The py-strsim library is a wrapper for the fabulous strsim Rust crate. This package extends the functionality marginally by enabling parallelized versions of each strsim function using rayon.

It is advised to use a virtual environment for most projects. The instructions below assume that python is the name of your Python-3 interpreter.

git clone https://github.com/knight9114/py-strsim.git
cd py-strsim
python setup.py install

The py-strsim package has two parts - single and vectorized. The single API matches the original strsim crate exactly. The vectorized versions of the functions look slightly different:

strsim.vectorized.<function>(n: int, a: str, bs: list[str]) -> list[int] | list[float]:
    ...

The first argument, n, specifies the number of threads to use during the computation. Each element in bs will be right-compared to the input a. The ordering in the output matches the ordering in the input bs.

import strsim

assert strsim.single.levenshtein('hello world', 'Hello, World') == 3
assert strsim.single.normalized_levenshtein('hello world', 'Hello, World') == 0.75
...

assert strsim.vectorized.levenshtein(2, 'hello world', ['Hello, World', 'hello world!']) == [3, 1]
...

My only contribution to this project is writing the Python bindings. All of the credit belongs to

• strsim
• rayon
• pyo3