hashish

@johnhenry/hashish implements Locality-Sensitive Hashing (LSH) for fast, scalable approximate nearest-neighbor / similarity search over text. Documents are shingled, MinHashed, and bucketed with LSH banding so that similar documents are cheap to find without comparing every pair.

Provenance: hashish is a modernized fork of agtabesh/lsh-js (previously published under that same name), rewritten in TypeScript with a fixed banding algorithm, no native dependencies, and a pluggable storage layer. See the repo’s CHANGELOG for what changed and why — including the banding bug this fork fixes.

• Zero native dependencies. MurmurHash3 is implemented in pure JS (no node-gyp), works in Node, browsers, and edge runtimes.
• Pluggable storage. In-memory by default; bring your own Redis-compatible client to share an index across processes or survive restarts.
• Correct LSH banding. Buckets are keyed by signature position, not just by value.
• TypeScript-first, ESM + CJS builds, no bundled dependencies.

Install

npm install @johnhenry/hashish

Quick example

import { Hashish } from '@johnhenry/hashish'

const hashish = new Hashish({
  shingleSize: 5,
  numberOfHashFunctions: 120,
  bucketSize: 4, // rows per LSH band; lower = more recall, higher = more precision
})

await hashish.addDocument(1, 'the quick brown fox jumps over the lazy dog')
await hashish.addDocument(2, 'the quick brown fox jumps over the lazy dog again')
await hashish.addDocument(3, 'quarterly revenue projections indicate a modest increase')

// find documents similar to document 1
const byId = await hashish.query({ id: 1 })

// re-rank LSH candidates by exact Jaccard similarity, and drop weak matches
const ranked = await hashish.query({ id: 1, rerank: true, minSimilarity: 0.3, limit: 10 })
// => [{ id: 2, similarity: 0.87 }]

All index/query methods are async, so the same code works whether storage is in-memory or a network-backed adapter like Redis.

The pages here

• API — constructor options, addDocument/query and the rest of the instance API, exporting/importing/migrating an index, and storage adapters
• How it works — shingling, MinHashing, and LSH banding, explained

Development

npm install
npm run lint
npm run typecheck
npm run test
npm run build

License

MIT

Source

github.com/johnhenry/hashish