node-symspell

JavaScript port of SymSpell 6.6 based on the original C# version by Wolf Garde and the Python version by mammothb.

Just like the Python version, this cuts out some of the C# memory optimisation which aren't really relevant in JavaScript. As a result, this port is not optimised for speed, though it's still quite fast.

This version also includes the additions of the Python version such as the ignoreToken and transferCasing options. The unit tests provided are those of the Python version which are much more comprehensive than the original ones.

This library uses the iter-tools and difflib modules which are Javascript ports of the Python modules with similar names. Because it uses async/await and async generators, it needs at least Node 12.x.

NOTE: this is still a work in progress and the API is likely to change

Basic Example

const SymSpell = require('node-symspell')

const maxEditDistance = 2
const prefixLength = 7
const symSpell = new SymSpell(maxEditDistance, prefixLength)
await symSpell.loadDictionary(dictionaryPath, 0, 1)
await symSpell.loadBigramDictionary(bigramPath, 0, 2)

const typo = 'Can yu readthis messa ge despite thehorible sppelingmsitakes'
const results = symSpell.lookupCompound(typo, maxEditDistance)

console.log(results[0])
// {
// 	term: 'can you read this message despite the horrible spelling mistakes',
// 	distance: 10,
// 	count: 0
// } 

Main API overview

constructor (maxDictionaryEditDistance = 2, prefixLength = 7, countThreshold = 1)

async loadDictionary (dictFile, termIndex, countIndex, separator = ' ')

async loadBigramDictionary (dictFile, termIndex, countIndex, separator = ' ')

lookup (input, verbosity, maxEditDistance = null, { includeUnknown, ignoreToken, transferCasing } = {})

lookupCompound (input, maxEditDistance = null, { ignoreNonWords, transferCasing } = {})

wordSegmentation (input, { maxEditDistance = null, maxSegmentationWordLength = null, ignoreToken } = {})

References

https://github.com/wolfgarbe/SymSpell
https://github.com/mammothb/symspellpy