参考链接：https://docs.langchain.com/oss/python/langchain/agents
https://reference.langchain.com/python/langchain/agents/?_gl=1*16rzxkd*_gcl_au*MTE0NTQ0OC4xNzY3ODQyNTI1*_ga*MTU3ODc0MjQ4NC4xNzY3ODQyNTI1*_ga_47WX3HKKY2*czE3Njc4NjIzNTYkbzQkZzEkdDE3Njc4NjQ0MTQkajYwJGwwJGgw#langchain.agents.create_agent

langchain使用create_agent函数将大模型和工具等结合起来，实现智能体开发
此处研究一下该函数的输入和含义
create_agent(
    model: str | BaseChatModel,
    tools: Sequence[BaseTool | Callable | dict[str, Any]] | None = None,
    *,
    system_prompt: str | SystemMessage | None = None,
    middleware: Sequence[AgentMiddleware[StateT_co, ContextT]] = (),
    response_format: ResponseFormat[ResponseT] | type[ResponseT] | None = None,
    state_schema: type[AgentState[ResponseT]] | None = None,
    context_schema: type[ContextT] | None = None,
    checkpointer: Checkpointer | None = None,
    store: BaseStore | None = None,
    interrupt_before: list[str] | None = None,
    interrupt_after: list[str] | None = None,
    debug: bool = False,
    name: str | None = None,
    cache: BaseCache | None = None,
) -> CompiledStateGraph[
    AgentState[ResponseT], ContextT, _InputAgentState, _OutputAgentState[ResponseT]

# model
指代智能体的大语言模型，可以是字符串或者模型实体
其中支持的字符串包含以下内容：（https://reference.langchain.com/python/langchain/models/#langchain.chat_models.init_chat_model(model_provider)）
- `openai` -> [`langchain-openai`](https://docs.langchain.com/oss/python/integrations/providers/openai)
- `anthropic` -> [`langchain-anthropic`](https://docs.langchain.com/oss/python/integrations/providers/anthropic)
- `azure_openai` -> [`langchain-openai`](https://docs.langchain.com/oss/python/integrations/providers/openai)
- `azure_ai` -> [`langchain-azure-ai`](https://docs.langchain.com/oss/python/integrations/providers/microsoft)
- `google_vertexai` -> [`langchain-google-vertexai`](https://docs.langchain.com/oss/python/integrations/providers/google)
- `google_genai` -> [`langchain-google-genai`](https://docs.langchain.com/oss/python/integrations/providers/google)
- `bedrock` -> [`langchain-aws`](https://docs.langchain.com/oss/python/integrations/providers/aws)
- `bedrock_converse` -> [`langchain-aws`](https://docs.langchain.com/oss/python/integrations/providers/aws)
- `cohere` -> [`langchain-cohere`](https://docs.langchain.com/oss/python/integrations/providers/cohere)
- `fireworks` -> [`langchain-fireworks`](https://docs.langchain.com/oss/python/integrations/providers/fireworks)
- `together` -> [`langchain-together`](https://docs.langchain.com/oss/python/integrations/providers/together)
- `mistralai` -> [`langchain-mistralai`](https://docs.langchain.com/oss/python/integrations/providers/mistralai)
- `huggingface` -> [`langchain-huggingface`](https://docs.langchain.com/oss/python/integrations/providers/huggingface)
- `groq` -> [`langchain-groq`](https://docs.langchain.com/oss/python/integrations/providers/groq)
- `ollama` -> [`langchain-ollama`](https://docs.langchain.com/oss/python/integrations/providers/ollama)
- `google_anthropic_vertex` -> [`langchain-google-vertexai`](https://docs.langchain.com/oss/python/integrations/providers/google)
- `deepseek` -> [`langchain-deepseek`](https://docs.langchain.com/oss/python/integrations/providers/deepseek)
- `ibm` -> [`langchain-ibm`](https://docs.langchain.com/oss/python/integrations/providers/ibm)
- `nvidia` -> [`langchain-nvidia-ai-endpoints`](https://docs.langchain.com/oss/python/integrations/providers/nvidia)
- `xai` -> [`langchain-xai`](https://docs.langchain.com/oss/python/integrations/providers/xai)
- `perplexity` -> [`langchain-perplexity`](https://docs.langchain.com/oss/python/integrations/providers/perplexity)
- `upstage` -> [`langchain-upstage`](https://docs.langchain.com/oss/python/integrations/providers/upstage)

尝试一下deepseek

from langchain.agents import create_agent
from langchain.chat_models import init_chat_model
import os


os.environ["DEEPSEEK_API_KEY"] = "sk-XXXX"
# model = init_chat_model("deepseek-chat")

agent=create_agent(
    model="deepseek-reasoner",
    tools=None
)


#运行agent，获得结果
result=agent.invoke(
    {
        "messages":[{"role":"user","content":"what is the weather in sf"}]
    }
)

print(result)

注意model="deepseek-reasoner"，此处是填的模型名称，而通过deepseek官网，支持两个模型deepseek-chat，deepseek-reasoner

参考链接：https://api-docs.deepseek.com/zh-cn/quick_start/pricing

另外在实验过程中报错，需要安装相关库