NGramTokenFilter

Generates n-grams from tokens at the token level (after tokenization).

Import

import NGramTokenFilter from 'dynamosearch/filters/NGramTokenFilter';

Constructor

new NGramTokenFilter(options?: { minGram?: number; maxGram?: number })

Parameters

• minGram (number, optional) - Minimum n-gram size (default: 1)
• maxGram (number, optional) - Maximum n-gram size (default: 2)

Example

const filter = new NGramTokenFilter({ minGram: 2, maxGram: 3 });
const tokens = filter.apply([
  { token: 'hello', startOffset: 0, endOffset: 5, position: 0 },
]);
// Generates n-grams: he, el, ll, lo, hel, ell, llo

Best For

• Partial token matching
• Autocomplete
• Fuzzy search

See Also

• NGramTokenizer - For character-level n-grams
• EdgeNGramTokenFilter - For edge n-grams