LangChain实现AgentExecutor调用tools

在我们使用LangChain进行我们自己功能开发时，同时想借助AI能力来帮我们处理数据并按照一定顺序去执行方法时，那么我们就会需要使用到Agent。

这里由于我的方法比较单一，所以使用的是OpenAI Functions。不过建议是更换为OPENAI Tools，它支持并行。

必不可少的模型：

from langchain_openai import ChatOpenAI

llm = ChatOpenAI(model="gpt-3.5-turbo", temperature=0)

首先我们需要定义调用方法：

from langchain.agents import tool


@tool
def get_word_length(word: str) -> int:
    """Returns the length of a word."""
    return len(word)

接下来定义我们这个AI具体是来做什么的，自定义System Prompt,并且当我们想用历史记录的时候那么就可以按照以下方式来定义：

from langchain.prompts import MessagesPlaceholder

prompt = ChatPromptTemplate.from_messages(
    [
        (
            "system",
            "You are very powerful assistant, but bad at calculating lengths of words.",
        ),
        MessagesPlaceholder(variable_name='chat_history'),
        ("user", "{input}"),
        MessagesPlaceholder(variable_name="agent_scratchpad"),
    ]
)

接下来将我们的方法进行绑定：

llm_with_tools = llm.bind_functions(tools)

如果你想直接返回工具生成的结果，可以再加上自定义输出,这部分可以自己打断点去看看内部实现步骤并且对于流程很有帮助。

def parse(output):
    # If no function was invoked, return to user
    if "function_call" not in output.additional_kwargs:
        return AgentFinish(return_values={"output": output.content}, log=output.content)

    # Parse out the function call
    function_call = output.additional_kwargs["function_call"]
    name = function_call["name"]
    inputs = json.loads(function_call["arguments"])

    # If the cusResponse function was invoked, return to the user with the function inputs
    if name == "cusResponse":
        return AgentFinish(return_values=inputs, log=str(function_call))
    # Otherwise, return an agent action
    else:
        return AgentActionMessageLog(
            tool=name, tool_input=inputs, log="", message_log=[output]
        )

创建代理：

agent = (
            {
                "input": lambda x: x["input"],
                "agent_scratchpad": lambda x: format_to_openai_function_messages(
                    x["intermediate_steps"]
                ),
                "chat_history": lambda x: x["chat_history"],
            }
            | prompt
            | llm_with_tools
            | parse
    )

代理执行：

  agent_executor = AgentExecutor(agent=agent, tools=[searchVideos], verbose=False).with_config(
        {"run_name": "Agent"}
    )

正常的运行：

result=agent_executor.astream({"input": content, "chat_history": chat_history})

这里我想要流式输出结果：

 async for event in agent_executor.astream_events(
            {"input": content, "chat_history": chat_history},
            version="v1"
    ):
         kind = event["event"]
        if kind == "on_chain_start":
            if (
                    event["name"] == "Agent"
            ):  # Was assigned when creating the agent with `.with_config({"run_name": "Agent"})`
                print(
                    f"Starting agent: {event['name']} with input: {event['data'].get('input')}"
                )
        elif kind == "on_chain_end":
            if (
                    event["name"] == "Agent"
            ):  # Was assigned when creating the agent with `.with_config({"run_name": "Agent"})`
                print()
                print("--")
                print(
                    f"Done agent: {event['name']} with output: {event['data'].get('output')['output']}"
                )
        if kind == "on_chat_model_stream":
            print(event)
            content = event["data"]["chunk"].content
            if content:
                # Empty content in the context of OpenAI means
                # that the model is asking for a tool to be invoked.
                # So we only print non-empty content
                print(content, end="|")
        elif kind == "on_tool_start":
            print("--")
            print(
                f"Starting tool: {event['name']} with inputs: {event['data'].get('input')}"
            )
        elif kind == "on_tool_end":
            print(f"Done tool: {event['name']}")
            print(f"Tool output was: {event['data'].get('output')}")
            print("--")

以上就是LangChain对于Tools的调用。如果你还有什么疑问可以加下面的公众号，联系我获取更多资源和解答。

如果你想关注更多的AI最新信息和获取到更多关于AI的资源，请关注我们的公众号：