LlamaIndex - Data Framework for LLM Applications

LlamaIndex - 面向LLM应用的数据框架

The leading framework for connecting LLMs with your data.

连接LLM与你的数据的领先框架。

When to use LlamaIndex

何时使用LlamaIndex

Use LlamaIndex when:

Building RAG (retrieval-augmented generation) applications
Need document question-answering over private data
Ingesting data from multiple sources (300+ connectors)
Creating knowledge bases for LLMs
Building chatbots with enterprise data
Need structured data extraction from documents

Metrics:

45,100+ GitHub stars
23,000+ repositories use LlamaIndex
300+ data connectors (LlamaHub)
1,715+ contributors
v0.14.7 (stable)

Use alternatives instead:

LangChain: More general-purpose, better for agents
Haystack: Production search pipelines
txtai: Lightweight semantic search
Chroma: Just need vector storage

在以下场景使用LlamaIndex：

构建RAG（检索增强生成）应用
需要针对私有数据的文档问答功能
从多源摄入数据（支持300+连接器）
为LLM创建知识库
构建基于企业数据的聊天机器人
需要从文档中提取结构化数据

相关数据:

45,100+ GitHub星标
23,000+ 仓库使用LlamaIndex
300+ 数据连接器（LlamaHub）
1,715+ 贡献者
v0.14.7（稳定版）

可选择替代方案的场景:

LangChain: 更通用，更适合Agent开发
Haystack: 生产级搜索流水线
txtai: 轻量级语义搜索
Chroma: 仅需向量存储时

Quick start

快速开始

Installation

安装

bash

undefined

bash

undefined

Starter package (recommended)

入门包（推荐）

pip install llama-index

Or minimal core + specific integrations

或最小核心包 + 特定集成

pip install llama-index-core pip install llama-index-llms-openai pip install llama-index-embeddings-openai

undefined

pip install llama-index-core pip install llama-index-llms-openai pip install llama-index-embeddings-openai

undefined

5-line RAG example

5行代码实现RAG示例

python

from llama_index.core import VectorStoreIndex, SimpleDirectoryReader

python

from llama_index.core import VectorStoreIndex, SimpleDirectoryReader

Load documents

加载文档

documents = SimpleDirectoryReader("data").load_data()

Create index

创建索引

index = VectorStoreIndex.from_documents(documents)

Query

查询

query_engine = index.as_query_engine() response = query_engine.query("What did the author do growing up?") print(response)

undefined

query_engine = index.as_query_engine() response = query_engine.query("What did the author do growing up?") print(response)

undefined

Core concepts

核心概念

1. Data connectors - Load documents

1. 数据连接器 - 加载文档

python

from llama_index.core import SimpleDirectoryReader, Document
from llama_index.readers.web import SimpleWebPageReader
from llama_index.readers.github import GithubRepositoryReader

python

from llama_index.core import SimpleDirectoryReader, Document
from llama_index.readers.web import SimpleWebPageReader
from llama_index.readers.github import GithubRepositoryReader

Directory of files

加载目录文件

documents = SimpleDirectoryReader("./data").load_data()

Web pages

加载网页

reader = SimpleWebPageReader() documents = reader.load_data(["https://example.com"])

GitHub repository

加载GitHub仓库

reader = GithubRepositoryReader(owner="user", repo="repo") documents = reader.load_data(branch="main")

Manual document creation

手动创建文档

doc = Document( text="This is the document content", metadata={"source": "manual", "date": "2025-01-01"} )

undefined

doc = Document( text="This is the document content", metadata={"source": "manual", "date": "2025-01-01"} )

undefined

2. Indices - Structure data

2. 索引 - 结构化数据

python

from llama_index.core import VectorStoreIndex, ListIndex, TreeIndex

python

from llama_index.core import VectorStoreIndex, ListIndex, TreeIndex

Vector index (most common - semantic search)

向量索引（最常用 - 语义搜索）

vector_index = VectorStoreIndex.from_documents(documents)

List index (sequential scan)

列表索引（顺序扫描）

list_index = ListIndex.from_documents(documents)

Tree index (hierarchical summary)

树形索引（分层摘要）

tree_index = TreeIndex.from_documents(documents)

Save index

保存索引

index.storage_context.persist(persist_dir="./storage")

Load index

加载索引

from llama_index.core import load_index_from_storage, StorageContext storage_context = StorageContext.from_defaults(persist_dir="./storage") index = load_index_from_storage(storage_context)

undefined

from llama_index.core import load_index_from_storage, StorageContext storage_context = StorageContext.from_defaults(persist_dir="./storage") index = load_index_from_storage(storage_context)

undefined

3. Query engines - Ask questions

3. 查询引擎 - 提问

python

undefined

python

undefined

Basic query

基础查询

query_engine = index.as_query_engine() response = query_engine.query("What is the main topic?") print(response)

Streaming response

流式响应

query_engine = index.as_query_engine(streaming=True) response = query_engine.query("Explain quantum computing") for text in response.response_gen: print(text, end="", flush=True)

Custom configuration

自定义配置

query_engine = index.as_query_engine( similarity_top_k=3, # Return top 3 chunks response_mode="compact", # Or "tree_summarize", "simple_summarize" verbose=True )

undefined

query_engine = index.as_query_engine( similarity_top_k=3, # 返回前3个片段 response_mode="compact", # 或 "tree_summarize"、"simple_summarize" verbose=True )

undefined

4. Retrievers - Find relevant chunks

4. 检索器 - 查找相关片段

python

undefined

python

undefined

Vector retriever

向量检索器

retriever = index.as_retriever(similarity_top_k=5) nodes = retriever.retrieve("machine learning")

With filtering

带过滤的检索器

retriever = index.as_retriever( similarity_top_k=3, filters={"metadata.category": "tutorial"} )

Custom retriever

自定义检索器

from llama_index.core.retrievers import BaseRetriever

class CustomRetriever(BaseRetriever): def _retrieve(self, query_bundle): # Your custom retrieval logic return nodes

undefined

from llama_index.core.retrievers import BaseRetriever

class CustomRetriever(BaseRetriever): def _retrieve(self, query_bundle): # 你的自定义检索逻辑 return nodes

undefined

Agents with tools

带工具的Agent

Basic agent

基础Agent

python

from llama_index.core.agent import FunctionAgent
from llama_index.llms.openai import OpenAI

python

from llama_index.core.agent import FunctionAgent
from llama_index.llms.openai import OpenAI

Define tools

定义工具

def multiply(a: int, b: int) -> int: """Multiply two numbers.""" return a * b

def add(a: int, b: int) -> int: """Add two numbers.""" return a + b

def multiply(a: int, b: int) -> int: """Multiply two numbers.""" return a * b

def add(a: int, b: int) -> int: """Add two numbers.""" return a + b

Create agent

创建Agent

llm = OpenAI(model="gpt-4o") agent = FunctionAgent.from_tools( tools=[multiply, add], llm=llm, verbose=True )

Use agent

使用Agent

response = agent.chat("What is 25 * 17 + 142?") print(response)

undefined

response = agent.chat("What is 25 * 17 + 142?") print(response)

undefined

RAG agent (document search + tools)

RAG Agent（文档搜索 + 工具）

python

from llama_index.core.tools import QueryEngineTool

python

from llama_index.core.tools import QueryEngineTool

Create index as before

如之前一样创建索引

index = VectorStoreIndex.from_documents(documents)

Wrap query engine as tool

将查询引擎包装为工具

query_tool = QueryEngineTool.from_defaults( query_engine=index.as_query_engine(), name="python_docs", description="Useful for answering questions about Python programming" )

Agent with document search + calculator

带文档搜索 + 计算器的Agent

agent = FunctionAgent.from_tools( tools=[query_tool, multiply, add], llm=llm )

Agent decides when to search docs vs calculate

Agent会自主决定何时搜索文档或进行计算

response = agent.chat("According to the docs, what is Python used for?")

undefined

response = agent.chat("According to the docs, what is Python used for?")

undefined

Advanced RAG patterns

高级RAG模式

Chat engine (conversational)

聊天引擎（会话式）

python

from llama_index.core.chat_engine import CondensePlusContextChatEngine

python

from llama_index.core.chat_engine import CondensePlusContextChatEngine

Chat with memory

带记忆的聊天

chat_engine = index.as_chat_engine( chat_mode="condense_plus_context", # Or "context", "react" verbose=True )

chat_engine = index.as_chat_engine( chat_mode="condense_plus_context", # 或 "context"、"react" verbose=True )

Multi-turn conversation

多轮对话

response1 = chat_engine.chat("What is Python?") response2 = chat_engine.chat("Can you give examples?") # Remembers context response3 = chat_engine.chat("What about web frameworks?")

undefined

response1 = chat_engine.chat("What is Python?") response2 = chat_engine.chat("Can you give examples?") # 记住上下文 response3 = chat_engine.chat("What about web frameworks?")

undefined

Metadata filtering

元数据过滤

python

from llama_index.core.vector_stores import MetadataFilters, ExactMatchFilter

python

from llama_index.core.vector_stores import MetadataFilters, ExactMatchFilter

Filter by metadata

按元数据过滤

filters = MetadataFilters( filters=[ ExactMatchFilter(key="category", value="tutorial"), ExactMatchFilter(key="difficulty", value="beginner") ] )

retriever = index.as_retriever( similarity_top_k=3, filters=filters )

query_engine = index.as_query_engine(filters=filters)

undefined

filters = MetadataFilters( filters=[ ExactMatchFilter(key="category", value="tutorial"), ExactMatchFilter(key="difficulty", value="beginner") ] )

retriever = index.as_retriever( similarity_top_k=3, filters=filters )

query_engine = index.as_query_engine(filters=filters)

undefined

Structured output

结构化输出

python

from pydantic import BaseModel
from llama_index.core.output_parsers import PydanticOutputParser

class Summary(BaseModel):
    title: str
    main_points: list[str]
    conclusion: str

python

from pydantic import BaseModel
from llama_index.core.output_parsers import PydanticOutputParser

class Summary(BaseModel):
    title: str
    main_points: list[str]
    conclusion: str

Get structured response

获取结构化响应

output_parser = PydanticOutputParser(output_cls=Summary) query_engine = index.as_query_engine(output_parser=output_parser)

response = query_engine.query("Summarize the document") summary = response # Pydantic model print(summary.title, summary.main_points)

undefined

output_parser = PydanticOutputParser(output_cls=Summary) query_engine = index.as_query_engine(output_parser=output_parser)

response = query_engine.query("Summarize the document") summary = response # Pydantic模型 print(summary.title, summary.main_points)

undefined

Data ingestion patterns

数据摄入模式

Multiple file types

多文件类型

python

undefined

python

undefined

Load all supported formats

加载所有支持的格式

documents = SimpleDirectoryReader( "./data", recursive=True, required_exts=[".pdf", ".docx", ".txt", ".md"] ).load_data()

undefined

documents = SimpleDirectoryReader( "./data", recursive=True, required_exts=[".pdf", ".docx", ".txt", ".md"] ).load_data()

undefined

Web scraping

网页爬取

python

from llama_index.readers.web import BeautifulSoupWebReader

reader = BeautifulSoupWebReader()
documents = reader.load_data(urls=[
    "https://docs.python.org/3/tutorial/",
    "https://docs.python.org/3/library/"
])

python

from llama_index.readers.web import BeautifulSoupWebReader

reader = BeautifulSoupWebReader()
documents = reader.load_data(urls=[
    "https://docs.python.org/3/tutorial/",
    "https://docs.python.org/3/library/"
])

Database

数据库

python

from llama_index.readers.database import DatabaseReader

reader = DatabaseReader(
    sql_database_uri="postgresql://user:pass@localhost/db"
)
documents = reader.load_data(query="SELECT * FROM articles")

python

from llama_index.readers.database import DatabaseReader

reader = DatabaseReader(
    sql_database_uri="postgresql://user:pass@localhost/db"
)
documents = reader.load_data(query="SELECT * FROM articles")

API endpoints

API端点

python

from llama_index.readers.json import JSONReader

reader = JSONReader()
documents = reader.load_data("https://api.example.com/data.json")

python

from llama_index.readers.json import JSONReader

reader = JSONReader()
documents = reader.load_data("https://api.example.com/data.json")

Vector store integrations

向量存储集成

Chroma (local)

Chroma（本地）

python

from llama_index.vector_stores.chroma import ChromaVectorStore
import chromadb

python

from llama_index.vector_stores.chroma import ChromaVectorStore
import chromadb

Initialize Chroma

初始化Chroma

db = chromadb.PersistentClient(path="./chroma_db") collection = db.get_or_create_collection("my_collection")

Create vector store

创建向量存储

vector_store = ChromaVectorStore(chroma_collection=collection)

Use in index

在索引中使用

from llama_index.core import StorageContext storage_context = StorageContext.from_defaults(vector_store=vector_store) index = VectorStoreIndex.from_documents(documents, storage_context=storage_context)

undefined

from llama_index.core import StorageContext storage_context = StorageContext.from_defaults(vector_store=vector_store) index = VectorStoreIndex.from_documents(documents, storage_context=storage_context)

undefined

Pinecone (cloud)

Pinecone（云端）

python

from llama_index.vector_stores.pinecone import PineconeVectorStore
import pinecone

python

from llama_index.vector_stores.pinecone import PineconeVectorStore
import pinecone

Initialize Pinecone

初始化Pinecone

pinecone.init(api_key="your-key", environment="us-west1-gcp") pinecone_index = pinecone.Index("my-index")

Create vector store

创建向量存储

vector_store = PineconeVectorStore(pinecone_index=pinecone_index) storage_context = StorageContext.from_defaults(vector_store=vector_store)

index = VectorStoreIndex.from_documents(documents, storage_context=storage_context)

undefined

vector_store = PineconeVectorStore(pinecone_index=pinecone_index) storage_context = StorageContext.from_defaults(vector_store=vector_store)

index = VectorStoreIndex.from_documents(documents, storage_context=storage_context)

undefined

FAISS (fast)

FAISS（快速）

python

from llama_index.vector_stores.faiss import FaissVectorStore
import faiss

python

from llama_index.vector_stores.faiss import FaissVectorStore
import faiss

Create FAISS index

创建FAISS索引

d = 1536 # Dimension of embeddings faiss_index = faiss.IndexFlatL2(d)

vector_store = FaissVectorStore(faiss_index=faiss_index) storage_context = StorageContext.from_defaults(vector_store=vector_store)

index = VectorStoreIndex.from_documents(documents, storage_context=storage_context)

undefined

d = 1536 # 嵌入向量维度 faiss_index = faiss.IndexFlatL2(d)

vector_store = FaissVectorStore(faiss_index=faiss_index) storage_context = StorageContext.from_defaults(vector_store=vector_store)

index = VectorStoreIndex.from_documents(documents, storage_context=storage_context)

undefined

Customization

自定义

Custom LLM

自定义LLM

python

from llama_index.llms.anthropic import Anthropic
from llama_index.core import Settings

python

from llama_index.llms.anthropic import Anthropic
from llama_index.core import Settings

Set global LLM

设置全局LLM

Settings.llm = Anthropic(model="claude-sonnet-4-5-20250929")

Now all queries use Anthropic

现在所有查询都使用Anthropic模型

query_engine = index.as_query_engine()

undefined

query_engine = index.as_query_engine()

undefined

Custom embeddings

自定义嵌入

python

from llama_index.embeddings.huggingface import HuggingFaceEmbedding

python

from llama_index.embeddings.huggingface import HuggingFaceEmbedding

Use HuggingFace embeddings

使用HuggingFace嵌入模型

Settings.embed_model = HuggingFaceEmbedding( model_name="sentence-transformers/all-mpnet-base-v2" )

index = VectorStoreIndex.from_documents(documents)

undefined

Settings.embed_model = HuggingFaceEmbedding( model_name="sentence-transformers/all-mpnet-base-v2" )

index = VectorStoreIndex.from_documents(documents)

undefined

Custom prompt templates

自定义提示模板

python

from llama_index.core import PromptTemplate

qa_prompt = PromptTemplate(
    "Context: {context_str}\n"
    "Question: {query_str}\n"
    "Answer the question based only on the context. "
    "If the answer is not in the context, say 'I don't know'.\n"
    "Answer: "
)

query_engine = index.as_query_engine(text_qa_template=qa_prompt)

python

from llama_index.core import PromptTemplate

qa_prompt = PromptTemplate(
    "Context: {context_str}\n"
    "Question: {query_str}\n"
    "Answer the question based only on the context. "
    "If the answer is not in the context, say 'I don't know'.\n"
    "Answer: "
)

query_engine = index.as_query_engine(text_qa_template=qa_prompt)

Multi-modal RAG

多模态RAG

Image + text

图片 + 文本

python

from llama_index.core import SimpleDirectoryReader
from llama_index.multi_modal_llms.openai import OpenAIMultiModal

python

from llama_index.core import SimpleDirectoryReader
from llama_index.multi_modal_llms.openai import OpenAIMultiModal

Load images and documents

加载图片和文档

documents = SimpleDirectoryReader( "./data", required_exts=[".jpg", ".png", ".pdf"] ).load_data()

Multi-modal index

多模态索引

index = VectorStoreIndex.from_documents(documents)

Query with multi-modal LLM

使用多模态LLM进行查询

multi_modal_llm = OpenAIMultiModal(model="gpt-4o") query_engine = index.as_query_engine(llm=multi_modal_llm)

response = query_engine.query("What is in the diagram on page 3?")

undefined

multi_modal_llm = OpenAIMultiModal(model="gpt-4o") query_engine = index.as_query_engine(llm=multi_modal_llm)

response = query_engine.query("What is in the diagram on page 3?")

undefined

Evaluation

评估

Response quality

响应质量

python

from llama_index.core.evaluation import RelevancyEvaluator, FaithfulnessEvaluator

python

from llama_index.core.evaluation import RelevancyEvaluator, FaithfulnessEvaluator

Evaluate relevance

评估相关性

relevancy = RelevancyEvaluator() result = relevancy.evaluate_response( query="What is Python?", response=response ) print(f"Relevancy: {result.passing}")

Evaluate faithfulness (no hallucination)

评估忠实度（无幻觉）

faithfulness = FaithfulnessEvaluator() result = faithfulness.evaluate_response( query="What is Python?", response=response ) print(f"Faithfulness: {result.passing}")

undefined

faithfulness = FaithfulnessEvaluator() result = faithfulness.evaluate_response( query="What is Python?", response=response ) print(f"Faithfulness: {result.passing}")

undefined

Best practices

最佳实践

Use vector indices for most cases - Best performance
Save indices to disk - Avoid re-indexing
Chunk documents properly - 512-1024 tokens optimal
Add metadata - Enables filtering and tracking
Use streaming - Better UX for long responses
Enable verbose during dev - See retrieval process
Evaluate responses - Check relevance and faithfulness
Use chat engine for conversations - Built-in memory
Persist storage - Don't lose your index
Monitor costs - Track embedding and LLM usage

大多数场景使用向量索引 - 性能最佳
将索引保存到磁盘 - 避免重复索引
合理拆分文档 - 512-1024个token为最优
添加元数据 - 支持过滤与追踪
使用流式响应 - 长响应时用户体验更好
开发阶段启用verbose - 查看检索过程
评估响应 - 检查相关性与忠实度
对话场景使用聊天引擎 - 内置记忆功能
持久化存储 - 不要丢失你的索引
监控成本 - 追踪嵌入与LLM使用量

Common patterns

常见模式

Document Q&A system

文档问答系统

python

undefined

python

undefined

Complete RAG pipeline

完整的RAG流水线

documents = SimpleDirectoryReader("docs").load_data() index = VectorStoreIndex.from_documents(documents) index.storage_context.persist(persist_dir="./storage")

Query

查询

query_engine = index.as_query_engine( similarity_top_k=3, response_mode="compact", verbose=True ) response = query_engine.query("What is the main topic?") print(response) print(f"Sources: {[node.metadata['file_name'] for node in response.source_nodes]}")

undefined

query_engine = index.as_query_engine( similarity_top_k=3, response_mode="compact", verbose=True ) response = query_engine.query("What is the main topic?") print(response) print(f"Sources: {[node.metadata['file_name'] for node in response.source_nodes]}")

undefined

Chatbot with memory

带记忆的聊天机器人

python

undefined

python

undefined

Conversational interface

会话式交互界面

chat_engine = index.as_chat_engine( chat_mode="condense_plus_context", verbose=True )

Multi-turn chat

多轮聊天

while True: user_input = input("You: ") if user_input.lower() == "quit": break response = chat_engine.chat(user_input) print(f"Bot: {response}")

undefined

while True: user_input = input("You: ") if user_input.lower() == "quit": break response = chat_engine.chat(user_input) print(f"Bot: {response}")

undefined

Performance benchmarks

性能基准

Operation	Latency	Notes
Index 100 docs	~10-30s	One-time, can persist
Query (vector)	~0.5-2s	Retrieval + LLM
Streaming query	~0.5s first token	Better UX
Agent with tools	~3-8s	Multiple tool calls

操作	延迟	说明
索引100个文档	~10-30秒	一次性操作，可持久化
向量查询	~0.5-2秒	检索 + LLM处理
流式查询	~0.5秒返回首个token	用户体验更好
带工具的Agent	~3-8秒	多工具调用

LlamaIndex vs LangChain

Feature	LlamaIndex	LangChain
Best for	RAG, document Q&A	Agents, general LLM apps
Data connectors	300+ (LlamaHub)	100+
RAG focus	Core feature	One of many
Learning curve	Easier for RAG	Steeper
Customization	High	Very high
Documentation	Excellent	Good

Use LlamaIndex when:

Your primary use case is RAG
Need many data connectors
Want simpler API for document Q&A
Building knowledge retrieval system

Use LangChain when:

Building complex agents
Need more general-purpose tools
Want more flexibility
Complex multi-step workflows

特性	LlamaIndex	LangChain
最适合场景	RAG、文档问答	Agent、通用LLM应用
数据连接器	300+（LlamaHub）	100+
RAG专注度	核心功能	众多功能之一
学习曲线	针对RAG更简单	更陡峭
自定义能力	高	极高
文档质量	优秀	良好

选择LlamaIndex的场景:

核心用例是RAG
需要大量数据连接器
想要更简洁的文档问答API
构建知识检索系统

选择LangChain的场景:

构建复杂Agent
需要更多通用工具
想要更高灵活性
复杂多步骤工作流

References

参考资料

Query Engines Guide - Query modes, customization, streaming
Agents Guide - Tool creation, RAG agents, multi-step reasoning
Data Connectors Guide - 300+ connectors, custom loaders

查询引擎指南 - 查询模式、自定义、流式响应
Agent指南 - 工具创建、RAG Agent、多步骤推理
数据连接器指南 - 300+连接器、自定义加载器

Resources

资源

GitHub: https://github.com/run-llama/llama_index ⭐ 45,100+
Docs: https://developers.llamaindex.ai/python/framework/
LlamaHub: https://llamahub.ai (data connectors)
LlamaCloud: https://cloud.llamaindex.ai (enterprise)
Discord: https://discord.gg/dGcwcsnxhU
Version: 0.14.7+
License: MIT

GitHub: https://github.com/run-llama/llama_index ⭐ 45,100+
文档: https://developers.llamaindex.ai/python/framework/
LlamaHub: https://llamahub.ai（数据连接器）
LlamaCloud: https://cloud.llamaindex.ai（企业版）
Discord: https://discord.gg/dGcwcsnxhU
版本: 0.14.7+
许可证: MIT

llamaindex

Original

Translation

LlamaIndex - Data Framework for LLM Applications

LlamaIndex - 面向LLM应用的数据框架

When to use LlamaIndex

何时使用LlamaIndex

Quick start

快速开始

Installation

安装

Starter package (recommended)

入门包（推荐）

Or minimal core + specific integrations

或最小核心包 + 特定集成

5-line RAG example

5行代码实现RAG示例

Load documents

加载文档

Create index

创建索引

Query

查询

Core concepts

核心概念

1. Data connectors - Load documents

1. 数据连接器 - 加载文档

Directory of files

加载目录文件

Web pages

加载网页

GitHub repository

加载GitHub仓库

Manual document creation

手动创建文档

2. Indices - Structure data

2. 索引 - 结构化数据

Vector index (most common - semantic search)

向量索引（最常用 - 语义搜索）

List index (sequential scan)

列表索引（顺序扫描）

Tree index (hierarchical summary)

树形索引（分层摘要）

Save index

保存索引

Load index

加载索引

3. Query engines - Ask questions

3. 查询引擎 - 提问

Basic query

基础查询

Streaming response

流式响应

Custom configuration

自定义配置

4. Retrievers - Find relevant chunks

4. 检索器 - 查找相关片段

Vector retriever

向量检索器

With filtering

带过滤的检索器

Custom retriever

自定义检索器

Agents with tools

带工具的Agent

Basic agent

基础Agent

Define tools

定义工具

Create agent

创建Agent

Use agent

使用Agent

RAG agent (document search + tools)

RAG Agent（文档搜索 + 工具）

Create index as before

如之前一样创建索引

Wrap query engine as tool

将查询引擎包装为工具

Agent with document search + calculator