构建一个检索增强生成（RAG）应用程序：第一部分

大型语言模型（LLM）实现的最强大应用之一是复杂的问答（Q&A）聊天机器人。这些应用程序可以回答关于特定源信息的问题。这些应用程序使用一种称为检索增强生成（Retrieval Augmented Generation），或 RAG 的技术。

这是一个多部分的教程

第一部分（本指南）介绍 RAG 并讲解一个最小的实现。
第二部分扩展了实现，以适应对话式交互和多步检索过程。

本教程将展示如何构建一个基于文本数据源的简单问答应用程序。在此过程中，我们将介绍典型的问答架构，并重点介绍用于更高级问答技术的额外资源。我们还将了解 LangSmith 如何帮助我们追踪和理解应用程序。随着应用程序复杂性的增加，LangSmith 将变得越来越有用。

如果您已经熟悉基本的检索，您可能也会对这篇关于不同检索技术的高级概述感兴趣。

注意：这里我们专注于非结构化数据的问答。如果您对结构化数据上的 RAG 感兴趣，请查看我们关于 SQL 数据问答的教程。

概述

一个典型的 RAG 应用程序包含两个主要组件

索引：一个用于从源数据摄取数据并进行索引的管道。这通常在离线状态下进行。

检索和生成：实际的 RAG 链，在运行时接收用户查询，从索引中检索相关数据，然后将其传递给模型。

注意：本教程的索引部分将主要遵循语义搜索教程。

从原始数据到答案的最常见完整序列如下所示

索引

加载：首先我们需要加载数据。这通过文档加载器（Document Loaders）完成。
切分：文本切分器（Text splitters）将大型 文档（Documents） 分解成更小的块。这对于索引数据和将其传递给模型都很有用，因为大型块难以搜索，并且无法适应模型的有限上下文窗口。
存储：我们需要一个地方来存储和索引我们的切分块，以便将来可以对其进行搜索。这通常通过使用向量存储（VectorStore）和嵌入（Embeddings）模型来完成。

index_diagram

检索与生成

检索：给定用户输入，使用检索器（Retriever）从存储中检索相关切分块。
生成：一个聊天模型（ChatModel） / LLM 使用包含问题和检索数据的提示来生成答案

retrieval_diagram

一旦我们索引了数据，我们将使用 LangGraph 作为我们的编排框架来实现检索和生成步骤。

设置

Jupyter Notebook

本教程及其他教程或许在 Jupyter Notebooks 中运行最为方便。在交互式环境中学习指南是更好地理解它们的好方法。有关安装说明，请参见此处。

安装

本教程需要以下 LangChain 依赖项

Pip
Conda

%pip install --quiet --upgrade langchain-text-splitters langchain-community langgraph

conda install langchain-text-splitters langchain-community langgraph -c conda-forge

更多详情，请参阅我们的安装指南。

LangSmith

您使用 LangChain 构建的许多应用程序将包含多个步骤，其中涉及对 LLM 的多次调用。随着这些应用程序变得越来越复杂，检查您的链或代理内部究竟发生了什么变得至关重要。最好的方法是使用 LangSmith。

在上述链接注册后，请务必设置您的环境变量以开始记录跟踪日志

export LANGSMITH_TRACING="true"
export LANGSMITH_API_KEY="..."

或者，如果在Notebook中，您可以这样设置

import getpass
import os

os.environ["LANGSMITH_TRACING"] = "true"
os.environ["LANGSMITH_API_KEY"] = getpass.getpass()

组件

我们需要从LangChain的集成套件中选择三个组件。

选择聊天模型

pip install -qU "langchain[google-genai]"

import getpass
import os

if not os.environ.get("GOOGLE_API_KEY"):
  os.environ["GOOGLE_API_KEY"] = getpass.getpass("Enter API key for Google Gemini: ")

from langchain.chat_models import init_chat_model

llm = init_chat_model("gemini-2.0-flash", model_provider="google_genai")

选择嵌入模型

pip install -qU langchain-openai

import getpass
import os

if not os.environ.get("OPENAI_API_KEY"):
  os.environ["OPENAI_API_KEY"] = getpass.getpass("Enter API key for OpenAI: ")

from langchain_openai import OpenAIEmbeddings

embeddings = OpenAIEmbeddings(model="text-embedding-3-large")

选择向量存储

pip install -qU langchain-core

from langchain_core.vectorstores import InMemoryVectorStore

vector_store = InMemoryVectorStore(embeddings)

预览

在本指南中，我们将构建一个能够回答关于网站内容问题的应用程序。我们将使用的特定网站是 Lilian Weng 的博客文章 LLM 驱动的自主智能体（LLM Powered Autonomous Agents），它允许我们询问关于文章内容的问题。

我们可以用大约 50 行代码创建一个简单的索引管道和 RAG 链来实现这一点。

import bs4
from langchain import hub
from langchain_community.document_loaders import WebBaseLoader
from langchain_core.documents import Document
from langchain_text_splitters import RecursiveCharacterTextSplitter
from langgraph.graph import START, StateGraph
from typing_extensions import List, TypedDict

# Load and chunk contents of the blog
loader = WebBaseLoader(
    web_paths=("https://lilianweng.github.io/posts/2023-06-23-agent/",),
    bs_kwargs=dict(
        parse_only=bs4.SoupStrainer(
            class_=("post-content", "post-title", "post-header")
        )
    ),
)
docs = loader.load()

text_splitter = RecursiveCharacterTextSplitter(chunk_size=1000, chunk_overlap=200)
all_splits = text_splitter.split_documents(docs)

# Index chunks
_ = vector_store.add_documents(documents=all_splits)

# Define prompt for question-answering
# N.B. for non-US LangSmith endpoints, you may need to specify
# api_url="https://api.smith.langchain.com" in hub.pull.
prompt = hub.pull("rlm/rag-prompt")


# Define state for application
class State(TypedDict):
    question: str
    context: List[Document]
    answer: str


# Define application steps
def retrieve(state: State):
    retrieved_docs = vector_store.similarity_search(state["question"])
    return {"context": retrieved_docs}


def generate(state: State):
    docs_content = "\n\n".join(doc.page_content for doc in state["context"])
    messages = prompt.invoke({"question": state["question"], "context": docs_content})
    response = llm.invoke(messages)
    return {"answer": response.content}


# Compile application and test
graph_builder = StateGraph(State).add_sequence([retrieve, generate])
graph_builder.add_edge(START, "retrieve")
graph = graph_builder.compile()

API 参考：hub | WebBaseLoader | Document | RecursiveCharacterTextSplitter | StateGraph

response = graph.invoke({"question": "What is Task Decomposition?"})
print(response["answer"])

Task Decomposition is the process of breaking down a complicated task into smaller, manageable steps to facilitate easier execution and understanding. Techniques like Chain of Thought (CoT) and Tree of Thoughts (ToT) guide models to think step-by-step, allowing them to explore multiple reasoning possibilities. This method enhances performance on complex tasks and provides insight into the model's thinking process.

查看 LangSmith 追踪。

详细演练

让我们一步一步地深入了解上面的代码，真正理解它的运作方式。

1. 索引

注意

本节是语义搜索教程内容的缩略版。如果您熟悉文档加载器、嵌入和向量存储，可以直接跳到下一节检索和生成。

加载文档

我们首先需要加载博客文章内容。我们可以为此使用文档加载器（DocumentLoaders），它们是从源加载数据并返回文档（Document）对象列表的对象。

在本例中，我们将使用 WebBaseLoader，它使用 urllib 从网页 URL 加载 HTML，并使用 BeautifulSoup 将其解析为文本。我们可以通过 bs_kwargs 将参数传递给 BeautifulSoup 解析器来定制 HTML 到文本的解析（请参阅 BeautifulSoup 文档）。在这种情况下，只有带有“post-content”、“post-title”或“post-header”类的 HTML 标签是相关的，因此我们将删除所有其他标签。

import bs4
from langchain_community.document_loaders import WebBaseLoader

# Only keep post title, headers, and content from the full HTML.
bs4_strainer = bs4.SoupStrainer(class_=("post-title", "post-header", "post-content"))
loader = WebBaseLoader(
    web_paths=("https://lilianweng.github.io/posts/2023-06-23-agent/",),
    bs_kwargs={"parse_only": bs4_strainer},
)
docs = loader.load()

assert len(docs) == 1
print(f"Total characters: {len(docs[0].page_content)}")

API 参考：WebBaseLoader

Total characters: 43131

print(docs[0].page_content[:500])

      LLM Powered Autonomous Agents
    
Date: June 23, 2023  |  Estimated Reading Time: 31 min  |  Author: Lilian Weng


Building agents with LLM (large language model) as its core controller is a cool concept. Several proof-of-concepts demos, such as AutoGPT, GPT-Engineer and BabyAGI, serve as inspiring examples. The potentiality of LLM extends beyond generating well-written copies, stories, essays and programs; it can be framed as a powerful general problem solver.
Agent System Overview#
In

深入了解

DocumentLoader：从源加载数据并作为 Document 列表返回的对象。

文档：关于如何使用 DocumentLoaders 的详细文档。
集成：160+ 种集成可供选择。
接口：基础接口的 API 参考。

切分文档

我们加载的文档超过 42k 字符，太长了，无法适应许多模型的上下文窗口。即使对于那些能够将整篇文章放入其上下文窗口的模型，也很难在非常长的输入中找到信息。

为了解决这个问题，我们将把 文档 分割成块，用于嵌入和向量存储。这应该有助于我们在运行时只检索博客文章中最相关的部分。

如语义搜索教程中所述，我们使用 RecursiveCharacterTextSplitter，它会使用换行符等常见分隔符递归地分割文档，直到每个块达到适当的大小。这是通用文本用例推荐的文本分割器。

from langchain_text_splitters import RecursiveCharacterTextSplitter

text_splitter = RecursiveCharacterTextSplitter(
    chunk_size=1000,  # chunk size (characters)
    chunk_overlap=200,  # chunk overlap (characters)
    add_start_index=True,  # track index in original document
)
all_splits = text_splitter.split_documents(docs)

print(f"Split blog post into {len(all_splits)} sub-documents.")

API 参考：RecursiveCharacterTextSplitter

Split blog post into 66 sub-documents.

深入了解

TextSplitter：将 Document 列表分割成更小块的对象。是 DocumentTransformer 的子类。

通过阅读操作指南文档了解更多关于使用不同方法分割文本的信息
代码 (py 或 js)
科学论文
接口：基础接口的 API 参考。

DocumentTransformer：对 Document 对象列表执行转换的对象。

文档：关于如何使用 DocumentTransformers 的详细文档。
集成
接口：基础接口的 API 参考。

存储文档

现在我们需要索引我们的 66 个文本块，以便在运行时可以对其进行搜索。遵循语义搜索教程，我们的方法是嵌入每个文档切分块的内容，并将这些嵌入插入到向量存储中。给定一个输入查询，我们就可以使用向量搜索来检索相关文档。

我们可以使用在本教程开始时选择的向量存储和嵌入模型，通过一条命令嵌入并存储所有文档切分块。

document_ids = vector_store.add_documents(documents=all_splits)

print(document_ids[:3])

['07c18af6-ad58-479a-bfb1-d508033f9c64', '9000bf8e-1993-446f-8d4d-f4e507ba4b8f', 'ba3b5d14-bed9-4f5f-88be-44c88aedc2e6']

深入了解

Embeddings：文本嵌入模型的封装，用于将文本转换为嵌入。

文档：关于如何使用嵌入的详细文档。
集成：30+ 种集成可供选择。
接口：基础接口的 API 参考。

VectorStore：向量数据库的封装，用于存储和查询嵌入。

文档：关于如何使用向量存储的详细文档。
集成：40+ 种集成可供选择。
接口：基础接口的 API 参考。

至此，我们完成了管道的索引部分。现在我们有了一个可查询的向量存储，其中包含我们博客文章的切分内容。给定用户问题，我们理想情况下应该能够返回回答该问题的博客文章片段。

2. 检索与生成

现在我们来编写实际的应用程序逻辑。我们希望创建一个简单的应用程序，它接收用户问题，搜索与该问题相关的文档，将检索到的文档和初始问题传递给模型，并返回答案。

对于生成，我们将使用在教程开始时选择的聊天模型。

我们将使用一个已提交到 LangChain 提示中心（此处）的 RAG 提示。

from langchain import hub

# N.B. for non-US LangSmith endpoints, you may need to specify
# api_url="https://api.smith.langchain.com" in hub.pull.
prompt = hub.pull("rlm/rag-prompt")

example_messages = prompt.invoke(
    {"context": "(context goes here)", "question": "(question goes here)"}
).to_messages()

assert len(example_messages) == 1
print(example_messages[0].content)

API 参考：hub

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 goes here) 
Context: (context goes here) 
Answer:

我们将使用 LangGraph 将检索和生成步骤整合到一个应用程序中。这将带来诸多好处：

我们可以一次定义应用程序逻辑，并自动支持多种调用模式，包括流式、异步和批量调用。
我们通过 LangGraph 平台获得简化的部署。
LangSmith 将自动追踪我们应用程序的各个步骤。
我们可以轻松地为应用程序添加关键功能，包括持久性和人工干预批准，只需最少的代码更改。

要使用 LangGraph，我们需要定义三样东西

应用程序的状态；
应用程序的节点（即应用程序步骤）；
应用程序的“控制流”（例如，步骤的顺序）。

状态：

应用程序的状态控制着输入到应用程序的数据、步骤之间传输的数据以及应用程序输出的数据。它通常是一个 TypedDict，但也可以是一个 Pydantic BaseModel。

对于一个简单的 RAG 应用程序，我们只需跟踪输入问题、检索到的上下文和生成的答案

from langchain_core.documents import Document
from typing_extensions import List, TypedDict


class State(TypedDict):
    question: str
    context: List[Document]
    answer: str

API 参考：Document

节点（应用程序步骤）

让我们从一个简单的两步序列开始：检索和生成。

def retrieve(state: State):
    retrieved_docs = vector_store.similarity_search(state["question"])
    return {"context": retrieved_docs}


def generate(state: State):
    docs_content = "\n\n".join(doc.page_content for doc in state["context"])
    messages = prompt.invoke({"question": state["question"], "context": docs_content})
    response = llm.invoke(messages)
    return {"answer": response.content}

我们的检索步骤只是使用输入问题运行相似性搜索，而生成步骤则将检索到的上下文和原始问题格式化为聊天模型的提示。

控制流

最后，我们将应用程序编译成一个单独的 graph 对象。在这种情况下，我们只是将检索和生成步骤连接成一个单一的序列。

from langgraph.graph import START, StateGraph

graph_builder = StateGraph(State).add_sequence([retrieve, generate])
graph_builder.add_edge(START, "retrieve")
graph = graph_builder.compile()

API 参考：StateGraph

LangGraph 还内置了用于可视化应用程序控制流的实用工具

from IPython.display import Image, display

display(Image(graph.get_graph().draw_mermaid_png()))

我需要使用 LangGraph 吗？

构建 RAG 应用程序并非必须使用 LangGraph。事实上，我们可以通过调用各个组件来实现相同的应用程序逻辑

question = "..."

retrieved_docs = vector_store.similarity_search(question)
docs_content = "\n\n".join(doc.page_content for doc in retrieved_docs)
prompt = prompt.invoke({"question": question, "context": docs_content})
answer = llm.invoke(prompt)

LangGraph 的优势包括

支持多种调用模式：如果我们要流式传输输出 token 或流式传输单个步骤的结果，则需要重写此逻辑；
通过 LangSmith 自动支持追踪，并通过 LangGraph 平台自动支持部署；
支持持久性、人工干预等功能。

许多用例需要在对话体验中进行 RAG，以便用户可以通过有状态的对话接收到上下文相关的答案。正如我们将在教程的第二部分中看到的那样，LangGraph 对状态的管理和持久化极大地简化了这些应用程序。

使用

让我们测试一下我们的应用程序！LangGraph 支持多种调用模式，包括同步、异步和流式传输。

调用

result = graph.invoke({"question": "What is Task Decomposition?"})

print(f"Context: {result['context']}\n\n")
print(f"Answer: {result['answer']}")

Context: [Document(id='a42dc78b-8f76-472a-9e25-180508af74f3', metadata={'source': 'https://lilianweng.github.io/posts/2023-06-23-agent/', 'start_index': 1585}, page_content='Fig. 1. Overview of a LLM-powered autonomous agent system.\nComponent One: Planning#\nA complicated task usually involves many steps. An agent needs to know what they are and plan ahead.\nTask Decomposition#\nChain of thought (CoT; Wei et al. 2022) has become a standard prompting technique for enhancing model performance on complex tasks. The model is instructed to “think step by step” to utilize more test-time computation to decompose hard tasks into smaller and simpler steps. CoT transforms big tasks into multiple manageable tasks and shed lights into an interpretation of the model’s thinking process.'), Document(id='c0e45887-d0b0-483d-821a-bb5d8316d51d', metadata={'source': 'https://lilianweng.github.io/posts/2023-06-23-agent/', 'start_index': 2192}, page_content='Tree of Thoughts (Yao et al. 2023) extends CoT by exploring multiple reasoning possibilities at each step. It first decomposes the problem into multiple thought steps and generates multiple thoughts per step, creating a tree structure. The search process can be BFS (breadth-first search) or DFS (depth-first search) with each state evaluated by a classifier (via a prompt) or majority vote.\nTask decomposition can be done (1) by LLM with simple prompting like "Steps for XYZ.\\n1.", "What are the subgoals for achieving XYZ?", (2) by using task-specific instructions; e.g. "Write a story outline." for writing a novel, or (3) with human inputs.'), Document(id='4cc7f318-35f5-440f-a4a4-145b5f0b918d', metadata={'source': 'https://lilianweng.github.io/posts/2023-06-23-agent/', 'start_index': 29630}, page_content='Resources:\n1. Internet access for searches and information gathering.\n2. Long Term memory management.\n3. GPT-3.5 powered Agents for delegation of simple tasks.\n4. File output.\n\nPerformance Evaluation:\n1. Continuously review and analyze your actions to ensure you are performing to the best of your abilities.\n2. Constructively self-criticize your big-picture behavior constantly.\n3. Reflect on past decisions and strategies to refine your approach.\n4. Every command has a cost, so be smart and efficient. Aim to complete tasks in the least number of steps.'), Document(id='f621ade4-9b0d-471f-a522-44eb5feeba0c', metadata={'source': 'https://lilianweng.github.io/posts/2023-06-23-agent/', 'start_index': 19373}, page_content="(3) Task execution: Expert models execute on the specific tasks and log results.\nInstruction:\n\nWith the input and the inference results, the AI assistant needs to describe the process and results. The previous stages can be formed as - User Input: {{ User Input }}, Task Planning: {{ Tasks }}, Model Selection: {{ Model Assignment }}, Task Execution: {{ Predictions }}. You must first answer the user's request in a straightforward manner. Then describe the task process and show your analysis and model inference results to the user in the first person. If inference results contain a file path, must tell the user the complete file path.")]

Answer: Task decomposition is a technique used to break down complex tasks into smaller, manageable steps, allowing for more efficient problem-solving. This can be achieved through methods like chain of thought prompting or the tree of thoughts approach, which explores multiple reasoning possibilities at each step. It can be initiated through simple prompts, task-specific instructions, or human inputs.

流式步骤

for step in graph.stream(
    {"question": "What is Task Decomposition?"}, stream_mode="updates"
):
    print(f"{step}\n\n----------------\n")

{'retrieve': {'context': [Document(id='a42dc78b-8f76-472a-9e25-180508af74f3', metadata={'source': 'https://lilianweng.github.io/posts/2023-06-23-agent/', 'start_index': 1585}, page_content='Fig. 1. Overview of a LLM-powered autonomous agent system.\nComponent One: Planning#\nA complicated task usually involves many steps. An agent needs to know what they are and plan ahead.\nTask Decomposition#\nChain of thought (CoT; Wei et al. 2022) has become a standard prompting technique for enhancing model performance on complex tasks. The model is instructed to “think step by step” to utilize more test-time computation to decompose hard tasks into smaller and simpler steps. CoT transforms big tasks into multiple manageable tasks and shed lights into an interpretation of the model’s thinking process.'), Document(id='c0e45887-d0b0-483d-821a-bb5d8316d51d', metadata={'source': 'https://lilianweng.github.io/posts/2023-06-23-agent/', 'start_index': 2192}, page_content='Tree of Thoughts (Yao et al. 2023) extends CoT by exploring multiple reasoning possibilities at each step. It first decomposes the problem into multiple thought steps and generates multiple thoughts per step, creating a tree structure. The search process can be BFS (breadth-first search) or DFS (depth-first search) with each state evaluated by a classifier (via a prompt) or majority vote.\nTask decomposition can be done (1) by LLM with simple prompting like "Steps for XYZ.\\n1.", "What are the subgoals for achieving XYZ?", (2) by using task-specific instructions; e.g. "Write a story outline." for writing a novel, or (3) with human inputs.'), Document(id='4cc7f318-35f5-440f-a4a4-145b5f0b918d', metadata={'source': 'https://lilianweng.github.io/posts/2023-06-23-agent/', 'start_index': 29630}, page_content='Resources:\n1. Internet access for searches and information gathering.\n2. Long Term memory management.\n3. GPT-3.5 powered Agents for delegation of simple tasks.\n4. File output.\n\nPerformance Evaluation:\n1. Continuously review and analyze your actions to ensure you are performing to the best of your abilities.\n2. Constructively self-criticize your big-picture behavior constantly.\n3. Reflect on past decisions and strategies to refine your approach.\n4. Every command has a cost, so be smart and efficient. Aim to complete tasks in the least number of steps.'), Document(id='f621ade4-9b0d-471f-a522-44eb5feeba0c', metadata={'source': 'https://lilianweng.github.io/posts/2023-06-23-agent/', 'start_index': 19373}, page_content="(3) Task execution: Expert models execute on the specific tasks and log results.\nInstruction:\n\nWith the input and the inference results, the AI assistant needs to describe the process and results. The previous stages can be formed as - User Input: {{ User Input }}, Task Planning: {{ Tasks }}, Model Selection: {{ Model Assignment }}, Task Execution: {{ Predictions }}. You must first answer the user's request in a straightforward manner. Then describe the task process and show your analysis and model inference results to the user in the first person. If inference results contain a file path, must tell the user the complete file path.")]}}

----------------

{'generate': {'answer': 'Task decomposition is the process of breaking down a complex task into smaller, more manageable steps. This technique, often enhanced by methods like Chain of Thought (CoT) or Tree of Thoughts, allows models to reason through tasks systematically and improves performance by clarifying the thought process. It can be achieved through simple prompts, task-specific instructions, or human inputs.'}}

----------------

流式 token

for message, metadata in graph.stream(
    {"question": "What is Task Decomposition?"}, stream_mode="messages"
):
    print(message.content, end="|")

|Task| decomposition| is| the| process| of| breaking| down| complex| tasks| into| smaller|,| more| manageable| steps|.| It| can| be| achieved| through| techniques| like| Chain| of| Thought| (|Co|T|)| prompting|,| which| encourages| the| model| to| think| step| by| step|,| or| through| more| structured| methods| like| the| Tree| of| Thoughts|.| This| approach| not| only| simplifies| task| execution| but| also| provides| insights| into| the| model|'s| reasoning| process|.||

提示

对于异步调用，请使用

result = await graph.ainvoke(...)

和

async for step in graph.astream(...):

返回来源

请注意，通过将检索到的上下文存储在图的状态中，我们可以在状态的 "context" 字段中恢复模型生成答案的来源。有关返回来源的更多详细信息，请参阅此指南。

深入了解

聊天模型接收一系列消息并返回一条消息。

文档
集成：25+ 种集成可供选择。
接口：基础接口的 API 参考。

定制提示

如上所示，我们可以从提示中心加载提示（例如，此 RAG 提示）。提示也可以轻松定制。例如

from langchain_core.prompts import PromptTemplate

template = """Use the following pieces of context to answer the question at the end.
If you don't know the answer, just say that you don't know, don't try to make up an answer.
Use three sentences maximum and keep the answer as concise as possible.
Always say "thanks for asking!" at the end of the answer.

{context}

Question: {question}

Helpful Answer:"""
custom_rag_prompt = PromptTemplate.from_template(template)

API 参考：PromptTemplate

查询分析

到目前为止，我们是使用原始输入查询执行检索的。然而，让模型为检索目的生成查询有一些优点。例如

除了语义搜索，我们还可以内置结构化过滤器（例如，“查找自 2020 年以来的文档”）；
模型可以将可能多方面或包含无关语言的用户查询重写为更有效的搜索查询。

查询分析利用模型从原始用户输入中转换或构建优化的搜索查询。我们可以轻松地将查询分析步骤整合到我们的应用程序中。为了说明目的，让我们向向量存储中的文档添加一些元数据。我们将向文档添加一些（人为设计的）部分，以便稍后进行过滤。

total_documents = len(all_splits)
third = total_documents // 3

for i, document in enumerate(all_splits):
    if i < third:
        document.metadata["section"] = "beginning"
    elif i < 2 * third:
        document.metadata["section"] = "middle"
    else:
        document.metadata["section"] = "end"


all_splits[0].metadata

{'source': 'https://lilianweng.github.io/posts/2023-06-23-agent/',
 'start_index': 8,
 'section': 'beginning'}

我们需要更新向量存储中的文档。我们将为此使用一个简单的 InMemoryVectorStore，因为我们将使用它的一些特定功能（即，元数据过滤）。有关您选择的向量存储的相关功能，请参阅向量存储的集成文档。

from langchain_core.vectorstores import InMemoryVectorStore

vector_store = InMemoryVectorStore(embeddings)
_ = vector_store.add_documents(all_splits)

API 参考：InMemoryVectorStore

接下来，让我们定义搜索查询的 schema。我们将为此目的使用结构化输出。在这里，我们定义一个查询包含一个字符串查询和一个文档部分（可以是“开头”、“中间”或“结尾”），但您可以根据需要进行定义。

from typing import Literal

from typing_extensions import Annotated


class Search(TypedDict):
    """Search query."""

    query: Annotated[str, ..., "Search query to run."]
    section: Annotated[
        Literal["beginning", "middle", "end"],
        ...,
        "Section to query.",
    ]

最后，我们向 LangGraph 应用程序添加一个步骤，以从用户的原始输入中生成查询

class State(TypedDict):
    question: str
    query: Search
    context: List[Document]
    answer: str


def analyze_query(state: State):
    structured_llm = llm.with_structured_output(Search)
    query = structured_llm.invoke(state["question"])
    return {"query": query}


def retrieve(state: State):
    query = state["query"]
    retrieved_docs = vector_store.similarity_search(
        query["query"],
        filter=lambda doc: doc.metadata.get("section") == query["section"],
    )
    return {"context": retrieved_docs}


def generate(state: State):
    docs_content = "\n\n".join(doc.page_content for doc in state["context"])
    messages = prompt.invoke({"question": state["question"], "context": docs_content})
    response = llm.invoke(messages)
    return {"answer": response.content}


graph_builder = StateGraph(State).add_sequence([analyze_query, retrieve, generate])
graph_builder.add_edge(START, "analyze_query")
graph = graph_builder.compile()

完整代码

from typing import Literal

import bs4
from langchain import hub
from langchain_community.document_loaders import WebBaseLoader
from langchain_core.documents import Document
from langchain_core.vectorstores import InMemoryVectorStore
from langchain_text_splitters import RecursiveCharacterTextSplitter
from langgraph.graph import START, StateGraph
from typing_extensions import Annotated, List, TypedDict

# Load and chunk contents of the blog
loader = WebBaseLoader(
    web_paths=("https://lilianweng.github.io/posts/2023-06-23-agent/",),
    bs_kwargs=dict(
        parse_only=bs4.SoupStrainer(
            class_=("post-content", "post-title", "post-header")
        )
    ),
)
docs = loader.load()

text_splitter = RecursiveCharacterTextSplitter(chunk_size=1000, chunk_overlap=200)
all_splits = text_splitter.split_documents(docs)


# Update metadata (illustration purposes)
total_documents = len(all_splits)
third = total_documents // 3

for i, document in enumerate(all_splits):
    if i < third:
        document.metadata["section"] = "beginning"
    elif i < 2 * third:
        document.metadata["section"] = "middle"
    else:
        document.metadata["section"] = "end"


# Index chunks
vector_store = InMemoryVectorStore(embeddings)
_ = vector_store.add_documents(all_splits)


# Define schema for search
class Search(TypedDict):
    """Search query."""

    query: Annotated[str, ..., "Search query to run."]
    section: Annotated[
        Literal["beginning", "middle", "end"],
        ...,
        "Section to query.",
    ]

# Define prompt for question-answering
prompt = hub.pull("rlm/rag-prompt")


# Define state for application
class State(TypedDict):
    question: str
    query: Search
    context: List[Document]
    answer: str


def analyze_query(state: State):
    structured_llm = llm.with_structured_output(Search)
    query = structured_llm.invoke(state["question"])
    return {"query": query}


def retrieve(state: State):
    query = state["query"]
    retrieved_docs = vector_store.similarity_search(
        query["query"],
        filter=lambda doc: doc.metadata.get("section") == query["section"],
    )
    return {"context": retrieved_docs}


def generate(state: State):
    docs_content = "\n\n".join(doc.page_content for doc in state["context"])
    messages = prompt.invoke({"question": state["question"], "context": docs_content})
    response = llm.invoke(messages)
    return {"answer": response.content}


graph_builder = StateGraph(State).add_sequence([analyze_query, retrieve, generate])
graph_builder.add_edge(START, "analyze_query")
graph = graph_builder.compile()

display(Image(graph.get_graph().draw_mermaid_png()))

我们可以通过专门询问文章末尾的上下文来测试我们的实现。请注意，模型在其答案中包含了不同的信息。

for step in graph.stream(
    {"question": "What does the end of the post say about Task Decomposition?"},
    stream_mode="updates",
):
    print(f"{step}\n\n----------------\n")

{'analyze_query': {'query': {'query': 'Task Decomposition', 'section': 'end'}}}

----------------

{'retrieve': {'context': [Document(id='d6cef137-e1e8-4ddc-91dc-b62bd33c6020', metadata={'source': 'https://lilianweng.github.io/posts/2023-06-23-agent/', 'start_index': 39221, 'section': 'end'}, page_content='Finite context length: The restricted context capacity limits the inclusion of historical information, detailed instructions, API call context, and responses. The design of the system has to work with this limited communication bandwidth, while mechanisms like self-reflection to learn from past mistakes would benefit a lot from long or infinite context windows. Although vector stores and retrieval can provide access to a larger knowledge pool, their representation power is not as powerful as full attention.\n\n\nChallenges in long-term planning and task decomposition: Planning over a lengthy history and effectively exploring the solution space remain challenging. LLMs struggle to adjust plans when faced with unexpected errors, making them less robust compared to humans who learn from trial and error.'), Document(id='d1834ae1-eb6a-43d7-a023-08dfa5028799', metadata={'source': 'https://lilianweng.github.io/posts/2023-06-23-agent/', 'start_index': 39086, 'section': 'end'}, page_content='}\n]\nChallenges#\nAfter going through key ideas and demos of building LLM-centered agents, I start to see a couple common limitations:'), Document(id='ca7f06e4-2c2e-4788-9a81-2418d82213d9', metadata={'source': 'https://lilianweng.github.io/posts/2023-06-23-agent/', 'start_index': 32942, 'section': 'end'}, page_content='}\n]\nThen after these clarification, the agent moved into the code writing mode with a different system message.\nSystem message:'), Document(id='1fcc2736-30f4-4ef6-90f2-c64af92118cb', metadata={'source': 'https://lilianweng.github.io/posts/2023-06-23-agent/', 'start_index': 35127, 'section': 'end'}, page_content='"content": "You will get instructions for code to write.\\nYou will write a very long answer. Make sure that every detail of the architecture is, in the end, implemented as code.\\nMake sure that every detail of the architecture is, in the end, implemented as code.\\n\\nThink step by step and reason yourself to the right decisions to make sure we get it right.\\nYou will first lay out the names of the core classes, functions, methods that will be necessary, as well as a quick comment on their purpose.\\n\\nThen you will output the content of each file including ALL code.\\nEach file must strictly follow a markdown code block format, where the following tokens must be replaced such that\\nFILENAME is the lowercase file name including the file extension,\\nLANG is the markup code block language for the code\'s language, and CODE is the code:\\n\\nFILENAME\\n\`\`\`LANG\\nCODE\\n\`\`\`\\n\\nYou will start with the \\"entrypoint\\" file, then go to the ones that are imported by that file, and so on.\\nPlease')]}}

----------------

{'generate': {'answer': 'The end of the post highlights that task decomposition faces challenges in long-term planning and adapting to unexpected errors. LLMs struggle with adjusting their plans, making them less robust compared to humans who learn from trial and error. This indicates a limitation in effectively exploring the solution space and handling complex tasks.'}}

----------------

在流式步骤和 LangSmith 追踪中，我们现在可以观察到输入到检索步骤的结构化查询。

查询分析是一个涵盖广泛方法的复杂问题。有关更多示例，请参阅操作指南。

下一步

我们已经介绍了构建一个基本数据问答应用程序的步骤

使用文档加载器（Document Loader）加载数据
使用文本切分器（Text Splitter）对索引数据进行分块，使其更易于模型使用
嵌入数据并将数据存储在向量存储中
响应传入问题检索之前存储的块
使用检索到的块作为上下文生成答案。

在教程的第二部分中，我们将扩展这里的实现，以适应对话式交互和多步检索过程。

进一步阅读

返回来源：了解如何返回源文档
流式传输：了解如何流式传输输出和中间步骤
添加聊天历史：了解如何将聊天历史添加到您的应用中
检索概念指南：特定检索技术的高级概述

概述​

索引​

检索与生成​

设置​

Jupyter Notebook​

安装​

LangSmith​

组件​

预览​

详细演练​

1. 索引​

加载文档​

深入了解​

切分文档​

深入了解​

存储文档​

深入了解​

2. 检索与生成​

状态：​

节点（应用程序步骤）​

控制流​

使用​

返回来源​

深入了解​

查询分析​

下一步​

概述

索引

检索与生成

设置

Jupyter Notebook

安装

LangSmith

组件

预览

详细演练

1. 索引

加载文档

深入了解

切分文档

深入了解

存储文档

深入了解

2. 检索与生成

状态：

节点（应用程序步骤）

控制流

使用

返回来源

深入了解

查询分析

下一步