@picosearch/trie v0.1.0

Trie

This package implements a simple Trie(-Map) data structure.

Installation

pnpm add @picosearch/trie

Usage

TrieMap

When using TrieMap you can pass one or more values for a key.

import { TrieMap } from '@picosearch/trie';

const trieMap = new TrieMap<number>();
trieMap.insert('abc', 1);
trieMap.insert('abd', 2);

console.log(trieMap.lookup('abc')); // [1]
console.log(trieMap.lookup('ae')); // null

Trie

When using Trie you cannot pass values for a key. Instead of lookup you can use has to check if a key exists.

import { Trie } from '@picosearch/trie';

const trie = new Trie();
trie.insert('abc');
trie.insert('abd');

console.log(trie.has('abc')); // true
console.log(trie.has('ae')); // false

toJSON + fromJSON

import { Trie } from '@picosearch/trie';

const trie = new Trie();
trie.insert('abc');
trie.insert('abd');

const jsonStr = trie.toJSON();
const newTrie = Trie.fromJSON(jsonStr);

console.log(newTrie.has('abc')); // true
console.log(newTrie.has('ae')); // false

getFuzzyMatches

import { Trie } from '@picosearch/trie';

const trie = new Trie();
trie.insert('abc');
trie.insert('abd');

console.log(trie.getFuzzyMatches('ab', { maxErrors: 1 })); // ['abc', 'abd']

Please note that a Trie is not optimized for fuzzy search and this function implements a dynamic programming approach to compute the edit distance for each path in the Trie. This approach is faster than a naive implementation, but for larger data sets you should consider using other data structures like a BK-Tree.