LangGraph 简单项目文档

项目概述

langgraph_simple_project 是一个基于 LangGraph 框架构建的演示项目，展示了多智能体协作、内存管理、工具调用和自定义状态处理等核心功能。项目结合了 langchain 库和 qwen-plus 模型进行开发。

项目结构

langgraph_simple_project/
├── basics/
│ ├── langgraph_addMemory.py
│ ├── langgraph_chatbot_simple.py
│ ├── langgraph_custome_state.simple.py
│ ├── langgraph_custom_state_2_simple.py
│ └── langgraph_human_in_the_loop.py
├── prebuilts_agents/
│ ├── evals/
│ │ └── langgraph_evals_simple.py
│ ├── memory/
│ │ ├── langgraph_memory_simple.py
│ │ ├── long-term-memory-read-simple.py
│ │ ├── long-term-memory-write-simple.py
│ │ └── memory-write-in-tools-simple.py
│ └── multi_agent/
│ ├── multi_agent_handoff_simple.py
│ ├── multi_agent_swarm2_simple.py
│ ├── multi_agent_swarm_simple.py
│ ├── multi_agent_supervisor2_simple.py
│ └── multi_agent_supervisor_simple.py
├── .env
├── .idea/
│ └── misc.xml
├── .gitignore
├── langgraph.json
├── pyproject.toml
└── requirements.txt

环境配置

依赖安装

使用 Poetry 进行依赖管理：

poetry install
环境变量配置
.env 文件配置：

plaintext
复制
OPEN_API_KEY=sk-xxx
LANGSMITH_API_KEY=lsv2_sxxx
PYTHON_VERSION=3.11
核心功能模块
1. 多智能体协作
​文件位置:

prebuilts_agents/multi_agent/multi_agent_handoff_simple.py prebuilts_agents/multi_agent/multi_agent_swarm_simple.py prebuilts_agents/multi_agent/multi_agent_supervisor_simple.py

​功能描述:

创建航班预订和酒店预订智能体
实现智能体间的任务切换和协作
处理包含多个子任务的用户请求
​代码示例:

```python
transfer_to_hotel_assistant = create_handoff_tool(
   agent_name="hotel_assistant",
   description="Transfer user to the hotel-booking assistant."
)

flight_assistant = create_react_agent(
   model=model,
   tools=[book_flight, transfer_to_hotel_assistant],
   prompt="You are a flight booking assistant",
   name="flight_assistant"
)

2. 内存管理 ​文件位置:

prebuilts_agents/memory/long-term-memory-read-simple.py prebuilts_agents/memory/long-term-memory-write-simple.py ​功能描述:

使用 InMemoryStore 实现长期内存读写 存储和查询用户信息 ​代码示例:

python

@tool
def save_user_info(user_info: UserInfo, config:RunnableConfig) -> str:
   """save user info """
   store = get_store()
   user_id = config["configurable"].get("user_id")
   store.put(("users",), user_id, user_info)
   return "Successfully saved user info"

3. 工具调用 ​文件位置:

prebuilts_agents/langgraph_tools_simple.py

​功能描述:

定义自定义工具（如问候工具） 直接返回工具调用结果 ​代码示例:

python

@tool(return_direct=True)
def greet(user_name: str) -> int:
   """Greet user."""
   return f"Hello {user_name}!"

4. 自定义状态处理 ​文件位置:

basics/langgraph_custome_state.simple.py
basics/langgraph_custom_state_2_simple.py

​功能描述:

定义包含用户信息的自定义状态类 实现状态更新和验证机制 ​代码示例:

python

class State(TypedDict):
   messages: Annotated[list, add_messages]
   name: str
   birthday: str

@tool
def human_assistance(
   name: str, 
   birthday: str, 
   tool_call_id: Annotated[str, InjectedToolCallId]
) -> str:
   """Request assistance from a human."""
   human_response = interrupt({
       "question": "Is this correct?",
       "name": name,
       "birthday": birthday,
   })

运行示例 多智能体协作

python prebuilts_agents/multi_agent/multi_agent_handoff_simple.py

内存读写

python prebuilts_agents/memory/long-term-memory-read-simple.py

工具调用

python prebuilts_agents/langgraph_tools_simple.py

技术栈

LangGraph 框架
LangChain 库
Qwen-plus 模型
Poetry 依赖管理
Python 3.11