simon-white-similarity

Detects similar strings using the Simon White's approximate string matching method, that computes the Sørensen–Dice similarity coefficient of sets of adjacent letter pairs (also called bigrams).

Based on "How to strike a match" by Simon White, originally at http://www.devarticles.com/c/a/Development-Cycles/How-to-Strike-a-Match Nowadays the original document seems missed, but it can still be found in http://www.catalysoft.com/articles/StrikeAMatch.html

Original Java implementation by Simon White. First refactored and translated to Pascal by palbcn. And then readapted to javascript also by palbcn.

installation

npm install simon-white-similarity    

usage

const computeSimilarity  = require('simon-white-similarity');  

// computes similarity between two strings
let similarity = computeSimilarity('Tàkë á sad song and mäke it bétter','take a bad song made better');    

// or save the letter pairs of the string 
let bigrams = computeSimilarity('Tàkë á sad song and mäke it bétter');

// that can later be compared with different strings   
similarity = computeSimilarity('take a sad song and make it better',bigrams);
similarity = computeSimilarity('take a bad song made better',bigrams);

// multiple string comparison can be simplified 
let candidates = ['take a sad song and make it better','take a bad song made better'];
let similarities = computeSimilarity(candidates,'Tàkë á sad song and mäke it bétter');

// that might be used to find the closest match
let maxVal = Math.max(...similarities);
let closestMatch = candidates[similarities.indexOf(maxVal)];