lancedb/nodejs

LanceDB JavaScript SDK

A JavaScript library for LanceDB.

Installation

npm install @lancedb/lancedb

This will download the appropriate native library for your platform. We currently support:

• Linux (x86_64 and aarch64 on glibc and musl)
• MacOS (Intel and ARM/M1/M2)
• Windows (x86_64 and aarch64)

Usage

Basic Example

import * as lancedb from "@lancedb/lancedb";
const db = await lancedb.connect("data/sample-lancedb");
const table = await db.createTable("my_table", [
  { id: 1, vector: [0.1, 1.0], item: "foo", price: 10.0 },
  { id: 2, vector: [3.9, 0.5], item: "bar", price: 20.0 },
]);
const results = await table.vectorSearch([0.1, 0.3]).limit(20).toArray();
console.log(results);

Use an Existing Table with LangChain

When wrapping an existing table with @langchain/community, open the table with LanceDB and pass the resulting table handle to LangChain. The LangChain uri and tableName options are used when creating a table; they do not open an existing table for search.

import { LanceDB as LangChainLanceDB } from "@langchain/community/vectorstores/lancedb";
import * as lancedb from "@lancedb/lancedb";

const db = await lancedb.connect("data/sample-lancedb");
const table = await db.openTable("my_table");
const vectorStore = new LangChainLanceDB(embeddings, {
  table,
  textKey: "item",
});

const documents = await vectorStore.similaritySearch("foo", 5);

Affected versions of @langchain/community do not map LanceDB's _distance column to the score returned by similaritySearchVectorWithScore. Query the table directly when you need the numeric vector distance:

const results = await table.vectorSearch(queryVector).limit(5).toArray();
console.log(results[0]._distance);

The quickstart contains more complete examples.

Development

See CONTRIBUTING.md for information on how to contribute to LanceDB.