Metadata-Version: 2.1
Name: robotgpt
Version: 0.0.10
Summary: RobotGPT LLM 支持Langchain
Home-page: https://src.cloudminds.com/ai-api/robotgptllm
Author: blaze.zhang
Author-email: blaze.zhang@cloudminds.com
License: UNKNOWN
Platform: UNKNOWN
Requires-Dist: langchain

=================
RobotGPT LLM
=================

    RobotGPT 支持langchain

=================
Quick Install
=================

    `pip install robotgpt`

=================
使用样例
=================

agent块式输出::

    from robotgpt.robotgpt import RobotGPTLLM
    from langchain.agents import AgentType, initialize_agent
    from langchain.tools import BaseTool, StructuredTool, Tool, tool
    class CoffeeMaking:
        def inference(self):
            return "Making coffee requires a coffee machine, coffee beans, sugar packets, and paper cups."

    class ImageObjectDetect:
        def inference(self, obj):
            if obj == "sugar packets":
                return "No "+obj
            return "There is a "+obj

    class AskCustomer:
        def inference(self,):
            return "Hello! We don’t have any sugar packets at the moment. Do you need to add milk?"

    robotgpt_api_url = "https://dataai.harix.iamidata.com/llm/api/ask"  #流式智能问答统一适配服务，从用户控制台购买https://console.openai.iamidata.com/api/apiList
    model_name = "openai/gpt-3.5-turbo-0613"
    robotgpt_api_token = "Your token" #https://dataai-doc.dataarobotics.com/docs/getting-started/authentication
    llm = RobotGPTLLM(temperature=0, model_name=model_name,robotgpt_api_token=robotgpt_api_token,robotgpt_api_url=robotgpt_api_url)

    imgObjDetect = ImageObjectDetect()
    tools = [
        Tool.from_function(
            func=CoffeeMaking.inference,
            name="Coffee making",
            description="useful for when the user needs to make coffee."
            # coroutine= ... <- you can specify an async method if desired as well
        ),
        Tool.from_function(
            func=imgObjDetect.inference,
            name="Determine whether the object exists in the picture",
            description="useful for when you want to know what is inside the photo. receives object as input. The input to this tool should be a string, representing the object. "
            # coroutine= ... <- you can specify an async method if desired as well
        ),
        Tool.from_function(
            func=AskCustomer.inference,
            name="Ask the customer whether to add milk",
            description="useful for when making coffee without sugar packets, you can ask the customer whether you need to add milk. The input to this tool should be a bool, represents whether there is a sugar packet."
            # coroutine= ... <- you can specify an async method if desired as well
        ),
    ]
    agent = initialize_agent(
        tools, llm, agent=AgentType.ZERO_SHOT_REACT_DESCRIPTION, verbose=True
    )
    agent.run(
        "给我做杯咖啡"
    )      

块式输出::

    from robotgpt.robotgpt import RobotGPTLLM
    from langchain.schema import HumanMessage

    robotgpt_api_url = "https://dataai.harix.iamidata.com/llm/api/ask"  #流式智能问答统一适配服务
    model_name = "openai/gpt-3.5-turbo-0613"
    robotgpt_api_token = "Your token" #https://dataai-doc.dataarobotics.com/docs/getting-started/authentication

    llm = RobotGPTLLM(temperature=0, model_name=model_name,robotgpt_api_token=robotgpt_api_token,robotgpt_api_url=robotgpt_api_url)
    resp = llm([HumanMessage(content="Write me a song about sparkling water.")])
    print(resp)

流式输出::

    from langchain.callbacks import StreamingStdOutCallbackHandler
    from langchain.schema import HumanMessage
    from robotgpt.robotgpt import RobotGPTLLM
    robotgpt_api_url = "https://dataai.harix.iamidata.com/llm/api/ask"  #流式智能问答统一适配服务
    model_name = "openai/gpt-3.5-turbo-0613"
    robotgpt_api_token = "Your token" #https://dataai-doc.dataarobotics.com/docs/getting-started/authentication
    chat = RobotGPTLLM(streaming=True, callbacks=[StreamingStdOutCallbackHandler()], temperature=0,model_name=model_name,robotgpt_api_token=robotgpt_api_token,robotgpt_api_url=robotgpt_api_url)
    resp = chat([HumanMessage(content="Write me a song about sparkling water.")])
    print(resp)

