantities from various images.

import numpy as np
np.random.seed(1)
n = 10
l = 256
sizes = ndimage.sum(mask, label_im, range(nb_labels + 1))

mask_size = sizes < 1000
remove_pixel = mask_size[label_im]
label_im[remove_pixel] = 0
labels = np.unique(label_im)
label_clean = np.searchsorted(labels, label_im)
plt.figure(figsize=(6 ,3))
plt.subplot(121)
plt.imshow(label_im, cmap=plt.cm.spectral)
plt.axis('off')
plt.subplot(122)
plt.imshow(label_clean, vmax=nb_labels, cmap=plt.cm.spectral)
plt.axis('off')

plt.show()

Download Python source code: plot_measure_data.py
Download Jupyter notebook: plot_measure_data.ipynb

12.8.14 Geometrical transformations
This examples demos some simple geometrical transformations on a Racoon face.
import numpy as np
import scipy.misc
lx, ly = face.shape
# Cropping
crop_face = face[lx//4:-lx//4, ly//4:-ly//4]
# up <-> down flip
flip_ud_face = np.flipud(face)
# rotation
rotate_face = ndimage.rotate(face, 45)
rotate_face_noreshape = ndimage.rotate(face, 45, reshape=False)
plt.subplot(151)
plt.imshow(face, cmap=plt.cm.gray)
plt.axis('off')
plt.subplot(152)
plt.imshow(crop_face, cmap=plt.cm.gray)
plt.axis('off')
plt.subplot(153)
plt.imshow(flip_ud_face, cmap=plt.cm.gray)
plt.axis('off')
plt.subplot(154)
plt.imshow(rotate_face, cmap=plt.cm.gray)
plt.axis('off')
plt.subplot(155)
plt.imshow(rotate_face_noreshape, cmap=plt.cm.gray)
plt.axis('off')
plt.subplots_adjust(wspace=0.02, hspace=0.3, top=1, bottom=0.1, left=0,

right=1)
plt.show()

Download Python source code: plot_geom_face.py
Download Jupyter notebook: plot_geom_face.ipynb

12.8.15 Denoising an image with the median filter
This example shows the original image, the noisy image, the denoised one (with the median filter) and the
difference between the two.
import numpy as np
im = np.zeros((20, 20))
im[5:-5, 5:-5] = 1
im = ndimage.distance_transform_bf(im)
im_noise = im + 0.2*np.random.randn(*im.shape)
im_med = ndimage.median_filter(im_noise, 3)
plt.subplot(141)
plt.imshow(im, interpolation='nearest')
plt.axis('off')
plt.title('Original image', fontsize=20)
plt.subplot(142)
plt.imshow(im_noise, interpolation='nearest', vmin=0, vmax=5)
plt.axis('off')
plt.title('Noisy image', fontsize=20)
plt.subplot(143)
plt.imshow(im_med, interpolation='nearest', vmin=0, vmax=5)
plt.axis('off')
plt.title('Median filter', fontsize=20)
plt.subplot(144)
plt.imshow(np.abs(im - im_med), cmap=plt.cm.hot, interpolation='nearest')
plt.axis('off')
plt.title('Error', fontsize=20)

right=1)
plt.show()

Download Python source code: plot_denoising.py
Download Jupyter notebook: plot_denoising.ipynb

12.8.16 Histogram segmentation
This example does simple histogram analysis to perform segmentation.
import numpy as np
np.random.seed(1)
n = 10
l = 256
mask = (im > im.mean()).astype(np.float)
mask += 0.1 * im
img = mask + 0.2*np.random.randn(*mask.shape)
hist, bin_edges = np.histogram(img, bins=60)

bin_centers = 0.5*(bin_edges[:-1] + bin_edges[1:])
binary_img = img > 0.5
plt.subplot(131)
plt.imshow(img)
plt.axis('off')
plt.subplot(132)
plt.plot(bin_centers, hist, lw=2)
plt.axvline(0.5, color='r', ls='--', lw=2)
plt.text(0.57, 0.8, 'histogram', fontsize=20, transform = plt.gca().transAxes)
plt.yticks([])
plt.subplot(133)
plt.imshow(binary_img, cmap=plt.cm.gray, interpolation='nearest')
plt.axis('off')
plt.subplots_adjust(wspace=0.02, hspace=0.3, top=1, bottom=0.1, left=0, right=1)

plt.show()

Download Python source code: plot_histo_segmentation.py

Download Jupyter notebook: plot_histo_segmentation.ipynb

12.8.17 Finding edges with Sobel filters
The Sobel filter is one of the simplest way of finding edges.
import numpy as np
im = np.zeros((256, 256))
im[64:-64, 64:-64] = 1
im = ndimage.rotate(im, 15, mode='constant')

im = ndimage.gaussian_filter(im, 8)
sx = ndimage.sobel(im, axis=0, mode='constant')

sy = ndimage.sobel(im, axis=1, mode='constant')
sob = np.hypot(sx, sy)
plt.subplot(141)
plt.imshow(im, cmap=plt.cm.gray)
plt.axis('off')
plt.title('square', fontsize=20)
plt.subplot(142)
plt.imshow(sx)
plt.axis('off')
plt.title('Sobel (x direction)', fontsize=20)
plt.subplot(143)
plt.imshow(sob)
plt.axis('off')
plt.title('Sobel filter', fontsize=20)
im += 0.07*np.random.random(im.shape)
sx = ndimage.sobel(im, axis=0, mode='constant')

sy = ndimage.sobel(im, axis=1, mode='constant')
sob = np.hypot(sx, sy)
plt.subplot(144)
plt.imshow(sob)
plt.axis('off')
plt.title('Sobel for noisy image', fontsize=20)

plt.subplots_adjust(wspace=0.02, hspace=0.02, top=1, bottom=0, left=0, right=0.9)
plt.show()

Download Python source code: plot_find_edges.py
Download Jupyter notebook: plot_find_edges.ipynb
12.8.18 Total Variation denoising
This example demoes Total-Variation (TV) denoising on a Racoon face.
import numpy as np
import scipy
import scipy.misc
try:
from skimage.restoration import denoise_tv_chambolle
except ImportError:
# skimage < 0.12
from skimage.filters import denoise_tv_chambolle
f = f[230:290, 220:320]
noisy = f + 0.4*f.std()*np.random.random(f.shape)
tv_denoised = denoise_tv_chambolle(noisy, weight=10)
plt.subplot(131)
plt.imshow(noisy, cmap=plt.cm.gray, vmin=40, vmax=220)
plt.axis('off')
plt.title('noisy', fontsize=20)
plt.subplot(132)
plt.imshow(tv_denoised, cmap=plt.cm.gray, vmin=40, vmax=220)
plt.axis('off')
plt.title('TV denoising', fontsize=20)
tv_denoised = denoise_tv_chambolle(noisy, weight=50)

plt.subplot(133)
plt.imshow(tv_denoised, cmap=plt.cm.gray, vmin=40, vmax=220)
plt.axis('off')
plt.title('(more) TV denoising', fontsize=20)

right=1)
plt.show()

Download Python source code: plot_face_tv_denoise.py
Download Jupyter notebook: plot_face_tv_denoise.ipynb
12.8.19 Greyscale dilation
This example illustrates greyscale mathematical morphology.
import numpy as np
im = np.zeros((64, 64))
np.random.seed(2)
x, y = (63*np.random.random((2, 8))).astype(np.int)
im[x, y] = np.arange(8)
bigger_points = ndimage.grey_dilation(im, size=(5, 5), structure=np.ones((5, 5)))
square = np.zeros((16, 16))

square[4:-4, 4:-4] = 1
dist = ndimage.distance_transform_bf(square)
dilate_dist = ndimage.grey_dilation(dist, size=(3, 3), \
structure=np.ones((3, 3)))
plt.figure(figsize=(12.5, 3))
plt.subplot(141)
plt.imshow(im, interpolation='nearest', cmap=plt.cm.spectral)
plt.axis('off')
plt.subplot(142)
plt.imshow(bigger_points, interpolation='nearest', cmap=plt.cm.spectral)
plt.axis('off')
plt.subplot(143)
plt.imshow(dist, interpolation='nearest', cmap=plt.cm.spectral)
plt.axis('off')
plt.subplot(144)
plt.imshow(dilate_dist, interpolation='nearest', cmap=plt.cm.spectral)
plt.axis('off')
plt.subplots_adjust(wspace=0, hspace=0.02, top=0.99, bottom=0.01, left=0.01, right=0.99)

plt.show()

Download Python source code: plot_greyscale_dilation.py

Download Jupyter notebook: plot_greyscale_dilation.ipynb

12.8.20 Cleaning segmentation with mathematical morphology
An example showing how to clean segmentation with mathematical morphology: removing small regions and
holes.
import numpy as np
np.random.seed(1)
n = 10
l = 256
binary_img = img > 0.5
# Remove small white regions

open_img = ndimage.binary_opening(binary_img)
# Remove small black hole
close_img = ndimage.binary_closing(open_img)
l = 128
plt.subplot(141)
plt.imshow(binary_img[:l, :l], cmap=plt.cm.gray)
plt.axis('off')
plt.subplot(142)
plt.imshow(open_img[:l, :l], cmap=plt.cm.gray)
plt.axis('off')
plt.subplot(143)
plt.imshow(close_img[:l, :l], cmap=plt.cm.gray)
plt.axis('off')
plt.subplot(144)
plt.imshow(mask[:l, :l], cmap=plt.cm.gray)
plt.contour(close_img[:l, :l], [0.5], linewidths=2, colors='r')
plt.axis('off')

plt.show()

Download Python source code: plot_clean_morpho.py
Download Jupyter notebook: plot_clean_morpho.ipynb
12.8.21 Segmentation with Gaussian mixture models
This example performs a Gaussian mixture model analysis of the image histogram to find the right thresholds
for separating foreground from background.
import numpy as np
from sklearn.mixture import GMM
np.random.seed(1)
n = 10
l = 256
hist, bin_edges = np.histogram(img, bins=60)

bin_centers = 0.5*(bin_edges[:-1] + bin_edges[1:])
classif = GMM(n_components=2)
classif.fit(img.reshape((img.size, 1)))
threshold = np.mean(classif.means_)
binary_img = img > threshold

plt.subplot(131)
plt.imshow(img)
plt.axis('off')
plt.subplot(132)
plt.plot(bin_centers, hist, lw=2)
plt.axvline(0.5, color='r', ls='--', lw=2)
plt.text(0.57, 0.8, 'histogram', fontsize=20, transform = plt.gca().transAxes)
plt.yticks([])
plt.subplot(133)
plt.imshow(binary_img, cmap=plt.cm.gray, interpolation='nearest')
plt.axis('off')

plt.show()

Download Python source code: plot_GMM.py
Download Jupyter notebook: plot_GMM.ipynb
12.8.22 Watershed segmentation
This example shows how to do segmentation with watershed.
import numpy as np
from skimage.morphology import watershed
from skimage.feature import peak_local_max
# Generate an initial image with two overlapping circles

x, y = np.indices((80, 80))
x1, y1, x2, y2 = 28, 28, 44, 52
r1, r2 = 16, 20
mask_circle1 = (x - x1) ** 2 + (y - y1) ** 2 < r1 ** 2
image = np.logical_or(mask_circle1, mask_circle2)
# Now we want to separate the two objects in image
# Generate the markers as local maxima of the distance
# to the background
distance = ndimage.distance_transform_edt(image)

local_maxi = peak_local_max(
distance, indices=False, footprint=np.ones((3, 3)), labels=image)
markers = ndimage.label(local_maxi)[0]
labels = watershed(-distance, markers, mask=image)
plt.subplot(131)
plt.imshow(image, cmap='gray', interpolation='nearest')
plt.axis('off')
plt.subplot(132)
plt.imshow(-distance, interpolation='nearest')
plt.axis('off')
plt.subplot(133)
plt.imshow(labels, cmap='spectral', interpolation='nearest')
plt.axis('off')
plt.subplots_adjust(hspace=0.01, wspace=0.01, top=1, bottom=0, left=0,

right=1)
plt.show()

Download Python source code: plot_watershed_segmentation.py
Download Jupyter notebook: plot_watershed_segmentation.ipynb
12.8.23 Granulometry
This example performs a simple granulometry analysis.
import numpy as np
def disk_structure(n):
struct = np.zeros((2 * n + 1, 2 * n + 1))
x, y = np.indices((2 * n + 1, 2 * n + 1))
mask = (x - n)**2 + (y - n)**2 <= n**2
struct[mask] = 1
return struct.astype(np.bool)
def granulometry(data, sizes=None):

s = max(data.shape)
if sizes is None:

sizes = range(1, s/2, 2)

granulo = [ndimage.binary_opening(data, \
structure=disk_structure(n)).sum() for n in sizes]
return granulo
np.random.seed(1)
n = 10
l = 256
granulo = granulometry(mask, sizes=np.arange(2, 19, 4))
plt.subplot(121)
opened = ndimage.binary_opening(mask, structure=disk_structure(10))
opened_more = ndimage.binary_opening(mask, structure=disk_structure(14))
plt.contour(opened, [0.5], colors='b', linewidths=2)
plt.contour(opened_more, [0.5], colors='r', linewidths=2)
plt.axis('off')
plt.subplot(122)
plt.plot(np.arange(2, 19, 4), granulo, 'ok', ms=8)
plt.subplots_adjust(wspace=0.02, hspace=0.15, top=0.95, bottom=0.15, left=0, right=0.95)

plt.show()

Download Python source code: plot_granulo.py
Download Jupyter notebook: plot_granulo.ipynb
12.8.24 Segmentation with spectral clustering
This example uses spectral clustering to do segmentation.
import numpy as np
from sklearn.feature_extraction import image

from sklearn.cluster import spectral_clustering
l = 100
x, y = np.indices((l, l))
center1 = (28, 24)

center2 = (40, 50)
center3 = (67, 58)
center4 = (24, 70)
radius1, radius2, radius3, radius4 = 16, 14, 15, 14

circle1 = (x - center1[0])**2 + (y - center1[1])**2 < radius1**2

4 circles
img = circle1 + circle2 + circle3 + circle4

mask = img.astype(bool)
img = img.astype(float)
img += 1 + 0.2*np.random.randn(*img.shape)
# Convert the image into a graph with the value of the gradient on the
# edges.
graph = image.img_to_graph(img, mask=mask)
# Take a decreasing function of the gradient: we take it weakly

# dependant from the gradient the segmentation is close to a voronoi
graph.data = np.exp(-graph.data / graph.data.std())
# Force the solver to be arpack, since amg is numerically

# unstable on this example
labels = spectral_clustering(graph, n_clusters=4)
label_im = -np.ones(mask.shape)
label_im[mask] = labels
plt.subplot(121)
plt.imshow(img, cmap=plt.cm.spectral, interpolation='nearest')
plt.axis('off')
plt.subplot(122)
plt.imshow(label_im, cmap=plt.cm.spectral, interpolation='nearest')
plt.axis('off')
plt.subplots_adjust(wspace=0, hspace=0., top=0.99, bottom=0.01, left=0.01, right=0.99)

plt.show()

Download Python source code: plot_spectral_clustering.py
Download Jupyter notebook: plot_spectral_clustering.ipynb

See also:
More on image-processing:
The chapter on Scikit-image
Other, more powerful and complete modules: OpenCV (Python bindings), CellProfiler, ITK with Python
bindings

CHAPTER 13
Mathematical optimization: finding
minima of functions
Authors: Gal Varoquaux

Mathematical optimization deals with the problem of finding numerically minimums (or maximums or zeros)
of a function. In this context, the function is called cost function, or objective function, or energy.
Here, we are interested in using scipy.optimize for black-box optimization: we do not rely on the mathe-
matical expression of the function that we are optimizing. Note that this expression can often be used for more
efficient, non black-box, optimization.
Prerequisites
Numpy
Scipy
Matplotlib
See also:
References
Mathematical optimization is very . . . mathematical. If you want performance, it really pays to read the books:
Convex Optimization by Boyd and Vandenberghe (pdf available free online).
Numerical Optimization, by Nocedal and Wright. Detailed reference on gradient descent methods.
Practical Methods of Optimization by Fletcher: good at hand-waving explanations.
397
Chapters contents
Knowing your problem

Convex versus non-convex optimization
Smooth and non-smooth problems
Noisy versus exact cost functions
Constraints
A review of the different optimizers
Getting started: 1D optimization
Gradient based methods
Newton and quasi-newton methods
Full code examples
Examples for the mathematical optimization chapter
Gradient-less methods
Global optimizers
Practical guide to optimization with scipy
Choosing a method
Making your optimizer faster
Computing gradients
Synthetic exercices
Special case: non-linear least-squares
Minimizing the norm of a vector function
Curve fitting
Optimization with constraints
Box bounds
General constraints
Full code examples
Examples for the mathematical optimization chapter
13.1 Knowing your problem
Not all optimization problems are equal. Knowing your problem enables you to choose the right tool.
Dimensionality of the problem
The scale of an optimization problem is pretty much set by the dimensionality of the problem, i.e. the num-
ber of scalar variables on which the search is performed.
13.1. Knowing your problem 398

13.1.1 Convex versus non-convex optimization
A convex function: A non-convex function

f is above all its tangents.
equivalently, for two point A, B, f(C) lies below
the segment [f(A), f(B])], if A < C < B
Optimizing convex functions is easy. Optimizing non-convex functions can be very hard.
Note: It can be proven that for a convex function a local minimum is also a global minimum. Then, in some
sense, the minimum is unique.
13.1.2 Smooth and non-smooth problems
A smooth function: A non-smooth function

The gradient is defined everywhere, and is a continuous function
Optimizing smooth functions is easier (true in the context of black-box optimization, otherwise Linear Pro-
gramming is an example of methods which deal very efficiently with piece-wise linear functions).
13.1. Knowing your problem 399

13.1.3 Noisy versus exact cost functions
Noisy (blue) and non-noisy (green) functions
Noisy gradients
Many optimization methods rely on gradients of the objective function. If the gradient function is not given,
they are computed numerically, which induces errors. In such situation, even if the objective function is not
noisy, a gradient-based optimization may be a noisy optimization.
13.1.4 Constraints
Optimizations under constraints

Here:
1 < x 1 < 1
1 < x 2 < 1
13.2 A review of the different optimizers
13.2.1 Getting started: 1D optimization
Lets get started by finding the minimum of the scalar function f (x) = exp[(x 0.7)2 ]. scipy.optimize.
minimize_scalar() uses Brents method to find the minimum of a function:
13.2. A review of the different optimizers 400

>>> from scipy import optimize

>>> def f(x):
... return -np.exp(-(x - 0.7)**2)
>>> result = optimize.minimize_scalar(f)
>>> result.success # check if solver was successful
True
>>> x_min = result.x
>>> x_min
0.699999999...
>>> x_min - 0.7
-2.160590595323697e-10
Brents method on a quadratic function: it converges in

3 iterations, as the quadratic approximation is then ex-
act.
Brents method on a non-convex function: note that the

fact that the optimizer avoided the local minimum is a
matter of luck.
Note: You can use different solvers using the parameter method.
Note: scipy.optimize.minimize_scalar() can also be used for optimization constrained to an interval

using the parameter bounds.
13.2.2 Gradient based methods
Some intuitions about gradient descent
Here we focus on intuitions, not code. Code will follow.

Gradient descent basically consists in taking small steps in the direction of the gradient, that is the direction of
the steepest descent.

Table 13.1: Fixed step gradient descent
A well-conditioned quadratic function.
An ill-conditioned quadratic function.

The core problem of gradient-methods on ill-conditioned prob-
lems is that the gradient tends not to point in the direction of the
minimum.
We can see that very anisotropic (ill-conditioned) functions are harder to optimize.
Take home message: conditioning number and preconditioning
If you know natural scaling for your variables, prescale them so that they behave similarly. This is related to
preconditioning.
Also, it clearly can be advantageous to take bigger steps. This is done in gradient descent code using a line
search.

Table 13.2: Adaptive step gradient descent
A well-conditioned quadratic
function.
An ill-conditioned quadratic
function.
An ill-conditioned non-
quadratic function.
An ill-conditioned very non-

quadratic function.
The more a function looks like a quadratic function (elliptic iso-curves), the easier it is to optimize.
Conjugate gradient descent
The gradient descent algorithms above are toys not to be used on real problems.
As can be seen from the above experiments, one of the problems of the simple gradient descent algorithms, is
that it tends to oscillate across a valley, each time following the direction of the gradient, that makes it cross
the valley. The conjugate gradient solves this problem by adding a friction term: each step depends on the two
last values of the gradient and sharp turns are reduced.

Table 13.3: Conjugate gradient descent
quadratic function.

quadratic function.
scipy provides scipy.optimize.minimize() to find the minimum of scalar functions of one or more vari-
ables. The simple conjugate gradient method can be used by setting the parameter method to CG
>>> def f(x): # The rosenbrock function

... return .5*(1 - x[0])**2 + (x[1] - x[0]**2)**2
>>> optimize.minimize(f, [2, -1], method="CG")
fun: 1.6503...e-11
jac: array([ -6.1534...e-06, 2.5380...e-07])
message: ...'Optimization terminated successfully.'
nfev: 108
nit: 13
njev: 27
status: 0
success: True
x: array([ 0.99999..., 0.99998...])
Gradient methods need the Jacobian (gradient) of the function. They can compute it numerically, but will
perform better if you can pass them the gradient:
>>> def jacobian(x):

... return np.array((-2*.5*(1 - x[0]) - 4*x[0]*(x[1] - x[0]**2), 2*(x[1] - x[0]**2)))
>>> optimize.minimize(f, [2, 1], method="CG", jac=jacobian)
fun: 2.957...e-14
jac: array([ 7.1825...e-07, -2.9903...e-07])
nfev: 16
nit: 8
njev: 16
status: 0
success: True
x: array([ 1.0000..., 1.0000...])
Note that the function has only been evaluated 27 times, compared to 108 without the gradient.

13.2.3 Newton and quasi-newton methods
Newton methods: using the Hessian (2nd differential)
Newton methods use a local quadratic approximation to compute the jump direction. For this purpose, they
rely on the 2 first derivative of the function: the gradient and the Hessian.
An ill-conditioned quadratic function:

Note that, as the quadratic approximation is exact, the Newton
method is blazing fast
An ill-conditioned non-quadratic function:

Here we are optimizing a Gaussian, which is always below its
quadratic approximation. As a result, the Newton method over-
shoots and leads to oscillations.
An ill-conditioned very non-quadratic function:
In scipy, you can use the Newton method by setting method to Newton-CG in scipy.optimize.minimize().
Here, CG refers to the fact that an internal inversion of the Hessian is performed by conjugate gradient

... return .5*(1 - x[0])**2 + (x[1] - x[0]**2)**2
>>> optimize.minimize(f, [2,-1], method="Newton-CG", jac=jacobian)
fun: 1.5601...e-15
jac: array([ 1.0575...e-07, -7.4832...e-08])
nfev: 11
nhev: 0
nit: 10
njev: 52
status: 0
success: True
x: array([ 0.99999..., 0.99999...])
Note that compared to a conjugate gradient (above), Newtons method has required less function evaluations,
but more gradient evaluations, as it uses it to approximate the Hessian. Lets compute the Hessian and pass it
to the algorithm:
>>> def hessian(x): # Computed with sympy

... return np.array(((1 - 4*x[1] + 12*x[0]**2, -4*x[0]), (-4*x[0], 2)))
>>> optimize.minimize(f, [2,-1], method="Newton-CG", jac=jacobian, hess=hessian)
fun: 1.6277...e-15
jac: array([ 1.1104...e-07, -7.7809...e-08])
nfev: 11
nhev: 10
nit: 10

njev: 20
status: 0
success: True
x: array([ 0.99999..., 0.99999...])
Note: At very high-dimension, the inversion of the Hessian can be costly and unstable (large scale > 250).
Note: Newton optimizers should not to be confused with Newtons root finding method, based on the same
principles, scipy.optimize.newton().
Quasi-Newton methods: approximating the Hessian on the fly
BFGS: BFGS (Broyden-Fletcher-Goldfarb-Shanno algorithm) refines at each step an approximation of the Hes-
sian.
13.4 Examples for the mathematical optimization chapter
13.4.1 Noisy optimization problem
Draws a figure explaining noisy vs non-noisy optimization
import numpy as np
import pylab as pl
np.random.seed(0)
x = np.linspace(-5, 5, 101)
x_ = np.linspace(-5, 5, 31)
def f(x):
return -np.exp(-x**2)
# A smooth function
pl.figure(1, figsize=(3, 2.5))

pl.clf()
pl.plot(x_, f(x_) + .2*np.random.normal(size=31), linewidth=2)

pl.plot(x, f(x), linewidth=2)
pl.ylim(ymin=-1.3)
pl.axis('off')
pl.tight_layout()
pl.show()

Download Python source code: plot_noisy.py
Download Jupyter notebook: plot_noisy.ipynb
13.4.2 Smooth vs non-smooth
Draws a figure to explain smooth versus non smooth optimization.
import numpy as np
import pylab as pl
x = np.linspace(-1.5, 1.5, 101)
# A smooth function
pl.clf()
pl.plot(x, np.sqrt(.2 + x**2), linewidth=2)

pl.text(-1, 0, '$f$', size=20)
pl.ylim(ymin=-.2)
pl.axis('off')
pl.tight_layout()
# A non-smooth function
pl.clf()
pl.plot(x, np.abs(x), linewidth=2)
pl.text(-1, 0, '$f$', size=20)
pl.ylim(ymin=-.2)
13.4. Examples for the mathematical optimization chapter 407

pl.axis('off')
pl.tight_layout()
pl.show()

Download Python source code: plot_smooth.py
Download Jupyter notebook: plot_smooth.ipynb
A curve fitting example
import numpy as np
import pylab as pl
np.random.seed(0)
# Our test function

def f(t, omega, phi):
return np.cos(omega * t + phi)
# Our x and y data

y = f(x, 1.5, 1) + .1*np.random.normal(size=50)
# Fit the model: the parameters omega and phi can be found in the

# `params` vector
params, params_cov = optimize.curve_fit(f, x, y)
# plot the data and the fitted curve

pl.figure(1)
pl.clf()
pl.plot(x, y, 'bx')
pl.plot(t, f(t, *params), 'r-')
pl.show()

Download Python source code: plot_curve_fit.py
Download Jupyter notebook: plot_curve_fit.ipynb
13.4.4 Convex function
A figure showing the definition of a convex function
import numpy as np
import pylab as pl
x = np.linspace(-1, 2)

pl.clf()
# A convex function
pl.plot(x, x**2, linewidth=2)
pl.text(-.7, -.6**2, '$f$', size=20)
# The tangent in one point

pl.plot(x, 2*x - 1)
pl.plot(1, 1, 'k+')
pl.text(.3, -.75, "Tangent to $f$", size=15)
pl.text(1, 1 - .5, 'C', size=15)
# Convexity as barycenter
pl.plot([.35, 1.85], [.35**2, 1.85**2])
pl.plot([.35, 1.85], [.35**2, 1.85**2], 'k+')

pl.text(.35 - .2, .35**2 + .1, 'A', size=15)

pl.text(1.85 - .2, 1.85**2, 'B', size=15)
pl.ylim(ymin=-1)
pl.axis('off')
pl.tight_layout()
# Convexity as barycenter
pl.clf()
pl.plot(x, x**2 + np.exp(-5*(x - .5)**2), linewidth=2)
pl.text(-.7, -.6**2, '$f$', size=20)
pl.ylim(ymin=-1)
pl.axis('off')
pl.tight_layout()
pl.show()

Download Python source code: plot_convex.py
Download Jupyter notebook: plot_convex.ipynb
13.4.5 Finding a minimum in a flat neighborhood
An excercise of finding minimum. This excercise is hard because the function is very flat around the minimum
(all its derivatives are zero). Thus gradient information is unreliable.
The function admits a minimum in [0, 0]. The challenge is to get within 1e-7 of this minimum, starting at x0 =
[1, 1].
The solution that we adopt here is to give up on using gradient or information based on local differences, and
to rely on the Powell algorithm. With 162 function evaluations, we get to 1e-8 of the solution.
import numpy as np
import pylab as pl
def f(x):
return np.exp(-1/(.01*x[0]**2 + x[1]**2))
# A well-conditionned version of f:
def g(x):
return f([10*x[0], x[1]])
# The gradient of g. We won't use it here for the optimization.

def g_prime(x):
r = np.sqrt(x[0]**2 + x[1]**2)
return 2/r**3*g(x)*x/r
result = optimize.minimize(g, [1, 1], method="Powell", tol=1e-10)

x_min = result.x
Some pretty plotting
pl.figure(0)
pl.clf()
t = np.linspace(-1.1, 1.1, 100)
pl.plot(t, f([0, t]))

pl.figure(1)
pl.clf()
X, Y = np.mgrid[-1.5:1.5:100j, -1.1:1.1:100j]
pl.imshow(f([X, Y]).T, cmap=pl.cm.gray_r, extent=[-1.5, 1.5, -1.1, 1.1],
origin='lower')
pl.contour(X, Y, f([X, Y]), cmap=pl.cm.gnuplot)
# Plot the gradient

dX, dY = g_prime([.1*X[::5, ::5], Y[::5, ::5]])
# Adjust for our preconditioning
dX *= .1
pl.quiver(X[::5, ::5], Y[::5, ::5], dX, dY, color='.5')
# Plot our solution

pl.plot(x_min[0], x_min[1], 'r+', markersize=15)
pl.show()

Download Python source code: plot_exercise_flat_minimum.py
Download Jupyter notebook: plot_exercise_flat_minimum.ipynb

13.4.6 Optimization with constraints
An example showing how to do optimization with general constraints using SLSQP and cobyla.
import numpy as np
import pylab as pl
x, y = np.mgrid[-2.03:4.2:.04, -1.6:3.2:.04]
x = x.T
y = y.T

pl.clf()
pl.axes([0, 0, 1, 1])
contours = pl.contour(np.sqrt((x - 3)**2 + (y - 2)**2),

extent=[-2.03, 4.2, -1.6, 3.2],
cmap=pl.cm.gnuplot)
pl.clabel(contours,
inline=1,
fmt='%1.1f ',
fontsize=14)
pl.plot([-1.5, 0, 1.5, 0, -1.5],
[ 0, 1.5, 0, -1.5, 0], 'k', linewidth=2)
pl.fill_between([ -1.5, 0, 1.5],
[ 0, -1.5, 0],
[ 0, 1.5, 0],
color='.8')
pl.axvline(0, color='k')
pl.axhline(0, color='k')
pl.text(-.9, 2.8, '$x_2$', size=20)

pl.text(3.6, -.6, '$x_1$', size=20)
pl.axis('tight')
pl.axis('off')
# And now plot the optimization path

accumulator = list()
def f(x):
# Store the list of function calls
accumulator.append(x)
return np.sqrt((x[0] - 3)**2 + (x[1] - 2)**2)

def constraint(x):
return np.atleast_1d(1.5 - np.sum(np.abs(x)))
optimize.minimize(f, np.array([0, 0]), method="SLSQP",

constraints={"fun": constraint, "type": "ineq"})
accumulated = np.array(accumulator)
pl.plot(accumulated[:, 0], accumulated[:, 1])
pl.show()

Download Python source code: plot_non_bounds_constraints.py
Download Jupyter notebook: plot_non_bounds_constraints.ipynb
13.4.7 Brents method
Illustration of 1D optimization: Brents method

Out:

Converged at 6
Converged at 23
import numpy as np
import pylab as pl
x_0 = np.exp(-1)
def f(x):
return (x - x_0)**2 + epsilon*np.exp(-5*(x - .5 - x_0)**2)
for epsilon in (0, 1):

pl.figure(figsize=(3, 2.5))
pl.axes([0, 0, 1, 1])
# A convex function
pl.plot(x, f(x), linewidth=2)
# Apply brent method. To have access to the iteration, do this in an

# artificial way: allow the algorithm to iter only once
all_x = list()
all_y = list()
for iter in range(30):
result = optimize.minimize_scalar(f, bracket=(-5, 2.9, 4.5), method="Brent",
options={"maxiter": iter}, tol=np.finfo(1.).eps)
if result.success:
print('Converged at ', iter)
break
this_x = result.x
all_x.append(this_x)
all_y.append(f(this_x))
if iter < 6:
pl.text(this_x - .05*np.sign(this_x) - .05,
f(this_x) + 1.2*(.3 - iter % 2), iter + 1,
size=12)
pl.plot(all_x[:10], all_y[:10], 'k+', markersize=12, markeredgewidth=2)
pl.plot(all_x[-1], all_y[-1], 'rx', markersize=12)

pl.axis('off')
pl.ylim(ymin=-1, ymax=8)
pl.figure(figsize=(4, 3))
pl.semilogy(np.abs(all_y - all_y[-1]), linewidth=2)
pl.ylabel('Error on f(x)')
pl.xlabel('Iteration')
pl.tight_layout()
pl.show()

Download Python source code: plot_1d_optim.py
Download Jupyter notebook: plot_1d_optim.ipynb

13.4.8 Constraint optimization: visualizing the geometry
A small figure explaining optimization with constraints
import numpy as np
import pylab as pl
x, y = np.mgrid[-2.9:5.8:.05, -2.5:5:.05]
x = x.T
y = y.T
for i in (1, 2):

# Create 2 figure: only the second one will have the optimization
# path
pl.figure(i, figsize=(3, 2.5))
pl.clf()
pl.axes([0, 0, 1, 1])
contours = pl.contour(np.sqrt((x - 3)**2 + (y - 2)**2),

extent=[-3, 6, -2.5, 5],
cmap=pl.cm.gnuplot)
pl.clabel(contours,
inline=1,
fmt='%1.1f ',
fontsize=14)
pl.plot([-1.5, -1.5, 1.5, 1.5, -1.5],
[-1.5, 1.5, 1.5, -1.5, -1.5], 'k', linewidth=2)
pl.fill_between([ -1.5, 1.5],
[ -1.5, -1.5],
[ 1.5, 1.5],
color='.8')
pl.axvline(0, color='k')
pl.axhline(0, color='k')
pl.text(-.9, 4.4, '$x_2$', size=20)

pl.text(5.6, -.6, '$x_1$', size=20)
pl.axis('equal')
pl.axis('off')
# And now plot the optimization path

accumulator = list()

def f(x):
# Store the list of function calls
accumulator.append(x)
return np.sqrt((x[0] - 3)**2 + (x[1] - 2)**2)
# We don't use the gradient, as with the gradient, L-BFGS is too fast,
# and finds the optimum without showing us a pretty path
def f_prime(x):
r = np.sqrt((x[0] - 3)**2 + (x[0] - 2)**2)
return np.array(((x[0] - 3)/r, (x[0] - 2)/r))
optimize.minimize(f, np.array([0, 0]), method="L-BFGS-B",

bounds=((-1.5, 1.5), (-1.5, 1.5)))
accumulated = np.array(accumulator)
pl.plot(accumulated[:, 0], accumulated[:, 1])
pl.show()

Download Python source code: plot_constraints.py
Download Jupyter notebook: plot_constraints.ipynb
13.4.9 Plotting the comparison of optimizers
Plots the results from the comparison of optimizers.
import pickle
import sys
import numpy as np
import pylab as pl
results = pickle.load(open(
'helper/compare_optimizers_py%s .pkl' % sys.version_info[0],
'rb'))
n_methods = len(list(results.values())[0]['Rosenbrock '])
n_dims = len(results)
symbols = 'o>*Ds'

pl.figure(1, figsize=(10, 4))

pl.clf()
colors = pl.cm.Spectral(np.linspace(0, 1, n_dims))[:, :3]
method_names = list(list(results.values())[0]['Rosenbrock '].keys())

method_names.sort(key=lambda x: x[::-1], reverse=True)
for n_dim_index, ((n_dim, n_dim_bench), color) in enumerate(

zip(sorted(results.items()), colors)):
for (cost_name, cost_bench), symbol in zip(sorted(n_dim_bench.items()),
symbols):
for method_index, method_name, in enumerate(method_names):
this_bench = cost_bench[method_name]
bench = np.mean(this_bench)
pl.semilogy([method_index + .1*n_dim_index, ], [bench, ],
marker=symbol, color=color)
# Create a legend for the problem type

for cost_name, symbol in zip(sorted(n_dim_bench.keys()),
symbols):
pl.semilogy([-10, ], [0, ], symbol, color='.5',
label=cost_name)
pl.xticks(np.arange(n_methods), method_names, size=11)

pl.xlim(-.2, n_methods - .5)
pl.legend(loc='best', numpoints=1, handletextpad=0, prop=dict(size=12),
frameon=False)
pl.ylabel('# function calls (a.u.)')
# Create a second legend for the problem dimensionality

pl.twinx()
for n_dim, color in zip(sorted(results.keys()), colors):

pl.plot([-10, ], [0, ], 'o', color=color,
label='# dim: %i ' % n_dim)
pl.legend(loc=(.47, .07), numpoints=1, handletextpad=0, prop=dict(size=12),
frameon=False, ncol=2)
pl.xlim(-.2, n_methods - .5)
pl.xticks(np.arange(n_methods), method_names)
pl.yticks(())
pl.tight_layout()
pl.show()

Download Python source code: plot_compare_optimizers.py
Download Jupyter notebook: plot_compare_optimizers.ipynb
13.4.10 Alternating optimization
The challenge here is that Hessian of the problem is a very ill-conditioned matrix. This can easily be seen, as
the Hessian of the first term in simply 2*np.dot(K.T, K). Thus the conditioning of the problem can be judged
from looking at the conditioning of K.

import time
import numpy as np
import pylab as pl
np.random.seed(0)
K = np.random.normal(size=(100, 100))
def f(x):
return np.sum((np.dot(K, x - 1))**2) + np.sum(x**2)**2
def f_prime(x):
return 2*np.dot(np.dot(K.T, K), x - 1) + 4*np.sum(x**2)*x
def hessian(x):
H = 2*np.dot(K.T, K) + 4*2*x*x[:, np.newaxis]
return H + 4*np.eye(H.shape[0])*np.sum(x**2)
Some pretty plotting
pl.figure(1)
pl.clf()
Z = X, Y = np.mgrid[-1.5:1.5:100j, -1.1:1.1:100j]
# Complete in the additional dimensions with zeros
Z = np.reshape(Z, (2, -1)).copy()
Z.resize((100, Z.shape[-1]))
Z = np.apply_along_axis(f, 0, Z)
Z = np.reshape(Z, X.shape)
pl.imshow(Z.T, cmap=pl.cm.gray_r, extent=[-1.5, 1.5, -1.1, 1.1],
origin='lower')
pl.contour(X, Y, Z, cmap=pl.cm.gnuplot)
# A reference but slow solution:

t0 = time.time()
x_ref = optimize.minimize(f, K[0], method="Powell").x
print(' Powell: time %.2f s' % (time.time() - t0))
f_ref = f(x_ref)
# Compare different approaches

t0 = time.time()
x_bfgs = optimize.minimize(f, K[0], method="BFGS").x
print(' BFGS: time %.2f s, x error %.2f , f error %.2f ' % (time.time() - t0,
np.sqrt(np.sum((x_bfgs - x_ref)**2)), f(x_bfgs) - f_ref))
t0 = time.time()
x_l_bfgs = optimize.minimize(f, K[0], method="L-BFGS-B").x
print(' L-BFGS: time %.2f s, x error %.2f , f error %.2f ' % (time.time() - t0,
np.sqrt(np.sum((x_l_bfgs - x_ref)**2)), f(x_l_bfgs) - f_ref))
t0 = time.time()
x_bfgs = optimize.minimize(f, K[0], jac=f_prime, method="BFGS").x
print(" BFGS w f': time %.2f s, x error %.2f , f error %.2f " % (
time.time() - t0, np.sqrt(np.sum((x_bfgs - x_ref)**2)),
f(x_bfgs) - f_ref))
t0 = time.time()
x_l_bfgs = optimize.minimize(f, K[0], jac=f_prime, method="L-BFGS-B").x
print("L-BFGS w f': time %.2f s, x error %.2f , f error %.2f " % (

time.time() - t0, np.sqrt(np.sum((x_l_bfgs - x_ref)**2)),

f(x_l_bfgs) - f_ref))
t0 = time.time()
x_newton = optimize.minimize(f, K[0], jac=f_prime, hess=hessian, method="Newton-CG").x
print(" Newton: time %.2f s, x error %.2f , f error %.2f " % (
time.time() - t0, np.sqrt(np.sum((x_newton - x_ref)**2)),
f(x_newton) - f_ref))
pl.show()
Out:
Powell: time 0.29s

BFGS: time 0.78s, x error 0.02, f error -0.02
L-BFGS: time 0.06s, x error 0.02, f error -0.02
BFGS w f': time 0.05s, x error 0.02, f error -0.02
L-BFGS w f': time 0.00s, x error 0.02, f error -0.02
Newton: time 0.01s, x error 0.02, f error -0.02

Download Python source code: plot_exercise_ill_conditioned.py
Download Jupyter notebook: plot_exercise_ill_conditioned.ipynb
13.4.11 Gradient descent
An example demoing gradient descent by creating figures that trace the evolution of the optimizer.

import numpy as np
import pylab as pl
import sys, os
sys.path.append(os.path.abspath('helper'))
from cost_functions import mk_quad, mk_gauss, rosenbrock,\
rosenbrock_prime, rosenbrock_hessian, LoggingFunction,\
CountingFunction
x_min, x_max = -1, 2

y_min, y_max = 2.25/3*x_min - .2, 2.25/3*x_max - .2
A formatter to print values on contours
def super_fmt(value):
if value > 1:
if np.abs(int(value) - value) < .1:
out = '$10^{%.1i }$' % value
else:
out = '$10^{%.1f }$' % value
else:
value = np.exp(value - .01)
if value > .1:
out = '%1.1f ' % value
elif value > .01:
out = '%.2f ' % value
else:
out = '%.2e ' % value
return out
A gradient descent algorithm do not use: its a toy, use scipys optimize.fmin_cg
def gradient_descent(x0, f, f_prime, hessian=None, adaptative=False):

x_i, y_i = x0
all_x_i = list()
all_y_i = list()
all_f_i = list()

all_x_i.append(x_i)
all_y_i.append(y_i)
all_f_i.append(f([x_i, y_i]))
dx_i, dy_i = f_prime(np.asarray([x_i, y_i]))
if adaptative:
# Compute a step size using a line_search to satisfy the Wolf
# conditions
step = optimize.line_search(f, f_prime,
np.r_[x_i, y_i], -np.r_[dx_i, dy_i],
np.r_[dx_i, dy_i], c2=.05)
step = step[0]
if step is None:
step = 0
else:
step = 1
x_i += - step*dx_i
y_i += - step*dy_i
if np.abs(all_f_i[-1]) < 1e-16:
break
return all_x_i, all_y_i, all_f_i
def gradient_descent_adaptative(x0, f, f_prime, hessian=None):

return gradient_descent(x0, f, f_prime, adaptative=True)
def conjugate_gradient(x0, f, f_prime, hessian=None):

all_x_i = [x0[0]]
all_y_i = [x0[1]]
all_f_i = [f(x0)]
def store(X):
x, y = X
all_x_i.append(x)
all_y_i.append(y)
all_f_i.append(f(X))
optimize.minimize(f, x0, jac=f_prime, method="CG", callback=store, options={"gtol": 1e-12})
def newton_cg(x0, f, f_prime, hessian):

all_x_i = [x0[0]]
all_y_i = [x0[1]]
all_f_i = [f(x0)]
def store(X):
x, y = X
all_x_i.append(x)
all_y_i.append(y)
optimize.minimize(f, x0, method="Newton-CG", jac=f_prime, hess=hessian, callback=store,
,options={"xtol": 1e-12})
def bfgs(x0, f, f_prime, hessian=None):

all_x_i = [x0[0]]
all_y_i = [x0[1]]
all_f_i = [f(x0)]
def store(X):
x, y = X
all_x_i.append(x)
all_y_i.append(y)
optimize.minimize(f, x0, method="BFGS", jac=f_prime, callback=store, options={"gtol": 1e-
,12})
def powell(x0, f, f_prime, hessian=None):

all_x_i = [x0[0]]
all_y_i = [x0[1]]
all_f_i = [f(x0)]
def store(X):
x, y = X
all_x_i.append(x)
all_y_i.append(y)
optimize.minimize(f, x0, method="Powell", callback=store, options={"ftol": 1e-12})
def nelder_mead(x0, f, f_prime, hessian=None):

all_x_i = [x0[0]]
all_y_i = [x0[1]]
all_f_i = [f(x0)]
def store(X):
x, y = X

all_x_i.append(x)
all_y_i.append(y)
optimize.minimize(f, x0, method="Nelder-Mead", callback=store, options={"ftol": 1e-12})
Run different optimizers on these problems
levels = dict()
for index, ((f, f_prime, hessian), optimizer) in enumerate((

(mk_quad(.7), gradient_descent),
(mk_quad(.7), gradient_descent_adaptative),
(mk_quad(.02), gradient_descent),
(mk_quad(.02), gradient_descent_adaptative),
(mk_gauss(.02), gradient_descent_adaptative),
((rosenbrock, rosenbrock_prime, rosenbrock_hessian),
gradient_descent_adaptative),
(mk_gauss(.02), conjugate_gradient),
conjugate_gradient),
(mk_quad(.02), newton_cg),
(mk_gauss(.02), newton_cg),
newton_cg),
(mk_quad(.02), bfgs),
(mk_gauss(.02), bfgs),
bfgs),
(mk_quad(.02), powell),
(mk_gauss(.02), powell),
powell),
(mk_gauss(.02), nelder_mead),
nelder_mead),
)):
# Compute a gradient-descent
x_i, y_i = 1.6, 1.1
counting_f_prime = CountingFunction(f_prime)
counting_hessian = CountingFunction(hessian)
logging_f = LoggingFunction(f, counter=counting_f_prime.counter)
all_x_i, all_y_i, all_f_i = optimizer(np.array([x_i, y_i]),
logging_f, counting_f_prime,
hessian=counting_hessian)
# Plot the contour plot

if not max(all_y_i) < y_max:
x_min *= 1.2
x_max *= 1.2
y_min *= 1.2
y_max *= 1.2
x, y = np.mgrid[x_min:x_max:100j, y_min:y_max:100j]
x = x.T
y = y.T
pl.figure(index, figsize=(3, 2.5))

pl.clf()
pl.axes([0, 0, 1, 1])
X = np.concatenate((x[np.newaxis, ...], y[np.newaxis, ...]), axis=0)

z = np.apply_along_axis(f, 0, X)

log_z = np.log(z + .01)

pl.imshow(log_z,
extent=[x_min, x_max, y_min, y_max],
cmap=pl.cm.gray_r, origin='lower',
vmax=log_z.min() + 1.5*log_z.ptp())
contours = pl.contour(log_z,
levels=levels.get(f, None),
extent=[x_min, x_max, y_min, y_max],
cmap=pl.cm.gnuplot, origin='lower')
levels[f] = contours.levels
pl.clabel(contours, inline=1,
fmt=super_fmt, fontsize=14)
pl.plot(all_x_i, all_y_i, 'b-', linewidth=2)

pl.plot(all_x_i, all_y_i, 'k+')
pl.plot(logging_f.all_x_i, logging_f.all_y_i, 'k.', markersize=2)
pl.plot([0], [0], 'rx', markersize=12)
pl.xticks(())
pl.yticks(())
pl.xlim(x_min, x_max)
pl.ylim(y_min, y_max)
pl.draw()
pl.figure(index + 100, figsize=(4, 3))

pl.clf()
pl.semilogy(np.maximum(np.abs(all_f_i), 1e-30), linewidth=2,
label='# iterations')
pl.ylabel('Error on f(x)')
pl.semilogy(logging_f.counts,
np.maximum(np.abs(logging_f.all_f_i), 1e-30),
linewidth=2, color='g', label='# function calls')
pl.legend(loc='upper right', frameon=True, prop=dict(size=11),
borderaxespad=0, handlelength=1.5, handletextpad=.5)
pl.tight_layout()
pl.draw()







Download Python source code: plot_gradient_descent.py
Download Jupyter notebook: plot_gradient_descent.ipynb


On a exactly quadratic function, BFGS is not as fast as Newtons
method, but still very fast.
An ill-conditioned non-quadratic function:

Here BFGS does better than Newton, as its empirical estimate
of the curvature is better than that given by the Hessian.

... return .5*(1 - x[0])**2 + (x[1] - x[0]**2)**2
>>> optimize.minimize(f, [2, -1], method="BFGS", jac=jacobian)
fun: 2.6306...e-16
hess_inv: array([[ 0.99986..., 2.0000...],
[ 2.0000..., 4.498...]])
jac: array([ 6.7089...e-08, -3.2222...e-08])
nfev: 10
nit: 8
njev: 10
status: 0
success: True
x: array([ 1. , 0.99999...])
L-BFGS: Limited-memory BFGS Sits between BFGS and conjugate gradient: in very high dimensions (> 250)
the Hessian matrix is too costly to compute and invert. L-BFGS keeps a low-rank version. In addition, box
bounds are also supported by L-BFGS-B:

... return .5*(1 - x[0])**2 + (x[1] - x[0]**2)**2
>>> optimize.minimize(f, [2, 2], method="L-BFGS-B", jac=jacobian)
fun: 1.4417...e-15
jac: array([ 1.0233...e-07, -2.5929...e-08])
nfev: 17
nit: 16
status: 0
success: True
x: array([ 1.0000..., 1.0000...])

13.4.12 Gradient-less methods
A shooting method: the Powell algorithm
Almost a gradient approach

Powells method isnt too sensitive to local ill-
conditionning in low dimensions
Simplex method: the Nelder-Mead
The Nelder-Mead algorithms is a generalization of dichotomy approaches to high-dimensional spaces. The

algorithm works by refining a simplex, the generalization of intervals and triangles to high-dimensional spaces,
to bracket the minimum.
Strong points: it is robust to noise, as it does not rely on computing gradients. Thus it can work on functions
that are not locally smooth such as experimental data points, as long as they display a large-scale bell-shape
behavior. However it is slower than gradient-based methods on smooth, non-noisy functions.
quadratic function:

quadratic function:
Using the Nelder-Mead solver in scipy.optimize.minimize():


... return .5*(1 - x[0])**2 + (x[1] - x[0]**2)**2
>>> optimize.minimize(f, [2, -1], method="Nelder-Mead")
final_simplex: (array([[ 1.0000..., 1.0000...],
[ 0.99998... , 0.99996... ],
[ 1.0000..., 1.0000... ]]), array([ 1.1152...e-10, 1.5367...e-10, 4.9883...e-10]))
fun: 1.1152...e-10
nfev: 111
nit: 58
status: 0
success: True
x: array([ 1.0000..., 1.0000...])
13.4.13 Global optimizers
If your problem does not admit a unique local minimum (which can be hard to test unless the function is
convex), and you do not have prior information to initialize the optimization close to the solution, you may
need a global optimizer.
Brute force: a grid search
scipy.optimize.brute() evaluates the function on a given grid of parameters and returns the parameters
corresponding to the minimum value. The parameters are specified with ranges given to numpy.mgrid. By
default, 20 steps are taken in each direction:

... return .5*(1 - x[0])**2 + (x[1] - x[0]**2)**2
>>> optimize.brute(f, ((-1, 2), (-1, 2)))
array([ 1.0000..., 1.0000...])
13.5 Practical guide to optimization with scipy
13.5.1 Choosing a method
All methods are exposed as the method argument of scipy.optimize.minimize().
Without knowledge of the gradient
13.5. Practical guide to optimization with scipy 432

In general, prefer BFGS or L-BFGS, even if you have to approximate numerically gra-
dients. These are also the default if you omit the parameter method - depending if the
problem has constraints or bounds
On well-conditioned problems, Powell and Nelder-Mead, both gradient-free methods,
work well in high dimension, but they collapse for ill-conditioned problems.
With knowledge of the gradient
BFGS or L-BFGS.
Computational overhead of BFGS is larger than that L-BFGS, itself larger than that of
conjugate gradient. On the other side, BFGS usually needs less function evaluations than
CG. Thus conjugate gradient method is better than BFGS at optimizing computationally
cheap functions.
With the Hessian
If you can compute the Hessian, prefer the Newton method (Newton-CG or TCG).
If you have noisy measurements
Use Nelder-Mead or Powell.
13.5.2 Making your optimizer faster
Choose the right method (see above), do compute analytically the gradient and Hessian, if you can.
Use preconditionning when possible.
Choose your initialization points wisely. For instance, if you are running many similar optimizations,
warm-restart one with the results of another.
Relax the tolerance if you dont need precision using the parameter tol.
13.5.3 Computing gradients
Computing gradients, and even more Hessians, is very tedious but worth the effort. Symbolic computation
with Sympy may come in handy.
Warning: A very common source of optimization not converging well is human error in the computation
of the gradient. You can use scipy.optimize.check_grad() to check that your gradient is correct. It
returns the norm of the different between the gradient given, and a gradient computed numerically:
>>> optimize.check_grad(f, jacobian, [2, -1])
2.384185791015625e-07
See also scipy.optimize.approx_fprime() to find your errors.

13.5.4 Synthetic exercices
Exercice: A simple (?) quadratic function
Optimize the following function, using K[0] as a starting point:

np.random.seed(0)
K = np.random.normal(size=(100, 100))
def f(x):
return np.sum((np.dot(K, x - 1))**2) + np.sum(x**2)**2
Time your approach. Find the fastest approach. Why is BFGS not working well?
Exercice: A locally flat minimum
Consider the function exp(-1/(.1*x**2 + y**2). This function admits a minimum in (0, 0). Starting from an
initialization at (1, 1), try to get within 1e-8 of this minimum point.

13.6 Special case: non-linear least-squares
13.6.1 Minimizing the norm of a vector function
Least square problems, minimizing the norm of a vector function, have a specific structure that can be used in
the LevenbergMarquardt algorithm implemented in scipy.optimize.leastsq().
Lets try to minimize the norm of the following vectorial function:
>>> def f(x):

... return np.arctan(x) - np.arctan(np.linspace(0, 1, len(x)))
>>> x0 = np.zeros(10)
>>> optimize.leastsq(f, x0)
(array([ 0. , 0.11111111, 0.22222222, 0.33333333, 0.44444444,
0.55555556, 0.66666667, 0.77777778, 0.88888889, 1. ]), 2)
This took 67 function evaluations (check it with full_output=1). What if we compute the norm ourselves and
use a good generic optimizer (BFGS):
>>> def g(x):

... return np.sum(f(x)**2)
>>> optimize.minimize(g, x0, method="BFGS")
fun: 2.6940...e-11
hess_inv: array([[...
...
...]])
jac: array([
...
...])
nfev: 144
nit: 11
njev: 12
status: 0
success: True
x: array([ -7.3845...e-09, 1.1111...e-01, 2.2222...e-01,
3.3333...e-01, 4.4444...e-01, 5.5555...e-01,
13.6. Special case: non-linear least-squares 435

6.6666...e-01, 7.7777...e-01, 8.8889...e-01,

1.0000...e+00])
BFGS needs more function calls, and gives a less precise result.
Note: leastsq is interesting compared to BFGS only if the dimensionality of the output vector is large, and
larger than the number of parameters to optimize.
Warning: If the function is linear, this is a linear-algebra problem, and should be solved with scipy.
linalg.lstsq().
Least square problems occur often when

fitting a non-linear to data. While it is possible to construct our optimization problem ourselves, scipy provides
a helper function for this purpose: scipy.optimize.curve_fit():
>>> def f(t, omega, phi):

... return np.cos(omega * t + phi)
>>> x = np.linspace(0, 3, 50)

>>> y = f(x, 1.5, 1) + .1*np.random.normal(size=50)
>>> optimize.curve_fit(f, x, y)
(array([ 1.5185..., 0.92665...]), array([[ 0.00037..., -0.00056...],
[-0.0005..., 0.00123...]]))
Exercise
Do the same with omega = 3. What is the difficulty?
13.6. Special case: non-linear least-squares 436

13.7 Optimization with constraints
13.7.1 Box bounds
Box bounds correspond to limiting each of the individual parameters of the optimization. Note that some
problems that are not originally written as box bounds can be rewritten as such via change of variables. Both
scipy.optimize.minimize_scalar() and scipy.optimize.minimize() support bound constraints with
the parameter bounds:
>>> def f(x):

... return np.sqrt((x[0] - 3)**2 + (x[1] - 2)**2)
>>> optimize.minimize(f, np.array([0, 0]), bounds=((-1.5, 1.5), (-1.5, 1.5)))
fun: 1.5811...
jac: array([-0.94868..., -0.31622...])
nfev: 9
nit: 2
status: 0
success: True
x: array([ 1.5, 1.5])
13.7.2 General constraints
Equality and inequality constraints specified as functions: f (x) = 0 and g (x) < 0.
scipy.optimize.fmin_slsqp() Sequential least square programming: equality and inequality con-
straints:
>>> def f(x):

... return np.sqrt((x[0] - 3)**2 + (x[1] - 2)**2)
>>> def constraint(x):

... return np.atleast_1d(1.5 - np.sum(np.abs(x)))
>>> x0 = np.array([0, 0])
13.7. Optimization with constraints 437

>>> optimize.minimize(f, x0, constraints={"fun": constraint, "type": "ineq"})

fun: 2.4748...
jac: array([-0.70708..., -0.70712...])
nfev: 20
nit: 5
njev: 5
status: 0
success: True
x: array([ 1.2500..., 0.2499...])
Warning: The above problem is known as the Lasso problem in statistics, and there exist very efficient
solvers for it (for instance in scikit-learn). In general do not use generic solvers when specific ones exist.
Lagrange multipliers
If you are ready to do a bit of math, many constrained optimization problems can be converted to non-
constrained optimization problems using a mathematical trick known as Lagrange multipliers.
13.9 Examples for the mathematical optimization chapter


CHAPTER 14
Interfacing with C
Author: Valentin Haenel

This chapter contains an introduction to the many different routes for making your native code (primarily
C/C++) available from Python, a process commonly referred to wrapping. The goal of this chapter is to give
you a flavour of what technologies exist and what their respective merits and shortcomings are, so that you
can select the appropriate one for your specific needs. In any case, once you do start wrapping, you almost
certainly will want to consult the respective documentation for your selected technique.
Chapters contents
Introduction
Python-C-Api
Ctypes
SWIG
Cython
Summary
Further Reading and References
Exercises
14.1 Introduction
This chapter covers the following techniques:

Python-C-Api
439
Ctypes
SWIG (Simplified Wrapper and Interface Generator)
Cython
These four techniques are perhaps the most well known ones, of which Cython is probably the most advanced
one and the one you should consider using first. The others are also important, if you want to understand the
wrapping problem from different angles. Having said that, there are other alternatives out there, but having
understood the basics of the ones above, you will be in a position to evaluate the technique of your choice to
see if it fits your needs.
The following criteria may be useful when evaluating a technology:
Are additional libraries required?
Is the code autogenerated?
Does it need to be compiled?
Is there good support for interacting with Numpy arrays?
Does it support C++?
Before you set out, you should consider your use case. When interfacing with native code, there are usually
two use-cases that come up:
Existing code in C/C++ that needs to be leveraged, either because it already exists, or because it is faster.
Python code too slow, push inner loops to native code
Each technology is demonstrated by wrapping the cos function from math.h. While this is a mostly a trivial
example, it should serve us well to demonstrate the basics of the wrapping solution. Since each technique
also includes some form of Numpy support, this is also demonstrated using an example where the cosine is
computed on some kind of array.
Last but not least, two small warnings:
All of these techniques may crash (segmentation fault) the Python interpreter, which is (usually) due to
bugs in the C code.
All the examples have been done on Linux, they should be possible on other operating systems.
You will need a C compiler for most of the examples.
14.2 Python-C-Api
The Python-C-API is the backbone of the standard Python interpreter (a.k.a CPython). Using this API it is
possible to write Python extension module in C and C++. Obviously, these extension modules can, by virtue of
language compatibility, call any function written in C or C++.
When using the Python-C-API, one usually writes much boilerplate code, first to parse the arguments that were
given to a function, and later to construct the return type.
Advantages
Requires no additional libraries
Lots of low-level control
Entirely usable from C++
Disadvantages
May require a substantial amount of effort
Much overhead in the code
Must be compiled
14.2. Python-C-Api 440

High maintenance cost

No forward compatibility across Python versions as C-API changes
Reference count bugs are easy to create and very hard to track down.
Note: The Python-C-Api example here serves mainly for didactic reasons. Many of the other techniques
actually depend on this, so it is good to have a high-level understanding of how it works. In 99% of the use-
cases you will be better off, using an alternative technique.
Note: Since refernce counting bugs are easy to create and hard to track down, anyone really needing to use
the Python C-API should read the section about objects, types and reference counts from the official python
documentation. Additionally, there is a tool by the name of cpychecker which can help discover common
errors with reference counting.
14.2.1 Example
The following C-extension module, make the cos function from the standard math library available to Python:
/* Example of wrapping cos function from math.h with the Python-C-API. */
# include <Python.h>
# include <math.h>
/* wrapped cosine function */

static PyObject* cos_func(PyObject* self, PyObject* args)
{
double value;
double answer;
/* parse the input, from python float to c double */

if (!PyArg_ParseTuple(args, "d", &value))
return NULL;
/* if the above function returns -1, an appropriate Python exception will
* have been set, and the function simply returns NULL
*/
/* call cos from libm */

answer = cos(value);
/* construct the output from cos, from c double to python float */

return Py_BuildValue("f", answer);
}
/* define functions in module */

static PyMethodDef CosMethods[] =
{
{"cos_func", cos_func, METH_VARARGS, "evaluate the cosine"},
{NULL, NULL, 0, NULL}
};
# if PY_MAJOR_VERSION >= 3
/* module initialization */
/* Python version 3*/
static struct PyModuleDef cModPyDem =
{
PyModuleDef_HEAD_INIT,
"cos_module", "Some documentation",
-1,

CosMethods
};
PyMODINIT_FUNC
PyInit_cos_module(void)
{
return PyModule_Create(&cModPyDem);
}
# else
/* Python version 2 */
PyMODINIT_FUNC
initcos_module(void)
{
(void) Py_InitModule("cos_module", CosMethods);
}
# endif
As you can see, there is much boilerplate, both to massage the arguments and return types into place and for
the module initialisation. Although some of this is amortised, as the extension grows, the boilerplate required
for each function(s) remains.
The standard python build system distutils supports compiling C-extensions from a setup.py, which is
rather convenient:
from distutils.core import setup, Extension
# define the extension module

cos_module = Extension('cos_module', sources=['cos_module.c'])
# run the setup

setup(ext_modules=[cos_module])
This can be compiled:
$ cd advanced/interfacing_with_c/python_c_api
$ ls
cos_module.c setup.py
$ python setup.py build_ext --inplace

running build_ext
building 'cos_module' extension
creating build
creating build/temp.linux-x86_64-2.7
gcc -pthread -fno-strict-aliasing -g -O2 -DNDEBUG -g -fwrapv -O3 -Wall -Wstrict-prototypes -
,fPIC -I/home/esc/anaconda/include/python2.7 -c cos_module.c -o build/temp.linux-x86_64-2.7/
,cos_module.o
gcc -pthread -shared build/temp.linux-x86_64-2.7/cos_module.o -L/home/esc/anaconda/lib -
,lpython2.7 -o /home/esc/git-working/scipy-lecture-notes/advanced/interfacing_with_c/python_c_
,api/cos_module.so
$ ls
build/ cos_module.c cos_module.so setup.py
build_ext is to build extension modules

--inplace will output the compiled extension module into the current directory
The file cos_module.so contains the compiled extension, which we can now load in the IPython interpreter:

Note: In Python 3, the filename for compiled modules includes metadata on the Python interpreter (see PEP
3149) and is thus longer. The import statement is not affected by this.
In [1]: import cos_module
In [2]: cos_module?
Type: module
String Form:<module 'cos_module' from 'cos_module.so'>
File: /home/esc/git-working/scipy-lecture-notes/advanced/interfacing_with_c/python_c_api/
,cos_module.so
Docstring: <no docstring>
In [3]: dir(cos_module)
Out[3]: ['__doc__', '__file__', '__name__', '__package__', 'cos_func']
In [4]: cos_module.cos_func(1.0)
Out[4]: 0.5403023058681398
Out[5]: 1.0
Out[7]: -1.0
Now lets see how robust this is:
In [10]: cos_module.cos_func('foo')
---------------------------------------------------------------------------
TypeError Traceback (most recent call last)
<ipython-input-10-11bee483665d> in <module>()
----> 1 cos_module.cos_func('foo')
TypeError: a float is required
14.2.2 Numpy Support
Analog to the Python-C-API, Numpy, which is itself implemented as a C-extension, comes with the Numpy-C-
API. This API can be used to create and manipulate Numpy arrays from C, when writing a custom C-extension.
See also: Advanced NumPy.
Note: If you do ever need to use the Numpy C-API refer to the documentation about Arrays and Iterators.
The following example shows how to pass Numpy arrays as arguments to functions and how to iterate over
Numpy arrays using the (old) Numpy-C-API. It simply takes an array as argument applies the cosine function
from the math.h and returns a resulting new array.
/* Example of wrapping the cos function from math.h using the Numpy-C-API. */
# include <Python.h>
# include <numpy/arrayobject.h>
# include <math.h>
/* wrapped cosine function */

static PyObject* cos_func_np(PyObject* self, PyObject* args)
{
PyArrayObject *in_array;

PyObject *out_array;
NpyIter *in_iter;
NpyIter *out_iter;
NpyIter_IterNextFunc *in_iternext;
NpyIter_IterNextFunc *out_iternext;
/* parse single numpy array argument */

if (!PyArg_ParseTuple(args, "O!", &PyArray_Type, &in_array))
return NULL;
/* construct the output array, like the input array */

out_array = PyArray_NewLikeArray(in_array, NPY_ANYORDER, NULL, 0);
if (out_array == NULL)
return NULL;
/* create the iterators */

in_iter = NpyIter_New(in_array, NPY_ITER_READONLY, NPY_KEEPORDER,
NPY_NO_CASTING, NULL);
if (in_iter == NULL)
goto fail;
out_iter = NpyIter_New((PyArrayObject *)out_array, NPY_ITER_READWRITE,

NPY_KEEPORDER, NPY_NO_CASTING, NULL);
if (out_iter == NULL) {
NpyIter_Deallocate(in_iter);
goto fail;
}
in_iternext = NpyIter_GetIterNext(in_iter, NULL);

out_iternext = NpyIter_GetIterNext(out_iter, NULL);
if (in_iternext == NULL || out_iternext == NULL) {
NpyIter_Deallocate(out_iter);
goto fail;
}
double ** in_dataptr = (double **) NpyIter_GetDataPtrArray(in_iter);
double ** out_dataptr = (double **) NpyIter_GetDataPtrArray(out_iter);
/* iterate over the arrays */

do {
**out_dataptr = cos(**in_dataptr);
} while(in_iternext(in_iter) && out_iternext(out_iter));
/* clean up and return the result */

NpyIter_Deallocate(out_iter);
Py_INCREF(out_array);
return out_array;
/* in case bad things happen */

fail:
Py_XDECREF(out_array);
return NULL;
}
/* define functions in module */

static PyMethodDef CosMethods[] =
{
{"cos_func_np", cos_func_np, METH_VARARGS,
"evaluate the cosine on a numpy array"},
{NULL, NULL, 0, NULL}
};

PyMODINIT_FUNC
initcos_module_np(void)
{
(void) Py_InitModule("cos_module_np", CosMethods);
/* IMPORTANT: this must be called */
import_array();
}
To compile this we can use distutils again. However we need to be sure to include the Numpy headers by using
:func:numpy.get_include.

import numpy
# define the extension module

cos_module_np = Extension('cos_module_np', sources=['cos_module_np.c'],
include_dirs=[numpy.get_include()])
# run the setup

setup(ext_modules=[cos_module_np])
To convince ourselves if this does actually works, we run the following test script:
import cos_module_np
import numpy as np
import pylab
x = np.arange(0, 2 * np.pi, 0.1)

y = cos_module_np.cos_func_np(x)
pylab.plot(x, y)
pylab.show()
And this should result in the following figure:
14.3 Ctypes
Ctypes is a foreign function library for Python. It provides C compatible data types, and allows calling functions
in DLLs or shared libraries. It can be used to wrap these libraries in pure Python.
Advantages
Part of the Python standard library
Does not need to be compiled
Wrapping code entirely in Python
Disadvantages
14.3. Ctypes 445

Requires code to be wrapped to be available as a shared library (roughly speaking *.dll in Windows
*.so in Linux and *.dylib in Mac OSX.)
No good support for C++
14.3.1 Example
As advertised, the wrapper code is in pure Python.
""" Example of wrapping cos function from math.h using ctypes. """
import ctypes
# find and load the library
# OSX or linux
from ctypes.util import find_library
libm = ctypes.cdll.LoadLibrary(find_library('m'))
# Windows
# from ctypes import windll
# libm = cdll.msvcrt
# set the argument type

libm.cos.argtypes = [ctypes.c_double]
# set the return type
libm.cos.restype = ctypes.c_double
def cos_func(arg):
''' Wrapper for cos from math.h '''
return libm.cos(arg)
Finding and loading the library may vary depending on your operating system, check the documentation
for details
This may be somewhat deceptive, since the math library exists in compiled form on the system already.
If you were to wrap a in-house library, you would have to compile it first, which may or may not require
some additional effort.
We may now use this, as before:
In [2]: cos_module?
Type: module
String Form:<module 'cos_module' from 'cos_module.py'>
File: /home/esc/git-working/scipy-lecture-notes/advanced/interfacing_with_c/ctypes/cos_
,module.py
Out[3]:
['__builtins__',
'__doc__',
'__file__',
'__name__',
'__package__',
'cos_func',
'ctypes',
'find_library',
'libm']
14.3. Ctypes 446

Out[4]: 0.5403023058681398
Out[5]: 1.0
Out[6]: -1.0
As with the previous example, this code is somewhat robust, although the error message is not quite as helpful,
since it does not tell us what the type should be.
---------------------------------------------------------------------------
ArgumentError Traceback (most recent call last)
/home/esc/git-working/scipy-lecture-notes/advanced/interfacing_with_c/ctypes/cos_module.py in
,cos_func(arg)
12 def cos_func(arg):
13 ''' Wrapper for cos from math.h '''
---> 14 return libm.cos(arg)
ArgumentError: argument 1: <type 'exceptions.TypeError'>: wrong type
Numpy contains some support for interfacing with ctypes. In particular there is support for exporting certain
attributes of a Numpy array as ctypes data-types and there are functions to convert from C arrays to Numpy
arrays and back.
For more information, consult the corresponding section in the Numpy Cookbook and the API documentation
for numpy.ndarray.ctypes and numpy.ctypeslib.
For the following example, lets consider a C function in a library that takes an input and an output array,
computes the cosine of the input array and stores the result in the output array.
The library consists of the following header file (although this is not strictly needed for this example, we list it
for completeness):
void cos_doubles(double * in_array, double * out_array, int size);
The function implementation resides in the following C source file:
# include <math.h>
/* Compute the cosine of each element in in_array, storing the result in

* out_array. */
void cos_doubles(double * in_array, double * out_array, int size){
int i;
for(i=0;i<size;i++){
out_array[i] = cos(in_array[i]);
}
}
And since the library is pure C, we cant use distutils to compile it, but must use a combination of make and
gcc:
14.3. Ctypes 447

m.PHONY : clean
libcos_doubles.so : cos_doubles.o
gcc -shared -Wl,-soname,libcos_doubles.so -o libcos_doubles.so cos_doubles.o
cos_doubles.o : cos_doubles.c
gcc -c -fPIC cos_doubles.c -o cos_doubles.o
clean :
-rm -vf libcos_doubles.so cos_doubles.o cos_doubles.pyc
We can then compile this (on Linux) into the shared library libcos_doubles.so:
$ ls
cos_doubles.c cos_doubles.h cos_doubles.py makefile test_cos_doubles.py
$ make
gcc -c -fPIC cos_doubles.c -o cos_doubles.o
gcc -shared -Wl,-soname,libcos_doubles.so -o libcos_doubles.so cos_doubles.o
$ ls
cos_doubles.c cos_doubles.o libcos_doubles.so* test_cos_doubles.py
cos_doubles.h cos_doubles.py makefile
Now we can proceed to wrap this library via ctypes with direct support for (certain kinds of) Numpy arrays:
""" Example of wrapping a C library function that accepts a C double array as

input using the numpy.ctypeslib. """
import numpy as np
import numpy.ctypeslib as npct
from ctypes import c_int
# input type for the cos_doubles function

# must be a double array, with single dimension that is contiguous
array_1d_double = npct.ndpointer(dtype=np.double, ndim=1, flags='CONTIGUOUS')
# load the library, using numpy mechanisms

libcd = npct.load_library("libcos_doubles", ".")
# setup the return types and argument types

libcd.cos_doubles.restype = None
libcd.cos_doubles.argtypes = [array_1d_double, array_1d_double, c_int]
def cos_doubles_func(in_array, out_array):

return libcd.cos_doubles(in_array, out_array, len(in_array))
Note the inherent limitation of contiguous single dimensional Numpy arrays, since the C functions re-
quires this kind of buffer.
Also note that the output array must be preallocated, for example with numpy.zeros() and the function
will write into its buffer.
Although the original signature of the cos_doubles function is ARRAY, ARRAY, int the final
cos_doubles_func takes only two Numpy arrays as arguments.
And, as before, we convince ourselves that it worked:
import numpy as np
import pylab
import cos_doubles

y = np.empty_like(x)
14.3. Ctypes 448

cos_doubles.cos_doubles_func(x, y)
pylab.plot(x, y)
pylab.show()
14.4 SWIG
SWIG, the Simplified Wrapper Interface Generator, is a software development tool that connects programs
written in C and C++ with a variety of high-level programming languages, including Python. The important
thing with SWIG is, that it can autogenerate the wrapper code for you. While this is an advantage in terms of
development time, it can also be a burden. The generated file tend to be quite large and may not be too human
readable and the multiple levels of indirection which are a result of the wrapping process, may be a bit tricky
to understand.
Note: The autogenerated C code uses the Python-C-Api.
Advantages
Can automatically wrap entire libraries given the headers
Works nicely with C++
Disadvantages
Autogenerates enormous files
Hard to debug if something goes wrong
Steep learning curve
14.4.1 Example
Lets imagine that our cos function lives in a cos_module which has been written in c and consists of the
source file cos_module.c:
# include <math.h>
double cos_func(double arg){

return cos(arg);
}
and the header file cos_module.h:
double cos_func(double arg);
14.4. SWIG 449

And our goal is to expose the cos_func to Python. To achieve this with SWIG, we must write an interface file
which contains the instructions for SWIG.
/* Example of wrapping cos function from math.h using SWIG. */
%module cos_module
%{
/* the resulting C file should be built as a python extension */
# define SWIG_FILE_WITH_INIT
/* Includes the header in the wrapper code */
# include "cos_module.h"
%}
/* Parse the header file to generate wrappers */
%include "cos_module.h"
As you can see, not too much code is needed here. For this simple example it is enough to simply include
the header file in the interface file, to expose the function to Python. However, SWIG does allow for more fine
grained inclusion/exclusion of functions found in header files, check the documentation for details.
Generating the compiled wrappers is a two stage process:
1. Run the swig executable on the interface file to generate the files cos_module_wrap.c, which is the
source file for the autogenerated Python C-extension and cos_module.py, which is the autogenerated
pure python module.
2. Compile the cos_module_wrap.c into the _cos_module.so. Luckily, distutils knows how to handle
SWIG interface files, so that our setup.py is simply:
setup(ext_modules=[Extension("_cos_module",
sources=["cos_module.c", "cos_module.i"])])
$ cd advanced/interfacing_with_c/swig
$ ls
cos_module.c cos_module.h cos_module.i setup.py

running build_ext
building '_cos_module' extension
swigging cos_module.i to cos_module_wrap.c
swig -python -o cos_module_wrap.c cos_module.i
creating build
,cos_module.o
,fPIC -I/home/esc/anaconda/include/python2.7 -c cos_module_wrap.c -o build/temp.linux-x86_64-
,2.7/cos_module_wrap.o
gcc -pthread -shared build/temp.linux-x86_64-2.7/cos_module.o build/temp.linux-x86_64-2.7/cos_
,module_wrap.o -L/home/esc/anaconda/lib -lpython2.7 -o /home/esc/git-working/scipy-lecture-
,notes/advanced/interfacing_with_c/swig/_cos_module.so
$ ls
build/ cos_module.c cos_module.h cos_module.i cos_module.py _cos_module.so* cos_module_
,wrap.c setup.py
We can now load and execute the cos_module as we have done in the previous examples:
14.4. SWIG 450

In [2]: cos_module?
Type: module
String Form:<module 'cos_module' from 'cos_module.py'>
File: /home/esc/git-working/scipy-lecture-notes/advanced/interfacing_with_c/swig/cos_
,module.py
Out[3]:
['__builtins__',
'__doc__',
'__file__',
'__name__',
'__package__',
'_cos_module',
'_newclass',
'_object',
'_swig_getattr',
'_swig_property',
'_swig_repr',
'_swig_setattr',
'_swig_setattr_nondynamic',
'cos_func']
Out[4]: 0.5403023058681398
Out[5]: 1.0
Out[6]: -1.0
Again we test for robustness, and we see that we get a better error message (although, strictly speaking in
Python there is no double type):
---------------------------------------------------------------------------
TypeError: in method 'cos_func', argument 1 of type 'double'
Numpy provides support for SWIG with the numpy.i file. This interface file defines various so-called typemaps
which support conversion between Numpy arrays and C-Arrays. In the following example we will take a quick
look at how such typemaps work in practice.
We have the same cos_doubles function as in the ctypes example:
# include <math.h>

* out_array. */
int i;
14.4. SWIG 451

}
}
This is wrapped as cos_doubles_func using the following SWIG interface file:
/* Example of wrapping a C function that takes a C double array as input using

* numpy typemaps for SWIG. */
%module cos_doubles
%{
/* the resulting C file should be built as a python extension */
# define SWIG_FILE_WITH_INIT
/* Includes the header in the wrapper code */
# include "cos_doubles.h"
%}
/* include the numpy typemaps */

%include "numpy.i"
/* need this for correct module initialization */
%init %{
import_array();
%}
/* typemaps for the two arrays, the second will be modified in-place */
%apply (double* IN_ARRAY1, int DIM1) {(double * in_array, int size_in)}
%apply (double* INPLACE_ARRAY1, int DIM1) {(double * out_array, int size_out)}
/* Wrapper for cos_doubles that massages the types */

%inline %{
/* takes as input two numpy arrays */
void cos_doubles_func(double * in_array, int size_in, double * out_array, int size_out) {
/* calls the original funcion, providing only the size of the first */
cos_doubles(in_array, out_array, size_in);
}
%}
To use the Numpy typemaps, we need include the numpy.i file.

Observe the call to import_array() which we encountered already in the Numpy-C-API example.
Since the type maps only support the signature ARRAY, SIZE we need to wrap the cos_doubles as
cos_doubles_func which takes two arrays including sizes as input.
As opposed to the simple SWIG example, we dont include the cos_doubles.h header, There is nothing
there that we wish to expose to Python since we expose the functionality through cos_doubles_func.
And, as before we can use distutils to wrap this:

import numpy
setup(ext_modules=[Extension("_cos_doubles",
sources=["cos_doubles.c", "cos_doubles.i"],
include_dirs=[numpy.get_include()])])
As previously, we need to use include_dirs to specify the location.
$ ls
cos_doubles.c cos_doubles.h cos_doubles.i numpy.i setup.py test_cos_doubles.py
$ python setup.py build_ext -i
running build_ext
building '_cos_doubles' extension
14.4. SWIG 452

swigging cos_doubles.i to cos_doubles_wrap.c

swig -python -o cos_doubles_wrap.c cos_doubles.i
cos_doubles.i:24: Warning(490): Fragment 'NumPy_Backward_Compatibility' not found.
creating build
,fPIC -I/home/esc/anaconda/lib/python2.7/site-packages/numpy/core/include -I/home/esc/
,anaconda/include/python2.7 -c cos_doubles.c -o build/temp.linux-x86_64-2.7/cos_doubles.o
,anaconda/include/python2.7 -c cos_doubles_wrap.c -o build/temp.linux-x86_64-2.7/cos_doubles_
,wrap.o
In file included from /home/esc/anaconda/lib/python2.7/site-packages/numpy/core/include/numpy/
,ndarraytypes.h:1722,
from /home/esc/anaconda/lib/python2.7/site-packages/numpy/core/include/numpy/
,ndarrayobject.h:17,
,arrayobject.h:15,
from cos_doubles_wrap.c:2706:
/home/esc/anaconda/lib/python2.7/site-packages/numpy/core/include/numpy/npy_deprecated_api.
,h:11:2: warning: #warning "Using deprecated NumPy API, disable it by #defining NPY_NO_
,DEPRECATED_API NPY_1_7_API_VERSION"
gcc -pthread -shared build/temp.linux-x86_64-2.7/cos_doubles.o build/temp.linux-x86_64-2.7/cos_
,doubles_wrap.o -L/home/esc/anaconda/lib -lpython2.7 -o /home/esc/git-working/scipy-lecture-
,notes/advanced/interfacing_with_c/swig_numpy/_cos_doubles.so
$ ls
build/ cos_doubles.h cos_doubles.py cos_doubles_wrap.c setup.py
cos_doubles.c cos_doubles.i _cos_doubles.so* numpy.i test_cos_doubles.py
import numpy as np
import pylab
import cos_doubles

pylab.plot(x, y)
pylab.show()
14.5 Cython
Cython is both a Python-like language for writing C-extensions and an advanced compiler for this language.
14.5. Cython 453

The Cython language is a superset of Python, which comes with additional constructs that allow you call C
functions and annotate variables and class attributes with c types. In this sense one could also call it a Python
with types.
In addition to the basic use case of wrapping native code, Cython supports an additional use-case, namely
interactive optimization. Basically, one starts out with a pure-Python script and incrementally adds Cython
types to the bottleneck code to optimize only those code paths that really matter.
In this sense it is quite similar to SWIG, since the code can be autogenerated but in a sense it also quite similar
to ctypes since the wrapping code can (almost) be written in Python.
While others solutions that autogenerate code can be quite difficult to debug (for example SWIG) Cython
comes with an extension to the GNU debugger that helps debug Python, Cython and C code.
Note: The autogenerated C code uses the Python-C-Api.
Advantages
Python like language for writing C-extensions
Autogenerated code
Supports incremental optimization
Includes a GNU debugger extension
Support for C++ (Since version 0.13)
Disadvantages
Must be compiled
Requires an additional library ( but only at build time, at this problem can be overcome by shipping the
generated C files)
14.5.1 Example
The main Cython code for our cos_module is contained in the file cos_module.pyx:
""" Example of wrapping cos function from math.h using Cython. """
cdef extern from "math.h":

double cos(double arg)
def cos_func(arg):
return cos(arg)
Note the additional keywords such as cdef and extern. Also the cos_func is then pure Python.
Again we can use the standard distutils module, but this time we need some additional pieces from the
Cython.Distutils:

from Cython.Distutils import build_ext
setup(
cmdclass={'build_ext': build_ext},
ext_modules=[Extension("cos_module", ["cos_module.pyx"])]
)
Compiling this:
14.5. Cython 454

$ cd advanced/interfacing_with_c/cython
$ ls
cos_module.pyx setup.py
running build_ext
cythoning cos_module.pyx to cos_module.c
building 'cos_module' extension
creating build
,cos_module.o
gcc -pthread -shared build/temp.linux-x86_64-2.7/cos_module.o -L/home/esc/anaconda/lib -
,lpython2.7 -o /home/esc/git-working/scipy-lecture-notes/advanced/interfacing_with_c/cython/
,cos_module.so
$ ls
build/ cos_module.c cos_module.pyx cos_module.so* setup.py
And running it:
In [2]: cos_module?
Type: module
String Form:<module 'cos_module' from 'cos_module.so'>
File: /home/esc/git-working/scipy-lecture-notes/advanced/interfacing_with_c/cython/cos_
,module.so
Out[3]:
['__builtins__',
'__doc__',
'__file__',
'__name__',
'__package__',
'__test__',
'cos_func']
Out[4]: 0.5403023058681398
Out[5]: 1.0
Out[6]: -1.0
And, testing a little for robustness, we can see that we get good error messages:
---------------------------------------------------------------------------
/home/esc/git-working/scipy-lecture-notes/advanced/interfacing_with_c/cython/cos_module.so in
,cos_module.cos_func (cos_module.c:506)()
TypeError: a float is required
Additionally, it is worth noting that Cython ships with complete declarations for the C math library, which
14.5. Cython 455

simplifies the code above to become:
""" Simpler example of wrapping cos function from math.h using Cython. """
from libc.math cimport cos
def cos_func(arg):
return cos(arg)
In this case the cimport statement is used to import the cos function.
Cython has support for Numpy via the numpy.pyx file which allows you to add the Numpy array type to your
Cython code. I.e. like specifying that variable i is of type int, you can specify that variable a is of type numpy.
ndarray with a given dtype. Also, certain optimizations such as bounds checking are supported. Look at the
corresponding section in the Cython documentation. In case you want to pass Numpy arrays as C arrays to
your Cython wrapped C functions, there is a section about this in the Cython wiki.
In the following example, we will show how to wrap the familiar cos_doubles function using Cython.
# include <math.h>

* out_array. */
int i;
}
}
This is wrapped as cos_doubles_func using the following Cython code:
""" Example of wrapping a C function that takes C double arrays as input using
the Numpy declarations from Cython """
# cimport the Cython declarations for numpy

cimport numpy as np
# if you want to use the Numpy-C-API from Cython

# (not strictly necessary for this example, but good practice)
np.import_array()
# cdefine the signature of our c function

cdef extern from "cos_doubles.h":
void cos_doubles (double * in_array, double * out_array, int size)
# create the wrapper code, with numpy type annotations

def cos_doubles_func(np.ndarray[double, ndim=1, mode="c"] in_array not None,
np.ndarray[double, ndim=1, mode="c"] out_array not None):
cos_doubles(<double*> np.PyArray_DATA(in_array),
<double*> np.PyArray_DATA(out_array),
in_array.shape[0])
And can be compiled using distutils:

import numpy
14.5. Cython 456

from Cython.Distutils import build_ext
setup(
cmdclass={'build_ext': build_ext},
ext_modules=[Extension("cos_doubles",
sources=["_cos_doubles.pyx", "cos_doubles.c"],
include_dirs=[numpy.get_include()])],
)
As with the previous compiled Numpy examples, we need the include_dirs option.
$ ls
cos_doubles.c cos_doubles.h _cos_doubles.pyx setup.py test_cos_doubles.py
$ python setup.py build_ext -i
running build_ext
cythoning _cos_doubles.pyx to _cos_doubles.c
building 'cos_doubles' extension
creating build
,anaconda/include/python2.7 -c _cos_doubles.c -o build/temp.linux-x86_64-2.7/_cos_doubles.o
In file included from /home/esc/anaconda/lib/python2.7/site-packages/numpy/core/include/numpy/
,ndarraytypes.h:1722,
,ndarrayobject.h:17,
,arrayobject.h:15,
from _cos_doubles.c:253:
/home/esc/anaconda/lib/python2.7/site-packages/numpy/core/include/numpy/npy_deprecated_api.
,h:11:2: warning: #warning "Using deprecated NumPy API, disable it by #defining NPY_NO_
,DEPRECATED_API NPY_1_7_API_VERSION"
/home/esc/anaconda/lib/python2.7/site-packages/numpy/core/include/numpy/__ufunc_api.h:236:
,warning: _import_umath defined but not used
,anaconda/include/python2.7 -c cos_doubles.c -o build/temp.linux-x86_64-2.7/cos_doubles.o
gcc -pthread -shared build/temp.linux-x86_64-2.7/_cos_doubles.o build/temp.linux-x86_64-2.7/
,cos_doubles.o -L/home/esc/anaconda/lib -lpython2.7 -o /home/esc/git-working/scipy-lecture-
,notes/advanced/interfacing_with_c/cython_numpy/cos_doubles.so
$ ls
build/ _cos_doubles.c cos_doubles.c cos_doubles.h _cos_doubles.pyx cos_doubles.so* setup.
,py test_cos_doubles.py
import numpy as np
import pylab
import cos_doubles

pylab.plot(x, y)
pylab.show()
14.5. Cython 457

14.6 Summary
In this section four different techniques for interfacing with native code have been presented. The table below
roughly summarizes some of the aspects of the techniques.
x Part of CPython Compiled Autogenerated Numpy Support

Python-C-API True True False True
Ctypes True False False True
Swig False True True True
Cython False True True True
Of all three presented techniques, Cython is the most modern and advanced. In particular, the ability to opti-
mize code incrementally by adding types to your Python code is unique.
14.7 Further Reading and References
Gal Varoquauxs blog post about avoiding data copies provides some insight on how to handle memory
management cleverly. If you ever run into issues with large datasets, this is a reference to come back to
for some inspiration.
14.8 Exercises
Since this is a brand new section, the exercises are considered more as pointers as to what to look at next, so
pick the ones that you find more interesting. If you have good ideas for exercises, please let us know!
1. Download the source code for each example and compile and run them on your machine.
2. Make trivial changes to each example and convince yourself that this works. ( E.g. change cos for sin.)
3. Most of the examples, especially the ones involving Numpy may still be fragile and respond badly to
input errors. Look for ways to crash the examples, figure what the problem is and devise a potential
solution. Here are some ideas:
(a) Numerical overflow.
(b) Input and output arrays that have different lengths.
(c) Multidimensional array.
(d) Empty array
(e) Arrays with non-double types
4. Use the %timeit IPython magic to measure the execution time of the various solutions
14.6. Summary 458

14.8.1 Python-C-API
1. Modify the Numpy example such that the function takes two input arguments, where the second is the
preallocated output array, making it similar to the other Numpy examples.
2. Modify the example such that the function only takes a single input array and modifies this in place.
3. Try to fix the example to use the new Numpy iterator protocol. If you manage to obtain a working solu-
tion, please submit a pull-request on github.
4. You may have noticed, that the Numpy-C-API example is the only Numpy example that does not wrap
cos_doubles but instead applies the cos function directly to the elements of the Numpy array. Does
this have any advantages over the other techniques.
5. Can you wrap cos_doubles using only the Numpy-C-API. You may need to ensure that the arrays have
the correct type, are one dimensional and contiguous in memory.
14.8.2 Ctypes
1. Modify the Numpy example such that cos_doubles_func handles the preallocation for you, thus mak-
ing it more like the Numpy-C-API example.
14.8.3 SWIG
1. Look at the code that SWIG autogenerates, how much of it do you understand?
3. Modify the cos_doubles C function so that it returns an allocated array. Can you wrap this using SWIG
typemaps? If not, why not? Is there a workaround for this specific situation? (Hint: you know the size of
the output array, so it may be possible to construct a Numpy array from the returned double *.)
14.8.4 Cython
1. Look at the code that Cython autogenerates. Take a closer look at some of the comments that Cython
inserts. What do you see?
2. Look at the section Working with Numpy from the Cython documentation to learn how to incrementally
optimize a pure python script that uses Numpy.
14.8. Exercises 459

Part III
Packages and applications
460
This part of the Scipy lecture notes is dedicated to various scientific packages useful for extended needs.
461
CHAPTER 15
Statistics in Python
Requirements
Standard scientific Python environment (numpy, scipy, matplotlib)

Pandas
Statsmodels
Seaborn
To install Python and these dependencies, we recommend that you download Anaconda Python or En-
thought Canopy, or preferably use the package manager if you are under Ubuntu or other linux.
See also:
Bayesian statistics in Python: This chapter does not cover tools for Bayesian statistics. Of particular
interest for Bayesian modelling is PyMC, which implements a probabilistic programming language in
Python.
Read a statistics book: The Think stats book is available as free PDF or in print and is a great introduction
to statistics.
Tip: Why Python for statistics?
462
R is a language dedicated to statistics. Python is a general-purpose language with statistics modules. R has
more statistical analysis features than Python, and specialized syntaxes. However, when it comes to building
complex analysis pipelines that mix statistics with e.g. image analysis, text mining, or control of a physical
experiment, the richness of Python is an invaluable asset.
Contents
Data representation and interaction

Data as a table
The pandas data-frame
Hypothesis testing: comparing two groups
Students t-test: the simplest statistical test
Paired tests: repeated measurements on the same indivuals
Linear models, multiple factors, and analysis of variance
formulas to specify statistical models in Python
Multiple Regression: including multiple factors
Post-hoc hypothesis testing: analysis of variance (ANOVA)
More visualization: seaborn for statistical exploration
Pairplot: scatter matrices
lmplot: plotting a univariate regression
Testing for interactions
Full code for the figures
Solutions to this chapters exercises
Tip: In this document, the Python inputs are represented with the sign >>>.
Disclaimer: Gender questions

Some of the examples of this tutorial are chosen around gender questions. The reason is that on such questions
controlling the truth of a claim actually matters to many people.
15.1 Data representation and interaction
15.1.1 Data as a table
The setting that we consider for statistical analysis is that of multiple observations or samples described by a
set of different attributes or features. The data can than be seen as a 2D table, or matrix, with columns giving
the different attributes of the data, and rows the observations. For instance, the data contained in examples/
brain_size.csv:
15.1. Data representation and interaction 463

"";"Gender";"FSIQ";"VIQ";"PIQ";"Weight";"Height";"MRI_Count"
"1";"Female";133;132;124;"118";"64.5";816932
"2";"Male";140;150;124;".";"72.5";1001121
"3";"Male";139;123;150;"143";"73.3";1038437
"4";"Male";133;129;128;"172";"68.8";965353
"5";"Female";137;132;134;"147";"65.0";951545
15.1.2 The pandas data-frame
Tip: We will store and manipulate this data in a pandas.DataFrame, from the pandas module. It is the Python
equivalent of the spreadsheet table. It is different from a 2D numpy array as it has named columns, can contain
a mixture of different data types by column, and has elaborate selection and pivotal mechanisms.
Creating dataframes: reading data files or converting arrays
Separator
It is a CSV file, but the separator is ;
Reading from a CSV file: Using the above CSV file that gives observations of brain size and weight and IQ
(Willerman et al. 1991), the data are a mixture of numerical and categorical values:
>>> import pandas

>>> data = pandas.read_csv('examples/brain_size.csv', sep=';', na_values=".")
>>> data
Unnamed: 0 Gender FSIQ VIQ PIQ Weight Height MRI_Count
0 1 Female 133 132 124 118.0 64.5 816932
1 2 Male 140 150 124 NaN 72.5 1001121
2 3 Male 139 123 150 143.0 73.3 1038437
3 4 Male 133 129 128 172.0 68.8 965353
4 5 Female 137 132 134 147.0 65.0 951545
...
Warning: Missing values

The weight of the second individual is missing in the CSV file. If we dont specify the missing value (NA =
not available) marker, we will not be able to do statistical analysis.
Creating from arrays: A pandas.DataFrame can also be seen as a dictionary of 1D series, eg arrays or lists. If
we have 3 numpy arrays:

>>> t = np.linspace(-6, 6, 20)
>>> sin_t = np.sin(t)
>>> cos_t = np.cos(t)
We can expose them as a pandas.DataFrame:

>>> pandas.DataFrame({'t': t, 'sin': sin_t, 'cos': cos_t})

cos sin t
0 0.960170 0.279415 -6.000000
1 0.609977 0.792419 -5.368421
2 0.024451 0.999701 -4.736842
3 -0.570509 0.821291 -4.105263
4 -0.945363 0.326021 -3.473684
5 -0.955488 -0.295030 -2.842105
6 -0.596979 -0.802257 -2.210526
7 -0.008151 -0.999967 -1.578947
8 0.583822 -0.811882 -0.947368
...
Other inputs: pandas can input data from SQL, excel files, or other formats. See the pandas documentation.
Manipulating data
data is a pandas.DataFrame, that resembles Rs dataframe:
>>> data.shape # 40 rows and 8 columns

(40, 8)
>>> data.columns # It has columns

Index([u'Unnamed: 0', u'Gender', u'FSIQ', u'VIQ', u'PIQ', u'Weight', u'Height', u'MRI_Count'],
,dtype='object')
>>> print(data['Gender']) # Columns can be addressed by name

0 Female
1 Male
2 Male
3 Male
4 Female
...
>>> # Simpler selector

>>> data[data['Gender'] == 'Female']['VIQ'].mean()
109.45
Note: For a quick view on a large dataframe, use its describe method: pandas.DataFrame.describe().
groupby: splitting a dataframe on values of categorical variables:
>>> groupby_gender = data.groupby('Gender')

>>> for gender, value in groupby_gender['VIQ']:
... print((gender, value.mean()))

('Female', 109.45)
('Male', 115.25)
groupby_gender is a powerful object that exposes many operations on the resulting group of dataframes:
>>> groupby_gender.mean()
Unnamed: 0 FSIQ VIQ PIQ Weight Height MRI_Count
Gender
Female 19.65 111.9 109.45 110.45 137.200000 65.765000 862654.6
Male 21.35 115.0 115.25 111.60 166.444444 71.431579 954855.4
Tip: Use tab-completion on groupby_gender to find more. Other common grouping functions are median,
count (useful for checking to see the amount of missing values in different subsets) or sum. Groupby evalua-
tion is lazy, no work is done until an aggregation function is applied.
Exercise
What is the mean value for VIQ for the full population?
How many males/females were included in this study?
Hint use tab completion to find out the methods that can be called, instead of mean in the above
example.
What is the average value of MRI counts expressed in log units, for males and females?
Note: groupby_gender.boxplot is used for the plots above (see this example).

Plotting data
Pandas comes with some plotting tools (pandas.tools.plotting, using matplotlib behind the scene) to dis-
play statistics of the data in dataframes:
Scatter matrices:
>>> from pandas.tools import plotting

>>> plotting.scatter_matrix(data[['Weight', 'Height', 'MRI_Count']])
>>> plotting.scatter_matrix(data[['PIQ', 'VIQ', 'FSIQ']])
Two populations
The IQ metrics are bimodal, as if there are 2 sub-populations.

Exercise
Plot the scatter matrix for males only, and for females only. Do you think that the 2 sub-populations corre-
spond to gender?
15.2 Hypothesis testing: comparing two groups
For simple statistical tests, we will use the scipy.stats sub-module of scipy:
>>> from scipy import stats
See also:
Scipy is a vast library. For a quick summary to the whole library, see the scipy chapter.
15.2. Hypothesis testing: comparing two groups 468

15.2.1 Students t-test: the simplest statistical test
1-sample t-test: testing the value of a population mean
scipy.stats.ttest_1samp() tests if the population mean of data is likely to be equal to a given value (tech-
nically if observations are drawn from a Gaussian distributions of given population mean). It returns the T
statistic, and the p-value (see the functions help):
>>> stats.ttest_1samp(data['VIQ'], 0)
Ttest_1sampResult(statistic=30.088099970..., pvalue=1.32891964...e-28)
Tip: With a p-value of 10^-28 we can claim that the population mean for the IQ (VIQ measure) is not 0.
2-sample t-test: testing for difference across populations
We have seen above that the mean VIQ in the male and female populations were different. To test if this is
significant, we do a 2-sample t-test with scipy.stats.ttest_ind():
>>> female_viq = data[data['Gender'] == 'Female']['VIQ']

>>> male_viq = data[data['Gender'] == 'Male']['VIQ']
>>> stats.ttest_ind(female_viq, male_viq)
Ttest_indResult(statistic=-0.77261617232..., pvalue=0.4445287677858...)

15.2.2 Paired tests: repeated measurements on the same indivuals
PIQ, VIQ, and FSIQ give 3 measures of IQ. Let us test if FISQ
and PIQ are significantly different. We can use a 2 sample test:
>>> stats.ttest_ind(data['FSIQ'], data['PIQ'])

Ttest_indResult(statistic=0.46563759638..., pvalue=0.64277250...)
The problem with this approach is that it forgets that there are links between observations: FSIQ and PIQ are
measured on the same individuals. Thus the variance due to inter-subject variability is confounding, and can
be removed, using a paired test, or repeated measures test:
>>> stats.ttest_rel(data['FSIQ'], data['PIQ'])

Ttest_relResult(statistic=1.784201940..., pvalue=0.082172638183...)
This is equivalent to a 1-sample test on the difference:
>>> stats.ttest_1samp(data['FSIQ'] - data['PIQ'], 0)

Ttest_1sampResult(statistic=1.784201940..., pvalue=0.082172638...)
T-tests assume Gaussian errors. We can use a Wilcoxon signed-rank test, that relaxes this assumption:
>>> stats.wilcoxon(data['FSIQ'], data['PIQ'])

WilcoxonResult(statistic=274.5, pvalue=0.106594927...)
Note: The corresponding test in the non paired case is the MannWhitney U test, scipy.stats.
mannwhitneyu().

Exercise
Test the difference between weights in males and females.

Use non parametric statistics to test the difference between VIQ in males and females.
Conclusion: we find that the data does not support the hypothesis that males and females have different
VIQ.
15.3 Linear models, multiple factors, and analysis of variance
15.3.1 formulas to specify statistical models in Python
A simple linear regression
Given two set of observations, x and y, we want to test the

hypothesis that y is a linear function of x. In other terms:
y = x coe f + i nt er cept + e
where e is observation noise. We will use the statsmodels module to:
1. Fit a linear model. We will use the simplest strategy, ordinary least squares (OLS).
2. Test that coef is non zero.
First, we generate simulated data according to the model:

>>> x = np.linspace(-5, 5, 20)
>>> # normal distributed noise
>>> y = -5 + 3*x + 4 * np.random.normal(size=x.shape)
>>> # Create a data frame containing all the relevant variables
>>> data = pandas.DataFrame({'x': x, 'y': y})
15.3. Linear models, multiple factors, and analysis of variance 471

formulas for statistics in Python
See the statsmodels documentation
Then we specify an OLS model and fit it:
>>> from statsmodels.formula.api import ols

>>> model = ols("y ~ x", data).fit()
We can inspect the various statistics derived from the fit:
>>> print(model.summary())
OLS Regression Results
==========================...
Dep. Variable: y R-squared: 0.804
Model: OLS Adj. R-squared: 0.794
Method: Least Squares F-statistic: 74.03
Date: ... Prob (F-statistic): 8.56e-08
Time: ... Log-Likelihood: -57.988
No. Observations: 20 AIC: 120.0
Df Residuals: 18 BIC: 122.0
Df Model: 1
Covariance Type: nonrobust
==========================...
coef std err t P>|t| [95.0% Conf. Int.]
------------------------------------------...
Intercept -5.5335 1.036 -5.342 0.000 -7.710 -3.357
x 2.9369 0.341 8.604 0.000 2.220 3.654
==========================...
Omnibus: 0.100 Durbin-Watson: 2.956
Prob(Omnibus): 0.951 Jarque-Bera (JB): 0.322
Skew: -0.058 Prob(JB): 0.851
Kurtosis: 2.390 Cond. No. 3.03
==========================...
Warnings:
[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.
Terminology:
Statsmodels uses a statistical terminology: the y variable in statsmodels is called endogenous while the x
variable is called exogenous. This is discussed in more detail here.
To simplify, y (endogenous) is the value you are trying to predict, while x (exogenous) represents the features
you are using to make the prediction.
Exercise
Retrieve the estimated parameters from the model above. Hint: use tab-completion to find the relevent
attribute.

Categorical variables: comparing groups or multiple categories
Let us go back the data on brain size:
>>> data = pandas.read_csv('examples/brain_size.csv', sep=';', na_values=".")
We can write a comparison between IQ of male and female using a linear model:
>>> model = ols("VIQ ~ Gender + 1", data).fit()

==========================...
Dep. Variable: VIQ R-squared: 0.015
Model: OLS Adj. R-squared: -0.010
Date: ... Prob (F-statistic): 0.445
Df Model: 1
==========================...
-----------------------------------------------------------------------...
Intercept 109.4500 5.308 20.619 0.000 98.704 120.196
Gender[T.Male] 5.8000 7.507 0.773 0.445 -9.397 20.997
==========================...
Skew: 0.010 Prob(JB): 0.157
==========================...
Warnings:
Tips on specifying model
Forcing categorical: the Gender is automatically detected as a categorical variable, and thus each of its
different values are treated as different entities.
An integer column can be forced to be treated as categorical using:
>>> model = ols('VIQ ~ C(Gender)', data).fit()
Intercept: We can remove the intercept using - 1 in the formula, or force the use of an intercept using + 1.
Tip: By default, statsmodels treats a categorical variable with K possible values as K-1 dummy boolean
variables (the last level being absorbed into the intercept term). This is almost always a good default
choice - however, it is possible to specify different encodings for categorical variables (http://statsmodels.
sourceforge.net/devel/contrasts.html).

Link to t-tests between different FSIQ and PIQ
To compare different types of IQ, we need to create a long-form table, listing IQs, where the type of IQ is
indicated by a categorical variable:
>>> data_fisq = pandas.DataFrame({'iq': data['FSIQ'], 'type': 'fsiq'})
>>> data_piq = pandas.DataFrame({'iq': data['PIQ'], 'type': 'piq'})
>>> data_long = pandas.concat((data_fisq, data_piq))
>>> print(data_long)
iq type
0 133 fsiq
1 140 fsiq
2 139 fsiq
...
31 137 piq
32 110 piq
33 86 piq
...
>>> model = ols("iq ~ type", data_long).fit()

...
==========================...
------------------------------------------...
Intercept 113.4500 3.683 30.807 0.000 106.119 120.781
type[T.piq] -2.4250 5.208 -0.466 0.643 -12.793 7.943
...
We can see that we retrieve the same values for t-test and corresponding p-values for the effect of the type
of iq than the previous t-test:
>>> stats.ttest_ind(data['FSIQ'], data['PIQ'])
Ttest_indResult(statistic=0.46563759638..., pvalue=0.64277250...)
15.3.2 Multiple Regression: including multiple factors
Consider a linear model explaining a variable z (the dependent variable) with 2 variables x and y:
z = x c1 + y c2 + i + e

Such a model can be seen in 3D as fitting a plane to a cloud of (x, y, z) points.
Example: the iris data (examples/iris.csv)
Tip: Sepal and petal size tend to be related: bigger flowers are bigger! But is there in addition a systematic
effect of species?
>>> data = pandas.read_csv('examples/iris.csv')

>>> model = ols('sepal_width ~ name + petal_length', data).fit()
==========================...
Dep. Variable: sepal_width R-squared: 0.478
Date: ... Prob (F-statistic): 1.58e-20
Df Model: 3
==========================...
------------------------------------------...

Intercept 2.9813 0.099 29.989 0.000 2.785 3.178

name[T.versicolor] -1.4821 0.181 -8.190 0.000 -1.840 -1.124
name[T.virginica] -1.6635 0.256 -6.502 0.000 -2.169 -1.158
petal_length 0.2983 0.061 4.920 0.000 0.178 0.418
==========================...
Skew: -0.082 Prob(JB): 0.236
==========================...
Warnings:
15.3.3 Post-hoc hypothesis testing: analysis of variance (ANOVA)
In the above iris example, we wish to test if the petal length is different between versicolor and virginica, after
removing the effect of sepal width. This can be formulated as testing the difference between the coefficient as-
sociated to versicolor and virginica in the linear model estimated above (it is an Analysis of Variance, ANOVA).
For this, we write a vector of contrast on the parameters estimated: we want to test "name[T.versicolor]
- name[T.virginica]", with an F-test:
>>> print(model.f_test([0, 1, -1, 0]))

<F test: F=array([[ 3.24533535]]), p=0.07369..., df_denom=146, df_num=1>
Is this difference significant?
Exercise
Going back to the brain size + IQ data, test if the VIQ of male and female are different after removing the
effect of brain size, height and weight.
15.4 More visualization: seaborn for statistical exploration
Seaborn combines simple statistical fits with plotting on pandas dataframes.

Let us consider a data giving wages and many other personal information on 500 individuals (Berndt, ER. The
Practice of Econometrics. 1991. NY: Addison-Wesley).
Tip: The full code loading and plotting of the wages data is found in corresponding example.
>>> print data

EDUCATION SOUTH SEX EXPERIENCE UNION WAGE AGE RACE \
15.4. More visualization: seaborn for statistical exploration 476

0 8 0 1 21 0 0.707570 35 2
1 9 0 1 42 0 0.694605 57 3
2 12 0 0 1 0 0.824126 19 3
3 12 0 0 4 0 0.602060 22 3
...
15.4.1 Pairplot: scatter matrices
We can easily have an intuition on the interactions between continuous variables using seaborn.pairplot()
to display a scatter matrix:
>>> import seaborn

>>> seaborn.pairplot(data, vars=['WAGE', 'AGE', 'EDUCATION'],
... kind='reg')
Categorical variables can be

plotted as the hue:
>>> seaborn.pairplot(data, vars=['WAGE', 'AGE', 'EDUCATION'],

... kind='reg', hue='SEX')

Look and feel and matplotlib settings
Seaborn changes the default of matplotlib figures to achieve a more modern, excel-like look. It does that
upon import. You can reset the default using:
>>> plt.rcdefaults()
Tip: To switch back to seaborn settings, or understand better styling in seaborn, see the relevent section of
the seaborn documentation.

15.4.2 lmplot: plotting a univariate regression
A regression capturing the relation between one variable

and another, eg wage and eduction, can be plotted using seaborn.lmplot():
>>> seaborn.lmplot(y='WAGE', x='EDUCATION', data=data)
Robust regression
Tip: Given that, in the above plot, there seems to be a couple of data points that are outside of the main
cloud to the right, they might be outliers, not representative of the population, but driving the regression.
To compute a regression that is less sentive to outliers, one must use a robust model. This is done in seaborn
using robust=True in the plotting functions, or in statsmodels by replacing the use of the OLS by a Robust
Linear Model, statsmodels.formula.api.rlm().

15.5 Testing for interactions
Do wages increase more with educa-

tion for males than females?
Tip: The plot above is made of two different fits. We need to formulate a single model that tests for a variance
of slope across the to population. This is done via an interaction.
>>> result = sm.ols(formula='wage ~ education + gender + education * gender',

... data=data).fit()
>>> print(result.summary())
...
------------------------------------------------------------------------------
Intercept 0.2998 0.072 4.173 0.000 0.159 0.441
gender[T.male] 0.2750 0.093 2.972 0.003 0.093 0.457
education 0.0415 0.005 7.647 0.000 0.031 0.052
education:gender[T.male] -0.0134 0.007 -1.919 0.056 -0.027 0.000
==========================...
...
Can we conclude that education benefits males more than females?
Take home messages
Hypothesis testing and p-value give you the significance of an effect / difference
Formulas (with categorical variables) enable you to express rich links in your data
Visualizing your data and simple model fits matters!
Conditionning (adding factors that can explain all or part of the variation) is important modeling
aspect that changes the interpretation.
15.5. Testing for interactions 480

15.6 Full code for the figures
Code examples for the statistics chapter.
15.6.1 Boxplots and paired differences
Plot boxplots for FSIQ, PIQ, and the paired difference between the two: while the spread (error bars) for FSIQ
and PIQ are very large, there is a systematic (common) effect due to the subjects. This effect is cancelled out
in the difference and the spread of the difference (paired by subject) is much smaller than the spread of the
individual measures.
import pandas
data = pandas.read_csv('brain_size.csv', sep=';', na_values='.')
# Box plot of FSIQ and PIQ (different measures od IQ)

data.boxplot(column=['FSIQ', 'PIQ'])
# Boxplot of the difference

plt.boxplot(data['FSIQ'] - data['PIQ'])
plt.xticks((1, ), ('FSIQ - PIQ', ))
plt.show()

Download Python source code: plot_paired_boxplots.py
Download Jupyter notebook: plot_paired_boxplots.ipynb
15.6. Full code for the figures 481

15.6.2 Plotting simple quantities of a pandas dataframe
This example loads from a CSV file data with mixed numerical and categorical entries, and plots a few quan-
tities, separately for females and males, thanks to the pandas integrated plotting tool (that uses matplotlib
behind the scene).
See http://pandas.pydata.org/pandas-docs/stable/visualization.html

import pandas
data = pandas.read_csv('brain_size.csv', sep=';', na_values='.')
# Box plots of different columns for each gender

groupby_gender = data.groupby('Gender')
groupby_gender.boxplot(column=['FSIQ', 'VIQ', 'PIQ'])
from pandas.tools import plotting
# Scatter matrices for different columns

plotting.scatter_matrix(data[['Weight', 'Height', 'MRI_Count']])
plotting.scatter_matrix(data[['PIQ', 'VIQ', 'FSIQ']])

plt.show()

Download Python source code: plot_pandas.py
Download Jupyter notebook: plot_pandas.ipynb
15.6.3 Analysis of Iris petal and sepal sizes
Ilustrate an analysis on a real dataset:

Visualizing the data to formulate intuitions
Fitting of a linear model
Hypothesis test of the effect of a categorical variable in the presence of a continuous confound
import pandas
from statsmodels.formula.api import ols
# Load the data

data = pandas.read_csv('iris.csv')

Plot a scatter matrix
# Express the names as categories

categories = pandas.Categorical(data['name'])
# The parameter 'c' is passed to plt.scatter and will control the color
plotting.scatter_matrix(data, c=categories.labels, marker='o')
fig = plt.gcf()
fig.suptitle("blue: setosa, green: versicolor, red: virginica", size=13)
Statistical analysis
# Let us try to explain the sepal length as a function of the petal

# width and the category of iris
model = ols('sepal_width ~ name + petal_length', data).fit()

print(model.summary())
# Now formulate a "contrast", to test if the offset for versicolor and

# virginica are identical
print('Testing the difference between effect of versicolor and virginica')

print(model.f_test([0, 1, -1, 0]))
plt.show()
Out:

==============================================================================
Dep. Variable: sepal_width R-squared: 0.478


Date: Tue, 03 Oct 2017 Prob (F-statistic): 1.58e-20
Time: 16:00:29 Log-Likelihood: -38.185
Df Model: 3
======================================================================================
--------------------------------------------------------------------------------------
Intercept 2.9813 0.099 29.989 0.000 2.785 3.178
name[T.versicolor] -1.4821 0.181 -8.190 0.000 -1.840 -1.124
name[T.virginica] -1.6635 0.256 -6.502 0.000 -2.169 -1.158
petal_length 0.2983 0.061 4.920 0.000 0.178 0.418
==============================================================================
Skew: -0.082 Prob(JB): 0.236
==============================================================================
Warnings:
Testing the difference between effect of versicolor and virginica
<F test: F=array([[ 3.24533535]]), p=0.07369058781700738, df_denom=146, df_num=1>

Download Python source code: plot_iris_analysis.py
Download Jupyter notebook: plot_iris_analysis.ipynb
15.6.4 Simple Regression
Fit a simple linear regression using statsmodels, compute corresponding p-values.
# Original author: Thomas Haslwanter
import numpy as np
import pandas
# For statistics. Requires statsmodels 5.0 or more

# Analysis of Variance (ANOVA) on linear models
from statsmodels.stats.anova import anova_lm
Generate and show the data
# To get reproducable values, provide a seed value

np.random.seed(1)
y = -5 + 3*x + 4 * np.random.normal(size=x.shape)
# Plot the data

plt.plot(x, y, 'o')

Multilinear regression model, calculating fit, P-values, confidence intervals etc.
# Convert the data into a Pandas DataFrame to use the formulas framework
# in statsmodels
data = pandas.DataFrame({'x': x, 'y': y})
# Fit the model

model = ols("y ~ x", data).fit()
# Print the summary

# Peform analysis of variance on fitted linear model

anova_results = anova_lm(model)
print('\nANOVA results')
print(anova_results)
Out:

==============================================================================
Dep. Variable: y R-squared: 0.804
Df Model: 1
==============================================================================
------------------------------------------------------------------------------
Intercept -5.5335 1.036 -5.342 0.000 -7.710 -3.357
x 2.9369 0.341 8.604 0.000 2.220 3.654

==============================================================================
Skew: -0.058 Prob(JB): 0.851
==============================================================================
Warnings:
ANOVA results
df sum_sq mean_sq F PR(>F)
x 1.0 1588.873443 1588.873443 74.029383 8.560649e-08
Residual 18.0 386.329330 21.462741 NaN NaN
Plot the fitted model
# Retrieve the parameter estimates

offset, coef = model._results.params
plt.plot(x, x*coef + offset)
plt.xlabel('x')
plt.ylabel('y')
plt.show()

Download Python source code: plot_regression.py
Download Jupyter notebook: plot_regression.ipynb

15.6.5 Multiple Regression
Calculate using statsmodels just the best fit, or all the corresponding statistical parameters.
Also shows how to make 3d plots.
# Original author: Thomas Haslwanter
import numpy as np
import pandas
# For 3d plots. This import is necessary to have 3D plotting below

# For statistics. Requires statsmodels 5.0 or more

# Analysis of Variance (ANOVA) on linear models
from statsmodels.stats.anova import anova_lm
Generate and show the data
# We generate a 2D grid
X, Y = np.meshgrid(x, x)
# To get reproducable values, provide a seed value

np.random.seed(1)
# Z is the elevation of this 2D grid

Z = -5 + 3*X - 0.5*Y + 8 * np.random.normal(size=X.shape)
# Plot the data

fig = plt.figure()
ax = fig.gca(projection='3d')
surf = ax.plot_surface(X, Y, Z, cmap=plt.cm.coolwarm,
rstride=1, cstride=1)
ax.view_init(20, -120)
ax.set_xlabel('X')
ax.set_ylabel('Y')
ax.set_zlabel('Z')

Multilinear regression model, calculating fit, P-values, confidence intervals etc.
# Convert the data into a Pandas DataFrame to use the formulas framework
# in statsmodels
# First we need to flatten the data: it's 2D layout is not relevent.

X = X.flatten()
Y = Y.flatten()
Z = Z.flatten()
data = pandas.DataFrame({'x': X, 'y': Y, 'z': Z})
# Fit the model

model = ols("z ~ x + y", data).fit()
# Print the summary

print("\nRetrieving manually the parameter estimates:")

print(model._results.params)
# should be array([-4.99754526, 3.00250049, -0.50514907])
# Peform analysis of variance on fitted linear model

anova_results = anova_lm(model)
print('\nANOVA results')
print(anova_results)
plt.show()
Out:


==============================================================================
Dep. Variable: z R-squared: 0.594
No. Observations: 441 AIC: 3081.
Df Residuals: 438 BIC: 3094.
Df Model: 2
==============================================================================
------------------------------------------------------------------------------
Intercept -4.5052 0.378 -11.924 0.000 -5.248 -3.763
x 3.1173 0.125 24.979 0.000 2.872 3.363
y -0.5109 0.125 -4.094 0.000 -0.756 -0.266
==============================================================================
Skew: -0.052 Prob(JB): 0.903
==============================================================================
Warnings:
Retrieving manually the parameter estimates:

[-4.50523303 3.11734237 -0.51091248]
ANOVA results
df sum_sq mean_sq F PR(>F)
x 1.0 39284.301219 39284.301219 623.962799 2.888238e-86
y 1.0 1055.220089 1055.220089 16.760336 5.050899e-05
Residual 438.0 27576.201607 62.959364 NaN NaN

Download Python source code: plot_regression_3d.py
Download Jupyter notebook: plot_regression_3d.ipynb
15.6.6 Test for an education/gender interaction in wages
Wages depend mostly on education. Here we investigate how this dependence is related to gender: not only
does gender create an offset in wages, it also seems that wages increase more with education for males than
females.
Does our data support this last hypothesis? We will test this using statsmodels formulas (http://statsmodels.
sourceforge.net/stable/example_formulas.html).
Load and massage the data
import pandas
import urllib
import os
if not os.path.exists('wages.txt'):
# Download the file if it is not present
urllib.urlretrieve('http://lib.stat.cmu.edu/datasets/CPS_85_Wages',

'wages.txt')
# EDUCATION: Number of years of education

# SEX: 1=Female, 0=Male
# WAGE: Wage (dollars per hour)
data = pandas.read_csv('wages.txt', skiprows=27, skipfooter=6, sep=None,
header=None, names=['education', 'gender', 'wage'],
usecols=[0, 2, 5],
)
# Convert genders to strings (this is particulary useful so that the

# statsmodels formulas detects that gender is a categorical variable)
import numpy as np
data['gender'] = np.choose(data.gender, ['male', 'female'])
# Log-transform the wages, because they typically are increased with

# multiplicative factors
data['wage'] = np.log10(data['wage'])
simple plotting
import seaborn
# Plot 2 linear fits for male and female.

seaborn.lmplot(y='wage', x='education', hue='gender', data=data)
statistical analysis

import statsmodels.formula.api as sm
# Note that this model is not the plot displayed above: it is one
# joined model for male and female, not separate models for male and
# female. The reason is that a single model enables statistical testing
result = sm.ols(formula='wage ~ education + gender', data=data).fit()
print(result.summary())
Out:

==============================================================================
Dep. Variable: wage R-squared: 0.193
Time: 16:00:42 Log-Likelihood: 86.654
No. Observations: 534 AIC: -167.3
Df Residuals: 531 BIC: -154.5
Df Model: 2
==================================================================================
----------------------------------------------------------------------------------
Intercept 0.4053 0.046 8.732 0.000 0.314 0.496
gender[T.male] 0.1008 0.018 5.625 0.000 0.066 0.136
education 0.0334 0.003 9.768 0.000 0.027 0.040
==============================================================================
Skew: -0.147 Prob(JB): 0.0873
==============================================================================
Warnings:
The plots above highlight that there is not only a different offset in wage but also a different slope
We need to model this using an interaction
result = sm.ols(formula='wage ~ education + gender + education * gender',

data=data).fit()
Out:

==============================================================================
Dep. Variable: wage R-squared: 0.198
Time: 16:00:42 Log-Likelihood: 88.503
No. Observations: 534 AIC: -169.0
Df Residuals: 530 BIC: -151.9
Df Model: 3
============================================================================================
--------------------------------------------------------------------------------------------
Intercept 0.2998 0.072 4.173 0.000 0.159 0.441
gender[T.male] 0.2750 0.093 2.972 0.003 0.093 0.457

education 0.0415 0.005 7.647 0.000 0.031 0.052

education:gender[T.male] -0.0134 0.007 -1.919 0.056 -0.027 0.000
==============================================================================
Skew: -0.156 Prob(JB): 0.0821
Kurtosis: 3.356 Cond. No. 194.
==============================================================================
Warnings:
Looking at the p-value of the interaction of gender and education, the data does not support the hypothesis
that education benefits males more than female (p-value > 0.05).

plt.show()

Download Python source code: plot_wage_education_gender.py
Download Jupyter notebook: plot_wage_education_gender.ipynb
15.6.7 Visualizing factors influencing wages
This example uses seaborn to quickly plot various factors relating wages, experience and eduction.
Seaborn (http://stanford.edu/~mwaskom/software/seaborn/) is a library that combines visualization and sta-
tistical fits to show trends in data.
Note that importing seaborn changes the matplotlib style to have an excel-like feeling. This changes affect
other matplotlib figures. To restore defaults once this example is run, we would need to call plt.rcdefaults().
# Standard library imports

import urllib
import os
Load the data
import pandas
if not os.path.exists('wages.txt'):
urllib.urlretrieve('http://lib.stat.cmu.edu/datasets/CPS_85_Wages',
'wages.txt')
# Give names to the columns

names = [
'EDUCATION: Number of years of education',
'SOUTH: 1=Person lives in South, 0=Person lives elsewhere',
'SEX: 1=Female, 0=Male',
'EXPERIENCE: Number of years of work experience',
'UNION: 1=Union member, 0=Not union member',
'WAGE: Wage (dollars per hour)',
'AGE: years',
'RACE: 1=Other, 2=Hispanic, 3=White',
'OCCUPATION: 1=Management, 2=Sales, 3=Clerical, 4=Service, 5=Professional, 6=Other',
'SECTOR: 0=Other, 1=Manufacturing, 2=Construction',

'MARR: 0=Unmarried, 1=Married',

]
short_names = [n.split(':')[0] for n in names]
data = pandas.read_csv('wages.txt', skiprows=27, skipfooter=6, sep=None,

header=None)
data.columns = short_names
# Log-transform the wages, because they typically are increased with

# multiplicative factors
import numpy as np
data['WAGE'] = np.log10(data['WAGE'])
Plot scatter matrices highlighting different aspects
import seaborn
seaborn.pairplot(data, vars=['WAGE', 'AGE', 'EDUCATION'],
kind='reg')

kind='reg', hue='SEX')
plt.suptitle('Effect of gender: 1=Female, 0=Male')

kind='reg', hue='RACE')
plt.suptitle('Effect of race: 1=Other, 2=Hispanic, 3=White')

kind='reg', hue='UNION')
plt.suptitle('Effect of union: 1=Union member, 0=Not union member')



Plot a simple regression
seaborn.lmplot(y='WAGE', x='EDUCATION', data=data)
plt.show()


Download Python source code: plot_wage_data.py
Download Jupyter notebook: plot_wage_data.ipynb
15.6.8 Air fares before and after 9/11
This is a business-intelligence (BI) like application.

What is interesting here is that we may want to study fares as a function of the year, paired accordingly to the
trips, or forgetting the year, only as a function of the trip endpoints.
Using statsmodels linear models, we find that both with an OLS (ordinary least square) and a robust fit, the
intercept and the slope are significantly non-zero: the air fares have decreased between 2000 and 2001, and
their dependence on distance travelled has also decreased
# Standard library imports

import urllib
import os
Load the data
import pandas
if not os.path.exists('airfares.txt'):
urllib.urlretrieve(
'http://www.stat.ufl.edu/~winner/data/airq4.dat',
'airfares.txt')
# As a seperator, ' +' is a regular expression that means 'one of more

# space'
data = pandas.read_csv('airfares.txt', sep=' +', header=0,
names=['city1', 'city2', 'pop1', 'pop2',
'dist', 'fare_2000', 'nb_passengers_2000',
'fare_2001', 'nb_passengers_2001'])
# we log-transform the number of passengers

import numpy as np
data['nb_passengers_2000'] = np.log10(data['nb_passengers_2000'])
data['nb_passengers_2001'] = np.log10(data['nb_passengers_2001'])
Make a dataframe whith the year as an attribute, instead of separate columns
# This involves a small danse in which we separate the dataframes in 2,

# one for year 2000, and one for 2001, before concatenating again.
# Make an index of each flight

data_flat = data.reset_index()
data_2000 = data_flat[['city1', 'city2', 'pop1', 'pop2',

'dist', 'fare_2000', 'nb_passengers_2000']]
# Rename the columns
data_2000.columns = ['city1', 'city2', 'pop1', 'pop2', 'dist', 'fare',
'nb_passengers']
# Add a column with the year
data_2000['year'] = 2000
data_2001 = data_flat[['city1', 'city2', 'pop1', 'pop2',

'dist', 'fare_2001', 'nb_passengers_2001']]

# Rename the columns

data_2001.columns = ['city1', 'city2', 'pop1', 'pop2', 'dist', 'fare',
'nb_passengers']
# Add a column with the year
data_2001['year'] = 2001
data_flat = pandas.concat([data_2000, data_2001])
Plot scatter matrices highlighting different aspects
import seaborn
seaborn.pairplot(data_flat, vars=['fare', 'dist', 'nb_passengers'],
kind='reg', markers='.')
# A second plot, to show the effect of the year (ie the 9/11 effect)
seaborn.pairplot(data_flat, vars=['fare', 'dist', 'nb_passengers'],
kind='reg', hue='year', markers='.')


Plot the difference in fare
seaborn.boxplot(data.fare_2001 - data.fare_2000)
plt.title('Fare: 2001 - 2000')
plt.subplots_adjust()
seaborn.boxplot(data.nb_passengers_2001 - data.nb_passengers_2000)
plt.title('NB passengers: 2001 - 2000')
plt.subplots_adjust()

Statistical testing: dependence of fare on distance and number of passengers
import statsmodels.formula.api as sm
result = sm.ols(formula='fare ~ 1 + dist + nb_passengers', data=data_flat).fit()

# Using a robust fit

result = sm.rlm(formula='fare ~ 1 + dist + nb_passengers', data=data_flat).fit()
Out:


==============================================================================
Dep. Variable: fare R-squared: 0.275
Method: Least Squares F-statistic: 1585.
Date: Tue, 03 Oct 2017 Prob (F-statistic): 0.00
Time: 16:00:28 Log-Likelihood: -45532.
No. Observations: 8352 AIC: 9.107e+04
Df Residuals: 8349 BIC: 9.109e+04
Df Model: 2
=================================================================================
---------------------------------------------------------------------------------
Intercept 211.2428 2.466 85.669 0.000 206.409 216.076
dist 0.0484 0.001 48.149 0.000 0.046 0.050
nb_passengers -32.8925 1.127 -29.191 0.000 -35.101 -30.684
==============================================================================
Skew: 0.710 Prob(JB): 1.42e-161
Kurtosis: 3.338 Cond. No. 5.23e+03
==============================================================================
Warnings:
[2] The condition number is large, 5.23e+03. This might indicate that there are
strong multicollinearity or other numerical problems.
Robust linear Model Regression Results
==============================================================================
Dep. Variable: fare No. Observations: 8352
Model: RLM Df Residuals: 8349
Method: IRLS Df Model: 2
Norm: HuberT
Scale Est.: mad
Cov Type: H1
Date: Tue, 03 Oct 2017
Time: 16:00:28
No. Iterations: 12
=================================================================================
coef std err z P>|z| [95.0% Conf. Int.]
---------------------------------------------------------------------------------
Intercept 215.0848 2.448 87.856 0.000 210.287 219.883
dist 0.0460 0.001 46.166 0.000 0.044 0.048
nb_passengers -35.2686 1.119 -31.526 0.000 -37.461 -33.076
=================================================================================
If the model instance has been used for another fit with different fit
parameters, then the fit options might not be the correct ones anymore .
Statistical testing: regression of fare on distance: 2001/2000 difference
result = sm.ols(formula='fare_2001 - fare_2000 ~ 1 + dist', data=data).fit()

# Plot the corresponding regression

data['fare_difference'] = data['fare_2001'] - data['fare_2000']
seaborn.lmplot(x='dist', y='fare_difference', data=data)
plt.show()

Out:

==============================================================================
Dep. Variable: fare_2001 R-squared: 0.159
Time: 16:00:28 Log-Likelihood: -22640.
No. Observations: 4176 AIC: 4.528e+04
Df Residuals: 4174 BIC: 4.530e+04
Df Model: 1
==============================================================================
------------------------------------------------------------------------------
Intercept 148.0279 1.673 88.480 0.000 144.748 151.308
dist 0.0388 0.001 28.136 0.000 0.036 0.041
==============================================================================
Skew: 0.462 Prob(JB): 3.23e-33
==============================================================================
Warnings:


Download Python source code: plot_airfare.py
Download Jupyter notebook: plot_airfare.ipynb
15.7 Solutions to this chapters exercises
15.7.1 Relating Gender and IQ
Going back to the brain size + IQ data, test if the VIQ of male and female are different after removing the effect
of brain size, height and weight.
Notice that here Gender is a categorical value. As it is a non-float data type, statsmodels is able to automati-
cally infer this.
import pandas
data = pandas.read_csv('../brain_size.csv', sep=';', na_values='.')
model = ols('VIQ ~ Gender + MRI_Count + Height', data).fit()

# Here, we don't need to define a contrast, as we are testing a single

# coefficient of our model, and not a combination of coefficients.
# However, defining a contrast, which would then be a 'unit contrast',
# will give us the same results
print(model.f_test([0, 1, 0, 0]))
Out:

==============================================================================
Dep. Variable: VIQ R-squared: 0.246
Date: Tue, 03 Oct 2017 Prob (F-statistic): 0.0184
Df Model: 3
==================================================================================
----------------------------------------------------------------------------------
Intercept 166.6258 88.824 1.876 0.069 -13.696 346.948
Gender[T.Male] 8.8524 10.710 0.827 0.414 -12.890 30.595
MRI_Count 0.0002 6.46e-05 2.615 0.013 3.78e-05 0.000
Height -3.0837 1.276 -2.417 0.021 -5.674 -0.494
==============================================================================
Skew: 0.005 Prob(JB): 0.324
==============================================================================
Warnings:
15.7. Solutions to this chapters exercises 502


<F test: F=array([[ 0.68319608]]), p=0.4140878441244694, df_denom=35, df_num=1>
Here we plot a scatter matrix to get intuitions on our results. This goes beyond what was asked in the exercise
# This plotting is useful to get an intuitions on the relationships between

# our different variables

# Fill in the missing values for Height for plotting

data['Height'].fillna(method='pad', inplace=True)
# The parameter 'c' is passed to plt.scatter and will control the color
# The same holds for parameters 'marker', 'alpha' and 'cmap', that
# control respectively the type of marker used, their transparency and
# the colormap
plotting.scatter_matrix(data[['VIQ', 'MRI_Count', 'Height']],
c=(data['Gender'] == 'Female'), marker='o',
alpha=1, cmap='winter')
fig = plt.gcf()
fig.suptitle("blue: male, green: female", size=13)
plt.show()

Download Python source code: plot_brain_size.py
Download Jupyter notebook: plot_brain_size.ipynb


CHAPTER 16
Sympy : Symbolic Mathematics in
Python
Author: Fabian Pedregosa
Objectives
1. Evaluate expressions with arbitrary precision.

2. Perform algebraic manipulations on symbolic expressions.
3. Perform basic calculus tasks (limits, differentiation and integration) with symbolic expressions.
4. Solve polynomial and transcendental equations.
5. Solve some differential equations.
What is SymPy? SymPy is a Python library for symbolic mathematics. It aims to be an alternative to systems
such as Mathematica or Maple while keeping the code as simple as possible and easily extensible. SymPy is
written entirely in Python and does not require any external libraries.
Sympy documentation and packages for installation can be found on http://www.sympy.org/
Chapters contents
First Steps with SymPy

Using SymPy as a calculator
Symbols
505
Algebraic manipulations
Expand
Simplify
Calculus
Limits
Differentiation
Series expansion
Integration
Equation solving
Linear Algebra
Matrices
Differential Equations
16.1 First Steps with SymPy
16.1.1 Using SymPy as a calculator
SymPy defines three numerical types: Real, Rational and Integer.

The Rational class represents a rational number as a pair of two Integers: the numerator and the denominator,
so Rational(1, 2) represents 1/2, Rational(5, 2) 5/2 and so on:
>>> import sympy as sym

>>> a = sym.Rational(1, 2)
>>> a
1/2
>>> a*2
1
SymPy uses mpmath in the background, which makes it possible to perform computations using arbitrary-
precision arithmetic. That way, some special constants, like e, pi , oo (Infinity), are treated as symbols and can
be evaluated with arbitrary precision:
>>> sym.pi**2
pi**2
>>> sym.pi.evalf()
3.14159265358979
>>> (sym.pi + sym.exp(1)).evalf()

5.85987448204884
as you see, evalf evaluates the expression to a floating-point number.

There is also a class representing mathematical infinity, called oo:
>>> sym.oo > 99999

True
>>> sym.oo + 1
oo
16.1. First Steps with SymPy 506

Exercises
p
1. Calculate 2 with 100 decimals.
2. Calculate 1/2 + 1/3 in rational arithmetic.
16.1.2 Symbols
In contrast to other Computer Algebra Systems, in SymPy you have to declare symbolic variables explicitly:
>>> x = sym.Symbol('x')
>>> y = sym.Symbol('y')
Then you can manipulate them:
>>> x + y + x - y
2*x
>>> (x + y) ** 2
(x + y)**2
Symbols can now be manipulated using some of python operators: +, -`, ``*, ** (arithmetic), &, |, ~ , >>, <<
(boolean).
Printing
Sympy allows for control of the display of the output. From here we use the following setting for printing:
>>> sym.init_printing(use_unicode=False, wrap_line=True)
16.2 Algebraic manipulations
SymPy is capable of performing powerful algebraic manipulations. Well take a look into some of the most
frequently used: expand and simplify.
16.2.1 Expand
Use this to expand an algebraic expression. It will try to denest powers and multiplications:
>>> sym.expand((x + y) ** 3)
3 2 2 3
x + 3*x *y + 3*x*y + y
>>> 3 * x * y ** 2 + 3 * y * x ** 2 + x ** 3 + y ** 3
3 2 2 3
x + 3*x *y + 3*x*y + y
Further options can be given in form on keywords:
>>> sym.expand(x + y, complex=True)

re(x) + re(y) + I*im(x) + I*im(y)
>>> sym.I * sym.im(x) + sym.I * sym.im(y) + sym.re(x) + sym.re(y)
re(x) + re(y) + I*im(x) + I*im(y)
>>> sym.expand(sym.cos(x + y), trig=True)
16.2. Algebraic manipulations 507

-sin(x)*sin(y) + cos(x)*cos(y)
>>> sym.cos(x) * sym.cos(y) - sym.sin(x) * sym.sin(y)
-sin(x)*sin(y) + cos(x)*cos(y)
16.2.2 Simplify
Use simplify if you would like to transform an expression into a simpler form:
>>> sym.simplify((x + x * y) / x)
y + 1
Simplification is a somewhat vague term, and more precises alternatives to simplify exists: powsimp (simplifi-
cation of exponents), trigsimp (for trigonometric expressions) , logcombine, radsimp, together.
Exercises
1. Calculate the expanded form of (x + y)6 .

2. Simplify the trigonometric expression sin(x)/ cos(x)
16.3 Calculus
16.3.1 Limits
Limits are easy to use in SymPy, they follow the syntax limit(function, variable, point), so to compute
the limit of f (x) as x 0, you would issue limit(f, x, 0):
>>> sym.limit(sym.sin(x) / x, x, 0)
1
you can also calculate the limit at infinity:
>>> sym.limit(x, x, sym.oo)

oo
>>> sym.limit(1 / x, x, sym.oo)

0
>>> sym.limit(x ** x, x, 0)
1
16.3.2 Differentiation
You can differentiate any SymPy expression using diff(func, var). Examples:
>>> sym.diff(sym.sin(x), x)
cos(x)
>>> sym.diff(sym.sin(2 * x), x)
2*cos(2*x)
>>> sym.diff(sym.tan(x), x)
2
tan (x) + 1
16.3. Calculus 508

You can check, that it is correct by:
>>> sym.limit((sym.tan(x + y) - sym.tan(x)) / y, y, 0)

2
tan (x) + 1
Higher derivatives can be calculated using the diff(func, var, n) method:
>>> sym.diff(sym.sin(2 * x), x, 1)

2*cos(2*x)

-4*sin(2*x)

-8*cos(2*x)
16.3.3 Series expansion
SymPy also knows how to compute the Taylor series of an expression at a point. Use series(expr, var):
>>> sym.series(sym.cos(x), x)
2 4
x x / 6\
1 - -- + -- + O\x /
2 24
>>> sym.series(1/sym.cos(x), x)
2 4
x 5*x / 6\
1 + -- + ---- + O\x /
2 24
Exercises
1. Calculate limx0 sin(x)/x

2. Calculate the derivative of l og (x) for x.
16.3.4 Integration
SymPy has support for indefinite and definite integration of transcendental elementary and special functions
via integrate() facility, which uses the powerful extended Risch-Norman algorithm and some heuristics and
pattern matching. You can integrate elementary functions:
>>> sym.integrate(6 * x ** 5, x)
6
x
>>> sym.integrate(sym.sin(x), x)
-cos(x)
>>> sym.integrate(sym.log(x), x)
x*log(x) - x
>>> sym.integrate(2 * x + sym.sinh(x), x)
2
x + cosh(x)
Also special functions are handled easily:
16.3. Calculus 509

>>> sym.integrate(sym.exp(-x ** 2) * sym.erf(x), x)

____ 2
\/ pi *erf (x)
--------------
4
It is possible to compute definite integral:
>>> sym.integrate(x**3, (x, -1, 1))

0
>>> sym.integrate(sym.sin(x), (x, 0, sym.pi / 2))
1
>>> sym.integrate(sym.cos(x), (x, -sym.pi / 2, sym.pi / 2))
2
Also improper integrals are supported as well:
>>> sym.integrate(sym.exp(-x), (x, 0, sym.oo))

1
>>> sym.integrate(sym.exp(-x ** 2), (x, -sym.oo, sym.oo))
____
\/ pi
16.4 Equation solving
SymPy is able to solve algebraic equations, in one and several variables using solveset():
>>> sym.solveset(x ** 4 - 1, x)
{-1, 1, -I, I}
As you can see it takes as first argument an expression that is supposed to be equaled to 0. It also has (limited)
support for transcendental equations:
>>> sym.solveset(sym.exp(x) + 1, x)
{I*(2*n*pi + pi) | n in Integers()}
Systems of linear equations
Sympy is able to solve a large part of polynomial equations, and is also capable of solving multiple equa-
tions with respect to multiple variables giving a tuple as second argument. To do this you use the solve()
command:
>>> solution = sym.solve((x + 5 * y - 2, -3 * x + 6 * y - 15), (x, y))
>>> solution[x], solution[y]
(-3, 1)
Another alternative in the case of polynomial equations is factor. factor returns the polynomial factorized into
irreducible terms, and is capable of computing the factorization over various domains:
>>> f = x ** 4 - 3 * x ** 2 + 1
>>> sym.factor(f)
/ 2 \ / 2 \
\x - x - 1/*\x + x - 1/
>>> sym.factor(f, modulus=5)

2 2
(x - 2) *(x + 2)
16.4. Equation solving 510

SymPy is also able to solve boolean equations, that is, to decide if a certain boolean expression is satisfiable or
not. For this, we use the function satisfiable:
>>> sym.satisfiable(x & y)

{x: True, y: True}
This tells us that (x & y) is True whenever x and y are both True. If an expression cannot be true, i.e. no
values of its arguments can make the expression True, it will return False:
>>> sym.satisfiable(x & ~x)

False
Exercises
1. Solve the system of equations x + y = 2, 2 x + y = 0

2. Are there boolean values x, y that make (~x | y) & (~y | x) true?
16.5 Linear Algebra
16.5.1 Matrices
Matrices are created as instances from the Matrix class:
>>> sym.Matrix([[1, 0], [0, 1]])

[1 0]
[ ]
[0 1]
unlike a NumPy array, you can also put Symbols in it:
>>> x, y = sym.symbols('x, y')

>>> A = sym.Matrix([[1, x], [y, 1]])
>>> A
[1 x]
[ ]
[y 1]
>>> A**2
[x*y + 1 2*x ]
[ ]
[ 2*y x*y + 1]
16.5.2 Differential Equations
SymPy is capable of solving (some) Ordinary Differential. To solve differential equations, use dsolve. First,
create an undefined function by passing cls=Function to the symbols function:
>>> f, g = sym.symbols('f g', cls=sym.Function)
f and g are now undefined functions. We can call f(x), and it will represent an unknown function:
>>> f(x)
f(x)
>>> f(x).diff(x, x) + f(x)
16.5. Linear Algebra 511

2
d
f(x) + ---(f(x))
2
dx
>>> sym.dsolve(f(x).diff(x, x) + f(x), f(x))

f(x) = C1*sin(x) + C2*cos(x)
Keyword arguments can be given to this function in order to help if find the best possible resolution system.
For example, if you know that it is a separable equations, you can use keyword hint='separable' to force
dsolve to resolve it as a separable equation:
>>> sym.dsolve(sym.sin(x) * sym.cos(f(x)) + sym.cos(x) * sym.sin(f(x)) * f(x).diff(x), f(x),

,hint='separable')
/ _________________\ / ________________
| / C1 | | / C1
[f(x) = - asin| / ----------- + 1 | + pi, f(x) = asin| / ----------- + 1
| / 2 | | / 2
\\/ sin (x) - 1 / \\/ sin (x) - 1
_\ / _________________\ / ______________
| | / C1 | | / C1
| + pi, f(x) = -asin| / ----------- + 1 |, f(x) = asin| / ----------- +
| | / 2 | | / 2
/ \\/ sin (x) - 1 / \\/ sin (x) - 1
___\
|
1 |]
|
/
Exercises
1. Solve the Bernoulli differential equation
d f (x)
x + f (x) f (x)2 = 0
x
2. Solve the same equation using hint='Bernoulli'. What do you observe ?
16.5. Linear Algebra 512

CHAPTER 17
Scikit-image: image processing
Author: Emmanuelle Gouillart

scikit-image is a Python package dedicated to image processing, and using natively NumPy arrays as image
objects. This chapter describes how to use scikit-image on various image processing tasks, and insists on
the link with other scientific Python modules such as NumPy and SciPy.
See also:
For basic image manipulation, such as image cropping or simple filtering, a large number of simple operations
can be realized with NumPy and SciPy only. See Image manipulation and processing using Numpy and Scipy.
Note that you should be familiar with the content of the previous chapter before reading the current one, as
basic operations such as masking and labeling are a prerequisite.
Chapters contents
Introduction and concepts

scikit-image and the SciPy ecosystem
Whats to be found in scikit-image
Input/output, data types and colorspaces
Data types
Colorspaces
Image preprocessing / enhancement
Local filters
Non-local filters
513
Mathematical morphology
Image segmentation
Binary segmentation: foreground + background
Marker based methods
Measuring regions properties
Data visualization and interaction
Feature extraction for computer vision
Full code examples
Examples for the scikit-image chapter
17.1 Introduction and concepts
Images are NumPys arrays np.ndarray

image np.ndarray
pixels array values: a[2, 3]
channels array dimensions
image encoding dtype (np.uint8, np.uint16, np.float)
filters functions (numpy, skimage, scipy)

>>> check = np.zeros((9, 9))
>>> check[::2, 1::2] = 1
>>> check[1::2, ::2] = 1
>>> plt.imshow(check, cmap='gray', interpolation='nearest')
17.1. Introduction and concepts 514

17.1.1 scikit-image and the SciPy ecosystem
Recent versions of scikit-image is packaged in most Scientific Python distributions, such as Anaconda or
Enthought Canopy. It is also packaged for Ubuntu/Debian.
>>> import skimage

>>> from skimage import data # most functions are in subpackages
Most scikit-image functions take NumPy ndarrays as arguments
>>> camera = data.camera()

>>> camera.dtype
dtype('uint8')
>>> camera.shape
(512, 512)
>>> from skimage import filters
>>> filtered_camera = filters.gaussian(camera, 1)
>>> type(filtered_camera)
<type 'numpy.ndarray'>
Other Python packages are available for image processing and work with NumPy arrays:
scipy.ndimage : for nd-arrays. Basic filtering, mathematical morphology, regions properties
Mahotas
Also, powerful image processing libraries have Python bindings:
OpenCV (computer vision)
ITK (3D images and registration)
and many others
(but they are less Pythonic and NumPy friendly, to a variable extent).
17.1.2 Whats to be found in scikit-image
Website: http://scikit-image.org/
Gallery of examples: http://scikit-image.org/docs/stable/auto_examples/
Different kinds of functions, from boilerplate utility functions to high-level recent algorithms.
Filters: functions transforming images into other images.
NumPy machinery
Common filtering algorithms
Data reduction functions: computation of image histogram, position of local maxima, of corners, etc.
Other actions: I/O, visualization, etc.
17.2 Input/output, data types and colorspaces

I/O: skimage.io
>>> from skimage import io
Reading from files: skimage.io.imread()
>>> import os
>>> filename = os.path.join(skimage.data_dir, 'camera.png')
>>> camera = io.imread(filename)
17.2. Input/output, data types and colorspaces 515

Works with all data formats supported by the Python

Imaging Library (or any other I/O plugin provided to imread with the plugin keyword argument).
Also works with URL image paths:
>>> logo = io.imread('http://scikit-image.org/_static/img/logo.png')
Saving to files:
>>> io.imsave('local_logo.png', logo)
(imsave also uses an external plugin such as PIL)
17.2.1 Data types
Image ndarrays can be represented either by integers

(signed or unsigned) or floats.
Careful with overflows with integer data types

>>> camera.dtype
dtype('uint8')
>>> camera_multiply = 3 * camera
Different integer sizes are possible: 8-, 16- or 32-bytes, signed or unsigned.
Warning: An important (if questionable) skimage convention: float images are supposed to lie in [-1, 1]
(in order to have comparable contrast for all float images)
17.2. Input/output, data types and colorspaces 516

>>> from skimage import img_as_float

>>> camera_float = img_as_float(camera)
>>> camera.max(), camera_float.max()
(255, 1.0)
Some image processing routines need to work with float arrays, and may hence output an array with a different
type and the data range from the input array
>>> from skimage import filters

>>> camera_sobel = filters.sobel(camera)
>>> camera_sobel.max()
0.591502...
Utility functions are provided in skimage to convert both the dtype and the data range, following skimages
conventions: util.img_as_float, util.img_as_ubyte, etc.
See the user guide for more details.
17.2.2 Colorspaces
Color images are of shape (N, M, 3) or (N, M, 4) (when an alpha channel encodes transparency)
>>> face = scipy.misc.face()

>>> face.shape
(768, 1024, 3)
Routines converting between different colorspaces (RGB, HSV, LAB etc.) are available in skimage.color :
color.rgb2hsv, color.lab2rgb, etc. Check the docstring for the expected dtype (and data range) of input
images.
3D images
Most functions of skimage can take 3D images as input arguments. Check the docstring to know if a func-
tion can be used on 3D images (for example MRI or CT images).
Exercise
Open a color image on your disk as a NumPy array.

Find a skimage function computing the histogram of an image and plot the histogram of each
color channel
Convert the image to grayscale and plot its histogram.
17.3 Image preprocessing / enhancement
Goals: denoising, feature (edges) extraction, . . .
17.3.1 Local filters
Local filters replace the value of pixels by a function of the values of neighboring pixels. The function can be
linear or non-linear.
17.3. Image preprocessing / enhancement 517

Neighbourhood: square (choose size), disk, or more complicated structuring element.
Example : horizontal Sobel filter
>>> text = data.text()

>>> hsobel_text = filters.sobel_h(text)
Uses the following linear kernel for computing horizontal gradients:
1 2 1
0 0 0
-1 -2 -1
17.3.2 Non-local filters
Non-local filters use a large region of the image (or all the image) to transform the value of one pixel:
>>> from skimage import exposure

>>> camera_equalized = exposure.equalize_hist(camera)
Enhances contrast in large almost uniform regions.
See wikipedia for an introduction on mathematical morphology.

Probe an image with a simple shape (a structuring element), and modify this image according to how the
shape locally fits or misses the image.

Default structuring element: 4-connectivity of a pixel
>>> from skimage import morphology

>>> morphology.diamond(1)
array([[0, 1, 0],
[1, 1, 1],
[0, 1, 0]], dtype=uint8)
Erosion = minimum filter. Replace the value of a pixel by the minimal value covered by the structuring ele-
ment.:
>>> a = np.zeros((7,7), dtype=np.uint8)

>>> a[1:6, 2:5] = 1
>>> a
array([[0, 0, 0, 0, 0, 0, 0],
[0, 0, 1, 1, 1, 0, 0],
[0, 0, 1, 1, 1, 0, 0],
[0, 0, 1, 1, 1, 0, 0],
[0, 0, 1, 1, 1, 0, 0],
[0, 0, 1, 1, 1, 0, 0],
[0, 0, 0, 0, 0, 0, 0]], dtype=uint8)
>>> morphology.binary_erosion(a, morphology.diamond(1)).astype(np.uint8)
array([[0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 1, 0, 0, 0],
[0, 0, 0, 1, 0, 0, 0],
[0, 0, 0, 1, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0]], dtype=uint8)
>>> morphology.binary_erosion(a, morphology.diamond(2)).astype(np.uint8)
array([[0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0]], dtype=uint8)
Dilation: maximum filter:
>>> a = np.zeros((5, 5))

>>> a[2, 2] = 1
>>> a
array([[ 0., 0., 0., 0., 0.],
[ 0., 0., 0., 0., 0.],
[ 0., 0., 1., 0., 0.],
[ 0., 0., 0., 0., 0.],
[ 0., 0., 0., 0., 0.]])

>>> morphology.binary_dilation(a, morphology.diamond(1)).astype(np.uint8)

array([[0, 0, 0, 0, 0],
[0, 0, 1, 0, 0],
[0, 1, 1, 1, 0],
[0, 0, 1, 0, 0],
[0, 0, 0, 0, 0]], dtype=uint8)
Opening: erosion + dilation:

>>> a[1:4, 1:4] = 1; a[4, 4] = 1
>>> a
array([[0, 0, 0, 0, 0],
[0, 1, 1, 1, 0],
[0, 1, 1, 1, 0],
[0, 1, 1, 1, 0],
[0, 0, 0, 0, 1]])
>>> morphology.binary_opening(a, morphology.diamond(1)).astype(np.uint8)
array([[0, 0, 0, 0, 0],
[0, 0, 1, 0, 0],
[0, 1, 1, 1, 0],
[0, 0, 1, 0, 0],
[0, 0, 0, 0, 0]], dtype=uint8)
Opening removes small objects and smoothes corners.
Grayscale mathematical morphology
Mathematical morphology operations are also available for (non-binary) grayscale images (int or float type).
Erosion and dilation correspond to minimum (resp. maximum) filters.
Higher-level mathematical morphology are available: tophat, skeletonization, etc.

See also:
Basic mathematical morphology is also implemented in scipy.ndimage.morphology. The scipy.ndimage
implementation works on arbitrary-dimensional arrays.
Example of filters comparison: image denoising
>>> from skimage.morphology import disk

>>> coins = data.coins()
>>> coins_zoom = coins[10:80, 300:370]
>>> median_coins = filters.median(coins_zoom, disk(1))
>>> from skimage import restoration
>>> tv_coins = restoration.denoise_tv_chambolle(coins_zoom, weight=0.1)
>>> gaussian_coins = filters.gaussian(coins, sigma=2)

17.4 Image segmentation
Image segmentation is the attribution of different labels to different regions of the image, for example in order
to extract the pixels of an object of interest.
17.4.1 Binary segmentation: foreground + background
Histogram-based method: Otsu thresholding
Tip: The Otsu method is a simple heuristic to find a threshold to separate the foreground from the background.
Earlier scikit-image versions
skimage.filters is called skimage.filter in earlier versions of scikit-image
from skimage import data

from skimage import filters
camera = data.camera()
val = filters.threshold_otsu(camera)
mask = camera < val
Labeling connected components of a discrete image
Tip: Once you have separated foreground objects, it is use to separate them from each other. For this, we can
assign a different integer labels to each one.
Synthetic data:
>>> n = 20
>>> l = 256
>>> points = l * np.random.random((2, n ** 2))
>>> im = filters.gaussian(im, sigma=l / (4. * n))
>>> blobs = im > im.mean()
Label all connected components:
17.4. Image segmentation 521

>>> from skimage import measure

>>> all_labels = measure.label(blobs)
Label only foreground connected components:
>>> blobs_labels = measure.label(blobs, background=0)
See also:
scipy.ndimage.find_objects() is useful to return slices on object in an image.
17.4.2 Marker based methods
If you have markers inside a set of regions, you can use these to segment the regions.
Watershed segmentation
The Watershed (skimage.morphology.watershed()) is a region-growing approach that fills basins in the

image
>>> from skimage.morphology import watershed

>>> from skimage.feature import peak_local_max
>>>
>>> # Generate an initial image with two overlapping circles
>>> x, y = np.indices((80, 80))
>>> x1, y1, x2, y2 = 28, 28, 44, 52
>>> r1, r2 = 16, 20
>>> mask_circle1 = (x - x1) ** 2 + (y - y1) ** 2 < r1 ** 2
>>> mask_circle2 = (x - x2) ** 2 + (y - y2) ** 2 < r2 ** 2
>>> image = np.logical_or(mask_circle1, mask_circle2)
>>> # Now we want to separate the two objects in image
>>> # Generate the markers as local maxima of the distance
>>> # to the background
>>> from scipy import ndimage
>>> distance = ndimage.distance_transform_edt(image)
>>> local_maxi = peak_local_max(distance, indices=False, footprint=np.ones((3, 3)),
,labels=image)
>>> markers = morphology.label(local_maxi)
>>> labels_ws = watershed(-distance, markers, mask=image)
Random walker segmentation
The random walker algorithm (skimage.segmentation.random_walker()) is similar to the Watershed, but

with a more probabilistic approach. It is based on the idea of the diffusion of labels in the image:
17.4. Image segmentation 522

>>> from skimage import segmentation

>>> # Transform markers image so that 0-valued pixels are to
>>> # be labelled, and -1-valued pixels represent background
>>> markers[~image] = -1
>>> labels_rw = segmentation.random_walker(image, markers)
Postprocessing label images
skimage provides several utility functions that can be used on label images (ie images where
different discrete values identify different regions). Functions names are often self-explaining:
skimage.segmentation.clear_border(), skimage.segmentation.relabel_from_one(), skimage.
morphology.remove_small_objects(), etc.
Exercise
Load the coins image from the data submodule.

Separate the coins from the background by testing several segmentation methods: Otsu thresholding,
adaptive thresholding, and watershed or random walker segmentation.
If necessary, use a postprocessing function to improve the coins / background segmentation.
17.5 Measuring regions properties
>>> from skimage import measure
Example: compute the size and perimeter of the two segmented regions:
>>> properties = measure.regionprops(labels_rw)

>>> [prop.area for prop in properties]
[770, 1168]
>>> [prop.perimeter for prop in properties]
[100.91..., 126.81...]
See also:
for some properties, functions are available as well in scipy.ndimage.measurements with a different API (a
list is returned).
Exercise (continued)
Use the binary image of the coins and background from the previous exercise.
Compute an image of labels for the different coins.
17.5. Measuring regions properties 523

Compute the size and eccentricity of all coins.
17.6 Data visualization and interaction
Meaningful visualizations are useful when testing a given processing pipeline.

Some image processing operations:
>>> coins = data.coins()

>>> mask = coins > filters.threshold_otsu(coins)
>>> clean_border = segmentation.clear_border(mask)
Visualize binary result:
>>> plt.figure()
<matplotlib.figure.Figure object at 0x...>
>>> plt.imshow(clean_border, cmap='gray')
Visualize contour
>>> plt.figure()
<matplotlib.figure.Figure object at 0x...>
>>> plt.imshow(coins, cmap='gray')
>>> plt.contour(clean_border, [0.5])
<matplotlib.contour.QuadContourSet ...>
Use skimage dedicated utility function:
>>> coins_edges = segmentation.mark_boundaries(coins, clean_border.astype(np.int))
The (ex-
perimental) scikit-image viewer
skimage.viewer = matplotlib-based canvas for displaying images + experimental Qt-based GUI-toolkit
>>> from skimage import viewer

>>> new_viewer = viewer.ImageViewer(coins)
>>> new_viewer.show()
Useful for displaying pixel values.

For more interaction, plugins can be added to the viewer:
17.6. Data visualization and interaction 524

>>> new_viewer = viewer.ImageViewer(coins)

>>> from skimage.viewer.plugins import lineprofile
>>> new_viewer += lineprofile.LineProfile()
>>> new_viewer.show()
17.7 Feature extraction for computer vision
Geometric or textural descriptor can be extracted from images in order to
17.7. Feature extraction for computer vision 525

classify parts of the image (e.g. sky vs. buildings)

match parts of different images (e.g. for object detection)
and many other applications of Computer Vision
>>> from skimage import feature
Example: detecting corners using Harris detector
from skimage.feature import corner_harris, corner_subpix, corner_peaks

from skimage.transform import warp, AffineTransform
tform = AffineTransform(scale=(1.3, 1.1), rotation=1, shear=0.7,

translation=(210, 50))
image = warp(data.checkerboard(), tform.inverse, output_shape=(350, 350))
coords = corner_peaks(corner_harris(image), min_distance=5)

coords_subpix = corner_subpix(image, coords, window_size=13)
(this ex-
ample is taken from the plot_corner example in scikit-image)
Points of interest such as corners can then be used to match objects in different images, as described in the
plot_matching example of scikit-image.
17.9 Examples for the scikit-image chapter
17.9.1 Creating an image
How to create an image with basic NumPy commands : np.zeros, slicing. . .

This examples show how to create a simple checkerboard.
import numpy as np
check = np.zeros((9, 9))

check[::2, 1::2] = 1
check[1::2, ::2] = 1
plt.matshow(check, cmap='gray')
plt.show()

Download Python source code: plot_check.py
Download Jupyter notebook: plot_check.ipynb
17.9.2 Displaying a simple image
Load and display an image
17.9. Examples for the scikit-image chapter 527


plt.imshow(camera, cmap='gray', interpolation='nearest')
plt.axis('off')
plt.tight_layout()
plt.show()

Download Python source code: plot_camera.py
Download Jupyter notebook: plot_camera.ipynb
17.9.3 Integers can overflow
An illustration of overflow problem arising when working with integers


camera_multiply = 3 * camera
plt.subplot(121)
plt.axis('off')
plt.subplot(122)
plt.imshow(camera_multiply, cmap='gray', interpolation='nearest')
plt.axis('off')
plt.tight_layout()
plt.show()

Download Python source code: plot_camera_uint.py
Download Jupyter notebook: plot_camera_uint.ipynb
17.9.4 Computing horizontal gradients with the Sobel filter
This example illustrates the use of the horizontal Sobel filter, to compute horizontal gradients.


text = data.text()
hsobel_text = filters.sobel_h(text)
plt.subplot(121)
plt.imshow(text, cmap='gray', interpolation='nearest')
plt.axis('off')
plt.subplot(122)
plt.imshow(hsobel_text, cmap='spectral', interpolation='nearest')
plt.axis('off')
plt.tight_layout()
plt.show()

Download Python source code: plot_sobel.py
Download Jupyter notebook: plot_sobel.ipynb
17.9.5 Equalizing the histogram of an image
Histogram equalizing makes images have a uniform histogram.
from skimage import data, exposure

camera_equalized = exposure.equalize_hist(camera)
plt.subplot(121)
plt.axis('off')
plt.subplot(122)
plt.imshow(camera_equalized, cmap='gray', interpolation='nearest')
plt.axis('off')

plt.tight_layout()
plt.show()

Download Python source code: plot_equalize_hist.py
Download Jupyter notebook: plot_equalize_hist.ipynb
17.9.6 Segmentation contours
Visualize segmentation contours on original grayscale image.
from skimage import data, segmentation

import numpy as np
coins = data.coins()
mask = coins > filters.threshold_otsu(coins)
clean_border = segmentation.clear_border(mask).astype(np.int)
coins_edges = segmentation.mark_boundaries(coins, clean_border)
plt.subplot(121)
plt.imshow(clean_border, cmap='gray')
plt.axis('off')
plt.subplot(122)
plt.imshow(coins_edges)
plt.axis('off')
plt.tight_layout()
plt.show()

Download Python source code: plot_boundaries.py
Download Jupyter notebook: plot_boundaries.ipynb

17.9.7 Otsu thresholding
This example illustrates automatic Otsu thresholding.

from skimage import exposure
val = filters.threshold_otsu(camera)
hist, bins_center = exposure.histogram(camera)
plt.subplot(131)
plt.axis('off')
plt.subplot(132)
plt.imshow(camera < val, cmap='gray', interpolation='nearest')
plt.axis('off')
plt.subplot(133)
plt.plot(bins_center, hist, lw=2)
plt.axvline(val, color='k', ls='--')
plt.tight_layout()
plt.show()

Download Python source code: plot_threshold.py
Download Jupyter notebook: plot_threshold.ipynb
17.9.8 Labelling connected components of an image
This example shows how to label connected components of a binary image, using the dedicated skim-
age.measure.label function.

from skimage import measure

import numpy as np
n = 12
l = 256
np.random.seed(1)
points = l * np.random.random((2, n ** 2))
im = filters.gaussian_filter(im, sigma= l / (4. * n))
blobs = im > 0.7 * im.mean()
all_labels = measure.label(blobs)
blobs_labels = measure.label(blobs, background=0)
plt.subplot(131)
plt.imshow(blobs, cmap='gray')
plt.axis('off')
plt.subplot(132)
plt.imshow(all_labels, cmap='spectral')
plt.axis('off')
plt.subplot(133)
plt.imshow(blobs_labels, cmap='spectral')
plt.axis('off')
plt.tight_layout()
plt.show()

Download Python source code: plot_labels.py
Download Jupyter notebook: plot_labels.ipynb
17.9.9 Affine transform
Warping and affine transforms of images.


from skimage.feature import corner_harris, corner_subpix, corner_peaks
from skimage.transform import warp, AffineTransform
tform = AffineTransform(scale=(1.3, 1.1), rotation=1, shear=0.7,

translation=(210, 50))
image = warp(data.checkerboard(), tform.inverse, output_shape=(350, 350))
coords = corner_peaks(corner_harris(image), min_distance=5)

coords_subpix = corner_subpix(image, coords, window_size=13)
plt.gray()
plt.imshow(image, interpolation='nearest')
plt.plot(coords_subpix[:, 1], coords_subpix[:, 0], '+r', markersize=15, mew=5)
plt.plot(coords[:, 1], coords[:, 0], '.b', markersize=7)
plt.axis('off')
plt.show()

Download Python source code: plot_features.py
Download Jupyter notebook: plot_features.ipynb

17.9.10 Various denoising filters
This example compares several denoising filters available in scikit-image: a Gaussian filter, a median filter, and
total variation denoising.
import numpy as np
from skimage import restoration
coins = data.coins()
gaussian_filter_coins = filters.gaussian(coins, sigma=2)
med_filter_coins = filters.median(coins, np.ones((3, 3)))
tv_filter_coins = restoration.denoise_tv_chambolle(coins, weight=0.1)
plt.subplot(141)
plt.imshow(coins[10:80, 300:370], cmap='gray', interpolation='nearest')
plt.axis('off')
plt.title('Image')
plt.subplot(142)
plt.imshow(gaussian_filter_coins[10:80, 300:370], cmap='gray',
interpolation='nearest')
plt.axis('off')
plt.title('Gaussian filter')
plt.subplot(143)
plt.imshow(med_filter_coins[10:80, 300:370], cmap='gray',
plt.axis('off')
plt.title('Median filter')
plt.subplot(144)
plt.imshow(tv_filter_coins[10:80, 300:370], cmap='gray',
plt.axis('off')
plt.title('TV filter')
plt.show()

Download Python source code: plot_filter_coins.py
Download Jupyter notebook: plot_filter_coins.ipynb
17.9.11 Watershed and random walker for segmentation
This example compares two segmentation methods in order to separate two connected disks: the watershed
algorithm, and the random walker algorithm.

Both segmentation methods require seeds, that are pixels belonging unambigusouly to a reagion. Here, local
maxima of the distance map to the background are used as seeds.
import numpy as np
from skimage.morphology import watershed
from skimage.feature import peak_local_max
from skimage import measure
from skimage.segmentation import random_walker
# Generate an initial image with two overlapping circles

x, y = np.indices((80, 80))
x1, y1, x2, y2 = 28, 28, 44, 52
r1, r2 = 16, 20
image = np.logical_or(mask_circle1, mask_circle2)
# Now we want to separate the two objects in image
# Generate the markers as local maxima of the distance
# to the background
distance = ndimage.distance_transform_edt(image)
local_maxi = peak_local_max(
distance, indices=False, footprint=np.ones((3, 3)), labels=image)
markers = measure.label(local_maxi)
labels_ws = watershed(-distance, markers, mask=image)
markers[~image] = -1
labels_rw = random_walker(image, markers)
plt.subplot(141)
plt.imshow(image, cmap='gray', interpolation='nearest')
plt.axis('off')
plt.title('image')
plt.subplot(142)
plt.imshow(-distance, interpolation='nearest')
plt.axis('off')
plt.title('distance map')
plt.subplot(143)
plt.imshow(labels_ws, cmap='spectral', interpolation='nearest')
plt.axis('off')
plt.title('watershed segmentation')
plt.subplot(144)
plt.imshow(labels_rw, cmap='spectral', interpolation='nearest')
plt.axis('off')
plt.title('random walker segmentation')
plt.tight_layout()
plt.show()


Download Python source code: plot_segmentations.py
Download Jupyter notebook: plot_segmentations.ipynb

CHAPTER 18
Traits: building interactive dialogs
Author: Didrik Pinte

The Traits project allows you to simply add validation, initialization, delegation, notification and a graphical
user interface to Python object attributes.
Tip: In this tutorial we will explore the Traits toolset and learn how to dramatically reduce the amount of
boilerplate code you write, do rapid GUI application development, and understand the ideas which underly
other parts of the Enthought Tool Suite.
Traits and the Enthought Tool Suite are open source projects licensed under a BSD-style license.
Intended Audience
Intermediate to advanced Python programmers
Requirements
Either wxPython, PyQt or PySide

Numpy and Scipy
Enthought Tool Suite
All required software can be obtained by installing the EPD Free
538
Tutorial content
Introduction
Example
What are Traits
Initialisation
Validation
Documentation
Visualization: opening a dialog
Deferral
Notification
Some more advanced traits
18.1 Introduction
Tip: The Enthought Tool Suite enable the construction of sophisticated application frameworks for data anal-
ysis, 2D plotting and 3D visualization. These powerful, reusable components are released under liberal BSD-
style licenses.
The main packages of the Enthought Tool Suite are:

Traits - component based approach to build our applications.
Kiva - 2D primitives supporting path based rendering, affine transforms, alpha blending and more.
18.1. Introduction 539

Enable - object based 2D drawing canvas.

Chaco - plotting toolkit for building complex interactive 2D plots.
Mayavi - 3D visualization of scientific data based on VTK.
Envisage - application plugin framework for building scriptable and extensible applications
In this tutorial, we will focus on Traits.
18.2 Example
Throughout this tutorial, we will use an example based on a water resource management simple case. We will
try to model a dam and reservoir system. The reservoir and the dams do have a set of parameters :
Name
Minimal and maximal capacity of the reservoir [hm3]
Height and length of the dam [m]
Catchment area [km2]
Hydraulic head [m]
Power of the turbines [MW]
Minimal and maximal release [m3/s]
Efficiency of the turbines
The reservoir has a known behaviour. One part is related to the energy production based on the water released.
A simple formula for approximating electric power production at a hydroelectric plant is P = hr g k, where:
P is Power in watts,
is the density of water (~1000 kg/m3),
h is height in meters,
r is flow rate in cubic meters per second,
g is acceleration due to gravity of 9.8 m/s2,
k is a coefficient of efficiency ranging from 0 to 1.
Tip: Annual electric energy production depends on the available water supply. In some installations the water
flow rate can vary by a factor of 10:1 over the course of a year.
The second part of the behaviour is the state of the storage that depends on controlled and uncontrolled pa-
rameters :
st or ag e t +1 = st or ag e t + i n f l ow s r el ease spi l l ag e i r r i g at i on
Warning: The data used in this tutorial are not real and might even not have sense in the reality.
18.3 What are Traits
A trait is a type definition that can be used for normal Python object attributes, giving the attributes some
additional characteristics:
Standardization:
18.2. Example 540

Initialization
Validation
Deferral
Notification
Visualization
Documentation
A class can freely mix trait-based attributes with normal Python attributes, or can opt to allow the use of only
a fixed or open set of trait attributes within the class. Trait attributes defined by a class are automatically
inherited by any subclass derived from the class.
The common way of creating a traits class is by extending from the HasTraits base class and defining class
traits :
from traits.api import HasTraits, Str, Float
class Reservoir(HasTraits):
name = Str
max_storage = Float
Warning: For Traits 3.x users

If using Traits 3.x, you need to adapt the namespace of the traits packages:
traits.api should be enthought.traits.api
traitsui.api should be enthought.traits.ui.api
Using a traits class like that is as simple as any other Python class. Note that the trait value are passed using
keyword arguments:
reservoir = Reservoir(name='Lac de Vouglans', max_storage=605)
18.3.1 Initialisation
All the traits do have a default value that initialise the variables. For example, the basic python types do have
the following trait equivalents:
Trait Python Type Built-in Default Value

Bool Boolean False
Complex Complex number 0+0j
Float Floating point number 0.0
Int Plain integer 0
Long Long integer 0L
Str String
Unicode Unicode u
A number of other predefined trait type do exist : Array, Enum, Range, Event, Dict, List, Color, Set, Expression,
Code, Callable, Type, Tuple, etc.
Custom default values can be defined in the code:
18.3. What are Traits 541

name = Str
max_storage = Float(100)
reservoir = Reservoir(name='Lac de Vouglans')
Complex initialisation
When a complex initialisation is required for a trait, a _XXX_default magic method can be implemented. It
will be lazily called when trying to access the XXX trait. For example:
def _name_default(self):
""" Complex initialisation of the reservoir name. """
return 'Undefined'
18.3.2 Validation
Every trait does validation when the user tries to set its content:
reservoir = Reservoir(name='Lac de Vouglans', max_storage=605)
reservoir.max_storage = '230'
---------------------------------------------------------------------------
TraitError Traceback (most recent call last)
.../scipy-lecture-notes/advanced/traits/<ipython-input-7-979bdff9974a> in <module>()
----> 1 reservoir.max_storage = '230'
.../traits/trait_handlers.pyc in error(self, object, name, value)

166 """
167 raise TraitError( object, name, self.full_info( object, name, value ),
--> 168 value )
169
170 def arg_error ( self, method, arg_num, object, name, value ):
TraitError: The 'max_storage' trait of a Reservoir instance must be a float, but a value of '23
,' <type 'str'> was specified.
18.3.3 Documentation
By essence, all the traits do provide documentation about the model itself. The declarative approach to the
creation of classes makes it self-descriptive:
name = Str
max_storage = Float(100)
The desc metadata of the traits can be used to provide a more descriptive information about the trait :

name = Str
max_storage = Float(100, desc='Maximal storage [hm3]')
Lets now define the complete reservoir class:
from traits.api import HasTraits, Str, Float, Range
name = Str
max_storage = Float(1e6, desc='Maximal storage [hm3]')
max_release = Float(10, desc='Maximal release [m3/s]')
head = Float(10, desc='Hydraulic head [m]')
efficiency = Range(0, 1.)
def energy_production(self, release):

''' Returns the energy production [Wh] for the given release [m3/s]
'''
power = 1000 * 9.81 * self.head * release * self.efficiency
return power * 3600
if __name__ == '__main__':
reservoir = Reservoir(
name = 'Project A',
max_storage = 30,
max_release = 100.0,
head = 60,
efficiency = 0.8
)
release = 80
print 'Releasing {} m3/s produces {} kWh'.format(
release, reservoir.energy_production(release)
)
18.3.4 Visualization: opening a dialog
The Traits library is also aware of user interfaces and can pop up a default view for the Reservoir class:
reservoir1 = Reservoir()
reservoir1.edit_traits()

TraitsUI simplifies the way user interfaces are created. Every trait on a HasTraits class has a default editor that
will manage the way the trait is rendered to the screen (e.g. the Range trait is displayed as a slider, etc.).
In the very same vein as the Traits declarative way of creating classes, TraitsUI provides a declarative interface
to build user interfaces code:
from traits.api import HasTraits, Str, Float, Range

from traitsui.api import View
name = Str
traits_view = View(
'name', 'max_storage', 'max_release', 'head', 'efficiency',
title = 'Reservoir',
resizable = True,
)

'''
return power * 3600
if __name__ == '__main__':
name = 'Project A',
max_storage = 30,
head = 60,
efficiency = 0.8
)
reservoir.configure_traits()
18.3.5 Deferral
Being able to defer the definition of a trait and its value to another object is a powerful feature of Traits.
from traits.api import HasTraits, Instance, DelegatesTo, Float, Range
from reservoir import Reservoir

class ReservoirState(HasTraits):
"""Keeps track of the reservoir state given the initial storage.
"""
reservoir = Instance(Reservoir, ())
min_storage = Float
max_storage = DelegatesTo('reservoir')
min_release = Float
max_release = DelegatesTo('reservoir')
# state attributes
storage = Range(low='min_storage', high='max_storage')
# control attributes
inflows = Float(desc='Inflows [hm3]')
release = Range(low='min_release', high='max_release')
spillage = Float(desc='Spillage [hm3]')
def print_state(self):
print 'Storage\tRelease\tInflows\tSpillage'
str_format = '\t'.join(['{:7.2f} 'for i in range(4)])
print str_format.format(self.storage, self.release, self.inflows,
self.spillage)
print '-' * 79
if __name__ == '__main__':
projectA = Reservoir(
name = 'Project A',
max_storage = 30,
hydraulic_head = 60,
efficiency = 0.8
)
state = ReservoirState(reservoir=projectA, storage=10)

state.release = 90
state.inflows = 0
state.print_state()
print 'How do we update the current storage ?'
A special trait allows to manage events and trigger function calls using the magic _xxxx_fired method:
from traits.api import HasTraits, Instance, DelegatesTo, Float, Range, Event
For the simplicity of the example, the release is considered in

hm3/timestep and not in m3/s.
"""
min_storage = Float
min_release = Float
# state attributes


update_storage = Event(desc='Updates the storage to the next time step')
def _update_storage_fired(self):
# update storage state
new_storage = self.storage - self.release + self.inflows
self.storage = min(new_storage, self.max_storage)
overflow = new_storage - self.max_storage
self.spillage = max(overflow, 0)
self.spillage)
print '-' * 79
if __name__ == '__main__':
name = 'Project A',
max_storage = 30,
max_release = 5.0,
efficiency = 0.8
)

state.release = 5
state.inflows = 0
# release the maximum amount of water during 3 time steps

state.update_storage = True
state.print_state()
state.print_state()
state.print_state()
Dependency between objects can be made automatic using the trait Property. The depends_on attribute
expresses the dependency between the property and other traits. When the other traits gets changed, the
property is invalidated. Again, Traits uses magic method names for the property :
_get_XXX for the getter of the XXX Property trait
_set_XXX for the setter of the XXX Property trait

from traits.api import Property

"""

min_release = Float
# state attributes
storage = Property(depends_on='inflows, release')
spillage = Property(
desc='Spillage [hm3]', depends_on=['storage', 'inflows', 'release']
)
### Private traits.

_storage = Float
### Traits property implementation.

def _get_storage(self):
new_storage = self._storage - self.release + self.inflows
return min(new_storage, self.max_storage)
def _set_storage(self, storage_value):

self._storage = storage_value
def _get_spillage(self):
return max(overflow, 0)
self.spillage)
print '-' * 79
if __name__ == '__main__':
name = 'Project A',
max_storage = 30,
max_release = 5,
efficiency = 0.8
)

state.release = 4
state.inflows = 0
state.print_state()
Note: Caching property

Heavy computation or long running computation might be a problem when accessing a property where the
inputs have not changed. The @cached_property decorator can be used to cache the value and only recompute
them once invalidated.
Lets extend the TraitsUI introduction with the ReservoirState example:
from traits.api import HasTraits, Instance, DelegatesTo, Float, Range, Property

from traitsui.api import View, Item, Group, VGroup


"""
name = DelegatesTo('reservoir')
min_release = Float
# state attributes
)
### Traits view

traits_view = View(
Group(
VGroup(Item('name'), Item('storage'), Item('spillage'),
label = 'State', style = 'readonly'
),
VGroup(Item('inflows'), Item('release'), label='Control'),
)
)
### Private traits.

_storage = Float


self.spillage)
print '-' * 79
if __name__ == '__main__':
name = 'Project A',
max_storage = 30,
max_release = 5,

efficiency = 0.8
)

state.release = 4
state.inflows = 0
state.print_state()
state.configure_traits()
Some use cases need the delegation mechanism to be broken by the user when setting the value of the trait.
The PrototypeFrom trait implements this behaviour.
from traits.api import HasTraits, Str, Float, Range, PrototypedFrom, Instance
class Turbine(HasTraits):
turbine_type = Str
power = Float(1.0, desc='Maximal power delivered by the turbine [Mw]')
name = Str
turbine = Instance(Turbine)
installed_capacity = PrototypedFrom('turbine', 'power')
if __name__ == '__main__':
turbine = Turbine(turbine_type='type1', power=5.0)
name = 'Project A',
max_storage = 30,
head = 60,
efficiency = 0.8,
turbine = turbine,
)

print 'installed capacity is initialised with turbine.power'

print reservoir.installed_capacity
print '-' * 15
print 'updating the turbine power updates the installed capacity'
turbine.power = 10
print reservoir.installed_capacity
print '-' * 15
print 'setting the installed capacity breaks the link between turbine.power'
print 'and the installed_capacity trait'
reservoir.installed_capacity = 8
print turbine.power, reservoir.installed_capacity
18.3.6 Notification
Traits implements a Listener pattern. For each trait a list of static and dynamic listeners can be fed with call-
backs. When the trait does change, all the listeners are called.
Static listeners are defined using the _XXX_changed magic methods:
"""
min_storage = Float
min_release = Float
# state attributes
self.spillage)
print '-' * 79
### Traits listeners ###########

def _release_changed(self, new):
"""When the release is higher than zero, warn all the inhabitants of
the valley.
"""
if new > 0:
print 'Warning, we are releasing {} hm3 of water'.format(new)

if __name__ == '__main__':
name = 'Project A',
max_storage = 30,
efficiency = 0.8
)

state.release = 90
state.inflows = 0
state.print_state()
The static trait notification signatures can be:

def _release_changed(self ): pass
def _release_changed(self, new): pass
def _release_changed(self, old, new): pass
def _release_changed(self, name, old, new pass
Listening to all the changes
To listen to all the changes on a HasTraits class, the magic _any_trait_changed method can be implemented.
In many situations, you do not know in advance what type of listeners need to be activated. Traits offers the
ability to register listeners on the fly with the dynamic listeners

from reservoir_state_property import ReservoirState
def wake_up_watchman_if_spillage(new_value):
if new_value > 0:
print 'Wake up watchman! Spilling {} hm3'.format(new_value)
if __name__ == '__main__':
name = 'Project A',
max_storage = 30,
efficiency = 0.8
)
#register the dynamic listener

state.on_trait_change(wake_up_watchman_if_spillage, name='spillage')
state.release = 90
state.inflows = 0
state.print_state()
print 'Forcing spillage'

state.inflows = 100
state.release = 0
print 'Why do we have two executions of the callback ?'

The dynamic trait notification signatures are not the same as the static ones :
def wake_up_watchman(): pass
def wake_up_watchman(new): pass
def wake_up_watchman(name, new): pass
def wake_up_watchman(object, name, new): pass
def wake_up_watchman(object, name, old, new): pass
Removing a dynamic listener can be done by:
calling the remove_trait_listener method on the trait with the listener method as argument,
calling the on_trait_change method with listener method and the keyword remove=True,
deleting the instance that holds the listener.
Listeners can also be added to classes using the on_trait_change decorator:

from traits.api import Property, on_trait_change

"""
min_release = Float
# state attributes
)
### Private traits. ##########

_storage = Float


@on_trait_change('storage')

self.spillage)
print '-' * 79
if __name__ == '__main__':
name = 'Project A',
max_storage = 30,
max_release = 5,
efficiency = 0.8
)

state.release = 4
state.inflows = 0
The patterns supported by the on_trait_change method and decorator are powerful. The reader should look at
the docstring of HasTraits.on_trait_change for the details.
18.3.7 Some more advanced traits
The following example demonstrate the usage of the Enum and List traits :
from traits.api import HasTraits, Str, Float, Range, Enum, List

from traitsui.api import View, Item
class IrrigationArea(HasTraits):
name = Str
surface = Float(desc='Surface [ha]')
crop = Enum('Alfalfa', 'Wheat', 'Cotton')
name = Str
irrigated_areas = List(IrrigationArea)

'''
return power * 3600
traits_view = View(
Item('name'),
Item('max_storage'),
Item('max_release'),
Item('head'),
Item('efficiency'),
Item('irrigated_areas'),
resizable = True
)
if __name__ == '__main__':
upper_block = IrrigationArea(name='Section C', surface=2000, crop='Wheat')

name='Project A',
max_storage=30,
max_release=100.0,
head=60,
efficiency=0.8,
irrigated_areas=[upper_block]
)
release = 80
)
Trait listeners can be used to listen to changes in the content of the list to e.g. keep track of the total crop
surface on linked to a given reservoir.
from traits.api import HasTraits, Str, Float, Range, Enum, List, Property
from traitsui.api import View, Item
class IrrigationArea(HasTraits):
name = Str
surface = Float(desc='Surface [ha]')
crop = Enum('Alfalfa', 'Wheat', 'Cotton')
name = Str
irrigated_areas = List(IrrigationArea)
total_crop_surface = Property(depends_on='irrigated_areas.surface')
def _get_total_crop_surface(self):
return sum([iarea.surface for iarea in self.irrigated_areas])

'''
return power * 3600
traits_view = View(
Item('name'),
Item('max_storage'),
Item('max_release'),
Item('head'),
Item('efficiency'),
Item('irrigated_areas'),
Item('total_crop_surface'),
resizable = True
)
if __name__ == '__main__':
upper_block = IrrigationArea(name='Section C', surface=2000, crop='Wheat')

name='Project A',
max_storage=30,
max_release=100.0,
head=60,
efficiency=0.8,
irrigated_areas=[upper_block],
)
release = 80
)
The next example shows how the Array trait can be used to feed a specialised TraitsUI Item, the ChacoPlotItem:
import numpy as np
from traits.api import HasTraits, Array, Instance, Float, Property

from traits.api import DelegatesTo
from traitsui.api import View, Item, Group
from chaco.chaco_plot_editor import ChacoPlotItem
class ReservoirEvolution(HasTraits):
reservoir = Instance(Reservoir)
name = DelegatesTo('reservoir')
inflows = Array(dtype=np.float64, shape=(None))

releass = Array(dtype=np.float64, shape=(None))
initial_stock = Float
stock = Property(depends_on='inflows, releases, initial_stock')
month = Property(depends_on='stock')
### Traits view ###########

traits_view = View(
Item('name'),
Group(
ChacoPlotItem('month', 'stock', show_label=False),
),
width = 500,
resizable = True
)
### Traits properties #####

def _get_stock(self):
"""
fixme: should handle cases where we go over the max storage
"""
return self.initial_stock + (self.inflows - self.releases).cumsum()
def _get_month(self):
return np.arange(self.stock.size)
if __name__ == '__main__':
name = 'Project A',
max_storage = 30,

head = 60,
efficiency = 0.8
)
initial_stock = 10.
inflows_ts = np.array([6., 6, 4, 4, 1, 2, 0, 0, 3, 1, 5, 3])
releases_ts = np.array([4., 5, 3, 5, 3, 5, 5, 3, 2, 1, 3, 3])
view = ReservoirEvolution(
reservoir = reservoir,
inflows = inflows_ts,
releases = releases_ts
)
view.configure_traits()
See also:
References
ETS repositories
Traits manual
Traits UI manual
Mailing list : enthought-dev@enthought.com

CHAPTER 19
3D plotting with Mayavi
Tip: Mayavi is an interactive 3D plotting package. matplotlib can also do simple 3D plotting, but Mayavi relies
on a more powerful engine ( VTK ) and is more suited to displaying large or complex data.
Chapters contents
Mlab: the scripting interface

3D plotting functions
* Points
* Lines
* Elevation surface
* Arbitrary regular mesh
* Volumetric data
Figures and decorations
* Figure management
* Changing plot properties
* Decorations
Interactive work
The pipeline dialog
557
The script recording button

Slicing and dicing data: sources, modules and filters
An example: inspecting magnetic fields
Different views on data: sources and modules
* Different sources: scatters and fields

* Transforming data: filters
* mlab.pipeline: the scripting layer
Animating the data
Making interactive dialogs
A simple dialog
Making it interactive
Putting it together
19.1 Mlab: the scripting interface
The mayavi.mlab module provides simple plotting functions to apply to numpy arrays, similar to matplotlib
or matlabs plotting interface. Try using them in IPython, by starting IPython with the switch --gui=wx.
19.1.1 3D plotting functions
Points
Hint: Points in 3D, represented with markers (or glyphs) and optionaly different sizes.
x, y, z, value = np.random.random((4, 40))

mlab.points3d(x, y, z, value)
19.1. Mlab: the scripting interface 558

Lines
Hint: A line connecting points in 3D, with optional thickness and varying color.
mlab.clf() # Clear the figure

mlab.plot3d(np.sin(t), np.cos(t), 0.1*t, t)
Elevation surface
Hint: A surface given by its elevation, coded as a 2D array
mlab.clf()
x, y = np.mgrid[-10:10:100j, -10:10:100j]
r = np.sqrt(x**2 + y**2)
z = np.sin(r)/r
mlab.surf(z, warp_scale='auto')

Arbitrary regular mesh
Hint: A surface mesh given by x, y, z positions of its node points
mlab.clf()
phi, theta = np.mgrid[0:np.pi:11j, 0:2*np.pi:11j]
x = np.sin(phi) * np.cos(theta)
y = np.sin(phi) * np.sin(theta)
z = np.cos(phi)
mlab.mesh(x, y, z)
mlab.mesh(x, y, z, representation='wireframe', color=(0, 0, 0))
Note: A surface is defined by points connected to form triangles or polygones. In mayavi.mlab.surf() and
mayavi.mlab.mesh(), the connectivity is implicity given by the layout of the arrays. See also mayavi.mlab.
triangular_mesh().
Our data is often more than points and values: it needs some connectivity information
Volumetric data
Hint: If your data is dense in 3D, it is more difficult to display. One option is to take iso-contours of the data.
mlab.clf()
x, y, z = np.mgrid[-5:5:64j, -5:5:64j, -5:5:64j]
values = x*x*0.5 + y*y + z*z*2.0
mlab.contour3d(values)

This function works with a regular orthogonal grid: the value array is a 3D array that gives the shape of the
grid.
19.1.2 Figures and decorations
Figure management
Tip: Here is a list of functions useful to control the current figure
Get the current figure: mlab.gcf()

Clear the current figure: mlab.clf()
Set the current figure: mlab.figure(1, bgcolor=(1, 1, 1), fgcolor=(0.5, 0.5, 0.5)
Save figure to image file: mlab.savefig(foo.png, size=(300, 300))
Change the view: mlab.view(azimuth=45, elevation=54, distance=1.)
Changing plot properties
Tip: In general, many properties of the various objects on the figure can be changed. If these visualization are
created via mlab functions, the easiest way to change them is to use the keyword arguments of these functions,
as described in the docstrings.
Example docstring: mlab.mesh
Plots a surface using grid-spaced data supplied as 2D arrays.

Function signatures:
mesh(x, y, z, ...)
x, y, z are 2D arrays, all of the same shape, giving the positions of the vertices of the surface. The connectivity
between these points is implied by the connectivity on the arrays.
For simple structures (such as orthogonal grids) prefer the surf function, as it will create more efficient data
structures.
Keyword arguments:
color the color of the vtk object. Overides the colormap, if any, when specified. This
is specified as a triplet of float ranging from 0 to 1, eg (1, 1, 1) for white.

colormap type of colormap to use.

extent [xmin, xmax, ymin, ymax, zmin, zmax] Default is the x, y, z arrays extents.
Use this to change the extent of the object created.
figure Figure to populate.
line_width The with of the lines, if any used. Must be a float. Default: 2.0
mask boolean mask array to suppress some data points.
mask_points If supplied, only one out of mask_points data point is displayed. This
option is usefull to reduce the number of points displayed on large datasets Must
be an integer or None.
mode the mode of the glyphs. Must be 2darrow or 2dcircle or 2dcross or
2ddash or 2ddiamond or 2dhooked_arrow or 2dsquare or 2dthick_arrow
or 2dthick_cross or 2dtriangle or 2dvertex or arrow or cone or cube or
cylinder or point or sphere. Default: sphere
name the name of the vtk object created.
representation the representation type used for the surface. Must be surface or
wireframe or points or mesh or fancymesh. Default: surface
resolution The resolution of the glyph created. For spheres, for instance, this is the
number of divisions along theta and phi. Must be an integer. Default: 8
scalars optional scalar data.
scale_factor scale factor of the glyphs used to represent the vertices, in fancy_mesh
mode. Must be a float. Default: 0.05
scale_mode the scaling mode for the glyphs (vector, scalar, or none).
transparent make the opacity of the actor depend on the scalar.
tube_radius radius of the tubes used to represent the lines, in mesh mode. If None,
simple lines are used.
tube_sides number of sides of the tubes used to represent the lines. Must be an
integer. Default: 6
vmax vmax is used to scale the colormap If None, the max of the data will be used
vmin vmin is used to scale the colormap If None, the min of the data will be used

Example:
In [2]: r, theta = np.mgrid[0:10, -np.pi:np.pi:10j]
In [3]: x = r * np.cos(theta)
In [4]: y = r * np.sin(theta)
In [5]: z = np.sin(r)/r
In [6]: from mayavi import mlab
In [7]: mlab.mesh(x, y, z, colormap='gist_earth', extent=[0, 1, 0, 1, 0, 1])

Out[7]: <mayavi.modules.surface.Surface object at 0xde6f08c>
In [8]: mlab.mesh(x, y, z, extent=[0, 1, 0, 1, 0, 1],

...: representation='wireframe', line_width=1, color=(0.5, 0.5, 0.5))
Out[8]: <mayavi.modules.surface.Surface object at 0xdd6a71c>
Decorations
Tip: Different items can be added to the figure to carry extra information, such as a colorbar or a title.
In [9]: mlab.colorbar(Out[7], orientation='vertical')

Out[9]: <tvtk_classes.scalar_bar_actor.ScalarBarActor object at 0xd897f8c>
In [10]: mlab.title('polar mesh')

Out[10]: <enthought.mayavi.modules.text.Text object at 0xd8ed38c>
In [11]: mlab.outline(Out[7])
Out[11]: <enthought.mayavi.modules.outline.Outline object at 0xdd21b6c>
In [12]: mlab.axes(Out[7])
Out[12]: <enthought.mayavi.modules.axes.Axes object at 0xd2e4bcc>

Warning: extent: If we specified extents for a plotting object, mlab.outline and mlab.axes dont get them
by default.
19.2 Interactive work
Tip: The quickest way to create beautiful visualization with Mayavi is probably to interactively tweak the
various settings.
19.2.1 The pipeline dialog
Click on the Mayavi button in the scene, and you can control properties of objects with dialogs.
19.2. Interactive work 564

Set the background of the figure in the Mayavi Scene node

Set the colormap in the Colors and legends node
Right click on the node to add modules or filters
19.2.2 The script recording button
To find out what code can be used to program these changes, click on the red button as you modify those
properties, and it will generate the corresponding lines of code.
19.3 Slicing and dicing data: sources, modules and filters
19.3.1 An example: inspecting magnetic fields
Suppose we are simulating the magnetic field generated by Helmholtz coils. The examples/compute_field.
py script does this computation and gives you a B array, that is (3 x n), where the first axis is the direction of
the field (Bx, By, Bz), and the second axis the index number of the point. Arrays X, Y and Z give the positions
of these data points.
Excercise
Visualize this field. Your goal is to make sure that the simulation code is correct.
19.3. Slicing and dicing data: sources, modules and filters 565
Suggestions
If you compute the norm of the vector field, you can apply an isosurface to it.
using mayavi.mlab.quiver3d() you can plot vectors. You can also use the masking options (in the
GUI) to make the plot a bit less dense.
19.3.2 Different views on data: sources and modules
Tip: As we see above, it may be desirable to look at the same data in different ways.
Mayavi visualization are created by loading the data in a data source and then displayed on the screen using
modules.
This can be seen by looking at the pipeline view. By right-clicking on the nodes of the pipeline, you can add
new modules.
Quiz
Why is it not possible to add a VectorCutPlane to the vectors created by mayavi.mlab.quiver3d()?
Different sources: scatters and fields
Tip: Data comes in different descriptions.
A 3D block of regularly-spaced value is structured: it is easy to know how one measurement is related to
another neighboring and how to continuously interpolate between these. We can call such data a field,
borrowing from terminology used in physics, as it is continuously defined in space.
A set of data points measured at random positions in a random order gives rise to much more difficult
and ill-posed interpolation problems: the data structure itself does not tell us what are the neighbors of
a data point. We call such data a scatter.
Unstructured and unconnected data: a scatter Structured and connected data: a field
mlab.points3d, mlab.quiver3d mlab.contour3d
Data sources corresponding to scatters can be created with mayavi.mlab.pipeline.scalar_scatter()

or mayavi.mlab.pipeline.vector_scatter(); field data sources can be created with mlab.pipeline.
scalar_field() or mlab.pipeline.vector_field().
Exercice:
1. Create a contour (for instance of the magnetic field norm) by using one of those functions and adding
the right module by clicking on the GUI dialog.
2. Create the right source to apply a vector_cut_plane and reproduce the picture of the magnetic field
shown previously.
Note that one of the difficulties is providing the data in the right form (number of arrays, shape) to the
functions. This is often the case with real-life data.
See also:
Sources are described in details in the Mayavi manual.
Transforming data: filters
If you create a vector field, you may want to visualize the iso-contours of its magnitude. But the isosurface
module can only be applied to scalar data, and not vector data. We can use a filter, ExtractVectorNorm to
add this scalar value to the vector field.
Filters apply a transformation to data, and can be added between sources and modules
Excercice
Using the GUI, add the ExtractVectorNorm filter to display iso-contours of the field magnitude.
mlab.pipeline: the scripting layer
The mlab scripting layer builds pipelines for you. You can reproduce these pipelines programmatically with
the mlab.pipeline interface: each step has a corresponding mlab.pipeline function (simply convert the
name of the step to lower-case underscore-separated: ExtractVectorNorm gives extract_vector_norm). This
function takes as an argument the node that it applies to, as well as optional parameters, and returns the new
node.
For example, iso-contours of the magnitude are coded as:
mlab.pipeline.iso_surface(mlab.pipeline.extract_vector_norm(field),
contours=[0.1*Bmax, 0.4*Bmax],
opacity=0.5)
Excercice
Using the mlab.pipeline interface, generate a complete visualization, with iso-contours of the field magni-
tude, and a vector cut plane.
(click on the figure for a solution)
19.4 Animating the data
Tip: To make movies, or interactive application, you may want to change the data represented on a given
visualization.
If you have built a visualization, using the mlab plotting functions, or the mlab.pipeline function, we can
update the data by assigning new values to the mlab_source attributes
x , y , z = np.ogrid[-5:5:100j ,-5:5:100j, -5:5:100j]

scalars = np.sin(x * y * z) / (x * y * z)
19.4. Animating the data 568

iso = mlab.contour3d(scalars, transparent=True, contours=[0.5])

scalars = np.sin(i * x * y * z) /(x * y * z)
iso.mlab_source.scalars = scalars
See also:
More details in the Mayavi documentation
Event loops
For the interaction with the user (for instance changing the view with the mouse), Mayavi needs some time
to process these events. The for loop above prevents this. The Mayavi documentation details a workaround
19.5 Making interactive dialogs
It is very simple to make interactive dialogs with Mayavi using the Traits library (see the dedicated chapter
Traits: building interactive dialogs).
19.5.1 A simple dialog
from traits.api import HasTraits, Instance

from traitsui.api import View, Item, HGroup
from mayavi.core.ui.api import SceneEditor, MlabSceneModel
def curve(n_turns):
"The function creating the x, y, z coordinates needed to plot"
phi = np.linspace(0, 2*np.pi, 2000)
return [np.cos(phi) * (1 + 0.5*np.cos(n_turns*phi)),
np.sin(phi) * (1 + 0.5*np.cos(n_turns*phi)),
0.5*np.sin(n_turns*phi)]
class Visualization(HasTraits):
"The class that contains the dialog"
scene = Instance(MlabSceneModel, ())
def __init__(self):
HasTraits.__init__(self)
x, y, z = curve(n_turns=2)
# Populating our plot
self.plot = self.scene.mlab.plot3d(x, y, z)
# Describe the dialog

view = View(Item('scene', height=300, show_label=False,
editor=SceneEditor()),
HGroup('n_turns'), resizable=True)
# Fire up the dialog

Visualization().configure_traits()
Tip: Let us read a bit the code above (examples/mlab_dialog.py).

First, the curve function is used to compute the coordinate of the curve we want to plot.
19.5. Making interactive dialogs 569

Second, the dialog is defined by an object inheriting from HasTraits, as it is done with Traits. The important
point here is that a Mayavi scene is added as a specific Traits attribute (Instance). This is important for
embedding it in the dialog. The view of this dialog is defined by the view attribute of the object. In the init of
this object, we populate the 3D scene with a curve.
Finally, the configure_traits method creates the dialog and starts the event loop.
See also:
There are a few things to be aware of when doing dialogs with Mayavi. Please read the Mayavi documentation
19.5.2 Making it interactive
We can combine the Traits events handler with the mlab_source to modify the visualization with the dialog.
We will enable the user to vary the n_turns parameter in the definition of the curve. For this, we need:
to define an n_turns attribute on our visualization object, so that it can appear in the dialog. We use a
Range type.
to wire modification of this attribute to a recomputation of the curve. For this, we use the
on_traits_change decorator.
from traits.api import Range, on_trait_change
class Visualization(HasTraits):
n_turns = Range(0, 30, 11)
scene = Instance(MlabSceneModel, ())
def __init__(self):
HasTraits.__init__(self)
x, y, z = curve(self.n_turns)
self.plot = self.scene.mlab.plot3d(x, y, z)
@on_trait_change('n_turns')
def update_plot(self):
x, y, z = curve(self.n_turns)
self.plot.mlab_source.set(x=x, y=y, z=z)
19.5. Making interactive dialogs 570

view = View(Item('scene', height=300, show_label=False,

editor=SceneEditor()),
HGroup('n_turns'), resizable=True)
# Fire up the dialog

Visualization().configure_traits()
Tip: Full code of the example: examples/mlab_dialog.py.
19.6 Putting it together
Exercise
Using the code from the magnetic field simulation, create a dialog that enable to move the 2 coils: change
their parameters.
Hint: to define a dialog entry for a vector of dimension 3
direction = Array(float, value=(0, 0, 1), cols=3, shape=(3,))
You can look at the example_coil_application.py to see a full-blown application for coil design in 270 lines of
code.
19.6. Putting it together 571

CHAPTER 20
scikit-learn: machine learning in Python
Authors: Gael Varoquaux
Prerequisites
numpy
scipy
matplotlib (optional)
ipython (the enhancements come handy)
Acknowledgements
This chapter is adapted from a tutorial given by Gal Varoquaux, Jake Vanderplas, Olivier Grisel.
See also:
Data science in Python
The Statistics in Python chapter may also be of interest for readers looking into machine learning.
The documentation of scikit-learn is very complete and didactic.
572
Chapters contents
Introduction: problem settings

Basic principles of machine learning with scikit-learn
Supervised Learning: Classification of Handwritten Digits
Supervised Learning: Regression of Housing Data
Measuring prediction performance
Unsupervised Learning: Dimensionality Reduction and Visualization
The eigenfaces example: chaining PCA and SVMs
Parameter selection, Validation, and Testing
Examples for the scikit-learn chapter
20.1 Introduction: problem settings
20.1.1 What is machine learning?
Tip: Machine Learning is about building programs with tunable parameters that are adjusted automatically
so as to improve their behavior by adapting to previously seen data.
Machine Learning can be considered a subfield of Artificial Intelligence since those algorithms can be seen
as building blocks to make computers learn to behave more intelligently by somehow generalizing rather that
just storing and retrieving data items like a database system would do.
Fig. 20.1: A classification problem
Well take a look at two very simple machine learning tasks here. The first is a classification task: the figure
shows a collection of two-dimensional data, colored according to two different class labels. A classification
algorithm may be used to draw a dividing boundary between the two clusters of points:
By drawing this separating line, we have learned a model which can generalize to new data: if you were to drop
another point onto the plane which is unlabeled, this algorithm could now predict whether its a blue or a red
20.1. Introduction: problem settings 573

point.
Fig. 20.2: A regression problem
The next simple task well look at is a regression task: a simple best-fit line to a set of data.
Again, this is an example of fitting a model to data, but our focus here is that the model can make general-
izations about new data. The model has been learned from the training data, and can be used to predict the
result of test data: here, we might be given an x-value, and the model would allow us to predict the y value.
20.1.2 Data in scikit-learn
The data matrix
Machine learning algorithms implemented in scikit-learn expect data to be stored in a two-dimensional array
or matrix. The arrays can be either numpy arrays, or in some cases scipy.sparse matrices. The size of the
array is expected to be [n_samples, n_features]
n_samples: The number of samples: each sample is an item to process (e.g. classify). A sample can be a
document, a picture, a sound, a video, an astronomical object, a row in database or CSV file, or whatever
you can describe with a fixed set of quantitative traits.
n_features: The number of features or distinct traits that can be used to describe each item in a quanti-
tative manner. Features are generally real-valued, but may be boolean or discrete-valued in some cases.
Tip: The number of features must be fixed in advance. However it can be very high dimensional (e.g. millions
of features) with most of them being zeros for a given sample. This is a case where scipy.sparse matrices
can be useful, in that they are much more memory-efficient than numpy arrays.
A Simple Example: the Iris Dataset
The application problem
As an example of a simple dataset, let us a look at the iris data stored by scikit-learn. Suppose we want to
recognize species of irises. The data consists of measurements of three different species of irises:

Setosa Iris Versicolor Iris Virginica Iris
Quick Question:
If we want to design an algorithm to recognize iris species, what might the data be?
Remember: we need a 2D array of size [n_samples x n_features].
What would the n_samples refer to?
What might the n_features refer to?
Remember that there must be a fixed number of features for each sample, and feature number i must be a
similar kind of quantity for each sample.
Loading the Iris Data with Scikit-learn
Scikit-learn has a very straightforward set of data on these iris species. The data consist of the following:
Features in the Iris dataset:
sepal length (cm)
sepal width (cm)
petal length (cm)
petal width (cm)
Target classes to predict:
Setosa
Versicolour
Virginica
scikit-learn embeds a copy of the iris CSV file along with a function to load it into numpy arrays:
>>> from sklearn.datasets import load_iris
>>> iris = load_iris()
Note: Import sklearn Note that scikit-learn is imported as sklearn
The features of each sample flower are stored in the data attribute of the dataset:
>>> print(iris.data.shape)
(150, 4)
>>> n_samples, n_features = iris.data.shape
>>> print(n_samples)

150
>>> print(n_features)
4
>>> print(iris.data[0])
[ 5.1 3.5 1.4 0.2]
The information about the class of each sample is stored in the target attribute of the dataset:
>>> print(iris.target.shape)
(150,)
>>> print(iris.target)
[0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2
2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
2 2]
The names of the classes are stored in the last attribute, namely target_names:
>>> print(iris.target_names)
['setosa' 'versicolor' 'virginica']
This data is four-dimensional, but we can visualize two of the dimensions at a time using a scatter plot:
Excercise:
Can you choose 2 features to find a plot where it is easier to seperate the different classes of irises?
Hint: click on the figure above to see the code that generates it, and modify this code.

20.2 Basic principles of machine learning with scikit-learn
20.2.1 Introducing the scikit-learn estimator object
Every algorithm is exposed in scikit-learn via an Estimator object. For instance a linear regression is:
sklearn.linear_model.LinearRegression
>>> from sklearn.linear_model import LinearRegression
Estimator parameters: All the parameters of an estimator can be set when it is instantiated:
>>> model = LinearRegression(normalize=True)

>>> print(model.normalize)
True
>>> print(model)
LinearRegression(copy_X=True, fit_intercept=True, n_jobs=1, normalize=True)
Fitting on data
Lets create some simple data with numpy:

>>> x = np.array([0, 1, 2])
>>> y = np.array([0, 1, 2])
>>> X = x[:, np.newaxis] # The input data for sklearn is 2D: (samples == 3 x features == 1)
>>> X
array([[0],
[1],
[2]])
>>> model.fit(X, y)
LinearRegression(copy_X=True, fit_intercept=True, n_jobs=1, normalize=True)
Estimated parameters: When data is fitted with an estimator, parameters are estimated from the data at hand.
All the estimated parameters are attributes of the estimator object ending by an underscore:
>>> model.coef_
array([ 1.])
20.2.2 Supervised Learning: Classification and regression
In Supervised Learning, we have a dataset consisting of both features and labels. The task is to construct an
estimator which is able to predict the label of an object given the set of features. A relatively simple example
is predicting the species of iris given a set of measurements of its flower. This is a relatively simple task. Some
more complicated examples are:
given a multicolor image of an object through a telescope, determine whether that object is a star, a
quasar, or a galaxy.
given a photograph of a person, identify the person in the photo.
given a list of movies a person has watched and their personal rating of the movie, recommend a list of
movies they would like (So-called recommender systems: a famous example is the Netflix Prize).
Tip: What these tasks have in common is that there is one or more unknown quantities associated with the
object which needs to be determined from other observed quantities.
20.2. Basic principles of machine learning with scikit-learn 577

Supervised learning is further broken down into two categories, classification and regression. In classifica-
tion, the label is discrete, while in regression, the label is continuous. For example, in astronomy, the task
of determining whether an object is a star, a galaxy, or a quasar is a classification problem: the label is from
three distinct categories. On the other hand, we might wish to estimate the age of an object based on such
observations: this would be a regression problem, because the label (age) is a continuous quantity.
Classification: K nearest neighbors (kNN) is one of the simplest learning strategies: given a new, unknown ob-
servation, look up in your reference database which ones have the closest features and assign the predominant
class. Lets try it out on our iris classification problem:
from sklearn import neighbors, datasets

iris = datasets.load_iris()
X, y = iris.data, iris.target
knn = neighbors.KNeighborsClassifier(n_neighbors=1)
knn.fit(X, y)
# What kind of iris has 3cm x 5cm sepal and 4cm x 2cm petal?
print(iris.target_names[knn.predict([[3, 5, 4, 2]])])
Fig. 20.3: A plot of the sepal space and the prediction of the KNN
Regression: The simplest possible regression setting is the linear regression one:
20.2.3 A recap on Scikit-learns estimator interface
Scikit-learn strives to have a uniform interface across all methods, and well see examples of these below. Given
a scikit-learn estimator object named model, the following methods are available:
In all Estimators

Fig. 20.4: A plot of a simple linear regression.
model.fit() : fit training data. For supervised learning applications, this accepts two
arguments: the data X and the labels y (e.g. model.fit(X, y)). For unsupervised learn-
ing applications, this accepts only a single argument, the data X (e.g. model.fit(X)).
In supervised estimators
model.predict() : given a trained model, predict the label of a new set of data. This
method accepts one argument, the new data X_new (e.g. model.predict(X_new)), and
returns the learned label for each object in the array.
model.predict_proba() : For classification problems, some estimators also provide
this method, which returns the probability that a new observation has each categor-
ical label. In this case, the label with the highest probability is returned by model.
predict().
model.score() : for classification or regression problems, most (all?) estimators imple-
ment a score method. Scores are between 0 and 1, with a larger score indicating a better
fit.
In unsupervised estimators
model.transform() : given an unsupervised model, transform new data into the new
basis. This also accepts one argument X_new, and returns the new representation of the
data based on the unsupervised model.
model.fit_transform() : some estimators implement this method, which more effi-
ciently performs a fit and a transform on the same input data.
20.2.4 Regularization: what it is and why it is necessary
Prefering simpler models
Train errors Suppose you are using a 1-nearest neighbor estimator. How many errors do you expect on your
train set?
Train set error is not a good measurement of prediction performance. You need to leave out a test set.
In general, we should accept errors on the train set.

An example of regularization The core idea behind regularization is that we are going to prefer models that
are simpler, for a certain definition of simpler, even if they lead to more errors on the train set.
As an example, lets generate with a 9th order polynomial, with noise:
And now, lets fit a 4th order and a 9th order polynomial to the data.
With your naked eyes, which model do you prefer, the 4th order one, or the 9th order one?
Lets look at the ground truth:
Tip: Regularization is ubiquitous in machine learning. Most scikit-learn estimators have a parameter to tune
the amount of regularization. For instance, with k-NN, it is k, the number of nearest neighbors used to make

the decision. k=1 amounts to no regularization: 0 error on the training set, whereas large k will push toward
smoother decision boundaries in the feature space.
Simple versus complex models for classification
A linear separation A non-linear separation
Tip: For classification models, the decision boundary, that separates the class expresses the complexity of the
model. For instance, a linear model, that makes a decision based on a linear combination of features, is more
complex than a non-linear one.

20.3 Supervised Learning: Classification of Handwritten Digits
20.3.1 The nature of the data
Code and notebook
Python code and Jupyter notebook for this section are found here
In this section well apply scikit-learn to the classification of handwritten digits. This will go a bit beyond
the iris classification we saw before: well discuss some of the metrics which can be used in evaluating the
effectiveness of a classification model.
>>> from sklearn.datasets import load_digits

>>> digits = load_digits()
Let
us visualize the data and remind us what were looking at (click on the figure for the full code):
# plot the digits: each image is 8x8 pixels

for i in range(64):
ax = fig.add_subplot(8, 8, i + 1, xticks=[], yticks=[])
20.3. Supervised Learning: Classification of Handwritten Digits 582

ax.imshow(digits.images[i], cmap=plt.cm.binary, interpolation='nearest')
20.3.2 Visualizing the Data on its principal components
A good first-step for many problems is to visualize the data using a Dimensionality Reduction technique. Well
start with the most straightforward one, Principal Component Analysis (PCA).
PCA seeks orthogonal linear combinations of the features which show the greatest variance, and as such, can
help give you a good idea of the structure of the data set.
>>> from sklearn.decomposition import PCA

>>> pca = PCA(n_components=2)
>>> proj = pca.fit_transform(digits.data)
>>> plt.scatter(proj[:, 0], proj[:, 1], c=digits.target)
<matplotlib.collections.PathCollection object at ...>
>>> plt.colorbar()
<matplotlib.colorbar.Colorbar object at ...>
Question
Given these projections of the data, which numbers do you think a classifier might have trouble distinguish-
ing?

20.3.3 Gaussian Naive Bayes Classification
For most classification problems, its nice to have a simple, fast method to provide a quick baseline classifi-
cation. If the simple and fast method is sufficient, then we dont have to waste CPU cycles on more complex
models. If not, we can use the results of the simple method to give us clues about our data.
One good method to keep in mind is Gaussian Naive Bayes (sklearn.naive_bayes.GaussianNB).
Old scikit-learn versions
train_test_split() is imported from sklearn.cross_validation
Tip: Gaussian Naive Bayes fits a Gaussian distribution to each training label independantly on each feature,
and uses this to quickly give a rough classification. It is generally not sufficiently accurate for real-world data,
but can perform surprisingly well, for instance on text data.
>>> from sklearn.naive_bayes import GaussianNB

>>> from sklearn.model_selection import train_test_split
>>> # split the data into training and validation sets

>>> X_train, X_test, y_train, y_test = train_test_split(digits.data, digits.target)
>>> # train the model

>>> clf = GaussianNB()
>>> clf.fit(X_train, y_train)
GaussianNB(priors=None)
>>> # use the model to predict the labels of the test data
>>> predicted = clf.predict(X_test)
>>> expected = y_test
>>> print(predicted)
[1 7 7 7 8 2 8 0 4 8 7 7 0 8 2 3 5 8 5 3 7 9 6 2 8 2 2 7 3 5...]
>>> print(expected)
[1 0 4 7 8 2 2 0 4 3 7 7 0 8 2 3 4 8 5 3 7 9 6 3 8 2 2 9 3 5...]
As above, we plot the digits with the predicted labels to get an idea of how well the classification is working.

Question
Why did we split the data into training and validation sets?
20.3.4 Quantitative Measurement of Performance
Wed like to measure the performance of our estimator without having to resort to plotting examples. A simple
method might be to simply compare the number of matches:
>>> matches = (predicted == expected)

>>> print(matches.sum())
367
>>> print(len(matches))
450
>>> matches.sum() / float(len(matches))
0.81555555555555559
We see that more than 80% of the 450 predictions match the input. But there are other more sophisticated
metrics that can be used to judge the performance of a classifier: several are available in the sklearn.metrics

submodule.
One of the most useful metrics is the classification_report, which combines several measures and prints
a table with the results:
>>> from sklearn import metrics

>>> print(metrics.classification_report(expected, predicted))
precision recall f1-score support
0 1.00 0.91 0.95 46

1 0.76 0.64 0.69 44
2 0.85 0.62 0.72 47
3 0.98 0.82 0.89 49
4 0.89 0.86 0.88 37
5 0.97 0.93 0.95 41
6 1.00 0.98 0.99 44
7 0.73 1.00 0.84 45
8 0.50 0.90 0.64 49
9 0.93 0.54 0.68 48
avg / total 0.86 0.82 0.82 450
Another enlightening metric for this sort of multi-label classification is a confusion matrix: it helps us visualize
which labels are being interchanged in the classification errors:
>>> print(metrics.confusion_matrix(expected, predicted))

[[42 0 0 0 3 0 0 1 0 0]
[ 0 28 0 0 0 0 0 1 13 2]
[ 0 3 29 0 0 0 0 0 15 0]
[ 0 0 2 40 0 0 0 2 5 0]
[ 0 0 1 0 32 1 0 3 0 0]
[ 0 0 0 0 0 38 0 2 1 0]
[ 0 0 1 0 0 0 43 0 0 0]
[ 0 0 0 0 0 0 0 45 0 0]
[ 0 3 1 0 0 0 0 1 44 0]
[ 0 3 0 1 1 0 0 7 10 26]]
We see here that in particular, the numbers 1, 2, 3, and 9 are often being labeled 8.
20.4 Supervised Learning: Regression of Housing Data
Here well do a short example of a regression problem: learning a continuous value from a set of features.
20.4.1 A quick look at the data
Code and notebook
Well use the simple Boston house prices set, available in scikit-learn. This records measurements of 13 at-
tributes of housing markets around Boston, as well as the median price. The question is: can you predict the
price of a new market given its attributes?:
>>> from sklearn.datasets import load_boston

>>> data = load_boston()
>>> print(data.data.shape)
(506, 13)
20.4. Supervised Learning: Regression of Housing Data 586

>>> print(data.target.shape)
(506,)
We can see that there are just over 500 data points.
The DESCR variable has a long description of the dataset:
>>> print(data.DESCR)
Boston House Prices dataset
===========================
Notes
------
Data Set Characteristics:
:Number of Instances: 506
:Number of Attributes: 13 numeric/categorical predictive
:Median Value (attribute 14) is usually the target
:Attribute Information (in order):

- CRIM per capita crime rate by town
- ZN proportion of residential land zoned for lots over 25,000 sq.ft.
- INDUS proportion of non-retail business acres per town
- CHAS Charles River dummy variable (= 1 if tract bounds river; 0 otherwise)
- NOX nitric oxides concentration (parts per 10 million)
- RM average number of rooms per dwelling
- AGE proportion of owner-occupied units built prior to 1940
- DIS weighted distances to five Boston employment centres
- RAD index of accessibility to radial highways
- TAX full-value property-tax rate per $10,000
- PTRATIO pupil-teacher ratio by town
- B 1000(Bk - 0.63)^2 where Bk is the proportion of blacks by town
- LSTAT % lower status of the population
- MEDV Median value of owner-occupied homes in $1000's
...
It often helps to quickly visualize pieces of the data using histograms, scatter plots, or other plot types. With
pylab, let us show a histogram of the target values: the median price in each neighborhood:
>>> plt.hist(data.target)
(array([...
Lets have a quick look to see if some features are more rele-
vant than others for our problem:
>>> for index, feature_name in enumerate(data.feature_names):

... plt.figure()
... plt.scatter(data.data[:, index], data.target)

<matplotlib.figure.Figure object...
This is a manual version of a technique called feature selection.
Tip: Sometimes, in Machine Learning it is useful to use feature selection to decide which features are the most
useful for a particular problem. Automated methods exist which quantify this sort of exercise of choosing the
most informative features.
20.4.2 Predicting Home Prices: a Simple Linear Regression
Now well use scikit-learn to perform a simple linear regression on the housing data. There are many
possibilities of regressors to use. A particularly simple one is LinearRegression: this is basically a wrapper
around an ordinary least squares calculation.
>>> from sklearn.model_selection import train_test_split

>>> X_train, X_test, y_train, y_test = train_test_split(data.data, data.target)
>>> clf = LinearRegression()
>>> clf.fit(X_train, y_train)
LinearRegression(copy_X=True, fit_intercept=True, n_jobs=1, normalize=False)
>>> predicted = clf.predict(X_test)
>>> expected = y_test
>>> print("RMS: %s " % np.sqrt(np.mean((predicted - expected) ** 2)))
RMS: 5.0059...
We can plot the error: expected as a

function of predicted:
>>> plt.scatter(expected, predicted)


Tip: The prediction at least correlates with the true price, though there are clearly some biases. We could
imagine evaluating the performance of the regressor by, say, computing the RMS residuals between the true
and predicted price. There are some subtleties in this, however, which well cover in a later section.
Exercise: Gradient Boosting Tree Regression
There are many other types of regressors available in scikit-learn: well try a more powerful one here.
Use the GradientBoostingRegressor class to fit the housing data.
hint You can copy and paste some of the above code, replacing LinearRegression with
GradientBoostingRegressor:
from sklearn.ensemble import GradientBoostingRegressor
# Instantiate the model, fit the results, and scatter in vs. out
Solution The solution is found in the code of this chapter
20.5 Measuring prediction performance
20.5.1 A quick test on the K-neighbors classifier
Here well continue to look at the digits data, but well switch to the K-Neighbors classifier. The K-neighbors
classifier is an instance-based classifier. The K-neighbors classifier predicts the label of an unknown point
based on the labels of the K nearest points in the parameter space.
>>> # Get the data

>>> X = digits.data
>>> y = digits.target
>>> # Instantiate and train the classifier

>>> from sklearn.neighbors import KNeighborsClassifier
>>> clf = KNeighborsClassifier(n_neighbors=1)
>>> clf.fit(X, y)
KNeighborsClassifier(...)
>>> # Check the results using metrics

>>> from sklearn import metrics
>>> y_pred = clf.predict(X)
>>> print(metrics.confusion_matrix(y_pred, y))

[[178 0 0 0 0 0 0 0 0 0]
[ 0 182 0 0 0 0 0 0 0 0]
[ 0 0 177 0 0 0 0 0 0 0]
[ 0 0 0 183 0 0 0 0 0 0]
[ 0 0 0 0 181 0 0 0 0 0]
[ 0 0 0 0 0 182 0 0 0 0]
[ 0 0 0 0 0 0 181 0 0 0]
[ 0 0 0 0 0 0 0 179 0 0]
[ 0 0 0 0 0 0 0 0 174 0]
[ 0 0 0 0 0 0 0 0 0 180]]
Apparently, weve found a perfect classifier! But this is misleading for the reasons we saw before: the classifier
essentially memorizes all the samples it has already seen. To really test how well this algorithm does, we need
to try some samples it hasnt yet seen.
20.5. Measuring prediction performance 589

This problem also occurs with regression models. In the following we fit an other instance-based model named
decision tree to the Boston Housing price dataset we introduced previously:
>>> from sklearn.datasets import load_boston

>>> from sklearn.tree import DecisionTreeRegressor
>>> data = load_boston()

>>> clf = DecisionTreeRegressor().fit(data.data, data.target)
>>> predicted = clf.predict(data.data)
>>> expected = data.target
>>> plt.scatter(expected, predicted)

>>> plt.plot([0, 50], [0, 50], '--k')
[<matplotlib.lines.Line2D object at ...]
Here again the predictions are seemingly perfect as the model was able to perfectly memorize the training set.
Warning: Performance on test set

Performance on test set does not measure overfit (as described above)
20.5.2 A correct approach: Using a validation set
Learning the parameters of a prediction function and testing it on the same data is a methodological mistake:
a model that would just repeat the labels of the samples that it has just seen would have a perfect score but
would fail to predict anything useful on yet-unseen data.
To avoid over-fitting, we have to define two different sets:
a training set X_train, y_train which is used for learning the parameters of a predictive model
a testing set X_test, y_test which is used for evaluating the fitted predictive model
In scikit-learn such a random split can be quickly computed with the train_test_split() function:
>>> from sklearn import model_selection

>>> X = digits.data
>>> X_train, X_test, y_train, y_test = model_selection.train_test_split(X, y,

... test_size=0.25, random_state=0)

>>> print("%r , %r , %r " % (X.shape, X_train.shape, X_test.shape))

(1797, 64), (1347, 64), (450, 64)
Now we train on the training data, and test on the testing data:
>>> clf = KNeighborsClassifier(n_neighbors=1).fit(X_train, y_train)

>>> y_pred = clf.predict(X_test)
>>> print(metrics.confusion_matrix(y_test, y_pred))

[[37 0 0 0 0 0 0 0 0 0]
[ 0 43 0 0 0 0 0 0 0 0]
[ 0 0 43 1 0 0 0 0 0 0]
[ 0 0 0 45 0 0 0 0 0 0]
[ 0 0 0 0 38 0 0 0 0 0]
[ 0 0 0 0 0 47 0 0 0 1]
[ 0 0 0 0 0 0 52 0 0 0]
[ 0 0 0 0 0 0 0 48 0 0]
[ 0 0 0 0 0 0 0 0 48 0]
[ 0 0 0 1 0 1 0 0 0 45]]
>>> print(metrics.classification_report(y_test, y_pred))
0 1.00 1.00 1.00 37

1 1.00 1.00 1.00 43
2 1.00 0.98 0.99 44
3 0.96 1.00 0.98 45
4 1.00 1.00 1.00 38
5 0.98 0.98 0.98 48
6 1.00 1.00 1.00 52
7 1.00 1.00 1.00 48
8 1.00 1.00 1.00 48
9 0.98 0.96 0.97 47
avg / total 0.99 0.99 0.99 450
The averaged f1-score is often used as a convenient measure of the overall performance of an algorithm. It
appears in the bottom row of the classification report; it can also be accessed directly:
>>> metrics.f1_score(y_test, y_pred, average="macro")

0.991367...
The over-fitting we saw previously can be quantified by computing the f1-score on the training data itself:
>>> metrics.f1_score(y_train, clf.predict(X_train), average="macro")

1.0
Note: Regression metrics In the case of regression models, we need to use different metrics, such as explained
variance.
20.5.3 Model Selection via Validation
Tip: We have applied Gaussian Naives, support vectors machines, and K-nearest neighbors classifiers to the
digits dataset. Now that we have these validation tools in place, we can ask quantitatively which of the three
estimators works best for this dataset.

With the default hyper-parameters for each estimator, which gives the best f1 score on the validation
set? Recall that hyperparameters are the parameters set when you instantiate the classifier: for example,
the n_neighbors in clf = KNeighborsClassifier(n_neighbors=1)
>>> from sklearn.naive_bayes import GaussianNB

>>> from sklearn.neighbors import KNeighborsClassifier
>>> from sklearn.svm import LinearSVC
>>> X = digits.data
>>> X_train, X_test, y_train, y_test = model_selection.train_test_split(X, y,
... test_size=0.25, random_state=0)
>>> for Model in [GaussianNB, KNeighborsClassifier, LinearSVC]:

... clf = Model().fit(X_train, y_train)
... y_pred = clf.predict(X_test)
... print('%s : %s ' %
... (Model.__name__, metrics.f1_score(y_test, y_pred, average="macro")))
GaussianNB: 0.8332741681...
KNeighborsClassifier: 0.9804562804...
LinearSVC: 0.93...
For each classifier, which value for the hyperparameters gives the best results for the digits data? For
LinearSVC, use loss='l2' and loss='l1'. For KNeighborsClassifier we use n_neighbors be-
tween 1 and 10. Note that GaussianNB does not have any adjustable hyperparameters.
LinearSVC(loss='l1'): 0.930570687535
LinearSVC(loss='l2'): 0.933068826918
-------------------
KNeighbors(n_neighbors=1): 0.991367521884
Solution: code source
20.5.4 Cross-validation
Cross-validation consists in repetively splitting the data in pairs of train and test sets, called folds. Scikit-learn
comes with a function to automatically compute score on all these folds. Here we do KFold with k=5.
>>> clf = KNeighborsClassifier()

>>> from sklearn.model_selection import cross_val_score
>>> cross_val_score(clf, X, y, cv=5)
array([ 0.9478022 , 0.9558011 , 0.96657382, 0.98039216, 0.96338028])
We can use different splitting strategies, such as random splitting:
>>> from sklearn.model_selection import ShuffleSplit

>>> cv = ShuffleSplit(n_splits=5)
>>> cross_val_score(clf, X, y, cv=cv)
array([...])
Tip: There exists many different cross-validation strategies in scikit-learn. They are often useful to take in

account non iid datasets.
20.5.5 Hyperparameter optimization with cross-validation
Consider regularized linear models, such as Ridge Regression, which uses l2 regularlization, and Lasso Regres-
sion, which uses l1 regularization. Choosing their regularization parameter is important.
Let us set these parameters on the Diabetes dataset, a simple regression problem. The diabetes data consists
of 10 physiological variables (age, sex, weight, blood pressure) measure on 442 patients, and an indication of
disease progression after one year:
>>> from sklearn.datasets import load_diabetes

>>> data = load_diabetes()
>>> X, y = data.data, data.target
>>> print(X.shape)
(442, 10)
With the default hyper-parameters: we compute the cross-validation score:
>>> from sklearn.linear_model import Ridge, Lasso
>>> for Model in [Ridge, Lasso]:

... model = Model()
... print('%s : %s ' % (Model.__name__, cross_val_score(model, X, y).mean()))
Ridge: 0.409427438303
Lasso: 0.353800083299
Basic Hyperparameter Optimization
We compute the cross-validation score as a function of alpha, the strength of the regularization for Lasso and
Ridge. We choose 20 values of alpha between 0.0001 and 1:
>>> alphas = np.logspace(-3, -1, 30)
>>> for Model in [Lasso, Ridge]:

... scores = [cross_val_score(Model(alpha), X, y, cv=3).mean()
... for alpha in alphas]
... plt.plot(alphas, scores, label=Model.__name__)
[<matplotlib.lines.Line2D object at ...
Question
Can we trust our results to be actually useful?

Automatically Performing Grid Search
sklearn.grid_search.GridSearchCV is constructed with an estimator, as well as a dictionary of parameter

values to be searched. We can find the optimal parameters this way:
>>> from sklearn.grid_search import GridSearchCV

>>> for Model in [Ridge, Lasso]:
... gscv = GridSearchCV(Model(), dict(alpha=alphas), cv=3).fit(X, y)
... print('%s : %s ' % (Model.__name__, gscv.best_params_))
Ridge: {'alpha': 0.062101694189156162}
Lasso: {'alpha': 0.01268961003167922}
Built-in Hyperparameter Search
For some models within scikit-learn, cross-validation can be performed more efficiently on large datasets. In
this case, a cross-validated version of the particular model is included. The cross-validated versions of Ridge
and Lasso are RidgeCV and LassoCV, respectively. Parameter search on these estimators can be performed as
follows:
>>> from sklearn.linear_model import RidgeCV, LassoCV

>>> for Model in [RidgeCV, LassoCV]:
... model = Model(alphas=alphas, cv=3).fit(X, y)
... print('%s : %s ' % (Model.__name__, model.alpha_))
RidgeCV: 0.0621016941892
LassoCV: 0.0126896100317
We see that the results match those returned by GridSearchCV
Nested cross-validation
How do we measure the performance of these estimators? We have used data to set the hyperparameters, so
we need to test on actually new data. We can do this by running cross_val_score() on our CV objects. Here
there are 2 cross-validation loops going on, this is called nested cross validation:
for Model in [RidgeCV, LassoCV]:

scores = cross_val_score(Model(alphas=alphas, cv=3), X, y, cv=3)
print(Model.__name__, np.mean(scores))
Note: Note that these results do not match the best results of our curves above, and LassoCV seems to under-
perform RidgeCV. The reason is that setting the hyper-parameter is harder for Lasso, thus the estimation error
on this hyper-parameter is larger.
20.6 Unsupervised Learning: Dimensionality Reduction and Visual-

ization
Unsupervised learning is applied on X without y: data without labels. A typical use case is to find hidden
structure in the data.
20.6.1 Dimensionality Reduction: PCA
Dimensionality reduction derives a set of new artificial features smaller than the original feature set. Here well
use Principal Component Analysis (PCA), a dimensionality reduction that strives to retain most of the variance
of the original data. Well use sklearn.decomposition.PCA on the iris dataset:
20.6. Unsupervised Learning: Dimensionality Reduction and Visualization 594

>>> X = iris.data
>>> y = iris.target
Tip: PCA computes linear combinations of the original features using a truncated Singular Value Decomposi-
tion of the matrix X, to project the data onto a base of the top singular vectors.
>>> from sklearn.decomposition import PCA

>>> pca = PCA(n_components=2, whiten=True)
>>> pca.fit(X)
PCA(..., n_components=2, ...)
Once fitted, PCA exposes the singular vectors in the components_ attribute:
>>> pca.components_
array([[ 0.36158..., -0.08226..., 0.85657..., 0.35884...],
[ 0.65653..., 0.72971..., -0.17576..., -0.07470...]])
Other attributes are available as well:
>>> pca.explained_variance_ratio_
array([ 0.92461..., 0.05301...])
Let us project the iris dataset along those first two dimensions::
>>> X_pca = pca.transform(X)

>>> X_pca.shape
(150, 2)
PCA normalizes and whitens the data, which means that the data is now centered on both components with
unit variance:
>>> X_pca.mean(axis=0)
array([ ...e-15, ...e-15])
>>> X_pca.std(axis=0)
array([ 1., 1.])
Furthermore, the samples components do no longer carry any linear correlation:
>>> np.corrcoef(X_pca.T)
array([[ 1.00000000e+00, ...],
[ ..., 1.00000000e+00]])
With a number of retained components 2 or 3, PCA is useful to visualize the dataset:
>>> target_ids = range(len(iris.target_names))

>>> for i, c, label in zip(target_ids, 'rgbcmykw', iris.target_names):
... plt.scatter(X_pca[y == i, 0], X_pca[y == i, 1],
... c=c, label=label)
<matplotlib.collections.PathCollection ...

Tip: Note that this projection was determined without any information about the labels (represented by the
colors): this is the sense in which the learning is unsupervised. Nevertheless, we see that the projection gives
us insight into the distribution of the different flowers in parameter space: notably, iris setosa is much more
distinct than the other two species.
20.6.2 Visualization with a non-linear embedding: tSNE
For visualization, more complex embeddings can be useful (for statistical analysis, they are harder to control).
sklearn.manifold.TSNE is such a powerful manifold learning method. We apply it to the digits dataset, as
the digits are vectors of dimension 8*8 = 64. Embedding them in 2D enables visualization:
>>> # Take the first 500 data points: it's hard to see 1500 points
>>> X = digits.data[:500]
>>> y = digits.target[:500]
>>> # Fit and transform with a TSNE

>>> from sklearn.manifold import TSNE
>>> tsne = TSNE(n_components=2, random_state=0)
>>> X_2d = tsne.fit_transform(X)
>>> # Visualize the data

>>> plt.scatter(X_2d[:, 0], X_2d[:, 1], c=y)

fit_transform
As TSNE cannot be applied to new data, we need to use its fit_transform method.
sklearn.manifold.TSNE separates quite well the different classes of digits eventhough it had no access to
the class information.
Exercise: Other dimension reduction of digits
sklearn.manifold has many other non-linear embeddings. Try them out on the digits dataset. Could you
judge their quality without knowing the labels y?
>>> # ...
20.7 The eigenfaces example: chaining PCA and SVMs
Code and notebook
The goal of this example is to show how an unsupervised method and a supervised one can be chained for
better prediction. It starts with a didactic but lengthy way of doing things, and finishes with the idiomatic
approach to pipelining in scikit-learn.
20.7. The eigenfaces example: chaining PCA and SVMs 597

Here well take a look at a simple facial recognition example. Ideally, we would use a dataset consisting of a
subset of the Labeled Faces in the Wild data that is available with sklearn.datasets.fetch_lfw_people().
However, this is a relatively large download (~200MB) so we will do the tutorial on a simpler, less rich dataset.
Feel free to explore the LFW dataset.
from sklearn import datasets

faces = datasets.fetch_olivetti_faces()
faces.data.shape
Lets visualize these faces to see what were working with

# plot several images
for i in range(15):
ax.imshow(faces.images[i], cmap=plt.cm.bone)
Tip: Note is that these faces have already been localized and scaled to a common size. This is an important
preprocessing piece for facial recognition, and is a process that can require a large collection of training data.
This can be done in scikit-learn, but the challenge is gathering a sufficient amount of training data for the
algorithm to work. Fortunately, this piece is common enough that it has been done. One good resource is
OpenCV, the Open Computer Vision Library.
Well perform a Support Vector classification of the images. Well do a typical train-test split on the images:
from sklearn.model_selection import train_test_split

X_train, X_test, y_train, y_test = train_test_split(faces.data,
faces.target, random_state=0)

print(X_train.shape, X_test.shape)
Out:
(300, 4096) (100, 4096)
20.7.1 Preprocessing: Principal Component Analysis
1850 dimensions is a lot for SVM. We can use PCA to reduce these 1850 features to a manageable size, while
maintaining most of the information in the dataset.
from sklearn import decomposition

pca = decomposition.PCA(n_components=150, whiten=True)
pca.fit(X_train)
One interesting part of PCA is that it computes the mean face, which can be interesting to examine:
plt.imshow(pca.mean_.reshape(faces.images[0].shape),
cmap=plt.cm.bone)
The principal components measure deviations about this mean along orthogonal axes.
print(pca.components_.shape)
Out:
(150, 4096)

It is also interesting to visualize these principal components:

for i in range(30):
ax.imshow(pca.components_[i].reshape(faces.images[0].shape),
cmap=plt.cm.bone)
The components (eigenfaces) are ordered by their importance from top-left to bottom-right. We see that the
first few components seem to primarily take care of lighting conditions; the remaining components pull out
certain identifying features: the nose, eyes, eyebrows, etc.
With this projection computed, we can now project our original training and test data onto the PCA basis:
X_train_pca = pca.transform(X_train)
X_test_pca = pca.transform(X_test)
print(X_train_pca.shape)
Out:
(300, 150)
print(X_test_pca.shape)
Out:
(100, 150)
These projected components correspond to factors in a linear combination of component images such that
the combination approaches the original face.
20.7.2 Doing the Learning: Support Vector Machines
Now well perform support-vector-machine classification on this reduced dataset:
from sklearn import svm

clf = svm.SVC(C=5., gamma=0.001)
clf.fit(X_train_pca, y_train)
Finally, we can evaluate how well this classification did. First, we might plot a few of the test-cases with the
labels learned from the training set:
import numpy as np
for i in range(15):


ax.imshow(X_test[i].reshape(faces.images[0].shape),
cmap=plt.cm.bone)
y_pred = clf.predict(X_test_pca[i, np.newaxis])[0]
color = ('black' if y_pred == y_test[i] else 'red')
ax.set_title(faces.target[y_pred],
fontsize='small', color=color)
The classifier is correct on an impressive number of images given the simplicity of its learning model! Using
a linear classifier on 150 features derived from the pixel-level data, the algorithm correctly identifies a large
number of the people in the images.
Again, we can quantify this effectiveness using one of several measures from sklearn.metrics. First we can
do the classification report, which shows the precision, recall and other measures of the goodness of the
classification:
from sklearn import metrics

y_pred = clf.predict(X_test_pca)
print(metrics.classification_report(y_test, y_pred))
Out:
0 1.00 0.67 0.80 6

1 1.00 1.00 1.00 4
2 0.50 1.00 0.67 2
3 1.00 1.00 1.00 1
4 1.00 1.00 1.00 1
5 1.00 1.00 1.00 5
6 1.00 1.00 1.00 4

7 1.00 0.67 0.80 3

9 1.00 1.00 1.00 1
10 1.00 1.00 1.00 4
11 1.00 1.00 1.00 1
12 0.67 1.00 0.80 2
13 1.00 1.00 1.00 3
14 1.00 1.00 1.00 5
15 1.00 1.00 1.00 3
17 1.00 1.00 1.00 6
19 1.00 1.00 1.00 4
20 1.00 1.00 1.00 1
21 1.00 1.00 1.00 1
22 1.00 1.00 1.00 2
23 0.50 1.00 0.67 1
24 1.00 1.00 1.00 2
25 1.00 0.50 0.67 2
26 1.00 0.75 0.86 4
27 1.00 1.00 1.00 1
28 0.67 1.00 0.80 2
29 1.00 1.00 1.00 3
30 1.00 1.00 1.00 4
31 1.00 1.00 1.00 3
32 1.00 1.00 1.00 3
33 1.00 1.00 1.00 2
34 1.00 1.00 1.00 3
35 1.00 1.00 1.00 1
36 1.00 1.00 1.00 3
37 1.00 1.00 1.00 3
38 1.00 1.00 1.00 1
39 1.00 1.00 1.00 3
avg / total 0.97 0.95 0.95 100
Another interesting metric is the confusion matrix, which indicates how often any two items are mixed-up.
The confusion matrix of a perfect classifier would only have nonzero entries on the diagonal, with zeros on the
off-diagonal:
print(metrics.confusion_matrix(y_test, y_pred))
Out:
[[4 0 0 ..., 0 0 0]
[0 4 0 ..., 0 0 0]
[0 0 2 ..., 0 0 0]
...,
[0 0 0 ..., 3 0 0]
[0 0 0 ..., 0 1 0]
[0 0 0 ..., 0 0 3]]
20.7.3 Pipelining
Above we used PCA as a pre-processing step before applying our support vector machine classifier. Plugging
the output of one estimator directly into the input of a second estimator is a commonly used pattern; for this
reason scikit-learn provides a Pipeline object which automates this process. The above problem can be re-
expressed as a pipeline as follows:
from sklearn.pipeline import Pipeline

clf = Pipeline([('pca', decomposition.PCA(n_components=150, whiten=True)),
('svm', svm.LinearSVC(C=1.0))])

clf.fit(X_train, y_train)
y_pred = clf.predict(X_test)
print(metrics.confusion_matrix(y_pred, y_test))
20.8 Parameter selection, Validation, and Testing
20.8.1 Hyperparameters, Over-fitting, and Under-fitting
See also:
This section is adapted from Andrew Ngs excellent Coursera course
The issues associated with validation and cross-validation are some of the most important aspects of the prac-
tice of machine learning. Selecting the optimal model for your data is vital, and is a piece of the problem that
is not often appreciated by machine learning practitioners.
The central question is: If our estimator is underperforming, how should we move forward?
Use simpler or more complicated model?
Add more features to each observed data point?
Add more training samples?
The answer is often counter-intuitive. In particular, Sometimes using a more complicated model will give
worse results. Also, Sometimes adding training data will not improve your results. The ability to determine
what steps will improve your model is what separates the successful machine learning practitioners from the
unsuccessful.
Bias-variance trade-off: illustration on a simple regression problem
Code and notebook
Let us start with a simple 1D regression problem. This will help us to easily visualize the data and the model,
and the results generalize easily to higher-dimensional datasets. Well explore a simple linear regression prob-
lem, with sklearn.linear_model.
X = np.c_[ .5, 1].T

y = [.5, 1]
X_test = np.c_[ 0, 2].T
Without noise, as linear regression fits the data perfectly
from sklearn import linear_model

regr = linear_model.LinearRegression()
regr.fit(X, y)
plt.plot(X, y, 'o')
plt.plot(X_test, regr.predict(X_test))
20.8. Parameter selection, Validation, and Testing 603

In real life situation, we have noise (e.g. measurement noise) in our data:
np.random.seed(0)
for _ in range(6):
noisy_X = X + np.random.normal(loc=0, scale=.1, size=X.shape)
plt.plot(noisy_X, y, 'o')
regr.fit(noisy_X, y)
As we can see, our linear model captures and amplifies the noise in the data. It displays a lot of variance.
We can use another linear estimator that uses regularization, the Ridge estimator. This estimator regularizes
the coefficients by shrinking them to zero, under the assumption that very high correlations are often spurious.
The alpha parameter controls the amount of shrinkage used.
regr = linear_model.Ridge(alpha=.1)
np.random.seed(0)
for _ in range(6):
plt.show()

As we can see, the estimator displays much less variance. However it systematically under-estimates the coef-
ficient. It displays a biased behavior.
This is a typical example of bias/variance tradeof: non-regularized estimator are not biased, but they can
display a lot of bias. Highly-regularized models have little variance, but high bias. This bias is not necessarily
a bad thing: what matters is choosing the tradeoff between bias and variance that leads to the best prediction
performance. For a specific dataset there is a sweet spot corresponding to the highest complexity that the data
can support, depending on the amount of noise and of observations available.
20.8.2 Visualizing the Bias/Variance Tradeoff
Tip: Given a particular dataset and a model (e.g. a polynomial), wed like to understand whether bias (underfit)
or variance limits prediction, and how to tune the hyperparameter (here d, the degree of the polynomial) to give
the best fit.
On a given data, let us fit a simple polynomial regression model with varying degrees:
Tip: In the above figure, we see fits for three different values of d. For d = 1, the data is under-fit. This means
that the model is too simplistic: no straight line will ever be a good fit to this data. In this case, we say that the
model suffers from high bias. The model itself is biased, and this will be reflected in the fact that the data is
poorly fit. At the other extreme, for d = 6 the data is over-fit. This means that the model has too many free
parameters (6 in this case) which can be adjusted to perfectly fit the training data. If we add a new point to this
plot, though, chances are it will be very far from the curve representing the degree-6 fit. In this case, we say
that the model suffers from high variance. The reason for the term high variance is that if any of the input
points are varied slightly, it could result in a very different model.
In the middle, for d = 2, we have found a good mid-point. It fits the data fairly well, and does not suffer from
the bias and variance problems seen in the figures on either side. What we would like is a way to quantitatively

identify bias and variance, and optimize the metaparameters (in this case, the polynomial degree d) in order
to determine the best algorithm.
Polynomial regression with scikit-learn
A polynomial regression is built by pipelining PolynomialFeatures and a LinearRegression:

>>> from sklearn.pipeline import make_pipeline
>>> from sklearn.preprocessing import PolynomialFeatures
>>> model = make_pipeline(PolynomialFeatures(degree=2), LinearRegression())
Validation Curves
Let us create a dataset like in the example above:
>>> def generating_func(x, err=0.5):

... return np.random.normal(10 - 1. / (x + 0.1), err)
>>> # randomly sample more data

>>> x = np.random.random(size=200)
>>> y = generating_func(x, err=1.)
Central to quantify bias and variance of a

model is to apply it on test data, sampled from the same distribution as the train, but that will capture in-
dependent noise:
>>> xtrain, xtest, ytrain, ytest = train_test_split(x, y, test_size=0.4)
Validation curve A validation curve consists in varying a model parameter that controls its complexity (here
the degree of the polynomial) and measures both error of the model on training data, and on test data (eg with
cross-validation). The model parameter is then adjusted so that the test error is minimized:
We use sklearn.model_selection.validation_curve() to compute train and test error, and plot it:
>>> from sklearn.model_selection import validation_curve
>>> degrees = np.arange(1, 21)
>>> model = make_pipeline(PolynomialFeatures(), LinearRegression())
>>> # Vary the "degrees" on the pipeline step "polynomialfeatures"

>>> train_scores, validation_scores = validation_curve(

... model, x[:, np.newaxis], y,

... param_name='polynomialfeatures__degree',
... param_range=degrees)
>>> # Plot the mean train score and validation score across folds
>>> plt.plot(degrees, validation_scores.mean(axis=1), label='cross-validation')
>>> plt.plot(degrees, train_scores.mean(axis=1), label='training')
>>> plt.legend(loc='best')
This figure shows why validation is important.

On the left side of the plot, we have very low-degree polynomial, which under-fit the data. This leads to a low
explained variance for both the training set and the validation set. On the far right side of the plot, we have a
very high degree polynomial, which over-fits the data. This can be seen in the fact that the training explained
variance is very high, while on the validation set, it is low. Choosing d around 4 or 5 gets us the best tradeoff.
Tip: The astute reader will realize that something is amiss here: in the above plot, d = 4 gives the best results.
But in the previous plot, we found that d = 6 vastly over-fits the data. Whats going on here? The difference
is the number of training points used. In the previous example, there were only eight training points. In this
example, we have 100. As a general rule of thumb, the more training points used, the more complicated model
can be used. But how can you determine for a given model whether more training points will be helpful? A
useful diagnostic for this are learning curves.
Learning Curves
A learning curve shows the training and validation score as a function of the number of training points. Note
that when we train on a subset of the training data, the training score is computed using this subset, not the
full training set. This curve gives a quantitative view into how beneficial it will be to add training samples.
Questions:
As the number of training samples are increased, what do you expect to see for the training score? For
the validation score?
Would you expect the training score to be higher or lower than the validation score? Would you ever
expect this to change?
scikit-learn provides sklearn.model_selection.learning_curve():

>>> from sklearn.model_selection import learning_curve

>>> train_sizes, train_scores, validation_scores = learning_curve(
... model, x[:, np.newaxis], y, train_sizes=np.logspace(-1, 0, 20))
>>> # Plot the mean train score and validation score across folds
>>> plt.plot(train_sizes, validation_scores.mean(axis=1), label='cross-validation')
>>> plt.plot(train_sizes, train_scores.mean(axis=1), label='training')
Fig. 20.5: For a degree=1 model
Note that the validation score generally increases with a growing training set, while the training score generally
decreases with a growing training set. As the training size increases, they will converge to a single value.
From the above discussion, we know that d = 1 is a high-bias estimator which under-fits the data. This is
indicated by the fact that both the training and validation scores are low. When confronted with this type of
learning curve, we can expect that adding more training data will not help: both lines converge to a relatively
low score.
When the learning curves have converged to a low score, we have a high bias model.
A high-bias model can be improved by:
Using a more sophisticated model (i.e. in this case, increase d)
Gather more features for each sample.
Decrease regularization in a regularized model.
Increasing the number of samples, however, does not improve a high-bias model.
Now lets look at a high-variance (i.e. over-fit) model:

Fig. 20.6: For a degree=15 model
Here we show the learning curve for d = 15. From the above discussion, we know that d = 15 is a high-
variance estimator which over-fits the data. This is indicated by the fact that the training score is much higher
than the validation score. As we add more samples to this training set, the training score will continue to
decrease, while the cross-validation error will continue to increase, until they meet in the middle.
Learning curves that have not yet converged with the full training set indicate a high-variance, over-fit
model.
A high-variance model can be improved by:
Gathering more training samples.
Using a less-sophisticated model (i.e. in this case, make d smaller)
Increasing regularization.
In particular, gathering more features for each sample will not help the results.
20.8.3 Summary on model selection
Weve seen above that an under-performing algorithm can be due to two possible situations: high bias (under-
fitting) and high variance (over-fitting). In order to evaluate our algorithm, we set aside a portion of our training
data for cross-validation. Using the technique of learning curves, we can train on progressively larger subsets
of the data, evaluating the training error and cross-validation error to determine whether our algorithm has
high variance or high bias. But what do we do with this information?
High Bias
If a model shows high bias, the following actions might help:

Add more features. In our example of predicting home prices, it may be helpful to make use of infor-
mation such as the neighborhood the house is in, the year the house was built, the size of the lot, etc.
Adding these features to the training and test sets can improve a high-bias estimator
Use a more sophisticated model. Adding complexity to the model can help improve on bias. For a
polynomial fit, this can be accomplished by increasing the degree d. Each learning technique has its
own methods of adding complexity.
Use fewer samples. Though this will not improve the classification, a high-bias algorithm can attain
nearly the same error with a smaller training sample. For algorithms which are computationally expen-
sive, reducing the training sample size can lead to very large improvements in speed.

Decrease regularization. Regularization is a technique used to impose simplicity in some machine

learning models, by adding a penalty term that depends on the characteristics of the parameters. If a
model has high bias, decreasing the effect of regularization can lead to better results.
High Variance
If a model shows high variance, the following actions might help:

Use fewer features. Using a feature selection technique may be useful, and decrease the over-fitting of
the estimator.
Use a simpler model. Model complexity and over-fitting go hand-in-hand.
Use more training samples. Adding training samples can reduce the effect of over-fitting, and lead to
improvements in a high variance estimator.
Increase Regularization. Regularization is designed to prevent over-fitting. In a high-variance model,
increasing regularization can lead to better results.
These choices become very important in real-world situations. For example, due to limited telescope time,
astronomers must seek a balance between observing a large number of objects, and observing a large number
of features for each object. Determining which is more important for a particular learning task can inform the
observing strategy that the astronomer employs.
20.8.4 A last word of caution: separate validation and test set
Using validation schemes to determine hyper-parameters means that we are fitting the hyper-parameters to
the particular validation set. In the same way that parameters can be over-fit to the training set, hyperpa-
rameters can be over-fit to the validation set. Because of this, the validation error tends to under-predict the
classification error of new data.
For this reason, it is recommended to split the data into three sets:
The training set, used to train the model (usually ~60% of the data)
The validation set, used to validate the model (usually ~20% of the data)
The test set, used to evaluate the expected error of the validated model (usually ~20% of the data)
Many machine learning practitioners do not separate test set and validation set. But if your goal is to gauge
the error of a model on unknown data, using an independent test set is vital.
20.9 Examples for the scikit-learn chapter
20.9.1 Measuring Decision Tree performance
Demonstrates overfit when testing on train set.

Get the data
from sklearn.datasets import load_boston

data = load_boston()
Train and test a model
20.9. Examples for the scikit-learn chapter 610

from sklearn.tree import DecisionTreeRegressor

clf = DecisionTreeRegressor().fit(data.data, data.target)
predicted = clf.predict(data.data)
expected = data.target
Plot predicted as a function of expected

plt.scatter(expected, predicted)
plt.plot([0, 50], [0, 50], '--k')
plt.axis('tight')
plt.xlabel('True price ($1000s)')
plt.ylabel('Predicted price ($1000s)')
plt.tight_layout()
Pretty much no errors!

This is too good to be true: we are testing the model on the train data, which is not a mesure of generalization.
The results are not valid
Download Python source code: plot_measuring_performance.py
Download Jupyter notebook: plot_measuring_performance.ipynb
20.9.2 Demo PCA in 2D
Load the iris data

X = iris.data
y = iris.target
Fit a PCA

from sklearn.decomposition import PCA

pca = PCA(n_components=2, whiten=True)
pca.fit(X)
Project the data in 2D
X_pca = pca.transform(X)
Visualize the data
target_ids = range(len(iris.target_names))

for i, c, label in zip(target_ids, 'rgbcmykw', iris.target_names):
plt.scatter(X_pca[y == i, 0], X_pca[y == i, 1],
c=c, label=label)
plt.legend()
plt.show()

Download Python source code: plot_pca.py
Download Jupyter notebook: plot_pca.ipynb

20.9.3 A simple linear regression
import numpy as np
from sklearn.linear_model import LinearRegression
# x from 0 to 30
x = 30 * np.random.random((20, 1))
# y = a*x + b with noise

y = 0.5 * x + 1.0 + np.random.normal(size=x.shape)
# create a linear regression model

model = LinearRegression()
model.fit(x, y)
# predict y from the data

x_new = np.linspace(0, 30, 100)
y_new = model.predict(x_new[:, np.newaxis])
# plot the results

ax = plt.axes()
ax.scatter(x, y)
ax.plot(x_new, y_new)
ax.set_xlabel('x')
ax.set_ylabel('y')
ax.axis('tight')
plt.show()

Download Python source code: plot_linear_regression.py
Download Jupyter notebook: plot_linear_regression.ipynb

20.9.4 Plot 2D views of the iris dataset
Plot a simple scatter plot of 2 features of the iris dataset.

Note that more elaborate visualization of this dataset is detailed in the Statistics in Python chapter.
# Load the data

from sklearn.datasets import load_iris
iris = load_iris()
# The indices of the features that we are plotting

x_index = 0
y_index = 1
# this formatter will label the colorbar with the correct target names
formatter = plt.FuncFormatter(lambda i, *args: iris.target_names[int(i)])
plt.scatter(iris.data[:, x_index], iris.data[:, y_index], c=iris.target)
plt.colorbar(ticks=[0, 1, 2], format=formatter)
plt.xlabel(iris.feature_names[x_index])
plt.ylabel(iris.feature_names[y_index])
plt.tight_layout()
plt.show()

Download Python source code: plot_iris_scatter.py
Download Jupyter notebook: plot_iris_scatter.ipynb

20.9.5 tSNE to visualize digits
Here we use sklearn.manifold.TSNE to visualize the digits datasets. Indeed, the digits are vectors in a 8*8
= 64 dimensional space. We want to project them in 2D for visualization. tSNE is often a good solution, as it
groups and separates data points based on their local relationship.
Load the iris data

digits = datasets.load_digits()
# Take the first 500 data points: it's hard to see 1500 points
X = digits.data[:500]
y = digits.target[:500]
Fit and transform with a TSNE
from sklearn.manifold import TSNE

tsne = TSNE(n_components=2, random_state=0)
Project the data in 2D
X_2d = tsne.fit_transform(X)
Visualize the data
target_ids = range(len(digits.target_names))

colors = 'r', 'g', 'b', 'c', 'm', 'y', 'k', 'w', 'orange', 'purple'
for i, c, label in zip(target_ids, colors, digits.target_names):
plt.scatter(X_2d[y == i, 0], X_2d[y == i, 1], c=c, label=label)
plt.legend()
plt.show()


Download Python source code: plot_tsne.py
Download Jupyter notebook: plot_tsne.ipynb
20.9.6 Use the RidgeCV and LassoCV to set the regularization parameter
Load the diabetes dataset
from sklearn.datasets import load_diabetes

data = load_diabetes()
X, y = data.data, data.target
print(X.shape)
Out:
(442, 10)
Compute the cross-validation score with the default hyper-parameters
from sklearn.model_selection import cross_val_score

from sklearn.linear_model import Ridge, Lasso
for Model in [Ridge, Lasso]:

model = Model()

print('%s : %s ' % (Model.__name__,

cross_val_score(model, X, y).mean()))
Out:
Ridge: 0.409427438303
Lasso: 0.353800083299
We compute the cross-validation score as a function of alpha, the strength of the regularization for Lasso and
Ridge
import numpy as np
alphas = np.logspace(-3, -1, 30)
for Model in [Lasso, Ridge]:

scores = [cross_val_score(Model(alpha), X, y, cv=3).mean()
for alpha in alphas]
plt.plot(alphas, scores, label=Model.__name__)
plt.legend(loc='lower left')
plt.xlabel('alpha')
plt.ylabel('cross validation score')
plt.tight_layout()
plt.show()

Download Python source code: plot_linear_model_cv.py
Download Jupyter notebook: plot_linear_model_cv.ipynb
20.9.7 Plot variance and regularization in linear models
import numpy as np

# Smaller figures
plt.rcParams['figure.figsize'] = (3, 2)
We consider the situation where we have only 2 data point
X = np.c_[ .5, 1].T

y = [.5, 1]
X_test = np.c_[ 0, 2].T
Without noise, as linear regression fits the data perfectly

regr.fit(X, y)
plt.plot(X, y, 'o')
In real life situation, we have noise (e.g. measurement noise) in our data:
np.random.seed(0)
for _ in range(6):
As we can see, our linear model captures and amplifies the noise in the data. It displays a lot of variance.
We can use another linear estimator that uses regularization, the Ridge estimator. This estimator regularizes
the coefficients by shrinking them to zero, under the assumption that very high correlations are often spurious.
The alpha parameter controls the amount of shrinkage used.
regr = linear_model.Ridge(alpha=.1)
np.random.seed(0)

for _ in range(6):
plt.show()

Download Python source code: plot_variance_linear_regr.py
Download Jupyter notebook: plot_variance_linear_regr.ipynb
20.9.8 Simple picture of the formal problem of machine learning
This example generates simple synthetic data ploints and shows a separating hyperplane on them.
import numpy as np
from sklearn.linear_model import SGDClassifier
from sklearn.datasets.samples_generator import make_blobs
# we create 50 separable synthetic points

X, Y = make_blobs(n_samples=50, centers=2,
random_state=0, cluster_std=0.60)

# fit the model

clf = SGDClassifier(loss="hinge", alpha=0.01,
n_iter=200, fit_intercept=True)
clf.fit(X, Y)
# plot the line, the points, and the nearest vectors to the plane
xx = np.linspace(-1, 5, 10)
yy = np.linspace(-1, 5, 10)
X1, X2 = np.meshgrid(xx, yy)

Z = np.empty(X1.shape)
for (i, j), val in np.ndenumerate(X1):
x1 = val
x2 = X2[i, j]
p = clf.decision_function([[x1, x2]])
Z[i, j] = p[0]
ax = plt.axes()
ax.contour(X1, X2, Z, [-1.0, 0.0, 1.0], colors='k',
linestyles=['dashed', 'solid', 'dashed'])
ax.scatter(X[:, 0], X[:, 1], c=Y, cmap=plt.cm.Paired)
ax.axis('tight')
plt.show()

Download Python source code: plot_separator.py
Download Jupyter notebook: plot_separator.ipynb
20.9.9 Compare classifiers on the digits data
Compare the performance of a variety of classifiers on a test set for the digits data.
Out:
LinearSVC: 0.927190026916
GaussianNB: 0.833274168101
KNeighborsClassifier: 0.980456280495
------------------
LinearSVC(loss='l1'): 0.9325901988044822
LinearSVC(loss='l2'): 0.9165874249889226
-------------------

from sklearn import model_selection, datasets, metrics

from sklearn.svm import LinearSVC
from sklearn.naive_bayes import GaussianNB
from sklearn.neighbors import KNeighborsClassifier
digits = datasets.load_digits()
X = digits.data
y = digits.target
X_train, X_test, y_train, y_test = model_selection.train_test_split(X, y,
test_size=0.25, random_state=0)
for Model in [LinearSVC, GaussianNB, KNeighborsClassifier]:

clf = Model().fit(X_train, y_train)
print('%s : %s ' %
(Model.__name__, metrics.f1_score(y_test, y_pred, average="macro")))
print('------------------')
# test SVC loss

for loss in ['l1', 'l2']:
clf = LinearSVC(loss=loss).fit(X_train, y_train)
print("LinearSVC(loss='{0}'): {1}".format(loss,
metrics.f1_score(y_test, y_pred, average="macro")))
print('-------------------')
# test the number of neighbors

for n_neighbors in range(1, 11):
clf = KNeighborsClassifier(n_neighbors=n_neighbors).fit(X_train, y_train)
print("KNeighbors(n_neighbors={0}): {1}".format(n_neighbors,
metrics.f1_score(y_test, y_pred, average="macro")))

Download Python source code: plot_compare_classifiers.py
Download Jupyter notebook: plot_compare_classifiers.ipynb
20.9.10 Plot fitting a 9th order polynomial
Fits data generated from a 9th order polynomial with model of 4th order and 9th order polynomials, to demon-
strate that often simpler models are to be prefered
import numpy as np
from matplotlib.colors import ListedColormap
# Create color maps for 3-class classification problem, as with iris

cmap_light = ListedColormap(['#FFAAAA', '#AAFFAA', '#AAAAFF'])
cmap_bold = ListedColormap(['#FF0000', '#00FF00', '#0000FF'])

rng = np.random.RandomState(0)
x = 2*rng.rand(100) - 1
f = lambda t: 1.2 * t**2 + .1 * t**3 - .4 * t **5 - .5 * t ** 9

y = f(x) + .4 * rng.normal(size=100)
x_test = np.linspace(-1, 1, 100)
The data
plt.scatter(x, y, s=4)
Fitting 4th and 9th order polynomials

For this we need to engineer features: the n_th powers of x:
X = np.array([x**i for i in range(5)]).T

X_test = np.array([x_test**i for i in range(5)]).T
regr.fit(X, y)
plt.plot(x_test, regr.predict(X_test), label='4th order')
X = np.array([x**i for i in range(10)]).T

X_test = np.array([x_test**i for i in range(10)]).T
regr.fit(X, y)
plt.plot(x_test, regr.predict(X_test), label='9th order')
plt.axis('tight')
plt.title('Fitting a 4th and a 9th order polynomial')

Ground truth
plt.plot(x_test, f(x_test), label="truth")
plt.axis('tight')
plt.title('Ground truth (9th order polynomial)')
plt.show()


Download Python source code: plot_polynomial_regression.py
Download Jupyter notebook: plot_polynomial_regression.ipynb
20.9.11 A simple regression analysis on the Boston housing data
Here we perform a simple regression analysis on the Boston housing data, exploring two types of regressors.
from sklearn.datasets import load_boston

data = load_boston()
Print a histogram of the quantity to predict: price

plt.hist(data.target)
plt.xlabel('price ($1000s)')
plt.ylabel('count')
plt.tight_layout()
Print the join histogram for each feature
for index, feature_name in enumerate(data.feature_names):

plt.scatter(data.data[:, index], data.target)
plt.ylabel('Price', size=15)
plt.xlabel(feature_name, size=15)
plt.tight_layout()



Simple prediction

X_train, X_test, y_train, y_test = train_test_split(data.data, data.target)


clf = LinearRegression()
predicted = clf.predict(X_test)
expected = y_test
plt.plot([0, 50], [0, 50], '--k')
plt.axis('tight')
plt.tight_layout()
Prediction with gradient boosted tree
from sklearn.ensemble import GradientBoostingRegressor
clf = GradientBoostingRegressor()
expected = y_test
plt.plot([0, 50], [0, 50], '--k')
plt.axis('tight')
plt.tight_layout()

Print the error rate
import numpy as np
print("RMS: %r " % np.sqrt(np.mean((predicted - expected) ** 2)))
plt.show()
Out:
RMS: 3.3940265988986305

Download Python source code: plot_boston_prediction.py
Download Jupyter notebook: plot_boston_prediction.ipynb
20.9.12 Nearest-neighbor prediction on iris
Plot the decision boundary of nearest neighbor decision on iris, first with a single nearest neighbor, and then
using 3 nearest neighbors.
import numpy as np
from sklearn import neighbors, datasets
from matplotlib.colors import ListedColormap
# Create color maps for 3-class classification problem, as with iris

cmap_light = ListedColormap(['#FFAAAA', '#AAFFAA', '#AAAAFF'])
cmap_bold = ListedColormap(['#FF0000', '#00FF00', '#0000FF'])
X = iris.data[:, :2] # we only take the first two features. We could
# avoid this ugly slicing by using a two-dim dataset
y = iris.target
knn.fit(X, y)
x_min, x_max = X[:, 0].min() - .1, X[:, 0].max() + .1

y_min, y_max = X[:, 1].min() - .1, X[:, 1].max() + .1

xx, yy = np.meshgrid(np.linspace(x_min, x_max, 100),
np.linspace(y_min, y_max, 100))
Z = knn.predict(np.c_[xx.ravel(), yy.ravel()])
Put the result into a color plot
Z = Z.reshape(xx.shape)
plt.figure()
plt.pcolormesh(xx, yy, Z, cmap=cmap_light)
# Plot also the training points

plt.scatter(X[:, 0], X[:, 1], c=y, cmap=cmap_bold)
plt.xlabel('sepal length (cm)')
plt.ylabel('sepal width (cm)')
plt.axis('tight')
And now, redo the analysis with 3 neighbors
knn.fit(X, y)
Z = knn.predict(np.c_[xx.ravel(), yy.ravel()])
# Put the result into a color plot

Z = Z.reshape(xx.shape)
plt.figure()
plt.pcolormesh(xx, yy, Z, cmap=cmap_light)
# Plot also the training points

plt.scatter(X[:, 0], X[:, 1], c=y, cmap=cmap_bold)

plt.xlabel('sepal length (cm)')

plt.ylabel('sepal width (cm)')
plt.axis('tight')
plt.show()

Download Python source code: plot_iris_knn.py
Download Jupyter notebook: plot_iris_knn.ipynb
20.9.13 Simple visualization and classification of the digits dataset
Plot the first few samples of the digits dataset and a 2D representation built using PCA, then do a simple clas-
sification
from sklearn.datasets import load_digits

digits = load_digits()
Plot the data: images of digits
Each data in a 8x8 image

fig = plt.figure(figsize=(6, 6)) # figure size in inches
fig.subplots_adjust(left=0, right=1, bottom=0, top=1, hspace=0.05, wspace=0.05)

for i in range(64):
ax.imshow(digits.images[i], cmap=plt.cm.binary, interpolation='nearest')
# label the image with the target value
ax.text(0, 7, str(digits.target[i]))
Plot a projection on the 2 first principal axis
plt.figure()
from sklearn.decomposition import PCA

pca = PCA(n_components=2)
proj = pca.fit_transform(digits.data)
plt.scatter(proj[:, 0], proj[:, 1], c=digits.target)
plt.colorbar()

Classify with Gaussian naive Bayes
from sklearn.naive_bayes import GaussianNB

# split the data into training and validation sets

X_train, X_test, y_train, y_test = train_test_split(digits.data, digits.target)
# train the model

clf = GaussianNB()
# use the model to predict the labels of the test data

expected = y_test
# Plot the prediction

fig = plt.figure(figsize=(6, 6)) # figure size in inches
fig.subplots_adjust(left=0, right=1, bottom=0, top=1, hspace=0.05, wspace=0.05)
# plot the digits: each image is 8x8 pixels

for i in range(64):
ax.imshow(X_test.reshape(-1, 8, 8)[i], cmap=plt.cm.binary,
# label the image with the target value

if predicted[i] == expected[i]:
ax.text(0, 7, str(predicted[i]), color='green')
else:

ax.text(0, 7, str(predicted[i]), color='red')
Quantify the performance
First print the number of correct matches
matches = (predicted == expected)

print(matches.sum())
Out:
384
The total number of data points
print(len(matches))
Out:
450

And now, the ration of correct predictions
matches.sum() / float(len(matches))
Print the classification report

print(metrics.classification_report(expected, predicted))
Out:
0 1.00 0.93 0.96 40

1 0.82 0.85 0.84 48
2 0.91 0.69 0.78 45
3 0.96 0.83 0.89 58
4 0.91 0.89 0.90 36
5 0.91 0.85 0.88 48
6 0.93 0.98 0.95 43
7 0.75 0.98 0.85 47
8 0.59 0.81 0.68 42
9 0.91 0.74 0.82 43
avg / total 0.87 0.85 0.86 450
Print the confusion matrix
print(metrics.confusion_matrix(expected, predicted))
plt.show()
Out:
[[37 0 0 0 2 0 0 1 0 0]
[ 0 41 0 0 0 0 1 0 4 2]
[ 0 2 31 0 0 0 1 0 11 0]
[ 0 0 2 48 0 1 0 2 4 1]
[ 0 1 1 0 32 0 0 1 1 0]
[ 0 1 0 1 1 41 0 4 0 0]
[ 0 0 0 0 0 1 42 0 0 0]
[ 0 0 0 0 0 1 0 46 0 0]
[ 0 4 0 0 0 1 0 3 34 0]
[ 0 1 0 1 0 0 1 4 4 32]]

Download Python source code: plot_digits_simple_classif.py
Download Jupyter notebook: plot_digits_simple_classif.ipynb
20.9.14 The eigenfaces example: chaining PCA and SVMs
The goal of this example is to show how an unsupervised method and a supervised one can be chained for
better prediction. It starts with a didactic but lengthy way of doing things, and finishes with the idiomatic
approach to pipelining in scikit-learn.
Here well take a look at a simple facial recognition example. Ideally, we would use a dataset consisting of a
subset of the Labeled Faces in the Wild data that is available with sklearn.datasets.fetch_lfw_people().
However, this is a relatively large download (~200MB) so we will do the tutorial on a simpler, less rich dataset.
Feel free to explore the LFW dataset.


faces = datasets.fetch_olivetti_faces()
faces.data.shape
Lets visualize these faces to see what were working with

# plot several images
for i in range(15):
ax.imshow(faces.images[i], cmap=plt.cm.bone)
Tip: Note is that these faces have already been localized and scaled to a common size. This is an important
preprocessing piece for facial recognition, and is a process that can require a large collection of training data.
This can be done in scikit-learn, but the challenge is gathering a sufficient amount of training data for the
algorithm to work. Fortunately, this piece is common enough that it has been done. One good resource is
OpenCV, the Open Computer Vision Library.
Well perform a Support Vector classification of the images. Well do a typical train-test split on the images:

X_train, X_test, y_train, y_test = train_test_split(faces.data,
faces.target, random_state=0)
print(X_train.shape, X_test.shape)
Out:

(300, 4096) (100, 4096)
Preprocessing: Principal Component Analysis
1850 dimensions is a lot for SVM. We can use PCA to reduce these 1850 features to a manageable size, while
maintaining most of the information in the dataset.
from sklearn import decomposition

pca = decomposition.PCA(n_components=150, whiten=True)
pca.fit(X_train)
One interesting part of PCA is that it computes the mean face, which can be interesting to examine:
plt.imshow(pca.mean_.reshape(faces.images[0].shape),
cmap=plt.cm.bone)
The principal components measure deviations about this mean along orthogonal axes.
print(pca.components_.shape)
Out:
(150, 4096)
It is also interesting to visualize these principal components:

for i in range(30):

ax.imshow(pca.components_[i].reshape(faces.images[0].shape),
cmap=plt.cm.bone)
The components (eigenfaces) are ordered by their importance from top-left to bottom-right. We see that the
first few components seem to primarily take care of lighting conditions; the remaining components pull out
certain identifying features: the nose, eyes, eyebrows, etc.
With this projection computed, we can now project our original training and test data onto the PCA basis:
X_train_pca = pca.transform(X_train)
X_test_pca = pca.transform(X_test)
print(X_train_pca.shape)
Out:
(300, 150)
print(X_test_pca.shape)
Out:
(100, 150)
These projected components correspond to factors in a linear combination of component images such that
the combination approaches the original face.
Doing the Learning: Support Vector Machines
Now well perform support-vector-machine classification on this reduced dataset:

clf = svm.SVC(C=5., gamma=0.001)
clf.fit(X_train_pca, y_train)
Finally, we can evaluate how well this classification did. First, we might plot a few of the test-cases with the
labels learned from the training set:
import numpy as np
for i in range(15):
ax.imshow(X_test[i].reshape(faces.images[0].shape),
cmap=plt.cm.bone)
y_pred = clf.predict(X_test_pca[i, np.newaxis])[0]
color = ('black' if y_pred == y_test[i] else 'red')

ax.set_title(faces.target[y_pred],
fontsize='small', color=color)
The classifier is correct on an impressive number of images given the simplicity of its learning model! Using
a linear classifier on 150 features derived from the pixel-level data, the algorithm correctly identifies a large
number of the people in the images.
Again, we can quantify this effectiveness using one of several measures from sklearn.metrics. First we can
do the classification report, which shows the precision, recall and other measures of the goodness of the
classification:

y_pred = clf.predict(X_test_pca)
print(metrics.classification_report(y_test, y_pred))
Out:
0 1.00 0.67 0.80 6

1 1.00 1.00 1.00 4
2 0.50 1.00 0.67 2
3 1.00 1.00 1.00 1
4 1.00 1.00 1.00 1
5 1.00 1.00 1.00 5
6 1.00 1.00 1.00 4
7 1.00 0.67 0.80 3
9 1.00 1.00 1.00 1
10 1.00 1.00 1.00 4
11 1.00 1.00 1.00 1
12 0.67 1.00 0.80 2

13 1.00 1.00 1.00 3

14 1.00 1.00 1.00 5
15 1.00 1.00 1.00 3
17 1.00 1.00 1.00 6
19 1.00 1.00 1.00 4
20 1.00 1.00 1.00 1
21 1.00 1.00 1.00 1
22 1.00 1.00 1.00 2
23 0.50 1.00 0.67 1
24 1.00 1.00 1.00 2
25 1.00 0.50 0.67 2
26 1.00 0.75 0.86 4
27 1.00 1.00 1.00 1
28 0.67 1.00 0.80 2
29 1.00 1.00 1.00 3
30 1.00 1.00 1.00 4
31 1.00 1.00 1.00 3
32 1.00 1.00 1.00 3
33 1.00 1.00 1.00 2
34 1.00 1.00 1.00 3
35 1.00 1.00 1.00 1
36 1.00 1.00 1.00 3
37 1.00 1.00 1.00 3
38 1.00 1.00 1.00 1
39 1.00 1.00 1.00 3
avg / total 0.97 0.95 0.95 100
Another interesting metric is the confusion matrix, which indicates how often any two items are mixed-up.
The confusion matrix of a perfect classifier would only have nonzero entries on the diagonal, with zeros on the
off-diagonal:
print(metrics.confusion_matrix(y_test, y_pred))
Out:
[[4 0 0 ..., 0 0 0]
[0 4 0 ..., 0 0 0]
[0 0 2 ..., 0 0 0]
...,
[0 0 0 ..., 3 0 0]
[0 0 0 ..., 0 1 0]
[0 0 0 ..., 0 0 3]]
Pipelining
Above we used PCA as a pre-processing step before applying our support vector machine classifier. Plugging
the output of one estimator directly into the input of a second estimator is a commonly used pattern; for this
reason scikit-learn provides a Pipeline object which automates this process. The above problem can be re-
expressed as a pipeline as follows:
from sklearn.pipeline import Pipeline

clf = Pipeline([('pca', decomposition.PCA(n_components=150, whiten=True)),
('svm', svm.LinearSVC(C=1.0))])
print(metrics.confusion_matrix(y_pred, y_test))
plt.show()

Out:
[[5 0 0 ..., 0 0 0]
[0 4 0 ..., 0 0 0]
[0 0 1 ..., 0 0 0]
...,
[0 0 0 ..., 3 0 0]
[0 0 0 ..., 0 1 0]
[0 0 0 ..., 0 0 3]]
A Note on Facial Recognition
Here we have used PCA eigenfaces as a pre-processing step for facial recognition. The reason we chose this
is because PCA is a broadly-applicable technique, which can be useful for a wide array of data types. Research
in the field of facial recognition in particular, however, has shown that other more specific feature extraction
methods are can be much more effective.
Download Python source code: plot_eigenfaces.py
Download Jupyter notebook: plot_eigenfaces.ipynb
20.9.15 Example of linear and non-linear models
This is an example plot from the tutorial which accompanies an explanation of the support vector machine
GUI.
import numpy as np
data that is linearly separable
def linear_model(rseed=42, n_samples=30):

" Generate data according to a linear model"
np.random.seed(rseed)
data = np.random.normal(0, 10, (n_samples, 2))

data[:n_samples // 2] -= 15
data[n_samples // 2:] += 15
labels = np.ones(n_samples)
labels[:n_samples // 2] = -1
return data, labels
X, y = linear_model()
clf = svm.SVC(kernel='linear')
clf.fit(X, y)
ax = plt.subplot(111, xticks=[], yticks=[])
ax.scatter(X[:, 0], X[:, 1], c=y, cmap=plt.cm.bone)
ax.scatter(clf.support_vectors_[:, 0],
clf.support_vectors_[:, 1],

s=80, edgecolors="k", facecolors="none")
delta = 1
y_min, y_max = -50, 50
x_min, x_max = -50, 50
x = np.arange(x_min, x_max + delta, delta)
y = np.arange(y_min, y_max + delta, delta)
X1, X2 = np.meshgrid(x, y)
Z = clf.decision_function(np.c_[X1.ravel(), X2.ravel()])
Z = Z.reshape(X1.shape)

linestyles=['dashed', 'solid', 'dashed'])
data with a non-linear separation
def nonlinear_model(rseed=42, n_samples=30):

radius = 40 * np.random.random(n_samples)
far_pts = radius > 20
radius[far_pts] *= 1.2
radius[~far_pts] *= 1.1
theta = np.random.random(n_samples) * np.pi * 2
data = np.empty((n_samples, 2))

data[:, 0] = radius * np.cos(theta)
data[:, 1] = radius * np.sin(theta)
labels = np.ones(n_samples)
labels[far_pts] = -1
return data, labels
X, y = nonlinear_model()
clf = svm.SVC(kernel='rbf', gamma=0.001, coef0=0, degree=3)

clf.fit(X, y)
ax = plt.subplot(1, 1, 1, xticks=[], yticks=[])
ax.scatter(X[:, 0], X[:, 1], c=y, cmap=plt.cm.bone, zorder=2)
ax.scatter(clf.support_vectors_[:, 0], clf.support_vectors_[:, 1],

s=80, edgecolors="k", facecolors="none")
delta = 1
y_min, y_max = -50, 50
x_min, x_max = -50, 50
x = np.arange(x_min, x_max + delta, delta)
y = np.arange(y_min, y_max + delta, delta)
X1, X2 = np.meshgrid(x, y)
Z = clf.decision_function(np.c_[X1.ravel(), X2.ravel()])
Z = Z.reshape(X1.shape)

linestyles=['dashed', 'solid', 'dashed'], zorder=1)
plt.show()

Download Python source code: plot_svm_non_linear.py
Download Jupyter notebook: plot_svm_non_linear.ipynb
20.9.16 Bias and variance of polynomial fit
Demo overfitting, underfitting, and validation and learning curves with polynomial regression.

Fit polynomes of different degrees to a dataset: for too small a degree, the model underfits, while for too large
a degree, it overfits.
import numpy as np
def generating_func(x, err=0.5):

return np.random.normal(10 - 1. / (x + 0.1), err)
A polynomial regression
from sklearn.pipeline import make_pipeline

from sklearn.preprocessing import PolynomialFeatures
A simple figure to illustrate the problem
n_samples = 8
np.random.seed(0)
x = 10 ** np.linspace(-2, 0, n_samples)
y = generating_func(x)
x_test = np.linspace(-0.2, 1.2, 1000)
titles = ['d = 1 (under-fit; high bias)',

'd = 2',
'd = 6 (over-fit; high variance)']
degrees = [1, 2, 6]
fig = plt.figure(figsize=(9, 3.5))

fig.subplots_adjust(left=0.06, right=0.98, bottom=0.15, top=0.85, wspace=0.05)
for i, d in enumerate(degrees):
ax = fig.add_subplot(131 + i, xticks=[], yticks=[])
ax.scatter(x, y, marker='x', c='k', s=50)
model = make_pipeline(PolynomialFeatures(d), LinearRegression())

model.fit(x[:, np.newaxis], y)
ax.plot(x_test, model.predict(x_test[:, np.newaxis]), '-b')
ax.set_xlim(-0.2, 1.2)
ax.set_ylim(0, 12)
ax.set_xlabel('house size')
if i == 0:
ax.set_ylabel('price')
ax.set_title(titles[i])

Generate a larger dataset
n_samples = 200
test_size = 0.4
error = 1.0
# randomly sample the data

np.random.seed(1)
x = np.random.random(n_samples)
y = generating_func(x, error)
# split into training, validation, and testing sets.

x_train, x_test, y_train, y_test = train_test_split(x, y, test_size=test_size)
# show the training and validation sets

plt.scatter(x_train, y_train, color='red', label='Training set')
plt.scatter(x_test, y_test, color='blue', label='Test set')
plt.title('The data')

Plot a validation curve
from sklearn.model_selection import validation_curve
degrees = np.arange(1, 21)
model = make_pipeline(PolynomialFeatures(), LinearRegression())
# The parameter to vary is the "degrees" on the pipeline step

# "polynomialfeatures"
train_scores, validation_scores = validation_curve(
model, x[:, np.newaxis], y,
param_name='polynomialfeatures__degree',
param_range=degrees)
# Plot the mean train error and validation error across folds
plt.plot(degrees, validation_scores.mean(axis=1), lw=2,
label='cross-validation')
plt.plot(degrees, train_scores.mean(axis=1), lw=2, label='training')
plt.xlabel('degree of fit')
plt.ylabel('explained variance')
plt.title('Validation curve')
plt.tight_layout()

Learning curves
Plot train and test error with an increasing number of samples
# A learning curve for d=1, 5, 15

for d in [1, 5, 15]:
model = make_pipeline(PolynomialFeatures(degree=d), LinearRegression())
from sklearn.model_selection import learning_curve

train_sizes, train_scores, validation_scores = learning_curve(
model, x[:, np.newaxis], y,
train_sizes=np.logspace(-1, 0, 20))
# Plot the mean train error and validation error across folds
plt.plot(train_sizes, validation_scores.mean(axis=1),
lw=2, label='cross-validation')
plt.plot(train_sizes, train_scores.mean(axis=1),
lw=2, label='training')
plt.ylim(ymin=-.1, ymax=1)
plt.xlabel('number of train samples')
plt.ylabel('explained variance')
plt.title('Learning curve (degree=%i )' % d)
plt.tight_layout()
plt.show()


Download Python source code: plot_bias_variance.py
Download Jupyter notebook: plot_bias_variance.ipynb
20.9.17 Tutorial Diagrams
This script plots the flow-charts used in the scikit-learn tutorials.

import numpy as np
import pylab as pl
from matplotlib.patches import Circle, Rectangle, Polygon, Arrow, FancyArrow
def create_base(box_bg = '#CCCCCC',

arrow1 = '#88CCFF',
arrow2 = '#88FF88',
supervised=True):
fig = pl.figure(figsize=(9, 6), facecolor='w')
ax = pl.axes((0, 0, 1, 1),
xticks=[], yticks=[], frameon=False)
ax.set_xlim(0, 9)
ax.set_ylim(0, 6)
patches = [Rectangle((0.3, 3.6), 1.5, 1.8, zorder=1, fc=box_bg),

Rectangle((0.5, 3.8), 1.5, 1.8, zorder=2, fc=box_bg),
Rectangle((2.9, 3.6), 0.2, 1.8, fc=box_bg),

Circle((5.5, 3.5), 1.0, fc=box_bg),
Polygon([[5.5, 1.7],
[6.1, 1.1],
[5.5, 0.5],
[4.9, 1.1]], fc=box_bg),
FancyArrow(2.3, 4.6, 0.35, 0, fc=arrow1,

width=0.25, head_width=0.5, head_length=0.2),
FancyArrow(3.75, 4.2, 0.5, -0.2, fc=arrow1,

FancyArrow(5.5, 2.4, 0, -0.4, fc=arrow1,




width=0.25, head_width=0.5, head_length=0.2)]
if supervised:
patches += [Rectangle((0.3, 2.4), 1.5, 0.5, zorder=1, fc=box_bg),
Rectangle((7.3, 0.85), 1.5, 0.5, fc=box_bg)]
else:
patches += [Rectangle((7.3, 0.2), 1.5, 1.8, fc=box_bg)]
for p in patches:
ax.add_patch(p)
pl.text(1.45, 4.9, "Training\nText,\nDocuments,\nImages,\netc.",

ha='center', va='center', fontsize=14)
pl.text(3.6, 4.9, "Feature\nVectors",

ha='left', va='center', fontsize=14)
pl.text(5.5, 3.5, "Machine\nLearning\nAlgorithm",

pl.text(1.05, 1.1, "New Text,\nDocument,\nImage,\netc.",

pl.text(3.3, 1.7, "Feature\nVector",

ha='left', va='center', fontsize=14)
pl.text(5.5, 1.1, "Predictive\nModel",

if supervised:
pl.text(1.45, 3.05, "Labels",
pl.text(8.05, 1.1, "Expected\nLabel",

pl.text(8.8, 5.8, "Supervised Learning Model",
ha='right', va='top', fontsize=18)
else:
pl.text(8.05, 1.1,
"Likelihood\nor Cluster ID\nor Better\nRepresentation",
pl.text(8.8, 5.8, "Unsupervised Learning Model",
ha='right', va='top', fontsize=18)
def plot_supervised_chart(annotate=False):
create_base(supervised=True)
if annotate:
fontdict = dict(color='r', weight='bold', size=14)
pl.text(1.9, 4.55, 'X = vec.fit_transform(input)',
fontdict=fontdict,
rotation=20, ha='left', va='bottom')
pl.text(3.7, 3.2, 'clf.fit(X, y)',
fontdict=fontdict,
pl.text(1.7, 1.5, 'X_new = vec.transform(input)',
fontdict=fontdict,
pl.text(6.1, 1.5, 'y_new = clf.predict(X_new)',
fontdict=fontdict,
def plot_unsupervised_chart():
create_base(supervised=False)
if __name__ == '__main__':
plot_supervised_chart(False)
plot_supervised_chart(True)
plot_unsupervised_chart()
pl.show()

Download Python source code: plot_ML_flow_chart.py

Download Jupyter notebook: plot_ML_flow_chart.ipynb

See also:
Going further
The documentation of scikit-learn is very complete and didactic.
Introduction to Machine Learning with Python, by Sarah Guido, Andreas Mller
(notebooks available here).

Index
D
diff, 508, 511
differentiation, 508
dsolve, 511
E
equations
algebraic, 510
differential, 511
I
integration, 509
M
Matrix, 511
P
Python Enhancement Proposals
PEP 255, 268
PEP 3118, 304
PEP 3129, 277
PEP 318, 270, 277
PEP 342, 268
PEP 343, 278
PEP 380, 269
PEP 380#id13, 269
PEP 8, 272
S
solve, 510
652

ScipyLectures Simple

Uploaded by

ScipyLectures Simple

Uploaded by

SciKits Numpy Matplotlib

Lecture Notes Emmanuelle Gouillart

Gal Varoquaux Emmanuelle Gouillart Olav Vahtras

I Getting started with Python for science 2

2 The Python language 12

3 NumPy: creating and manipulating numerical data 47

5 Scipy : high-level scientific computing 184

6 Getting help and finding documentation 259

II Advanced topics 262

8 Advanced NumPy 281

9 Debugging code 314

10 Optimizing code 325

11 Sparse Matrices in SciPy 332

12 Image manipulation and processing using Numpy and Scipy 352

13 Mathematical optimization: finding minima of functions 397

14 Interfacing with C 439

III Packages and applications 460

16 Sympy : Symbolic Mathematics in Python 505

17 Scikit-image: image processing 513

18 Traits: building interactive dialogs 538

19 3D plotting with Mayavi 557

20 scikit-learn: machine learning in Python 572

Getting started with Python for science

Authors: Fernando Perez, Emmanuelle Gouillart, Gal Varoquaux, Valentin Haenel

1.1 Why Python?

1.1.1 The scientists needs

Get data (simulation, experiment control),

1.1.2 Pythons strengths

1.1.3 How does Python compare to other solutions?

Compiled languages: C, C++, Fortran. . .

Matlab scripting language

Other scripting languages: Scilab, Octave, R, IDL, etc.

1.1. Why Python? 5

1.2 The Scientific Python ecosystem

Python, a generic and modern computing language

Core numeric libraries

1.2. The Scientific Python ecosystem 6

Advanced interactive environments:

1.2. The Scientific Python ecosystem 7

1.3 Before starting: Installing a working environment

1.4 The workflow: interactive environments and text editors

1.4.1 Interactive work

1.3. Before starting: Installing a working environment 8

Under the notebook

To execute code, press shift enter

In [1]: print('Hello world')

Getting help by using the ? operator after an object:

Prints the values to a stream, or to sys.stdout by default.

1.4.2 Elaboration of the work in an editor

In [1]: %run my_file.py

1.4. The workflow: interactive environments and text editors 9

From a script to functions

1.4.3 IPython and Jupyter Tips and Tricks

1.4. The workflow: interactive environments and text editors 10

The built-in cheat-sheet is accessible via the %quickref magic function.

1.4. The workflow: interactive environments and text editors 11

Authors: Chris Burns, Christophe Combelles, Emmanuelle Gouillart, Gal Varoquaux

Python for scientific computing

2.1 First steps

Start the Ipython shell (an enhanced interactive Python shell):

Once you have started the interpreter, type

>>> print("Hello, world!")

To get yourself started, type the following stack of instructions

2.1. First steps 13

2.2 Basic types

2.2.1 Numerical types

Tip: Python supports the following numerical, scalar types:

>>> a = 1.5 + 0.5j

2.2. Basic types 14

Type conversion (casting):

Warning: Integer division

To be safe: use floats:

Future behavior: to always get the behavior of Python3

Tip: If you explicitly want integer division use //:

Note: The behaviour of the division operator has changed in Python 3.

>>> colors = ['red', 'blue', 'green', 'black', 'white']