fastest-levenshtein 🚀

Fastest JavaScript implementation of the Levenshtein distance algorithm - (up to 10x+ faster than closest competitor). This algorithm can be used to measure the similarity of two strings.

Note: if you need to take the edit distance of one string against many strings (maybe billions), I have created a GPU-implementation that will be atleast 1000x faster than this. Contact me, if you’re interested.

$ npm i fastest-levenshtein

Usage

const {distance, closest} = require('fastest-levenshtein')

// Print levenshtein-distance between 'fast' and 'faster' 
console.log(distance('fast', 'faster'))
//=> 2

// Print string from array with lowest edit-distance to 'fast'
console.log(closest('fast', ['slow', 'faster', 'fastest']))
//=> 'faster'

Benchmark

I generated 500 pairs of strings with length N. I measured the ops/sec each library achieves to process all the given pairs. Higher is better. fastest-levenshtein is a lot faster in all cases.

Test Target	N=4	N=8	N=16	N=32	N=64	N=128	N=256	N=512	N=1024
fastest-levenshtein	44423	23702	10764	4595	1049	291.5	86.64	22.24	5.473
js-levenshtein	21261	10030	2939	824	223	57.62	14.77	3.717	0.934
leven	19688	6884	1606	436	117	30.34	7.604	1.929	0.478
fast-levenshtein	18577	6112	1265	345	89.41	22.70	5.676	1.428	0.348
levenshtein-edit-distance	22968	7445	1493	409	109	28.07	7.095	1.789	0.445

Relative Performance

This image shows the relative performance between fastest-levenshtein and js-levenshtein (the 2nd fastest). fastest-levenshtein is always a lot faster. x-axis shows “times faster”.

Benchmark

License

This project is licensed under the MIT License - see the LICENSE.md file for details