Upstash Vector 是一个基于 REST 的无服务器向量数据库，专为处理向量嵌入而设计。 本指南提供了 Upstash 向量存储的快速入门概述。有关所有 UpstashVectorStore 功能和配置的详细文档，请参阅 API 参考。
集成详情
要使用 Upstash 向量存储，您需要创建一个 Upstash 账户，创建一个索引，并安装 @langchain/community 集成包。您还需要安装 @upstash/vector 包作为对等依赖项。 本指南还将使用 OpenAI 嵌入，这需要您安装 @langchain/openai 集成包。您也可以使用 其他受支持的嵌入模型。npm install @langchain/community @langchain/core @upstash/vector @langchain/openai
要使用内置的 Upstash 嵌入，您需要在创建索引时选择一个嵌入模型。
process.env.UPSTASH_VECTOR_REST_URL = "your-rest-url";
process.env.UPSTASH_VECTOR_REST_TOKEN = "your-rest-token";
process.env.OPENAI_API_KEY = "YOUR_API_KEY";
// process.env.LANGSMITH_TRACING="true"
// process.env.LANGSMITH_API_KEY="your-api-key"
实例化
确保您的索引具有与您的嵌入相同的维度计数。OpenAI text-embedding-3-small 的默认值为 1536。
import { UpstashVectorStore } from "@langchain/community/vectorstores/upstash";
import { OpenAIEmbeddings } from "@langchain/openai";
import { Index } from "@upstash/vector";
const embeddings = new OpenAIEmbeddings({
model: "text-embedding-3-small",
});
const indexWithCredentials = new Index({
url: process.env.UPSTASH_VECTOR_REST_URL,
token: process.env.UPSTASH_VECTOR_REST_TOKEN,
});
const vectorStore = new UpstashVectorStore(embeddings, {
index: indexWithCredentials,
// You can use namespaces to partition your data in an index
// namespace: "test-namespace",
});
与内置嵌入的使用
要使用内置的 Upstash 嵌入，您可以将一个 FakeEmbeddings 实例传递给 UpstashVectorStore 构造函数。这将使 UpstashVectorStore 使用您在创建索引时选择的内置嵌入。
import { UpstashVectorStore } from "@langchain/community/vectorstores/upstash";
import { FakeEmbeddings } from "@langchain/core/utils/testing";
import { Index } from "@upstash/vector";
const indexWithEmbeddings = new Index({
url: process.env.UPSTASH_VECTOR_REST_URL,
token: process.env.UPSTASH_VECTOR_REST_TOKEN,
});
const vectorStore = new UpstashVectorStore(new FakeEmbeddings(), {
index: indexWithEmbeddings,
});
管理向量存储
向向量存储添加项目
import type { Document } from "@langchain/core/documents";
const document1: Document = {
pageContent: "The powerhouse of the cell is the mitochondria",
metadata: { source: "https://example.com" }
};
const document2: Document = {
pageContent: "Buildings are made out of brick",
metadata: { source: "https://example.com" }
};
const document3: Document = {
pageContent: "Mitochondria are made out of lipids",
metadata: { source: "https://example.com" }
};
const document4: Document = {
pageContent: "The 2024 Olympics are in Paris",
metadata: { source: "https://example.com" }
}
const documents = [document1, document2, document3, document4];
await vectorStore.addDocuments(documents, { ids: ["1", "2", "3", "4"] });
从向量存储中删除项目
await vectorStore.delete({ ids: ["4"] });
查询向量存储
一旦您的向量存储被创建并添加了相关文档，您很可能希望在链或代理运行期间查询它。
直接查询
执行简单的相似性搜索可以按如下方式完成
const filter = "source = 'https://example.com'";
const similaritySearchResults = await vectorStore.similaritySearch("biology", 2, filter);
for (const doc of similaritySearchResults) {
console.log(`* ${doc.pageContent} [${JSON.stringify(doc.metadata, null)}]`);
}
* The powerhouse of the cell is the mitochondria [{"source":"https://example.com"}]
* Mitochondria are made out of lipids [{"source":"https://example.com"}]
const similaritySearchWithScoreResults = await vectorStore.similaritySearchWithScore("biology", 2, filter)
for (const [doc, score] of similaritySearchWithScoreResults) {
console.log(`* [SIM=${score.toFixed(3)}] ${doc.pageContent} [${JSON.stringify(doc.metadata)}]`);
}
* [SIM=0.576] The powerhouse of the cell is the mitochondria [{"source":"https://example.com"}]
* [SIM=0.557] Mitochondria are made out of lipids [{"source":"https://example.com"}]
通过转换为检索器进行查询
您还可以将向量存储转换为检索器，以便在您的链中更轻松地使用。
const retriever = vectorStore.asRetriever({
// Optional filter
filter: filter,
k: 2,
});
await retriever.invoke("biology");
[
Document {
pageContent: 'The powerhouse of the cell is the mitochondria',
metadata: { source: 'https://example.com' },
id: undefined
},
Document {
pageContent: 'Mitochondria are made out of lipids',
metadata: { source: 'https://example.com' },
id: undefined
}
]
用于检索增强生成的使用
有关如何将此向量存储用于检索增强生成 (RAG) 的指南，请参阅以下部分
API 参考
有关所有 UpstashVectorStore 功能和配置的详细文档，请参阅 API 参考。
