pystrsim

Python wrapper for Rust's strsim library

Github repo: https://github.com/subpath/pystrsim

Installation:

pip install pystrsim

Usage:

import pystrsim

print(f"hamming: {pystrsim.hamming('hamming', 'hammers')} should be 3")
print(f"levenshtein: {pystrsim.levenshtein('kitten', 'sitting')} should be 3")
print(
    f"normalized_levenshtein: {pystrsim.normalized_levenshtein('kitten', 'sitting')} should be ~0.571"
)
print(f"osa_distance: {pystrsim.osa_distance('ac', 'cba')} should be 3")
print(f"damerau_levenshtein: {pystrsim.damerau_levenshtein('ac', 'cba')} should be 2")
print(
    f"normalized_damerau_levenshtein: {pystrsim.normalized_damerau_levenshtein('levenshtein', 'löwenbräu')} should be ~0.272"
)
print(
    f"jaro: {pystrsim.jaro('Friedrich Nietzsche', 'Jean-Paul Sartre')} should be ~0.392"
)
print(
    f"jaro_winkler: {pystrsim.jaro_winkler('cheeseburger', 'cheese fries')} should be ~0.911"
)
print(
    f"sorensen_dice: {pystrsim.sorensen_dice('web applications', 'applications of the web')} should be ~0.7878787878787878"
)

Is it blazingly fast?

Well, no : ) Jellyfish and Levenshtein are faster.

See the benchmark/benchmark.py file.

algorithm	library	function	time
DamerauLevenshtein	jellyfish	damerau_levenshtein_distance	0.00593378
Hamming	Levenshtein	hamming	0.000683438
Hamming	jellyfish	hamming_distance	0.00112426
Jaro	jellyfish	jaro_similarity	0.00206124
JaroWinkler	jellyfish	jaro_winkler_similarity	0.00221943
Levenshtein	Levenshtein	distance	0.00115115
Levenshtein	jellyfish	levenshtein_distance	0.00257007
damerau_levenshtein	pystrsim	damerau_levenshtein	0.380067
hamming	pystrsim	hamming	0.0116847
jaro	pystrsim	jaro	0.0547281
jaro_winkler	pystrsim	jaro_winkler	0.057244
levenshtein	pystrsim	levenshtein	0.102525
normalized_damerau_levenshtein	pystrsim	normalized_damerau_levenshtein	0.389092
normalized_levenshtein	pystrsim	normalized_levenshtein	0.107314
osa_distance	pystrsim	osa_distance	0.15746
sorensen_dice	pystrsim	sorensen_dice	0.0973786