使用 LangGraph 构建自定义 RAG 代理

概述

在本教程中，我们将使用 LangGraph 构建一个检索代理。 LangChain 提供了内置的代理实现，这些实现是使用 LangGraph 原语构建的。如果需要更深入的定制，可以直接在 LangGraph 中实现代理。本指南演示了检索代理的一个示例实现。检索代理在您希望 LLM 决定是否从向量存储中检索上下文或直接响应用户时非常有用。在本教程结束时，我们将完成以下操作：

获取并预处理将用于检索的文档。
索引这些文档以进行语义搜索并为代理创建检索器工具。
构建一个可以决定何时使用检索器工具的代理式 RAG 系统。

概念

我们将涵盖以下概念：

使用文档加载器、文本分割器、嵌入和向量存储进行检索
LangGraph 图 API，包括状态、节点、边和条件边。

设置

让我们下载所需的包并设置我们的 API 密钥：

npm install @langchain/langgraph @langchain/openai @langchain/community @langchain/textsplitters

1. 预处理文档

获取将在我们的 RAG 系统中使用的文档。我们将使用 Lilian Weng 的优秀博客上最新的三个页面。我们首先使用 CheerioWebBaseLoader 获取页面的内容：

import { CheerioWebBaseLoader } from "@langchain/community/document_loaders/web/cheerio";

const urls = [
  "https://lilianweng.github.io/posts/2023-06-23-agent/",
  "https://lilianweng.github.io/posts/2023-03-15-prompt-engineering/",
  "https://lilianweng.github.io/posts/2023-10-25-adv-attack-llm/",
];

const docs = await Promise.all(
  urls.map((url) => new CheerioWebBaseLoader(url).load()),
);

将获取的文档拆分为较小的块以便索引到我们的向量存储中：

import { RecursiveCharacterTextSplitter } from "@langchain/textsplitters";

const docsList = docs.flat();

const textSplitter = new RecursiveCharacterTextSplitter({
  chunkSize: 500,
  chunkOverlap: 50,
});
const docSplits = await textSplitter.splitDocuments(docsList);

2. 创建检索器工具

现在我们已经有了拆分好的文档，我们可以将它们索引到一个向量存储中，我们将用它来进行语义搜索。

使用内存中的向量存储和 OpenAI 嵌入：

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

const vectorStore = await MemoryVectorStore.fromDocuments(
  docSplits,
  new OpenAIEmbeddings(),
);

const retriever = vectorStore.asRetriever();

使用 LangChain 预建的 createRetrieverTool 创建检索器工具：

import { createRetrieverTool } from "@langchain/classic/tools/retriever";

const tool = createRetrieverTool(
  retriever,
  {
    name: "retrieve_blog_posts",
    description:
      "Search and return information about Lilian Weng blog posts on LLM agents, prompt engineering, and adversarial attacks on LLMs.",
  },
);
const tools = [tool];

3. 生成查询

现在我们将开始为我们的代理式 RAG 图构建组件（节点和边）。

构建 generateQueryOrRespond 节点。它将调用 LLM 基于当前图状态（消息列表）生成响应。给定输入消息，它将决定使用检索器工具进行检索，或直接响应用户。请注意，我们通过 .bindTools 给予聊天模型访问之前创建的 tools 的权限：

import { ChatOpenAI } from "@langchain/openai";
import { GraphNode } from "@langchain/langgraph";

const generateQueryOrRespond: GraphNode<typeof State> = async (state) => {
  const model = new ChatOpenAI({
    model: "gpt-4.1",
    temperature: 0,
  }).bindTools(tools);

  const response = await model.invoke(state.messages);
  return {
    messages: [response],
  };
}

尝试随机输入：

import { HumanMessage } from "@langchain/core/messages";

const input = { messages: [new HumanMessage("hello!")] };
const result = await generateQueryOrRespond(input);
console.log(result.messages[0]);

输出：

AIMessage {
  content: "Hello! How can I help you today?",
  tool_calls: []
}

询问需要语义搜索的问题：

const input = {
  messages: [
    new HumanMessage("What does Lilian Weng say about types of reward hacking?")
  ]
};
const result = await generateQueryOrRespond(input);
console.log(result.messages[0]);

输出：

AIMessage {
  content: "",
  tool_calls: [
    {
      name: "retrieve_blog_posts",
      args: { query: "types of reward hacking" },
      id: "call_...",
      type: "tool_call"
    }
  ]
}

4. 评估文档

添加一个节点—gradeDocuments—以确定检索到的文档是否与问题相关。我们将使用具有结构化输出的模型（使用 Zod）进行文档评估。我们还将添加一个条件边—checkRelevance—检查评估结果并返回要前往的节点名称（generate 或 rewrite）：

import * as z from "zod";
import { ChatPromptTemplate } from "@langchain/core/prompts";
import { ChatOpenAI } from "@langchain/openai";
import { GraphNode } from "@langchain/langgraph";
import { AIMessage } from "@langchain/core/messages";

const prompt = ChatPromptTemplate.fromTemplate(
  `You are a grader assessing relevance of retrieved docs to a user question.
  Here are the retrieved docs:
  \n ------- \n
  {context}
  \n ------- \n
  Here is the user question: {question}
  If the content of the docs are relevant to the users question, score them as relevant.
  Give a binary score 'yes' or 'no' score to indicate whether the docs are relevant to the question.
  Yes: The docs are relevant to the question.
  No: The docs are not relevant to the question.`,
);

const gradeDocumentsSchema = z.object({
  binaryScore: z.string().describe("Relevance score 'yes' or 'no'"),
})

const gradeDocuments: GraphNode<typeof State> = async (state) => {
  const model = new ChatOpenAI({
    model: "gpt-4.1",
    temperature: 0,
  }).withStructuredOutput(gradeDocumentsSchema);

  const score = await prompt.pipe(model).invoke({
    question: state.messages.at(0)?.content,
    context: state.messages.at(-1)?.content,
  });

  if (score.binaryScore === "yes") {
    return "generate";
  }
  return "rewrite";
}

在工具响应中使用不相关的文档运行此代码：

import { ToolMessage } from "@langchain/core/messages";

const input = {
  messages: [
    new HumanMessage("What does Lilian Weng say about types of reward hacking?"),
    new AIMessage({
      tool_calls: [
        {
          type: "tool_call",
          name: "retrieve_blog_posts",
          args: { query: "types of reward hacking" },
          id: "1",
        }
      ]
    }),
    new ToolMessage({
      content: "meow",
      tool_call_id: "1",
    })
  ]
}
const result = await gradeDocuments(input);

确认相关文档被分类为相关：

const input = {
  messages: [
    new HumanMessage("What does Lilian Weng say about types of reward hacking?"),
    new AIMessage({
      tool_calls: [
        {
          type: "tool_call",
          name: "retrieve_blog_posts",
          args: { query: "types of reward hacking" },
          id: "1",
        }
      ]
    }),
    new ToolMessage({
      content: "reward hacking can be categorized into two types: environment or goal misspecification, and reward tampering",
      tool_call_id: "1",
    })
  ]
}
const result = await gradeDocuments(input);

5. 重写问题

构建 rewrite 节点。检索器工具可能会返回可能不相关的文档，这表明需要改进原始用户问题。为此，我们将调用 rewrite 节点：

import { ChatPromptTemplate } from "@langchain/core/prompts";
import { ChatOpenAI } from "@langchain/openai";
import { GraphNode } from "@langchain/langgraph";

const rewritePrompt = ChatPromptTemplate.fromTemplate(
  `Look at the input and try to reason about the underlying semantic intent / meaning. \n
  Here is the initial question:
  \n ------- \n
  {question}
  \n ------- \n
  Formulate an improved question:`,
);

const rewrite: GraphNode<typeof State> = async (state) => {
  const question = state.messages.at(0)?.content;

  const model = new ChatOpenAI({
    model: "gpt-4.1",
    temperature: 0,
  });

  const response = await rewritePrompt.pipe(model).invoke({ question });
  return {
    messages: [response],
  };
}

尝试一下：

import { HumanMessage, AIMessage, ToolMessage } from "@langchain/core/messages";

const input = {
  messages: [
    new HumanMessage("What does Lilian Weng say about types of reward hacking?"),
    new AIMessage({
      content: "",
      tool_calls: [
        {
          id: "1",
          name: "retrieve_blog_posts",
          args: { query: "types of reward hacking" },
          type: "tool_call"
        }
      ]
    }),
    new ToolMessage({ content: "meow", tool_call_id: "1" })
  ]
};

const response = await rewrite(input);
console.log(response.messages[0].content);

输出：

What are the different types of reward hacking described by Lilian Weng, and how does she explain them?

6. 生成答案

构建 generate 节点：如果我们通过了评估器的检查，我们可以基于原始问题和检索到的上下文生成最终答案：

import { ChatPromptTemplate } from "@langchain/core/prompts";
import { ChatOpenAI } from "@langchain/openai";
import { GraphNode } from "@langchain/langgraph";

const generate: GraphNode<typeof State> = async (state) => {
  const question = state.messages.at(0)?.content;
  const context = state.messages.at(-1)?.content;

  const prompt = ChatPromptTemplate.fromTemplate(
  `You are an assistant for question-answering tasks.
      Use the following pieces of retrieved context to answer the question.
      If you don't know the answer, just say that you don't know.
      Use three sentences maximum and keep the answer concise.
      Question: {question}
      Context: {context}`
  );

  const llm = new ChatOpenAI({
    model: "gpt-4.1",
    temperature: 0,
  });

  const ragChain = prompt.pipe(llm);

  const response = await ragChain.invoke({
    context,
    question,
  });

  return {
    messages: [response],
  };
}

尝试一下：

import { HumanMessage, AIMessage, ToolMessage } from "@langchain/core/messages";

const input = {
  messages: [
    new HumanMessage("What does Lilian Weng say about types of reward hacking?"),
    new AIMessage({
      content: "",
      tool_calls: [
        {
          id: "1",
          name: "retrieve_blog_posts",
          args: { query: "types of reward hacking" },
          type: "tool_call"
        }
      ]
    }),
    new ToolMessage({
      content: "reward hacking can be categorized into two types: environment or goal misspecification, and reward tampering",
      tool_call_id: "1"
    })
  ]
};

const response = await generate(input);
console.log(response.messages[0].content);

输出：

Lilian Weng categorizes reward hacking into two types: environment or goal misspecification, and reward tampering. She considers reward hacking as a broad concept that includes both of these categories. Reward hacking occurs when an agent exploits flaws or ambiguities in the reward function to achieve high rewards without performing the intended behaviors.

7. 组装图

现在我们将所有节点和边组装成一个完整的图：

从 generateQueryOrRespond 开始，确定是否需要调用检索器工具
使用条件边路由到下一步：
- 如果 generateQueryOrRespond 返回了 tool_calls，调用检索器工具检索上下文
- 否则，直接响应用户
评估检索到的文档内容与问题的相关性（gradeDocuments）并路由到下一步：
- 如果不相关，使用 rewrite 重写问题，然后再次调用 generateQueryOrRespond
- 如果相关，继续执行 generate 并使用带检索文档上下文的 ToolMessage 生成最终响应

import { StateGraph, START, END, ConditionalEdgeRouter } from "@langchain/langgraph";
import { ToolNode } from "@langchain/langgraph/prebuilt";
import { AIMessage } from "langchain";

// Create a ToolNode for the retriever
const toolNode = new ToolNode(tools);

// Helper function to determine if we should retrieve
const shouldRetrieve: ConditionalEdgeRouter<typeof State, "retrieve"> = (state) => {
  const lastMessage = state.messages.at(-1);
  if (AIMessage.isInstance(lastMessage) && lastMessage.tool_calls.length) {
    return "retrieve";
  }
  return END;
}

// Define the graph
const builder = new StateGraph(GraphState)
  .addNode("generateQueryOrRespond", generateQueryOrRespond)
  .addNode("retrieve", toolNode)
  .addNode("gradeDocuments", gradeDocuments)
  .addNode("rewrite", rewrite)
  .addNode("generate", generate)
  // Add edges
  .addEdge(START, "generateQueryOrRespond")
  // Decide whether to retrieve
  .addConditionalEdges("generateQueryOrRespond", shouldRetrieve)
  .addEdge("retrieve", "gradeDocuments")
  // Edges taken after grading documents
  .addConditionalEdges(
    "gradeDocuments",
    // Route based on grading decision
    (state) => {
      // The gradeDocuments function returns either "generate" or "rewrite"
      const lastMessage = state.messages.at(-1);
      return lastMessage.content === "generate" ? "generate" : "rewrite";
    }
  )
  .addEdge("generate", END)
  .addEdge("rewrite", "generateQueryOrRespond");

// Compile
const graph = builder.compile();

8. 运行代理式 RAG

现在让我们通过用问题运行它来测试完整的图：

import { HumanMessage } from "@langchain/core/messages";

const inputs = {
  messages: [
    new HumanMessage("What does Lilian Weng say about types of reward hacking?")
  ]
};

for await (const output of await graph.stream(inputs)) {
  for (const [key, value] of Object.entries(output)) {
    const lastMsg = output[key].messages[output[key].messages.length - 1];
    console.log(`Output from node: '${key}'`);
    console.log({
      type: lastMsg._getType(),
      content: lastMsg.content,
      tool_calls: lastMsg.tool_calls,
    });
    console.log("---\n");
  }
}

输出：

Output from node: 'generateQueryOrRespond'
{
  type: 'ai',
  content: '',
  tool_calls: [
    {
      name: 'retrieve_blog_posts',
      args: { query: 'types of reward hacking' },
      id: 'call_...',
      type: 'tool_call'
    }
  ]
}
---

Output from node: 'retrieve'
{
  type: 'tool',
  content: '(Note: Some work defines reward tampering as a distinct category...\n' +
    'At a high level, reward hacking can be categorized into two types: environment or goal misspecification, and reward tampering.\n' +
    '...',
  tool_calls: undefined
}
---

Output from node: 'generate'
{
  type: 'ai',
  content: 'Lilian Weng categorizes reward hacking into two types: environment or goal misspecification, and reward tampering. She considers reward hacking as a broad concept that includes both of these categories. Reward hacking occurs when an agent exploits flaws or ambiguities in the reward function to achieve high rewards without performing the intended behaviors.',
  tool_calls: []
}
---

Edit this page on GitHub or file an issue.

Connect these docs to Claude, VSCode, and more via MCP for real-time answers.

Tutorials

Conceptual overviews

LangChain Academy

Additional resources

使用 LangGraph 构建自定义 RAG 代理

概述

概念

设置

1. 预处理文档

2. 创建检索器工具

3. 生成查询

4. 评估文档

5. 重写问题

6. 生成答案

7. 组装图

8. 运行代理式 RAG

Tutorials

Conceptual overviews

LangChain Academy

Additional resources

Documentation Index

​概述

​概念

​设置

​1. 预处理文档

​2. 创建检索器工具

​3. 生成查询

​4. 评估文档

​5. 重写问题

​6. 生成答案

​7. 组装图

​8. 运行代理式 RAG

概述

概念

设置

1. 预处理文档

2. 创建检索器工具

3. 生成查询

4. 评估文档

5. 重写问题

6. 生成答案

7. 组装图

8. 运行代理式 RAG