@nura-js/plugin-fuzzy

Fuzzy, phonetic, and token-based matching utilities shared by Nura.js and its voice tooling.

Installation

pnpm add @nura-js/plugin-fuzzy

Usage Example

import { matchFuzzy, tokenizeAndScore, compareWakeWord } from '@nura-js/plugin-fuzzy';

const brands = ['Nura', 'Núria', 'Nero'];
const match = matchFuzzy('nura', brands, { locale: 'es' });

const tokens = tokenizeAndScore('abre el modo noche', ['modo noche', 'modo día']);

const wake = compareWakeWord('hey nura', {
  canonical: 'hey nura',
  aliases: ['hola nura'],
});

Key APIs

• matchFuzzy scores candidates using the selected strategy and locale heuristics.
• tokenizeAndScore evaluates each token against candidate phrases.
• compareWakeWord checks wake-word candidates for hybrid phonetic similarity.
• damerauLevenshteinSimilarity exposes the raw edit-distance helper.

Type References

• FuzzyMatchOpts — strategy, locale, and threshold options.
• MatchResult — output from matchFuzzy including the winning candidate.
• TokenScore — granular token scoring result used by tokenizeAndScore.
• FuzzyStrategy — supported strategies ('hybrid', 'damerau', ...).

Additional Resources

• Repository: https://github.com/nura-ia/nurajs
• Issues: https://github.com/nura-ia/nurajs/issues