跳至主要内容

LanceDB

LanceDB 是一个用于 AI 应用程序的嵌入式向量数据库。它是开源的，并以 Apache-2.0 许可证分发。

LanceDB 数据集会持久化到磁盘，并且可以在 Node.js 和 Python 之间共享。

安装​

安装 LanceDB Node.js 绑定

npm install -S vectordb
提示

查看 有关安装集成包的通用说明的此部分。

npm install @langchain/openai @langchain/community @langchain/core

用法​

从文本创建新索引​

import { LanceDB } from "@langchain/community/vectorstores/lancedb";
import { OpenAIEmbeddings } from "@langchain/openai";
import * as fs from "node:fs/promises";
import * as path from "node:path";
import os from "node:os";

export const run = async () => {
  const vectorStore = await LanceDB.fromTexts(
    ["Hello world", "Bye bye", "hello nice world"],
    [{ id: 2 }, { id: 1 }, { id: 3 }],
    new OpenAIEmbeddings()
  );

  const resultOne = await vectorStore.similaritySearch("hello world", 1);
  console.log(resultOne);
  // [ Document { pageContent: 'hello nice world', metadata: { id: 3 } } ]
};

export const run_with_existing_table = async () => {
  const dir = await fs.mkdtemp(path.join(os.tmpdir(), "lancedb-"));
  const vectorStore = await LanceDB.fromTexts(
    ["Hello world", "Bye bye", "hello nice world"],
    [{ id: 2 }, { id: 1 }, { id: 3 }],
    new OpenAIEmbeddings()
  );

  const resultOne = await vectorStore.similaritySearch("hello world", 1);
  console.log(resultOne);
  // [ Document { pageContent: 'hello nice world', metadata: { id: 3 } } ]
};

API 参考

从加载器创建新索引​

import { LanceDB } from "@langchain/community/vectorstores/lancedb";
import { OpenAIEmbeddings } from "@langchain/openai";
import { TextLoader } from "langchain/document_loaders/fs/text";
import fs from "node:fs/promises";
import path from "node:path";
import os from "node:os";

// Create docs with a loader
const loader = new TextLoader("src/document_loaders/example_data/example.txt");
const docs = await loader.load();

export const run = async () => {
  const vectorStore = await LanceDB.fromDocuments(docs, new OpenAIEmbeddings());

  const resultOne = await vectorStore.similaritySearch("hello world", 1);
  console.log(resultOne);

  // [
  //   Document {
  //     pageContent: 'Foo\nBar\nBaz\n\n',
  //     metadata: { source: 'src/document_loaders/example_data/example.txt' }
  //   }
  // ]
};

export const run_with_existing_table = async () => {
  const dir = await fs.mkdtemp(path.join(os.tmpdir(), "lancedb-"));

  const vectorStore = await LanceDB.fromDocuments(docs, new OpenAIEmbeddings());

  const resultOne = await vectorStore.similaritySearch("hello world", 1);
  console.log(resultOne);

  // [
  //   Document {
  //     pageContent: 'Foo\nBar\nBaz\n\n',
  //     metadata: { source: 'src/document_loaders/example_data/example.txt' }
  //   }
  // ]
};

API 参考

打开现有数据集​

import { LanceDB } from "@langchain/community/vectorstores/lancedb";
import { OpenAIEmbeddings } from "@langchain/openai";
import { connect } from "vectordb";
import * as fs from "node:fs/promises";
import * as path from "node:path";
import os from "node:os";

//
//  You can open a LanceDB dataset created elsewhere, such as LangChain Python, by opening
//     an existing table
//
export const run = async () => {
  const uri = await createdTestDb();
  const db = await connect(uri);
  const table = await db.openTable("vectors");

  const vectorStore = new LanceDB(new OpenAIEmbeddings(), { table });

  const resultOne = await vectorStore.similaritySearch("hello world", 1);
  console.log(resultOne);
  // [ Document { pageContent: 'Hello world', metadata: { id: 1 } } ]
};

async function createdTestDb(): Promise<string> {
  const dir = await fs.mkdtemp(path.join(os.tmpdir(), "lancedb-"));
  const db = await connect(dir);
  await db.createTable("vectors", [
    { vector: Array(1536), text: "Hello world", id: 1 },
    { vector: Array(1536), text: "Bye bye", id: 2 },
    { vector: Array(1536), text: "hello nice world", id: 3 },
  ]);
  return dir;
}

API 参考

此页面是否有帮助？


您也可以在 GitHub 上留下详细的反馈 GitHub.