Download the Full Version of textbook for Fast Typing at textbookfull.

com

Advanced Data Analytics Using Python: With Machine

Learning, Deep Learning and NLP Examples
Mukhopadhyay

https://textbookfull.com/product/advanced-data-analytics-
using-python-with-machine-learning-deep-learning-and-nlp-
examples-mukhopadhyay/

OR CLICK BUTTON

DOWNLOAD NOW

Download More textbook Instantly Today - Get Yours Now at textbookfull.com

 Recommended digital products (PDF, EPUB, MOBI) that
you can download immediately if you are interested.

Deep Learning with Python Develop Deep Learning Models on

Theano and TensorFLow Using Keras Jason Brownlee

https://textbookfull.com/product/deep-learning-with-python-develop-
deep-learning-models-on-theano-and-tensorflow-using-keras-jason-
brownlee/
textboxfull.com

Mastering Machine Learning with Python in Six Steps: A

Practical Implementation Guide to Predictive Data
Analytics Using Python 1st Edition Manohar Swamynathan
(Auth.)
https://textbookfull.com/product/mastering-machine-learning-with-
python-in-six-steps-a-practical-implementation-guide-to-predictive-
data-analytics-using-python-1st-edition-manohar-swamynathan-auth/
textboxfull.com

Python Natural Language Processing Advanced machine

learning and deep learning techniques for natural language
processing 1st Edition Jalaj Thanaki
https://textbookfull.com/product/python-natural-language-processing-
advanced-machine-learning-and-deep-learning-techniques-for-natural-
language-processing-1st-edition-jalaj-thanaki/
textboxfull.com

Data Processing with Optimus: Supercharge big data

preparation tasks for analytics and machine learning with
Optimus using Dask and PySpark Leon
https://textbookfull.com/product/data-processing-with-optimus-
supercharge-big-data-preparation-tasks-for-analytics-and-machine-
learning-with-optimus-using-dask-and-pyspark-leon/
textboxfull.com
Practical Machine Learning for Data Analysis Using Python
1st Edition Abdulhamit Subasi

https://textbookfull.com/product/practical-machine-learning-for-data-
analysis-using-python-1st-edition-abdulhamit-subasi/

textboxfull.com

Data Science and Machine Learning Interview Questions

Using Python Second Edition Vishwanathan Narayanan

https://textbookfull.com/product/data-science-and-machine-learning-
interview-questions-using-python-second-edition-vishwanathan-
narayanan/
textboxfull.com

Introducing Data Science Big Data Machine Learning and

more using Python tools 1st Edition Davy Cielen

https://textbookfull.com/product/introducing-data-science-big-data-
machine-learning-and-more-using-python-tools-1st-edition-davy-cielen/

textboxfull.com

Beginning Anomaly Detection Using Python-Based Deep

Learning: With Keras and PyTorch Sridhar Alla

https://textbookfull.com/product/beginning-anomaly-detection-using-
python-based-deep-learning-with-keras-and-pytorch-sridhar-alla/

textboxfull.com

Learn PySpark: Build python-based machine learning and

deep learning models 1st Edition Pramod Singh

https://textbookfull.com/product/learn-pyspark-build-python-based-
machine-learning-and-deep-learning-models-1st-edition-pramod-singh/

textboxfull.com
Advanced
Data Analytics
Using Python
With Machine Learning,
Deep Learning and NLP Examples
—
Sayan Mukhopadhyay
 Advanced Data
Analytics Using
Python
With Machine Learning, Deep
Learning and NLP Examples

Sayan Mukhopadhyay
Advanced Data Analytics Using Python
Sayan Mukhopadhyay
Kolkata, West Bengal, India

ISBN-13 (pbk): 978-1-4842-3449-5 ISBN-13 (electronic): 978-1-4842-3450-1

https://doi.org/10.1007/978-1-4842-3450-1
Library of Congress Control Number: 2018937906

Copyright © 2018 by Sayan Mukhopadhyay

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole
or part of the material is concerned, specifically the rights of translation, reprinting, reuse of
illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical
way, and transmission or information storage and retrieval, electronic adaptation,
computer software, or by similar or dissimilar methodology now known or hereafter
developed.
Trademarked names, logos, and images may appear in this book. Rather than use a
trademark symbol with every occurrence of a trademarked name, logo, or image we use the
names, logos, and images only in an editorial fashion and to the benefit of the trademark
owner, with no intention of infringement of the trademark.
The use in this publication of trade names, trademarks, service marks, and similar terms,
even if they are not identified as such, is not to be taken as an expression of opinion as to
whether or not they are subject to proprietary rights.
While the advice and information in this book are believed to be true and accurate at the
date of publication, neither the authors nor the editors nor the publisher can accept any
legal responsibility for any errors or omissions that may be made. The publisher makes no
warranty, express or implied, with respect to the material contained herein.
Managing Director, Apress Media LLC: Welmoed Spahr
Acquisitions Editor: Celestin
Development Editor: Matthew Moodie
Coordinating Editor: Divya Modi
Cover designed by eStudioCalamar
Cover image designed by Freepik (www.freepik.com)
Distributed to the book trade worldwide by Springer Science+Business Media New York,
233 Spring Street, 6th Floor, New York, NY 10013. Phone 1-800-SPRINGER, fax (201)
348-4505, e-mail orders-ny@springer-sbm.com, or visit www.springeronline.com. Apress
Media, LLC is a California LLC and the sole member (owner) is Springer Science + Business
Media Finance Inc (SSBM Finance Inc). SSBM Finance Inc is a Delaware corporation.
For information on translations, please e-mail rights@apress.com, or visit www.apress.
com/rights-permissions.
Apress titles may be purchased in bulk for academic, corporate, or promotional use.
eBook versions and licenses are also available for most titles. For more information,
reference our Print and eBook Bulk Sales web page at www.apress.com/bulk-sales.
Any source code or other supplementary material referenced by the author in this
book is available to readers on GitHub via the book’s product page, located at
www.apress.com/978-1-4842-3449-5. For more detailed information, please visit
www.apress.com/source-code.
Printed on acid-free paper
This is dedicated to all my math teachers,
especially to Kalyan Chakraborty.
Table of Contents
About the Author���������������������������������������������������������������������������������xi
About the Technical Reviewer�����������������������������������������������������������xiii
Acknowledgments������������������������������������������������������������������������������xv

Chapter 1: Introduction������������������������������������������������������������������������1
Why Python?���������������������������������������������������������������������������������������������������������1
When to Avoid Using Python���������������������������������������������������������������������������������2
OOP in Python�������������������������������������������������������������������������������������������������������3
Calling Other Languages in Python���������������������������������������������������������������������12
Exposing the Python Model as a Microservice���������������������������������������������������14
High-Performance API and Concurrent Programming����������������������������������������17

Chapter 2: ETL with Python (Structured Data)������������������������������������23

MySQL�����������������������������������������������������������������������������������������������������������������23
How to Install MySQLdb?�������������������������������������������������������������������������������23
Database Connection�������������������������������������������������������������������������������������24
INSERT Operation������������������������������������������������������������������������������������������24
READ Operation���������������������������������������������������������������������������������������������25
DELETE Operation������������������������������������������������������������������������������������������26
UPDATE Operation�����������������������������������������������������������������������������������������27
COMMIT Operation�����������������������������������������������������������������������������������������28
ROLL-BACK Operation�����������������������������������������������������������������������������������28

v
Table of Contents

Elasticsearch�������������������������������������������������������������������������������������������������������31
Connection Layer API�������������������������������������������������������������������������������������33
Neo4j Python Driver��������������������������������������������������������������������������������������������34
neo4j-rest-client�������������������������������������������������������������������������������������������������35
In-Memory Database������������������������������������������������������������������������������������������35
MongoDB (Python Edition)����������������������������������������������������������������������������������36
Import Data into the Collection����������������������������������������������������������������������36
Create a Connection Using pymongo�������������������������������������������������������������37
Access Database Objects������������������������������������������������������������������������������37
Insert Data�����������������������������������������������������������������������������������������������������38
Update Data���������������������������������������������������������������������������������������������������38
Remove Data�������������������������������������������������������������������������������������������������38
Pandas����������������������������������������������������������������������������������������������������������������38
ETL with Python (Unstructured Data)������������������������������������������������������������������40
E-mail Parsing�����������������������������������������������������������������������������������������������40
Topical Crawling��������������������������������������������������������������������������������������������42

Chapter 3: Supervised Learning Using Python�����������������������������������49

Dimensionality Reduction with Python���������������������������������������������������������������49
Correlation Analysis���������������������������������������������������������������������������������������50
Principal Component Analysis�����������������������������������������������������������������������53
Mutual Information����������������������������������������������������������������������������������������56
Classifications with Python���������������������������������������������������������������������������������57
Semisupervised Learning�����������������������������������������������������������������������������������58
Decision Tree�������������������������������������������������������������������������������������������������������59
Which Attribute Comes First?������������������������������������������������������������������������59
Random Forest Classifier������������������������������������������������������������������������������60

vi
 Table of Contents

Naive Bayes Classifier�����������������������������������������������������������������������������������������61

Support Vector Machine��������������������������������������������������������������������������������������62
Nearest Neighbor Classifier��������������������������������������������������������������������������������64
Sentiment Analysis���������������������������������������������������������������������������������������������65
Image Recognition����������������������������������������������������������������������������������������������67
Regression with Python��������������������������������������������������������������������������������������67
Least Square Estimation��������������������������������������������������������������������������������68
Logistic Regression���������������������������������������������������������������������������������������69
Classification and Regression�����������������������������������������������������������������������������70
Intentionally Bias the Model to Over-Fit or Under-Fit������������������������������������������71
Dealing with Categorical Data�����������������������������������������������������������������������������73

Chapter 4: Unsupervised Learning: Clustering�����������������������������������77

K-Means Clustering��������������������������������������������������������������������������������������������78
Choosing K: The Elbow Method���������������������������������������������������������������������������82
Distance or Similarity Measure���������������������������������������������������������������������������82
Properties������������������������������������������������������������������������������������������������������82
General and Euclidean Distance��������������������������������������������������������������������83
Squared Euclidean Distance��������������������������������������������������������������������������84
Distance Between String-Edit Distance��������������������������������������������������������85
Similarity in the Context of Document����������������������������������������������������������������87
Types of Similarity�����������������������������������������������������������������������������������������87
What Is Hierarchical Clustering?�������������������������������������������������������������������������88
Bottom-Up Approach�������������������������������������������������������������������������������������89
Distance Between Clusters���������������������������������������������������������������������������90
Top-Down Approach��������������������������������������������������������������������������������������92
Graph Theoretical Approach��������������������������������������������������������������������������97
How Do You Know If the Clustering Result Is Good?�������������������������������������97

vii
Table of Contents

Chapter 5: Deep Learning and Neural Networks���������������������������������99

Backpropagation�����������������������������������������������������������������������������������������������100
Backpropagation Approach�������������������������������������������������������������������������100
Generalized Delta Rule��������������������������������������������������������������������������������100
Update of Output Layer Weights������������������������������������������������������������������101
Update of Hidden Layer Weights�����������������������������������������������������������������102
BPN Summary���������������������������������������������������������������������������������������������103
Backpropagation Algorithm�������������������������������������������������������������������������������104
Other Algorithms�����������������������������������������������������������������������������������������������106
TensorFlow��������������������������������������������������������������������������������������������������������106
Recurrent Neural Network��������������������������������������������������������������������������������113

Chapter 6: Time Series���������������������������������������������������������������������121

Classification of Variation���������������������������������������������������������������������������������121
Analyzing a Series Containing a Trend��������������������������������������������������������������121
Curve Fitting������������������������������������������������������������������������������������������������122
Removing Trends from a Time Series����������������������������������������������������������123
Analyzing a Series Containing Seasonality�������������������������������������������������������124
Removing Seasonality from a Time Series�������������������������������������������������������125
By Filtering��������������������������������������������������������������������������������������������������125
By Differencing��������������������������������������������������������������������������������������������126
Transformation��������������������������������������������������������������������������������������������������126
To Stabilize the Variance�����������������������������������������������������������������������������126
To Make the Seasonal Effect Additive���������������������������������������������������������127
To Make the Data Distribution Normal���������������������������������������������������������127
Stationary Time Series��������������������������������������������������������������������������������������128
Stationary Process��������������������������������������������������������������������������������������128
Autocorrelation and the Correlogram����������������������������������������������������������129
Estimating Autocovariance and Autocorrelation Functions�������������������������129

viii
 Table of Contents

Time-Series Analysis with Python���������������������������������������������������������������������130

Useful Methods��������������������������������������������������������������������������������������������131
Autoregressive Processes���������������������������������������������������������������������������133
Estimating Parameters of an AR Process����������������������������������������������������134
Mixed ARMA Models�����������������������������������������������������������������������������������������137
Integrated ARMA Models�����������������������������������������������������������������������������������138
The Fourier Transform���������������������������������������������������������������������������������������140
An Exceptional Scenario�����������������������������������������������������������������������������������141
Missing Data�����������������������������������������������������������������������������������������������������143

Chapter 7: Analytics at Scale�����������������������������������������������������������145

Hadoop��������������������������������������������������������������������������������������������������������������145
MapReduce Programming���������������������������������������������������������������������������145
Partitioning Function�����������������������������������������������������������������������������������146
Combiner Function��������������������������������������������������������������������������������������147
HDFS File System����������������������������������������������������������������������������������������159
MapReduce Design Pattern�������������������������������������������������������������������������159
Spark�����������������������������������������������������������������������������������������������������������������166
Analytics in the Cloud���������������������������������������������������������������������������������������168
Internet of Things����������������������������������������������������������������������������������������������179

Index�������������������������������������������������������������������������������������������������181

ix
About the Author
Sayan Mukhopadhyay has more than
13 years of industry experience and has been
associated with companies such as Credit
Suisse, PayPal, CA Technologies, CSC, and
Mphasis. He has a deep understanding of
applications for data analysis in domains such
as investment banking, online payments,
online advertisement, IT infrastructure, and
retail. His area of expertise is in applying
high-performance computing in distributed
and data-driven environments such as real-time analysis, high-frequency
trading, and so on.
He earned his engineering degree in electronics and instrumentation
from Jadavpur University and his master’s degree in research in
computational and data science from IISc in Bangalore.

xi
About the Technical Reviewer
Sundar Rajan Raman has more than 14 years
of full stack IT experience in machine
learning, deep learning, and natural
language processing. He has six years
of big data development and architect
experience, including working with Hadoop
and its ecosystems as well as other NoSQL
technologies such as MongoDB and
Cassandra. In fact, he has been the technical
reviewer of several books on these topics.
He is also interested in strategizing using Design Thinking principles
and in coaching and mentoring people.

xiii
Acknowledgments
Thanks to Labonic Chakraborty (Ripa) and Kusumika Mukherjee.

xv
CHAPTER 1

Introduction
In this book, I assume that you are familiar with Python programming.
In this introductory chapter, I explain why a data scientist should choose
Python as a programming language. Then I highlight some situations
where Python is not a good choice. Finally, I describe some good practices
in application development and give some coding examples that a data
scientist needs in their day-to-day job.

W
 hy Python?
So, why should you choose Python?

• It has versatile libraries. You always have a ready-­

made library in Python for any kind of application.
From statistical programming to deep learning to
network application to web crawling to embedded
systems, you will always have a ready-made library in
Python. If you learn this language, you do not have to
stick to a specific use case. R has a rich set of analytics
libraries, but if you are working on an Internet of Things
(IoT) application and need to code in a device-side
embedded system, it will be difficult in R.

© Sayan Mukhopadhyay 2018 1

S. Mukhopadhyay, Advanced Data Analytics Using Python,
https://doi.org/10.1007/978-1-4842-3450-1_1
Chapter 1 Introduction

• It is very high performance. Java is also a versatile

language and has lots of libraries, but Java code runs
on a Java virtual machine, which adds an extra layer
of latency. Python uses high-performance libraries
built in other languages. For example, SciPy uses
LAPACK, which is a Fortran library for linear algebra
applications. TensorFlow uses CUDA, which is a C
library for parallel GPU processing.

• It is simple and gives you a lot of freedom to code.

Python syntax is just like a natural language. It is easy to
remember, and it does not have constraints in variables
(like constants or public/private).

When to Avoid Using Python

Python has some downsides too.

• When you are writing very specific code, Python may

not always be the best choice. For example, if you are
writing code that deals only with statistics, R is a better
choice. If you are writing MapReduce code only, Java is
a better choice than Python.

• Python gives you a lot of freedom in coding. So, when

many developers are working on a large application,
Java/C++ is a better choice so that one developer/
architect can put constraints on another developer’s
code using public/private and constant keywords.

• For extremely high-performance applications, there is

no alternative to C/C++.

2
 Chapter 1 Introduction

O
 OP in Python
Before proceeding, I will explain some features of object-oriented
programming (OOP) in a Python context.
The most basic element of any modern application is an object. To
a programmer or architect, the world is a collection of objects. Objects
consist of two types of members: attributes and methods. Members can be
private, public, or protected. Classes are data types of objects. Every object
is an instance of a class. A class can be inherited in child classes. Two
classes can be associated using composition.
In a Python context, Python has no keywords for public, private, or
protected, so encapsulation (hiding a member from the outside world)
is not implicit in Python. Like C++, it supports multilevel and multiple
inheritance. Like Java, it has an abstract keyword. Classes and methods
both can be abstract.
The following code is an example of a generic web crawler that is
implemented as an airline’s web crawler on the Skytrax site and as a retail
crawler for the Mouthshut.com site. I’ll return to the topic of web crawling
in Chapter 2.

from abc import ABCMeta, abstractmethod

import BeautifulSoup
import urllib
import sys
import bleach
#################### Root Class (Abstract) ####################
class SkyThoughtCollector(object):
       __metaclass__ = ABCMeta

       baseURLString = "base_url"
       airlinesString = "air_lines"
       limitString = "limits"

3
Chapter 1 Introduction

       baseURl = ""
       airlines = []
       limit = 10

       @abstractmethod
       def collectThoughts(self):
             print "Something Wrong!! You're calling
an abstract method"

       @classmethod
       def getConfig(self, configpath):
             #print "In get Config"
             config = {}
             conf = open(configpath)
             for line in conf:
                    if ("#" not in line):
                          words = line.strip().split('=')
                          config[words[0].strip()] = words[1].
strip()
             #print config
             self.baseURl = config[self.baseURLString]
             if config.has_key(self.airlinesString):
                    self.airlines = config[self.
airlinesString].split(',')
             if config.has_key(self.limitString):
                    self.limit = int(config[self.limitString])
             #print self.airlines

       def downloadURL(self, url):

             #print "downloading url"
             pageFile = urllib.urlopen(url)

4
 Chapter 1 Introduction

             if pageFile.getcode() != 200:

                    return "Problem in URL"
             pageHtml = pageFile.read()
             pageFile.close()
             return "".join(pageHtml)

       def remove_junk(self, arg):

             f = open('junk.txt')
             for line in f:
                    arg.replace(line.strip(),'')
             return arg

       def print_args(self, args):

             out =''
             last = 0
             for arg in args:
                    if args.index(arg) == len(args) -1:
                          last = 1
                    reload(sys)
                    sys.setdefaultencoding("utf-8")
                    
arg = arg.decode('utf8','ignore').
encode('ascii','ignore').strip()
                    arg = arg.replace('\n',' ')
                    arg = arg.replace('\r','')
                    arg = self.remove_junk(arg)
                    if last == 0:
                          out = out + arg + '\t'
                    else:
                          out = out + arg
             print out

5
Chapter 1 Introduction

####################### Airlines Chield #######################

class AirLineReviewCollector(SkyThoughtCollector):

      months = ['January', 'February', 'March', 'April', 'May',

'June', 'July', 'August', 'September', 'October', 'November',
'December' ]

       def __init__(self, configpath):

             #print "In Config"
       super(AirLineReviewCollector,self).getConfig(configpath)

       def parseSoupHeader(self, header):

             #print "parsing header"
             name = surname = year = month = date = country =''
             txt = header.find("h9")
             words = str(txt).strip().split(' ')
             for j in range(len(words)-1):
                    if words[j] in self.months:
                          date = words[j-1]
                          month= words[j]
                          year = words[j+1]
                          name = words[j+3]
                          surname = words[j+4]
             if ")" in words[-1]:
                    country = words[-1].split(')')[0]
             if "(" in country:
                    country = country.split('(')[1]
             else:
                    country = words[-2].split('(')[1] + country
             return (name, surname, year, month, date, country)

6
 Chapter 1 Introduction

       def parseSoupTable(self, table):

             #print "parsing table"
             images = table.findAll("img")
             over_all = str(images[0]).split("grn_bar_")[1].
split(".gif")[0]
             money_value = str(images[1]).split("SCORE_")[1].
split(".gif")[0]
             seat_comfort = str(images[2]).split("SCORE_")[1].
split(".gif")[0]
             staff_service = str(images[3]).split("SCORE_")[1].
split(".gif")[0]
             catering = str(images[4]).split("SCORE_")[1].
split(".gif")[0]
             entertainment = str(images[4]).split("SCORE_")[1].
split(".gif")[0]
             if 'YES' in str(images[6]):
                    recommend = 'YES'
             else:
                    recommend = 'NO'
             status = table.findAll("p", {"class":"text25"})
             stat = str(status[2]).split(">")[1].split("<")[0]
             return (stat, over_all, money_value, seat_comfort,
staff_service, catering, entertainment, recomend)

       def collectThoughts(self):
             #print "Collecting Thoughts"
             for al in AirLineReviewCollector.airlines:
                    count = 0
                    while count < AirLineReviewCollector.limit:
                          count = count + 1
                          url = ''

7
Chapter 1 Introduction

                          if count == 1:
                                 url = AirLineReviewCollector.
baseURl + al + ".htm"
                          else:
                                 url = AirLineReviewCollector.
baseURl + al + "_"+str(count)+
".htm"
                          soup = BeautifulSoup.BeautifulSoup
(super(AirLineReviewCollector,self).
downloadURL(url))
                          blogs = soup.findAll("p",
{"class":"text2"})
                          tables = soup.findAll("table",
{"width":"192"})
                          review_headers = soup.findAll("td",
{"class":"airport"})
                          for i in range(len(tables)-1):
                                 (name, surname, year, month,
date, country) = self.parse
SoupHeader(review_headers[i])
                                 (stat, over_all, money_value,
seat_comfort, staff_service,
catering, entertainment,
recomend) = self.parseSoup
Table(tables[i])
                                 blog = str(blogs[i]).
split(">")[1].split("<")[0]
                                 args = [al, name, surname,
year, month, date, country,
stat, over_all, money_value,
seat_comfort, staff_service,
catering, entertainment,
recomend, blog]
8
 Chapter 1 Introduction

                                        
super(AirLineReviewCo
llector,self).print_
args(args)

######################## Retail Chield ########################

class RetailReviewCollector(SkyThoughtCollector):
       def __init__(self, configpath):
             #print "In Config"
       super(RetailReviewCollector,self).getConfig(configpath)

       def collectThoughts(self):
             soup = BeautifulSoup.BeautifulSoup(super(RetailRev
iewCollector,self).downloadURL(RetailReviewCollect
or.baseURl))
             lines = soup.findAll("a",{"style":
"font-­size:15px;"})
             links = []
             for line in lines:
                     if ("review" in str(line)) & ("target" in
str(line)):
                          ln = str(line)
                           link = ln.split("href=")[-1].split
("target=")[0].replace("\"","").
strip()
                          links.append(link)

             for link in links:

                    
soup = BeautifulSoup.BeautifulSoup(
super(RetailReviewCollector,self).
downloadURL(link))

9
Chapter 1 Introduction

                    comment = bleach.clean(str(soup.findAll("di
v",{"itemprop":"description"})[0]),tags=[],
strip=True)
                    tables = soup.findAll("table",
{"class":"smallfont space0 pad2"})
                    parking = ambience = range = economy =
product = 0
                    for table in tables:
                          if "Parking:" in str(table):
                                 rows = table.findAll("tbody")
[0].findAll("tr")
                                 for row in rows:
                                       if "Parking:" in
str(row):
                                              parking =
str(row).
count("read-
barfull")
                                       if "Ambience" in
str(row):
                                              ambience =
str(row).
count("read-
barfull")
                                       if "Store" in str(row):
                                              range = str(row).
count("read-
barfull")

10
 Chapter 1 Introduction

                                       if "Value" in str(row):

                                              
economy =
str(row).
count("read-
barfull")
                                       if "Product" in str(row):
                                              
product =
str(row).count
("smallratefull")

                    
author = bleach.clean(soup.findAll("spa
n",{"itemprop":"author"})[0], tags=[],
strip=True)
                    
date = soup.findAll("meta",{"itemprop":"dat
ePublished"})[0]["content"]
                    
args = [date, author,str(parking),
str(ambience),str(range), str(economy),
str(product), comment]
                                        
super(RetailReview
Collector,self).print_
args(args)

######################## Main Function ########################

if __name__ == "__main__":
       if sys.argv[1] == 'airline':
             instance = AirLineReviewCollector(sys.argv[2])
             instance.collectThoughts()
       else:
             if sys.argv[1] == 'retail':
                    
instance = RetailReviewCollector(sys.argv[2])
                    instance.collectThoughts()

11
Chapter 1 Introduction

             else:
                    print "Usage is"
                    print sys.argv[0], '<airline/retail>',
"<Config File Path>"

The configuration for the previous code is shown here:

base_url = http://www.airlinequality.com/Forum/
#base_url = http://www.mouthshut.com/product-reviews/Mega-Mart-­
Bangalore-reviews-925103466
#base_url = http://www.mouthshut.com/product-reviews/Megamart-­
Chennai-­reviews-925104102
air_lines = emrts,brit_awys,ual,biman,flydubai
limits = 10

I’ll now discuss the previous code in brief. It has a root class that is an
abstract class. It contains essential attributes such as a base URL and a
page limit; these are essential for all child classes. It also contains common
logic in class method functions such as the download URL, print output,
and read configuration. It also has an abstract method collectThoughts,
which must be implemented in child classes. This abstract method is
passing on a common behavior to every child class that all of them must
collect thoughts from the Web. Implementations of this thought collection
are child specific.

Calling Other Languages in Python

Now I will describe how to use other languages’ code in Python. There are
two examples here; one is calling R code from Python. R code is required
for some use cases. For example, if you want a ready-made function for the
Holt-Winter method in a time series, it is difficult to do in Python. But it is

12
 Chapter 1 Introduction

available in R. So, you can call R code from Python using the rpy2 module,
as shown here:

import rpy2.robjects as ro
ro.r('data(input)')
ro.r('x <-HoltWinters(input)')

Sometimes you need to call Java code from Python. For example,
say you are working on a name entity recognition problem in the field of
natural language processing (NLP); some text is given as input, and you
have to recognize the names in the text. Python’s NLTK package does have
a name entity recognition function, but its accuracy is not good. Stanford
NLP is a better choice here, which is written in Java. You can solve this
problem in two ways.

• You can call Java at the command line using

Python code.

import subprocess

subprocess.call(['java','-cp','*','edu.
stanford.nlp.sentiment.SentimentPipeline',
'-file','foo.txt'])

• You can expose Stanford NLP as a web service and call

it as a service.

nlp = StanfordCoreNLP('http://127.0.0.1:9000')
output = nlp.annotate(sentence, properties={
"annotators": "tokenize,ssplit,parse,sentiment",
"outputFormat": "json",
# Only split the sentence at End Of Line.
We assume that this method only takes in one
single sentence.
"ssplit.eolonly": "true",

13
Chapter 1 Introduction

# Setting enforceRequirements to skip some

annotators and make the process faster
"enforceRequirements": "false"
})

E xposing the Python Model

as a Microservice
You can expose the Python model as a microservice in the same way as
your Python model can be used by others to write their own code. The best
way to do this is to expose your model as a web service. As an example, the
following code exposes a deep learning model using Flask:

from flask import Flask, request, g

from flask_cors import CORS
import tensorflow as tf
from sqlalchemy import *
from sqlalchemy.orm import sessionmaker
import pygeoip
from pymongo import MongoClient
import json
import datetime as dt
import ipaddress
import math

app = Flask(__name__)
CORS(app)

@app.before_request
def before():
       db = create_engine('sqlite:///score.db')
       metadata = MetaData(db)

14
 Chapter 1 Introduction

       g.scores = Table('scores', metadata, autoload=True)

       Session = sessionmaker(bind=db)
       g.session = Session()

       client = MongoClient()
       g.db = client.frequency

       g.gi = pygeoip.GeoIP('GeoIP.dat')

       sess = tf.Session()
       new_saver = tf.train.import_meta_graph('model.obj.meta')
       new_saver.restore(sess, tf.train.latest_checkpoint('./'))
       all_vars = tf.get_collection('vars')

       g.dropped_features = str(sess.run(all_vars[0]))
       g.b = sess.run(all_vars[1])[0]
       return

def get_hour(timestamp):
       return dt.datetime.utcfromtimestamp(timestamp / 1e3).hour

def get_value(session, scores, feature_name, feature_value):

       s = scores.select((scores.c.feature_name == feature_
name) & (scores.c.feature_value == feature_value))
       rs = s.execute()
       row = rs.fetchone()
       if row is not None:
             return float(row['score'])
       else:
             return 0.0

15
Chapter 1 Introduction

@app.route('/predict', methods=['POST'])
def predict():
       input_json = request.get_json(force=True)

       features = ['size','domain','client_time','device',
'ad_position','client_size', 'ip','root']
       predicted = 0
       feature_value = ''
       for f in features:
             if f not in g.dropped_features:
                    if f == 'ip':
                          feature_value = str(ipaddress.
IPv4Address(ipaddress.ip_address
(unicode(request.remote_addr))))
                    else:
                          feature_value = input_json.get(f)
                    if f == 'ip':
                          if 'geo' not in g.dropped_features:
                                 geo = g.gi.country_name_by_
addr(feature_value)
                                 predicted = predicted + get_
value(g.session, g.scores,
'geo', geo)
                          if 'frequency' not in g.dropped_
features:
                                 res = g.db.frequency.find_
one({"ip" : feature_value})
                                 freq = 1
                                 if res is not None:
                                       freq = res['frequency']
                                 predicted = predicted + get_
value(g.session, g.scores,
'frequency', str(freq))
16
 Chapter 1 Introduction

                    if f == 'client_time':
                           feature_value = get_
hour(int(feature_value))
                    predicted = predicted + get_value(g.
session, g.scores, f, feature_value)
       return str(math.exp(predicted + g.b)-1)
app.run(debug = True, host ='0.0.0.0')

This code exposes a deep learning model as a Flask web service.

A JavaScript client will send the request with web user parameters such
as the IP address, ad size, ad position, and so on, and it will return the
price of the ad as a response. The features are categorical. You will learn
how to convert them into numerical scores in Chapter 3. These scores
are stored in an in-memory database. The service fetches the score from
the database, sums the result, and replies to the client. This score will be
updated real time in each iteration of training of a deep learning model. It
is using MongoDB to store the frequency of that IP address in that site. It is
an important parameter because a user coming to a site for the first time
is really searching for something, which is not true for a user where the
frequency is greater than 5. The number of IP addresses is huge, so they
are stored in a distributed MongoDB database.

 igh-Performance API and Concurrent

H
Programming
Flask is a good choice when you are building a general solution that is
also a graphical user interface (GUI). But if high performance is the most
critical requirement of your application, then Falcon is the best choice. The
following code is an example of the same model shown previously exposed
by the Falcon framework. Another improvement I made in this code is that
I implemented multithreading, so the code will be executed in parallel.

17
Random documents with unrelated
content Scribd suggests to you:
the wall ended. In fact here the ground rose a few feet to a grass
lined meadow.
As I looked at the mass of trees and shrubs, the more I began to
wonder what any one had been doing on the wall. It was not the
way to leave the grounds, for the person would have been forced to
climb the smooth stones. And then there was nothing but the distant
woods and the field. I wondered if the boy had been right when he
said the man had thrown something away.
My eyes came again to the slight hollow below me. It was a
tangled mass of shrubbery, with high grass, only hidden by the small
trees which grew to the height of about eight or nine feet. To try
and find anything in that tangle seemed impossible. But after a
moment I dropped from the wall and started to make my way
through the underbrush.
It was even worse than I had expected. The ground was soft
because of a spring which must have been near by. Not only were
the trees thick, but there was a tangle of wild rosebushes. Their
thorns clung to me as I tried to push my way through the bushes.
Long before I reached the other side I realized that it was too late to
expect to make any kind of a search. It had begun to be dusk by the
time I reached the wall, and when I plunged into the underbrush,
the thick foliage made it impossible to see.
Floundering at every step, my clothes pulled and twisted by the
thorns, and with the branches whipping across my face, I stumbled
through the swamp. Luckily it was not very wet, for if it had been, I
would have been unable to have gotten through. But long before I
reached the more solid ground I swore at myself for what I was
doing. And then, just when I began to wonder if I would ever reach
solid ground, I plunged out of the thicket into the high grass of the
field. With a sigh of relief I flung myself down on the grass for
breath.
My excursion had been of little value. True, I had discovered that
the boy could have seen some one on the wall, for I could see its
dim shape a few hundred yards away. I glanced at it a moment,
then looked behind me at the distant woods, and then glanced back
at the wall. And as I looked, to my great surprise I saw a figure
slowly drop from the same branch I had been on. A dim, indistinct
figure, which seemed to be covered with a long coat—a figure of a
man who, as I looked, dropped lightly on the top of the wall.
It was now dark, but not so dark that I could not make out the
man. The features I could not see, for he was too far away and the
darkness too dense. But there in front of me, on the other side of
the swamp, was a man—a man who remained but a moment on the
wall and then dropped to the ground. For a second I lost him from
sight, and then I saw the darkness split by the glare of a flashlight.
Like myself, he was going into the underbrush, throwing the light
carefully before him and going slowly, step by step, as if he was
looking for something.
For a second I wondered what I had better do. There was no
doubt he was searching for something, and searching very carefully.
I could hear the sound of the underbrush as it broke under his feet.
Once even I heard a muttered exclamation as the man half fell. The
small trees and vines hid him from my sight, but I could catch the
circle of flame from the flashlight as he swept it to and fro. He was
looking for something, but what? And then it dawned on me that the
boy had been right after all. Something had been thrown from the
wall into the small swamp.
There was little danger of my being seen. Not only was I in the
shadow, but it was also dark. The man was about a hundred feet
away, directly opposite from where I sat, crouched on the grass. He
seemed to be searching over a rather small circle of ground, and
evidently the search was not very successful. Then suddenly he
threw the light from off the ground at his feet and directed it
through the undergrowth in my direction. This caused me to move
quickly a few feet to where the grass was higher.
After all there was no necessity of my moving, for in a moment
the flame again swept the ground in a circle at the man's feet.
Wondering what he might find, and knowing that I could not be
seen from where he was, I rose and walked to the edge of the
swamp, but I could see nothing there. Then, remembering that the
ground rose to a pasture at my left, I carefully started in that
direction. And then, after I had taken one step, my foot slipped upon
a branch, and as the branch rolled over, I fell into the bushes with a
crash.
I was not hurt, but at the sound of my fall the flashlight was
suddenly extinguished. For a second I lay still, then, knowing there
was no use to keep silent, arose. And at that second I heard the
man blundering out of the underbrush, and the sound of his running
feet. Instinctively I started to run in the direction which would the
soonest bring me into the open—along the edge of the swamp, up
the little grassy hill, and then I paused an instant to listen.
For a moment I heard nothing, and then down below me, in the
direction of the lake, I heard the sound of running feet. Whoever
had been on the other side of the swamp was taking no chance of
being seen. As this thought struck me, I started down the hillside for
the water. The shore was not so far away, but before I reached it I
heard the man half tumble into a boat, and there came the sound of
oars splashing in the water.
As I reached the shore, almost crashing into the stone wall of the
Warren estate, I heard the faint click of the oars as the man began
to row down the lake. My canoe was on the other side of the wall.
To reach it, I had to stumble as softly as I could around the wall,
which stopped at the water. Lucky for me, though I had to wade, the
water was not above my shoe tops, and it took only a moment to
reach the canoe. I knew I had one advantage. It was almost
impossible for the man to row the boat without making some sound
with the oars; but I could paddle the canoe silently.
As I half fell over the canoe, I paused a second, waiting to catch
the faint sound of the oars; then silently I pushed the canoe into the
water, climbed in, and started to paddle as softly as possible in the
direction of the boat. It was far ahead of me, and the night was so
dark I could see but a few feet in front of me. But once in a while I
could hear the oars click, for the man was rowing with all his speed,
evidently not knowing that I had a canoe.
It had been my idea that as I could paddle faster than he could
row, I could soon reach him. My plan was to let him go ashore, and
then as he stepped out of the boat, to reach his side. One thing
above all I wished to do—see who it might be. After that there were
a few questions I wished to ask.
But either I misjudged the distance he was ahead of me or else
he could row far better than I expected. In my haste I had paddled
straight out into the lake, forgetting for a moment or so that the
shore took a wide curve. When I remembered that, it was to
discover the man had rowed straight across this half circle, and I
had to turn to the shore. This gave him several extra minutes, and
though, when he ran his boat up on the shore, I was only a few feet
away, yet it gave the time he needed.
He must have discovered just before he beached his boat that I
was behind him. In fact I was so close that when he rose and
jumped to the shore, I could partly see his figure—an indistinct mass
in the darkness. He half turned, jumped from his boat, partly
stumbled, only to gather himself and run away in the darkness. The
next second my canoe crashed on the beach, half throwing me to
my knees.
Jumping from it, I set out in the direction the man had taken. We
were on some one's lawn, for I saw the figure darting ahead of me,
thrown into reflection for a moment by the lights of a house. Then it
went through a hedge and was lost to sight. I ran over the grass,
through the opening in the hedge, and then discovered that I was
on Carter's land. As I glimpsed the running figure ahead, I heard
Trouble, who was locked in the garage, give a sudden bark, and
then commence to bark incessantly. I wished with all my heart the
dog was loose, but there was not time to open the garage door.
When I ran around the front of Carter's house, the man was not
in sight. In front of me was another hedge, to my left the lake, and
to my right the lawn ran to the street. In this direction the street
lights made it possible to see that there was no one there. Deciding
the man had gone through the hedge, I ran down the path, and the
next second was on the lawn which belonged to the minister. But I
could see no one, and though there was a light in the rectory, no
running figure crossed the reflection. The man was gone, but
where?
 I stopped running, to walk up the slight incline which led to the
church. Frankly I was puzzled. There had not been time for him to
have gone very far. The street lights gave enough illumination to at
least have allowed me to have seen if any running figure was ahead
of me. But no one was in sight.
Puzzled, I went rather slowly across the grass and came to the
dark shadow made by the wall of the church. As I stood silent a
moment, I cast my eyes at the dense mass of the tower, which
divided the rectory from the church proper. It was not a very high
tower, and its top was only a few feet above the roof of the church.
Then I went toward the street, keeping in the shadow, and when I
reached the tower itself, found to my surprise that the door at the
foot was open.
I peered within, only to have my eyes met by the blackness.
Finding a match, I struck it. For the shortest space of time the
darkness was lighted sufficiently for me to see a very narrow
winding pair of stairs which were lost to sight above my head.
Before the match was out I had time to see an iron railing which ran
along the side of the steps. As darkness came again, I once more
wondered what had happened to the man. There seemed no chance
now of finding him, for by this time he must have been far away.
And then, for some unknown reason, I decided to climb the stairs
in the tower and see if I could perceive anything from the roof.
Thinking it over later, I saw how absurd the idea must have been,
for the night was dark and even if I reached the roof, I would not
have been able to see very far. But at the time the only thing I
remembered was that when I had looked at the tower late in the
afternoon I had noticed an iron railing which ran around its top.
I reached in my pocket for another match, and then discovered I
had just struck the last one. A search of all my pockets gave no
results. Hesitating a moment, I finally stepped through the open
door, groping my way until my foot plunged against the stone stairs.
Then finding the railing, I began slowly to climb the steps.
The darkness was intense, and the stairway was not only very
narrow, but it wound around and around in a bewildering manner. I
groped my way from step to step, my hand firmly grasping the rail.
Just about the time I began to wonder if the steps would never end
I came out in a small room at the top of the tower. Directly in front
of me was an open door, and as my eyes became more accustomed
to the darkness, I saw that it led to the platform.
When I went through this door, I found myself on a little balcony,
which ran around the top of the tower. In my first quick glance I saw
below me the lights of the town, and, turning my head, I caught the
reflection of the village streets far away in the distance. I leaned
over the iron railing, saw the dark shadow of the roof a few feet
below me, and began to wonder why under heaven I had ever
climbed the steps.
After walking completely around the balcony, I came back to my
place near the open door. Though I had peered over the edge on all
sides, I had seen nothing. Only the lights of the town, the street
lamps and the reflections from the houses had broken the blackness.
I had just about made up my mind to descend when a sudden sound
behind me caused me to stiffen to attention. It was only a slight
sound, and what it was I could not tell. But as I started to turn,
suddenly from behind I felt something move, and the next moment
two hands suddenly gripped my body.
The attack was so sudden and unexpected that, taken off my
guard as I was, I did not move until the man had secured my arms
with a firm grip. I tried to throw them aside, but I might just as well
have tried to break a steel rope. Whoever it was had terrific
strength, for they held me in such a position that I was unable to
move. Throwing my body to this side and that, I tried to break the
grip, but it was impossible. And silently, save only for the man's
quick breathing, I was being borne to the iron rail.
The man had seized me from the rear, and as I was unable to
turn I could not see him. Even if I had been able to turn, I could not
have distinguished him, for the night was too dark. Silently, without
a word, he pushed me against the rail. Then as its iron began to cut
into my back, he tried to lift me from the platform. In horror it came
over me that he was trying to throw me from the tower.
With a sudden burst of strength I managed to free myself from
one of the clutching hands and half turned. But the effort was in
vain. Again the arms closed around me, and as I slipped on the
stone floor, I felt myself being raised and the iron rail pressing
against my back. I tried to regain my feet—tried to throw off the
clutching hands, but could not; and then, as if my weight was of no
avail, the man lifted me slowly but surely until my feet were off the
floor. Then he gave a sudden push, and with a cry I went off the
platform.
 Chapter VII.
The Chief and I—Make a Call
It is said that a person drowning has all the events of his life pass
before him. I do not know if that is true or not; but in the brief
second in which I was falling through space I had but one thought—
how foolish I had been to have ever climbed those tower stairs. The
thought must have come and gone in a mere flash of time, but it
seemed long enough for me to reflect over. Then, with a crash and a
jar which seemed to shake every bone in my body, I struck the roof.
Struck the roof and started to roll down its side.
Luckily, there was not a very steep pitch to the roof, and I
managed to stop my slide just as I reached the eaves. Aching in
every part of my body, I pulled myself into a sitting position and
gave a cautious look around. Above me was the tower, and I
thanked my stars that the height was only about six feet. If I had
fallen several feet from where I did, I would not have stopped until I
reached the ground. And there is no doubt I would have been killed.
As it was, it seemed a miracle that I had not broken any bones. For
though I was pretty sore from the fall, I decided there was nothing
else the matter.
But my position was far from enviable. I was sitting on the very
edge of the roof between the church and the rectory, about thirty
feet from the ground. The pitch of the roof was not so steep but
what I could climb it, but that would do me no particular good.
Above me was the roof of the house, and I remembered that I had
seen a light in the rectory as I came across the lawn. There must be
some sort of an arrangement to reach the roof through the rectory
attic. I decided to try and find it.
 Carefully I edged myself along and began to climb over the
smooth expanse of the rectory shingles. Though they were slippery,
yet the pitch was the same as that of the church, and I had no
difficulty. Halfway up the roof I found what I had expected—an
opening into the attic. But when I had seated myself on the wooden
hatch, I was no better off than before. It was locked, and I was
unable to open it.
Sitting there for a few moments, I tried to puzzle out what I
should do next. I did not intend to spend the night on the roof if I
could help it, and yet there did not seem to be any way down to the
ground. Then I began to bang on the covered opening—banging
with my feet, pounding and pounding. I would do this for a moment,
then wait, only to bang again. Just as I was about ready to give up
hope I heard footsteps in the attic, and a very surprised voice cried
in a muffled tone:
“What's that?”
I yelled out that I was on the roof and climbed off the cover of
the trap door. In a moment it was slowly lifted, and, looking through
the opening, I saw the rather startled face of the minister. I drew
myself to the edge and dropped at his surprised feet on the attic
floor. As I straightened up, I saw his eyes sweep over me, and they
grew very wide as he gasped:
“Good heavens! it's Mr. Pelt.”
I must have been a rather sorry-looking object. The branches of
the trees through which I had pushed my way while in the swamp
had torn my clothes. One leg of my trousers flapped to and fro as I
moved. I was covered with dirt, while my shoes were a mass of
mud. My face, which his eyes gazed upon in amazement, must have
been a sight, for after one look he exclaimed:
“What has happened to you, Mr. Pelt? And—and, why were you
on my roof?”
I started to explain, when, remembering his duties, he begged
me to come down in his library. I followed him down the attic steps
to the first floor, where we went into a room lined with books. He
motioned to a chair, excused himself and left the room. In a moment
he returned, bringing with him a bottle and a glass. As he poured
out the wine, he said he thought I needed it, and then, dropping
into a chair, watched me while I drank.
His eyes were very curious, yet he was too well bred to ask me
what had happened. Placing the glass on the near-by table, I told
him briefly of what had taken place. My visit to the swamp I passed
over lightly, but did tell of seeing a man run through Carter's hedge,
and how I had followed him. As I told of the sudden attack upon the
roof of his tower, and how I was thrown over the side, he gave a
gesture of horror.
“Heavens, man!” he said, “some one tried to kill you.”
I agreed to this, and there fell a silence. He broke it to say that
there was nothing unusual in the door of the tower being open. In
fact it was never locked, the village people using it for a place from
which to give their visitors a view of the lake. But to hear that which
had happened to me should have taken place upon his property
shocked him. His dark eyes never left my face as he expressed again
and again his horror. I could see that from his point of view the
whole thing was unexplainable.
He told me he had been reading, and he pointed out where he
had thrown his book when he heard the sound of my pounding on
the roof. He half laughed as he said that for a moment he was a
very astonished man. He could hear some kind of a banging
overhead, and at last went up the steps to the attic. There he was
startled to discover that the sound came from the roof. When he
heard my voice, he thought for a moment the boys of the village
were playing some sort of a prank.
The clock suddenly striking eleven roused me to the lateness of
the hour. With a little exclamation of pain at the stiffness of my
muscles I rose to my feet saying that I had better be going. He went
with me to the door, then out upon the lawn. There we both turned
and looked at the tower, which loomed above our heads. Saying
“good night,” I went rather slowly across the grass, through the
opening in the hedge, and over Carter's lawn.
The large living room was a blaze of light, and I went up the
steps and into the house. Carter and Ranville were sitting in two
large chairs talking very earnestly. It was the Englishman who saw
me first, and I half laughed as I saw his eyes sweep over my
disheveled figure and his jaw drop. Catching Ranville's expression,
Carter turned to see what he was looking at. Giving me one amazed
look, he rose to his feet and came over in my direction.
“In the name of God, Pelt, what have you been doing?” was his
shocked question.
I dropped into a chair and told him a drink of Scotch would be
very agreeable. With one look at my face he hurried out of the
room. In a second he returned with a bottle and, placing it on a
stand, took three glasses from his pocket. Ranville's eyes wore a
questioning look, but he made no inquiries until I had finished my
drink. Even then it was Carter who broke out:
“What happened, Pelt.”
Going into a long account of my experiences, I told them of the
visit of the boy and my trip to the swamp. They said nothing until I
spoke of the attack which had been made on me in the tower, and
how the man had tried to throw me off the top. Here Ranville broke
in to say dryly that from my looks he judged the man had
succeeded. I nodded, and then told how I had struck the roof, and
that the minister had let me into his house through the opening in
his attic.
When I had finished, Carter was silent a moment, then jumped
to his feet and left the room. Ranville poured me out another drink,
took a small one for himself and said:
“You struck something to-night, Pelt, but what it was I would
give a good deal to know.”
Shaking my head, I sank back in the chair. All at once I realized
how tired I was, and that I was terribly sore. Every bone in my body
seemed to be aching, and each time I moved I found a new sore
spot. My eyes fell on my suit—stained by the grass and smeared
with dirt. But for the time being I was too tired to go to my room
and change.
When Carter returned, he came to my side and said:
“I just got hold of the chief, Pelt. It struck me that if there is
anything in the swamp, the chief better see that no one gets it
before the police do. I told him what the boy said to you—how he
saw some one throw something away. He says that he has not been
to the station to-night, and so does not know if the boy went down,
as you told him to.”
He paused and went on. “I told him just a wee bit about what
you saw, suggesting he have a couple of men by the swamp to-night
in case the man goes back. He is going to do that.”
He gave me another look, and at my expression half laughed.
Then he said:
“What you need, Pelt, is a warm bath and a good night's rest.
You are pretty lucky not to have been killed; and after this we'd
better not let you go around alone.”
His suggestion of a bath was a good one, and they followed me
to my room and bade me good night. After I had turned on the
water in the tub I glanced at myself in the wall mirror. I was a sorry
object, and when I saw my face, I could not blame Carter for
smiling. There was a great streak of dirt across my cheek, and a
large bruise over one eye; my hair was a tangled mass. When I
turned away from the glass, I remembered my hat was still on the
roof of the church.
I looked myself over pretty well after I undressed and decided
that, save for several bruises, I was unharmed, True, I was very
sore, but the warm water would aid that. After staying in the water
for a long while, I rubbed myself well with alcohol, and crept
between the sheets of the bed. I gave a sigh of contentment as my
tired body felt the cool linen. For a while I tried to puzzle out all that
had taken place—tried, only to give it up as sleep crept down upon
me. Then with a sigh I turned over and knew no more.
It was long after ten when I went down to the dining room the
next morning. As I took my place at the breakfast table, the
housekeeper handed me a note saying that Carter had left it. It
simply announced that both he and Ranville had gone to Warren's
funeral, and they thought they would let me sleep. Breakfast over, I
strolled out on the lawn and for a while amused myself by playing
ball with the dog. Tiring of this after a while, I went up on the
veranda to read a magazine.
 I had been reading perhaps about ten minutes when I was hailed
by a voice from the walk and, glancing down, I saw the heavy, thick-
set figure of the chief of police. He came up the steps and dropped
into a near-by chair, wiping his face with his handkerchief as if he
was warm. Under one arm he had a small package which he was
handling very carefully. For a while we talked in a general way, until
I began to wonder what it might be that had caused him to come to
the house.
At length, after taking a pipe from his pocket and which he took
some time in filling, he said:
“Well, Mr. Pelt, I understand that you are with one of the best
detectives in the country.”
I agreed to this, and he went on:
“I have something here might interest you.”
He took the package from under his arm and slowly but very
carefully unwrapped it. There seemed to be far more paper used in
wrapping than was necessary, but at length he reached the object
he was after. Giving it one rather strange look, he handed it to me
without a word. It was a dagger of a style and workmanship which I
had never seen before. The blade was very long and extremely thin,
and it came to a decidedly sharp point—a point which not only was
sharp, but also very clean; the steel of which it was made was
glazed a dark green, and the handle had several figures carved upon
it.
After I had looked it over I gave him a questioning glance. His
response was quick:
“After Mr. Carter called me last night I sent two men round to
Warren's. Had them watch that swamp all night. They saw no one.
This morning I got a lot of boys and set them to work. They went all
over the place—that's what they found,” and he waved one hand at
the dagger.
I cast another glance at the dagger, wondering if it could have
been the thing the boy said he saw the man on the wall throwing
away. As if reading my thoughts, the chief said:
“The doc saw this awhile ago. He thinks it's the knife that did the
killing.”
 Realizing that the chief did not know all of my adventure of the
night before, I told him of what had taken place. He seemed very
startled, but then I knew that the murder itself had been a bit more
than he was accustomed to. He listened, however, without a word,
though he did slowly shake his head when I told of seeing the man
who was searching in the swamp. But when I mentioned of being
thrown from the tower of the church, he gave a sudden start. As I
ended, he said quickly:
“You know, Mr. Pelt, this murder, of course, upsets me. We don't
have many crimes that amount to anything here. My work is mostly
small stuff. But this murder sort of gets me. To start with, Warren
was the biggest man in this place. I can't for the life of me think of
any reason why any one should bump him off.”
I made no reply to this, studying the man before me. There was
something likable about the chief—something about the humorous
smile he had that caused me to warm to him. But his face was
puzzled, and it was clear to see he was, as he put it, upset. I could
readily understand that the murder of the most important man in his
town would cause him to realize the responsibilities resting upon his
shoulders. Our eyes suddenly met and, throwing out his hands in a
gesture of appeal, he said: “You know, Mr. Pelt, this murder is a little
too deep for me. I wish you could see your way clear to help me a
bit.”
I assured him I would be glad to do anything within my power,
but warned him that most of the work upon criminal cases which I
had been in had been done by Bartley. He listened with a serious air,
and when I finished remarked:
“That's all right; but you can help me, I know. Now I have been
thinking. You heard what Warren's housekeeper said yesterday
about that girl. How she said he ought to be killed.”
He paused as if refreshing his memory, then burst forth:
“That's a hell of a thing for a nice girl to say. But the thing I have
got to do now is to find her. I had two of the cops out, but they can't
find any one who has even seen her. Now I was thinking, perhaps
you would go with me to see her aunt. She ought to know
something about the girl. One of the boys asked the aunt last night
if she knew where she was, and she said, ‘No.’ But she might give us
some idea. What do you say?”
There was an appealing look on his face as he glanced at me,
and I was willing to agree that the aunt should be questioned. Upon
asking him when he wanted to go, he said: “At once,” and I told him
that if he would wait until I found my hat, I would go with him. The
hat found, we went down the steps, and he started for the road,
saying that his Ford was standing there. I told him I thought we
might go in more comfort in my car and, going to the garage, drove
it out to the road. The Airedale jumped into the seat as I stopped for
the chief, and as there was plenty of room, I let him stay.
It took us at least fifteen minutes to reach the house before
which we stopped. It was a small white place set far back from the
street, with a white fence before it. The path we went up was lined
with flowers, and roses climbed over the doorway of the house. It
was a very pleasant old lady who answered our ringing of the bell—
an old lady whose smile turned into an anxious look as she
perceived the chief.
She led us into an old-fashioned parlor of the sort I had not seen
for years. The haircloth furniture was of a beautiful design, and the
old prints upon the wall took me back to my boyhood days in New
England. Motioning us to a chair, she seated herself, with a very
anxious look upon her face. Then, without waiting for the chief to
tell his errand, she asked in a trembling voice:
“I hope you have not heard anything bad about Florence?”
The chief shook his head. “No, Mrs. West,” he replied. “I don't
know anything bad about her. But I wanted to ask you some
questions. You see we don't know where Florence is, and so far I
can't find any one who does. Now, did she come to your house
yesterday afternoon?”
The woman was rocking back and forth in her chair. When she
answered, I could tell she was worried.
“No,” came the reply. “I have not seen her for several days. You
know when she got that position with Mr. Warren, she had to stay at
his house. She said he often had her work at night, and it was too
far to come back here. She did run in every day or so, but I have not
seen her for at least three days. And then—” she paused as if
shocked by what she was going to add, then said: “And then I heard
about her saying that Mr. Warren ought to be killed.”
The face of the old lady was flushed, but the expression was
more one of perplexity than anything else. Her eyes did not leave
the chief's face.
“You know, of course, that, though Florence was a very impulsive
girl, she was a good girl,” she added proudly.
The chief nodded, and the aunt continued:
“Now I thought Florence just said that—that about Mr. Warren
being killed, because she was mad over something. She would not
have meant it, would she?”
Her eyes met the chief's imploringly. He shifted his heavy weight
and then with a sheepish expression remarked: “No—but then, Mrs.
West, why did she take her suit case and say she was not coming
back? Where did she go? That's what I want to know.”
To this the woman could give no reply. There was no doubt she
was very much puzzled over what had happened to her niece. But at
the same time she was sure the girl was all right. As to any place
that she might have gone, she did not know. She added that her
niece was a girl who did not go out, as she put it, very much, and
that she had no idea where she might be.
We were just on the verge of leaving when I asked her if she
knew of any young man she went around with. Her reply was that
Florence did not care much for young men; but in a moment she
mentioned a Robert Hunt with whom she once in a while went to
dances. At the name I shot a glance at the chief to see if he knew
the man.
He nodded. “I know him. That's the young man who is studying
law in Judge Thompson's law office. A fine young fellow.”
The aunt informed us that occasionally the young man came to
the house to see Florence and took her to dances. But he seemed to
be the only young man with whom the girl had anything to do, for
the old lady could not think of any other person the girl had been
interested in, and with this we left.
 I dropped the chief in front of Carter's. Leaving the car in the
drive, I went into the house to be greeted by the two men with the
news that lunch was ready. As we went into the dining room, they
asked me how I felt after my night's escapade, and grinned a little
when I told them I was a little sore. Then as we seated ourselves at
the table, I told them of the visit of the chief and of the finding of
the dagger. They made hardly any comment, though Ranville said he
wanted to have a look at the weapon.
During the lunch they told me about the funeral. Not only had it
been largely attended, but there had never been so many famous
men in the village before. From the various colleges there had come
the scientists to pay their last respects to the famous man who had
worked with them. And then, after mentioning the names of some
who had been at the funeral—names known the world over in
intellectual circles—Ranville said:
“You know, Pelt, we wondered what would be done with the
notes of Warren's trip, and how the world had been waiting for his
account of his discoveries. We mentioned that the book must be
finished. Well, we heard this morning that Niles Patton, who was
with him on the trip, will finish the book.”
I turned to Carter, asking him if he knew Patton, only to be told
that he knew him very well. He was a young assistant professor at
Harvard, working under Warren in anthropology, and had been with
him in China. And then Carter said that Patton would take up his
task at once, and would be at work in the Warren library to-morrow.
Lunch was almost over, and we were just drinking our coffee
when there came the sound of the door bell and the loud barking of
the dog upon the lawn. The housekeeper passed through the room
to the door, and in a moment returned. As she stood by Carter's
side, he asked:
“Who is it, Mrs. Hart?”
She shook her head saying: “I don't know, Mr. Carter. The man
says he must see you. I never saw him before.” She paused a
second, then finished: “It's a Chinaman.”
 Chapter VIII.
We Have a Visitor—and—I Take a
Ride
For a moment no one spoke. As our eyes met, the look we gave
each other was a wondering one. I saw a half smile play around
Ranville's lips, and then Carter rose and started for the living room,
we after him. The same thought must have been in the minds of
each of us. Was this man at the door the Chinaman Warren's
housekeeper had said came to the door around the time he was
killed?
The housekeeper had gone to the front door again, and by the
time we reached the living room was ushering the man into our
presence. He was the tallest man of his race I have seen, with a
well-developed body and a very lean face. He was wearing a white
silk suit, and as he came across the room to our side, bore himself
with easy grace. His cold black eyes swept over the three of us, and
then in the most perfect English and in a very cultured voice he
asked:
“And which is Mr. Carter?”
Carter replied, and the man half smiled—a smile which lit up his
yellow face. Then he said:
“No doubt you do not know me, Mr. Carter, but you have been
pointed out to me in Washington. My name is Lee Kong, and I have
at times been connected with our London Embassy.”
The English was perfect, and there was no doubt he was one of
those well-educated Chinamen which one meets so often in the
various capitals of the world. His general air was that of breeding
and of culture, and one could feel back of it all the many thousand
years of racial blood. Carter introduced us and motioned to a chair.
After seating himself the man said:
“I saw your name in the paper, Mr. Carter. And I thought that
perhaps you might be the best man to whom to tell my story.”
Carter took his cigar case and offered it to him, but he begged to
be excused. Taking a gold case from his pocket, he lighted a long,
straw-colored cigarette and leaned back in his chair. His shrewd eyes
took us all in, and then with a little laugh he leaned forward to say:
“Perhaps I had better tell you just who I am. I have been
connected with both our London and Washington Embassies. I knew
Dr. Warren very well. It is in regard to my visit to his house
yesterday that I wish to speak. You know, of course, I was there.”
I was rather surprised at the cool manner in which the oriental
admitted he was the man the housekeeper had seen. Carter replied
he knew some one of his race had been there, but that was all. Over
the yellow face came a half smile, and then the words:
“Yes, I was the man. I knew inquiries would be made, and
thought I had better tell my story to some one who could
understand. I have been staying at Cooperstown. When I read in the
papers about the inquest, I saw your name, and I decided you were
the person to see.”
There was an air of sureness about the man which was annoying.
That the police were very keen to question him did not seem to
enter his head, or if it did, he was not interested. I believed he was
the last and, for that matter, the only person who had seen Warren
around the time of his death. As I looked at the impassive, yellow
face, with the cold, black eyes, I wondered if he held the solution of
the mystery. As if reading my thoughts, his suave glance met mine,
and he said:
“I can assure you I did not kill Mr. Warren. He was dead when I
went into his library.”
We gave a start, but the face of the man did not change a
particle. Coolly, he repeated the statement: “He was dead when I
saw him.”
Slowly lighting another of his thin yellow cigarettes, he carefully
inhaled the smoke, then added:
 “Mr. Warren, as you know, spent two years in the heart of my
country. In fact, I aided him in some of his arrangements before he
started. China, like all other nations, has citizens who are not law-
abiding. You may remember that when Mr. Warren did not return at
the expected time, it was said he had been killed by outlaws. That,
of course, was but a rumor. He returned safely, but his leaving China
made some little comment among our people.”
He paused, then continued: “You perhaps know the story of
Buddha's death—how his body was burnt, the ashes placed in seven
boxes and buried in seven places.”
As I looked at him, I wondered what he was driving at, but he
went on:
“Of course, many of the Christian missionaries have said that the
story was a myth.” A little smile passed over his lips and, shrugging
his shoulders, he continued:
“Just what the missionaries tell their converts now I cannot say;
but only a few years ago there was discovered not only one of the
shrines of Buddha in India, but they found one of the boxes with his
ashes. That is now a matter of history.”
He glanced at our interested faces, and then said half
apologetically:
“You, perhaps, wonder what this has to do with Warren. I will tell
you. Buddhism is not the religion of China, though there are a good
many Buddhists. But tradition has said that in one of our temples far
out in the desert, there was a shrine which contained the ashes of
the great teacher. Three years ago outlaws ravaged the temple, and
among their spoils was the box which our tradition said contained
the ashes of the Blessed One.”
He paused to light another cigarette, and then came his smooth
voice:
“To get at the heart of the matter, that box in some manner fell
into Warren's hands. I had been in correspondence with him for
several months asking him to return it to me. It was of no value to
him, and it did mean a good deal to some of my countrymen. He
was willing to give it up, and it was for that purpose I came to his
house two days ago.”
 “And you found him dead?” came Carter's question.
The dark head bowed, repeating slowly: “I found him dead.”
He threw out his hands in a gesture of resignation.
“His soul had joined his ancestors.” He paused, then went on
rather quickly:
“I went up the path as the servant directed me. The library door
was open, and after knocking several times I took the liberty of
entering. Mr. Warren was lying back of his desk upon the floor; there
was a knife protruding from his chest.”
“You say that the knife was in his body?” came Carter's startled
voice.
As if surprised at his tone, the Chinaman studied him coolly. “Yes.
You seem surprised.”
Carter made no reply to the question, but asked him if he had
seen any one, and if he closed the door after he left. To both he
answered: “No”; then seeing our amazed faces, he added:
“You perhaps wonder why I did not report his death to the
police.”
He was silent a moment, studying the gloss of his nails; then,
raising his head, he said:
“That, of course, was not my affair. Then again I failed to see
why I should become—what is it you say—mixed up in a violent
death. I had nothing to do with it, and it was not my affair in the
least.”
He paused and there fell a silence—a silence broken by Ranville
speaking for the first time:
“You wrote to Mr. Warren, you say—telling him you were to visit
him?”
The yellow figure in the chair slightly inclined his head in assent.
Then Ranville asked:
“Did you type your letter?”
The voice was cold as the oriental replied:
“As it was a personal letter, naturally I did him the courtesy of
writing it with my own hand.”
Again silence, in which I wondered what might be in the brain of
the inscrutable figure in the chair. His eyes went over the three of
us, and I thought I saw a little show of contempt in the glance. It
was Carter who broke the silence—a silence which was becoming
awkward.
“Did you find the box?”
The man rose, and once more I thought that he was the tallest
man of his race I had ever seen. As he bent to take his hat in his
hand, I noticed how long his fingers were and how flexible they
appeared. Then as he straightened up, he answered the question:
“No, I did not find it. If I had, I would have taken it with me. I
left everything as it was and departed. The door was still open when
I went away.”
He turned as if he thought the interview was over. Behind the
man's back Ranville's eyes met Carter's, and there was a little
amused look in them. He reached the door before Carter spoke:
“Will you go with me and tell your story to the chief of police?”
The man turned quickly. There seemed to be a surprised tone in
his voice as he replied:
“The police? Why, no, why should I bother with them. You can
tell them what I have said.”
He smiled and added sarcastically: “They can find me. I am in
Cooperstown—at the Inn.” Then bowing very low, he begged us not
to go to the door, and went out into the air.
For a moment after his steps died away no one spoke. The man
had gone as quietly as he had arrived. His story, to say the least,
was odd. Then, with a decided shake of his head, Carter burst forth:
“Well, what do you think of that yarn? For my part I don't believe
it.”
“It is pretty hard to say, Carter,” Ranville laughed in reply. “I have
had a good deal of experience with the Chinese, and it has been my
idea that it is almost impossible to say what goes on back of their
almond-shaped eyes. Yet he may have told the truth, at that.”
“I don't believe it,” was Carter's energetic protest. “That
Chinaman was about as smooth an article as I have ever seen. It's
my idea he knew it was only a question of a day or so before he
would have been traced. That's why he came around and told that
yarn. There is no way of saying it is not so; his word is as good as
ours.”
He brought his fist down on the arm of his chair, then added:
“But there is one thing I can do. I will wire my chief in Washington
and get all the information there is about him. If he has been around
the Embassy, the chief will have the dope on him.”
“His story of Warren being dead, with the knife still in the body
and the door open, makes the affair look a bit different than what
we found,” I commented.
Ranville carefully lighted a cigar, watched the first smoke ring
circle above his head, and then said:
“If we presume that his story is true, it opens up several
interesting theories. Who took the knife from the body? Who closed
the door?” and he paused.
“And who killed him?” broke in Carter's disgusted voice.
Rising to his feet, he stood before our chairs, his hands in his
pocket, and said earnestly:
“There's that story of his about the box with the ashes of Buddha
in it. There were a number of boxes in the room. Who knows if he
did not take it? For that matter, who can tell if anything was taken
from Warren's library? We won't get very far as to discovering why
Warren was killed until we find a motive of some sort. At present
there is no one who can tell us if anything was taken from the
library. No one knows anything about its contents anyway.”
“Only the secretary,” was Ranville's dry comment, “and no one
knows where she is.”
As they both looked at me, I told them of my visit to the girl's
aunt, and of what the woman had said. They listened until I was
through, and then said that the most important thing the chief had
to do was to find the secretary. Not only would she have to explain
her statement, that Warren ought to be killed, but there was
something else she could do. As she had worked with Mr. Warren,
she must be familiar with the contents of his library. She alone could
tell us if anything was missing. Until we knew that fact there was
little we could go on.
 We discussed this for a while, and in the midst of our talk the
telephone rang. Carter went out into the hall to answer it. He
returned in a moment to tell me I was wanted on the phone. Taking
the receiver, I found the chief of police on the other end of the wire.
He had received—as he put it—a good tip as to where the missing
secretary was. He mentioned that it was about twenty miles away,
over the mountains, and asked if I would be willing to take my car
and go with him. I told him I would be at the station in a few
moments and rang off.
Returning to the living room, I told them what the chief had said,
and Carter tried to figure out where we might be going. He said
there were any number of small lakes in the near-by mountains
which were as inaccessible as though they were hundreds of miles
away. Some of these lakes were camping spots, and many of them
did not have even a telephone. He followed me out to the car to
remind me that it might not be a bad idea for me to take my chains
along. Then as he wished me good-by, I drove out of the yard.
The police station was a small squatty brick building, set directly
back of a white church. I stopped at the curb and, climbing down,
went through the open door into the large room within. Behind a
desk, with his feet higher than his head, was a policeman reading a
newspaper. He dropped the paper long enough to ask me what I
wanted. At my request for the chief, he motioned to a door at the
left and then went back to his reading. Walking across the floor, I
pushed open the door and passed into the next room.
Evidently it was the private office of the chief. A small room with
two chairs and a desk for furniture. Upon the walls were several
gaudy calendars and two or three placards announcing a reward for
various wanted men. At the desk, busily engaged in looking over a
large scale map, was the chief. He motioned to the other chair, and
as I pulled it up to the desk, he started to thank me for my
willingness to go with him.
As my eyes dropped to the map, he saw my glance and, pushing
back his chair, said:
“Well, I think maybe I have got hold of a good tip about that girl.
I got hold of the young law student she goes around with. He said
that he saw her about half-past four the afternoon of the murder.”
“He did!” was my comment.
“Yep, he did,” was the retort. “And he saw her again at five. She
was climbing into the old Ford which belongs to the camp at Lake
Pleasant. That's a small girls' camp over Bald Mountain. It's twenty-
five miles from here, and about as bad a road as you can find. Once
a week the Ford comes into town for provisions and the mail.”
A twenty-five-mile ride over a bad road did not appear very
pleasing. I asked the chief why he did not telephone and see if the
girl was there.
“For the simple reason there is no telephone in camp. The
nearest one is over five miles away, on the side of another hill.
Besides I had an idea it might be best to see the girl ourselves. The
young man said she was in the Ford when it went out of town. And
if she was, then she must be at the camp.”
I agreed it might be best to see the girl ourselves, and in a short
time we went out to the car. I stopped at a filling station to make
sure the tank was full and that I had enough oil. Then we headed
for the hills, which were but a few miles away. For a while we drove
on a fine state highway, with summer estates and small farms for
company. But after we had gone a few miles we turned off the main
road to follow a narrow, dirt highway—a road which grew narrower
as we went along—with a brook by the side, which was running
swiftly down to the lake.
Not only was the road narrow, but it was in horrible condition.
Deep ruts ran along its sides and across its surface. The scenery
became more desolate, and the farms fewer. In a short while the
pasture land vanished, and on each side of us were deep woods.
The road began to climb, growing steeper with every foot, and soon
I was forced to put the car into second gear. Though it was a very
well-built machine, yet we bounded from side to side upon the seat.
Once as we struck an extra deep rut, I heard the chief mutter an
oath as he came down against the side of the car.
We reached the top of the mountain at last, and found that there
was a clear space ahead of us. We could look far down into the
valley below and across to another range of hills. A wild country
useless for farming, and without a house in sight. But as I swept
around a bend in the road, I saw far below me the shining waters of
several small lakes. One of these the chief pointed out as being our
destination.
If anything, the descent of the hill was even worse than coming
up. I could well see why those who camped in the valley did not
make many trips to town. The road became so narrow that there
was not room for two cars to pass, and I wondered at times if I
could keep on it. Not only that, but it suddenly started to sweep
around abrupt heights, in sharp, sudden curves—curves which took
all one's skill to make. It was with a little sigh of relief that I came
off the mountain and began to follow the road which ran through
two rows of trees.
Suddenly we came out of the wood, to find ourselves by the
shore of a small lake—a lake which formed almost a complete
glimmering circle, with trees coming down to the water's edge. The
road followed the shore of the lake for a few hundred yards, then
took an abrupt bend, to again come back to the water. And then, as
we came around the bend, I saw three small buildings ahead of us—
buildings placed in a cleared spot of ground, and with a little wharf
in front which ran into the lake. And it was here we stopped.
The buildings were the sort of summer cottages that one finds so
often, built more for cheapness than for looks or even comfort. But
when we climbed out of the car and went across the high grass, we
found that there was no one at home. The cottages were occupied,
for the doors were not locked. Hammocks were swung under the
trees, and on the verandas were chairs with magazines flung upon
them. But there was no one at home; and when we went down to
the wharf, we found the boats were not in. Evidently we would have
to wait until they returned.
We seated ourselves upon the edge of the wharf and let our legs
dangle over the water. Lighting my cigar as the chief lighted his pipe,
I gazed at the scene before. On all sides the woods circled the lake,
the trees green and restful to the eyes. Above in a blue sky little
fleecy clouds floated by, only to be reflected in the darker blue of the
lake. And on every side were the hills, with the lake below, like the
bottom to a cup.
For a while we sat there, and the chief told me the camp was run
by one of the churches in the village. But after a time conversation
ceased, and we sat and smoked, dreaming in the sunlight. Once in a
while a fish would jump from the water—a gleaming mass of silver—
and fall back with a splash; then would come the scream of a bird
far down the lake. But save for these, the air was very still. And just
as I was beginning to grow sleepy, the chief cried, pointing down the
lake:
“I guess they are coming.”
I gazed in the direction he pointed and saw three small boats far
down the lake. They were moving very slowly, but once in a while I
could catch the echo of laughter. We watched them approach nearer,
until the people in the boats could be seen—a number of girls in
bathing suits, girls laughing and talking.
They had almost reached the dock before they saw us and for a
moment their voices suddenly ceased. Then, recognizing the chief,
they waved their hands at him. In a moment the row boats were by
the dock, and I started to offer to aid them with their ropes; but
instead of tying to the dock they ran their boats upon the shore and
soon ten girls in bathing suits came running down the planks of the
wharf. I could tell they were curious and that the presence of the
chief puzzled them.
From out of the ten figures there stepped a girl who evidently
had been in the water only a few moments before. Her one-piece
suit clung to the perfect figure, and I noticed that her shapely legs
were whiter than the other girls, as though she had not been in the
sun very long. Her face was keen, and her laughing eyes turned
upon the chief. She advanced almost in front of him and in a
laughing voice said:
“Did you come after me, chief?”
The chief rose slowly to his feet and his face was serious as he
turned and faced the girl. As their eyes met something about his
expression drove the laughter from her face and her cheeks flushed.
For a moment they stood there, their eyes meeting, and then the
chief said very simply:
“I am afraid I did, Miss Harlan.”
 Chapter IX.
In Which Bartley Arrives
The girl's question had been asked in a laughing voice and her
eyes had danced as she spoke to the chief. But the rather curt reply
of the police officer and the tone of seriousness in his voice caused
the smile to slowly fade from her face. For a moment she looked at
him, and then in a trembling voice asked:
“Why—what is the matter?”
The other girls had crowded around. Their eyes were bright with
wonder, curious to know what was the trouble. The chief gave them
an uncertain look, then said that there was not much the matter, but
he wished to speak to the secretary alone. Reluctantly they started
toward the cottages, casting back many wondering glances. As they
reached the piazza of the central cottage they broke into excited
conversation.
He turned to the secretary, who stood, with a grave face, in front
of him. Just what was in his mind regarding the girl would have
been rather hard to say. But I knew he was remembering the
statement of the housekeeper—that she had said Warren ought to
be killed. Yet, as I looked at the beautiful girl before us, with the
bathing suit showing every line of her figure, I decided that though
she might have made the statement it had meant nothing. Her face
was frank and the gaze which she gave us both was fearless.
“Florence,” said the chief slowly, “Mr. Warren was found dead the
evening you left.”
Her eyes opened wide in astonishment and she gasped:
“Dead?”
He nodded gravely, then added: “Worse—murdered.”
 The girl's face whitened, and then slowly the color flushed back
into her cheeks until they were a vivid red. For a moment she looked
at us as if not believing what had been said. Then slowly she went
to a near-by box and sank down upon it. There was no doubt she
was surprised and also horrified at the news she had heard.
As she did not speak, the chief nervously shifted his weight and
threw a puzzled glance in my direction. Purposely I turned my eyes
away, and in a moment, in an embarrassed voice, he said:
“You see, Florence, it's a bit awkward. You went away from Mr.
Warren's very suddenly.”
For the first time the girl showed a bit of temper, as though it
had just dawned upon her that the chief would not have taken the
long trip from the village just to tell her Mr. Warren was dead. She
spoke in a voice a little nervous and at the same time sharp.
“Well, Suppose I did; that's my business, is it not?”
The chief shook his head. “I am afraid not,” was the slow
response. “In a sense it's mine—now. You see the remark you made
to Mr. Warren's housekeeper made it necessary to find you.”
She looked at him as if not understanding and then half
stammered: “What remark?”
“That Mr. Warren ought to be killed,” was the cold reply.
As if realizing the seriousness of the chief's tone, the girl's face
went very white. Slowly her fingers opened and closed; her eyes
studied the water for a moment as if she was trying to fix in her
mind the distant shore. Then she slowly raised her head and in a
look which included us both said:
“I did say that. But it was a very silly thing to say and it meant
nothing. Mr. Warren was alive when I left him and, of course, I know
nothing about his death.”
As the girl hesitated, the chief broke in:
“I am not saying you did know anything about his death. I only
want to make you see how you placed yourself in a bad position by
going away and by the remark you made. Why did you say it?”
The secretary gave a half laugh, though there was a tone of
disgust in it.
Welcome to our website – the ideal destination for book lovers and
knowledge seekers. With a mission to inspire endlessly, we offer a
vast collection of books, ranging from classic literary works to
specialized publications, self-development books, and children's
literature. Each book is a new journey of discovery, expanding
knowledge and enriching the soul of the reade

Our website is not just a platform for buying books, but a bridge
connecting readers to the timeless values of culture and wisdom. With
an elegant, user-friendly interface and an intelligent search system,
we are committed to providing a quick and convenient shopping
experience. Additionally, our special promotions and home delivery
services ensure that you save time and fully enjoy the joy of reading.

Let us accompany you on the journey of exploring knowledge and

personal growth!

textbookfull.com