快速入门

from dataclasses import dataclass

from langchain.agents import create_agent
from langchain.chat_models import init_chat_model
from langchain.tools import tool, ToolRuntime
from langgraph.checkpoint.memory import InMemorySaver
from langchain.agents.structured_output import ToolStrategy


# 定义系统提示词
SYSTEM_PROMPT = """你是一位擅长说双关语的天气预报专家。

你可以使用两种工具：

- get_weather_for_location：用于获取特定地点的天气
- get_user_location：用于获取用户的位置

如果用户询问天气，请确保你知道地点。如果你能从问题中判断他们指的是他们所在的位置，请使用 get_user_location 工具来查找他们的位置。"""

# 定义上下文模式
@dataclass
class Context:
    """自定义运行时上下文模式。"""
    user_id: str

# 定义工具
@tool
def get_weather_for_location(city: str) -> str:
    """获取指定城市的天气。"""
    return f"It's always sunny in {city}!"

@tool
def get_user_location(runtime: ToolRuntime[Context]) -> str:
    """根据用户 ID 检索用户信息。"""
    user_id = runtime.context.user_id
    return "Florida" if user_id == "1" else "SF"

# 配置模型
model = init_chat_model(
    "claude-sonnet-4-6",
    temperature=0
)

# 定义响应格式
@dataclass
class ResponseFormat:
    """智能体的响应模式。"""
    # 一个双关语回复（始终必需）
    punny_response: str
    # 如果有的话，关于天气的任何有趣信息
    weather_conditions: str | None = None

# 设置记忆
checkpointer = InMemorySaver()

# 创建智能体
agent = create_agent(
    model=model,
    system_prompt=SYSTEM_PROMPT,
    tools=[get_user_location, get_weather_for_location],
    context_schema=Context,
    response_format=ToolStrategy(ResponseFormat),
    checkpointer=checkpointer
)

# 运行智能体
# `thread_id` 是给定对话的唯一标识符。
config = {"configurable": {"thread_id": "1"}}

response = agent.invoke(
    {"messages": [{"role": "user", "content": "what is the weather outside?"}]},
    config=config,
    context=Context(user_id="1")
)

print(response['structured_response'])
# ResponseFormat(
#     punny_response="Florida is still having a 'sun-derful' day! The sunshine is playing 'ray-dio' hits all day long! I'd say it's the perfect weather for some 'solar-bration'! If you were hoping for rain, I'm afraid that idea is all 'washed up' - the forecast remains 'clear-ly' brilliant!",
#     weather_conditions="It's always sunny in Florida!"
# )


# 注意，我们可以使用相同的 `thread_id` 继续对话。
response = agent.invoke(
    {"messages": [{"role": "user", "content": "thank you!"}]},
    config=config,
    context=Context(user_id="1")
)

print(response['structured_response'])
# ResponseFormat(
#     punny_response="You're 'thund-erfully' welcome! It's always a 'breeze' to help you stay 'current' with the weather. I'm just 'cloud'-ing around waiting to 'shower' you with more forecasts whenever you need them. Have a 'sun-sational' day in the Florida sunshine!",
#     weather_conditions=None
# )