LangChain 官方中文文档

概述 (Overview)

向量存储 (vector store) 用于存储嵌入 (embedded)数据并执行相似性搜索。

mermaid

flowchart LR

    subgraph "📥 索引阶段 (Indexing phase)"
        A[📄 文档] --> B[🔢 嵌入模型]
        B --> C[🔘 嵌入向量]
        C --> D[(向量存储)]
    end

    subgraph "📤 查询阶段 (Query phase)"
        E[❓ 查询文本] --> F[🔢 嵌入模型]
        F --> G[🔘 查询向量]
        G --> H[🔍 相似性搜索]
        H --> D
        D --> I[📄 Top-k 结果]
    end

接口 (Interface)

LangChain 为向量存储提供了统一的接口，允许您：

addDocuments - 向存储中添加文档。
delete - 根据 ID 删除已存储的文档。
similaritySearch - 查询语义相似的文档。

这种抽象让您可以在不同的实现之间切换，而无需更改应用程序逻辑。

初始化 (Initialization)

LangChain 中的大多数向量存储在初始化时接受一个嵌入模型作为参数。

typescript

import { OpenAIEmbeddings } from "@langchain/openai";
import { MemoryVectorStore } from "@langchain/classic/vectorstores/memory";

const embeddings = new OpenAIEmbeddings({
  model: "text-embedding-3-small",
});
const vectorStore = new MemoryVectorStore(embeddings);

添加文档 (Adding documents)

您可以使用 addDocuments 函数向向量存储添加文档。

typescript

import { Document } from "@langchain/core/documents";
const document = new Document({
  pageContent: "Hello world",
});
await vectorStore.addDocuments([document]);

删除文档 (Deleting documents)

您可以使用 delete 函数从向量存储中删除文档。

typescript

await vectorStore.delete({
  filter: {
    pageContent: "Hello world",
  },
});

相似性搜索 (Similarity search)

使用 similaritySearch 发出语义查询，它将返回最接近的嵌入文档：

typescript

const results = await vectorStore.similaritySearch("Hello world", 10);

许多向量存储支持以下参数：

k — 要返回的结果数量
filter — 基于元数据的条件过滤

相似性度量与索引 (Similarity metrics & indexing)

嵌入相似性可以使用以下方法计算：

余弦相似度 (Cosine similarity)
欧几里得距离 (Euclidean distance)
点积 (Dot product)

高效的搜索通常使用索引方法，例如 HNSW（分层可导航小世界），但具体细节取决于向量存储。

元数据过滤 (Metadata filtering)

通过元数据（例如来源、日期）进行过滤可以优化搜索结果：

typescript

vectorStore.similaritySearch("query", 2, { source: "tweets" });

主要集成 (Top integrations)

选择嵌入模型：

OpenAI

安装依赖项：

npmyarnpnpm

bash

npm i @langchain/openai

bash

yarn add @langchain/openai

bash

pnpm add @langchain/openai

添加环境变量：

bash

OPENAI_API_KEY=your-api-key

实例化模型：

typescript

import { OpenAIEmbeddings } from "@langchain/openai";

const embeddings = new OpenAIEmbeddings({
  model: "text-embedding-3-large"
});

Azure

安装依赖项：

npmyarnpnpm

bash

npm i @langchain/openai

bash

yarn add @langchain/openai

bash

pnpm add @langchain/openai

添加环境变量：

bash

AZURE_OPENAI_API_INSTANCE_NAME=<YOUR_INSTANCE_NAME>
AZURE_OPENAI_API_KEY=<YOUR_KEY>
AZURE_OPENAI_API_VERSION="2024-02-01"

实例化模型：

typescript

import { AzureOpenAIEmbeddings } from "@langchain/openai";

const embeddings = new AzureOpenAIEmbeddings({
  azureOpenAIApiEmbeddingsDeploymentName: "text-embedding-ada-002"
});

AWS

安装依赖项：

npmyarnpnpm

bash

npm i @langchain/aws

bash

yarn add @langchain/aws

bash

pnpm add @langchain/aws

添加环境变量：

bash

BEDROCK_AWS_REGION=your-region

实例化模型：

typescript

import { BedrockEmbeddings } from "@langchain/aws";

const embeddings = new BedrockEmbeddings({
  model: "amazon.titan-embed-text-v1"
});

Google Gemini

安装依赖项：

npmyarnpnpm

bash

npm i @langchain/google-genai

bash

yarn add @langchain/google-genai

bash

pnpm add @langchain/google-genai

添加环境变量：

bash

GOOGLE_API_KEY=your-api-key

实例化模型：

typescript

import { GoogleGenerativeAIEmbeddings } from "@langchain/google-genai";

const embeddings = new GoogleGenerativeAIEmbeddings({
  model: "text-embedding-004"
});

Google Vertex

安装依赖项：

npmyarnpnpm

bash

npm i @langchain/google-vertexai

bash

yarn add @langchain/google-vertexai

bash

pnpm add @langchain/google-vertexai

添加环境变量：

bash

GOOGLE_APPLICATION_CREDENTIALS=credentials.json

实例化模型：

typescript

import { VertexAIEmbeddings } from "@langchain/google-vertexai";

const embeddings = new VertexAIEmbeddings({
  model: "gemini-embedding-001"
});

MistralAI

安装依赖项：

npmyarnpnpm

bash

npm i @langchain/mistralai

bash

yarn add @langchain/mistralai

bash

pnpm add @langchain/mistralai

添加环境变量：

bash

MISTRAL_API_KEY=your-api-key

实例化模型：

typescript

import { MistralAIEmbeddings } from "@langchain/mistralai";

const embeddings = new MistralAIEmbeddings({
  model: "mistral-embed"
});

Cohere

安装依赖项：

npmyarnpnpm

bash

npm i @langchain/cohere

bash

yarn add @langchain/cohere

bash

pnpm add @langchain/cohere

添加环境变量：

bash

COHERE_API_KEY=your-api-key

实例化模型：

typescript

import { CohereEmbeddings } from "@langchain/cohere";

const embeddings = new CohereEmbeddings({
  model: "embed-english-v3.0"
});

Ollama

安装依赖项：

npmyarnpnpm

bash

npm i @langchain/ollama

bash

yarn add @langchain/ollama

bash

pnpm add @langchain/ollama

实例化模型：

typescript

import { OllamaEmbeddings } from "@langchain/ollama";

const embeddings = new OllamaEmbeddings({
  model: "llama2",
  baseUrl: "http://localhost:11434", // Default value
});

选择向量存储：

Memory

bashyarnpnpm

bash

npm i langchain

bash

yarn add langchain

bash

pnpm add langchain

typescript

import { MemoryVectorStore } from "@langchain/classic/vectorstores/memory";

const vectorStore = new MemoryVectorStore(embeddings);

Chroma

npmyarnpnpm

bash

npm i @langchain/community

bash

yarn add @langchain/community

bash

pnpm add @langchain/community

typescript

import { Chroma } from "@langchain/community/vectorstores/chroma";

const vectorStore = new Chroma(embeddings, {
  collectionName: "a-test-collection",
});

FAISS

npmyarnbash

bash

npm i @langchain/community

bash

yarn add @langchain/community

bash

pnpm add @langchain/community

typescript

import { FaissStore } from "@langchain/community/vectorstores/faiss";

const vectorStore = new FaissStore(embeddings, {});

MongoDB

npmyarnpnpm

bash

npm i @langchain/mongodb

bash

yarn add @langchain/mongodb

bash

pnpm add @langchain/mongodb

typescript

import { MongoDBAtlasVectorSearch } from "@langchain/mongodb"
import { MongoClient } from "mongodb";

const client = new MongoClient(process.env.MONGODB_ATLAS_URI || "");
const collection = client
  .db(process.env.MONGODB_ATLAS_DB_NAME)
  .collection(process.env.MONGODB_ATLAS_COLLECTION_NAME);

const vectorStore = new MongoDBAtlasVectorSearch(embeddings, {
  collection,
  indexName: "vector_index",
  textKey: "text",
  embeddingKey: "embedding",
});

PGVector

npmyarnpnpm

bash

npm i @langchain/community

bash

yarn add @langchain/community

bash

pnpm add @langchain/community

typescript

import { PGVectorStore } from "@langchain/community/vectorstores/pgvector";

const vectorStore = await PGVectorStore.initialize(embeddings, {});

Pinecone

npmyarnpnpm

bash

npm i @langchain/pinecone

bash

yarn add @langchain/pinecone

bash

pnpm add @langchain/pinecone

typescript

import { PineconeStore } from "@langchain/pinecone";
import { Pinecone as PineconeClient } from "@pinecone-database/pinecone";

const pinecone = new PineconeClient();
const vectorStore = new PineconeStore(embeddings, {
  pineconeIndex,
  maxConcurrency: 5,
});

Qdrant

npmyarnpnpm

bash

npm i @langchain/qdrant

bash

yarn add @langchain/qdrant

bash

pnpm add @langchain/qdrant

typescript

import { QdrantVectorStore } from "@langchain/qdrant";

const vectorStore = await QdrantVectorStore.fromExistingCollection(embeddings, {
  url: process.env.QDRANT_URL,
  collectionName: "langchainjs-testing",
});

Weaviate

npmyarnpnpm

bash

npm i @langchain/weaviate

bash

yarn add @langchain/weaviate

bash

pnpm add @langchain/weaviate

typescript

import { WeaviateStore } from "@langchain/weaviate";

const vectorStore = new WeaviateStore(embeddings, {
    client: weaviateClient,
    indexName: "Langchainjs_test",
});

LangChain.js 集成了多种向量存储。您可以在下方查看完整列表：

所有向量存储 (All vector stores)

向量存储 (Vector stores)

概述 (Overview) ​

接口 (Interface) ​

初始化 (Initialization) ​

添加文档 (Adding documents) ​

删除文档 (Deleting documents) ​

相似性搜索 (Similarity search) ​

相似性度量与索引 (Similarity metrics & indexing) ​

元数据过滤 (Metadata filtering) ​

主要集成 (Top integrations) ​

所有向量存储 (All vector stores) ​

AnalyticDB

Astra DB

Azion EdgeSQL

Azure AI Search

Azure Cosmos DB for MongoDB vCore

Azure Cosmos DB for NoSQL

Cassandra

Chroma

ClickHouse

CloseVector

Cloudflare Vectorize

Convex

Couchbase Query

Couchbase Search

Elasticsearch

Faiss

Google Cloud SQL for PostgreSQL

Google Vertex AI Matching Engine

SAP HANA Cloud Vector Engine

HNSWLib

LanceDB

libSQL

MariaDB

In-memory

Milvus

Momento Vector Index (MVI)

MongoDB Atlas

MyScale

Neo4j Vector Index

Neon Postgres

OpenSearch

PGVector

Pinecone

Prisma

Qdrant

Redis

Rockset

SingleStore

Supabase

Tigris

Turbopuffer

TypeORM

Typesense

Upstash Vector

USearch

Vectara

Vercel Postgres

Voy

Weaviate

Xata

Zep Open Source

Zep Cloud

概述 (Overview)

接口 (Interface)

初始化 (Initialization)

添加文档 (Adding documents)

删除文档 (Deleting documents)

相似性搜索 (Similarity search)

相似性度量与索引 (Similarity metrics & indexing)

元数据过滤 (Metadata filtering)

主要集成 (Top integrations)

所有向量存储 (All vector stores)