Hierarchical Planner - Observation: requests_put is not a valid tool, try one of [requests_get, requests_post].

codebanesr commented 1 year ago

System Info

version 0.300

\nObservation: requests_put is not a valid tool, try one of [requests_get, requests_post].\nThought:I need to use requests_post to update the pet\'s ship date.\n\nAction:

This keep happening time and time again!

Who can help?

hierarchical planner, requests agent

https://github.com/openchatai/OpenCopilot/blob/ecd2f51bfc1f989133a44c4cff15d858b0d2fd30/llm-server/api_caller/planner.py#L319

Information

[ ] The official example notebooks/scripts
[ ] My own modified scripts

Related Components

[ ] LLMs/Chat Models
[ ] Embedding Models
[ ] Prompts / Prompt Templates / Prompt Selectors
[ ] Output Parsers
[ ] Document Loaders
[ ] Vector Stores / Retrievers
[ ] Memory
[X] Agents / Agent Executors
[X] Tools / Toolkits
[ ] Chains
[ ] Callbacks/Tracing
[ ] Async

Reproduction

--

Expected behavior

The agent should be able to use the put tool when it needs to use it

dosubot[bot] commented 1 year ago

🤖

Hello,

Thank you for bringing this to our attention. The issue you're experiencing is due to the fact that the "requests_put" tool is not included in the list of allowed tools for the agent. The function _create_api_controller_agent creates an agent with a list of allowed tools, which currently only includes the "requests_get" and "requests_post" tools.

To allow the "requests_put" tool, you would need to include it in the list of allowed tools when creating the agent. Here's how you could modify the function to include the "requests_put" tool:

put_llm_chain = LLMChain(llm=llm, prompt=PARSING_PUT_PROMPT)
tools: List[BaseTool] = [
    RequestsGetToolWithParsing(
        requests_wrapper=requests_wrapper, llm_chain=get_llm_chain
    ),
    RequestsPostToolWithParsing(
        requests_wrapper=requests_wrapper, llm_chain=post_llm_chain
    ),
    RequestsPutToolWithParsing(
        requests_wrapper=requests_wrapper, llm_chain=put_llm_chain
    ),
]

This modification creates an instance of the "requests_put" tool and adds it to the list of allowed tools for the agent. The Hierarchical Planner should then recognize "requests_put" as a valid tool.

Please note that the "requests_put" tool is implemented in the LangChain codebase. It is a class named "RequestsPutTool" that inherits from the "BaseRequestsTool" and "BaseTool" classes. This tool is used for making a PUT request to an API endpoint. The tool's functionality is tested in the unit test "test_requests_put_tool" in the file "libs/langchain/tests/unit_tests/tools/requests/test_tool.py". The test ensures that the tool correctly makes PUT requests and handles input data.

I hope this helps! If you have any other questions or need further clarification, please don't hesitate to ask.

Best, Dosu

Sources

This response is meant to be useful and save you time. It is not meant to be a precise solution, but rather a starting point for your own research.

Help me be more useful! Please leave a 👍 if this is helpful and 👎 if it is irrelevant.

If you want to continue the conversation, start your reply with @dosu-bot.

codebanesr commented 1 year ago

I found it even before you replied. Thanks bot. I made the changes

codebanesr commented 1 year ago

🤖

Hello,

Thank you for bringing this to our attention. The issue you're experiencing is due to the fact that the "requests_put" tool is not included in the list of allowed tools for the agent. The function _create_api_controller_agent creates an agent with a list of allowed tools, which currently only includes the "requests_get" and "requests_post" tools.

To allow the "requests_put" tool, you would need to include it in the list of allowed tools when creating the agent. Here's how you could modify the function to include the "requests_put" tool:
put_llm_chain = LLMChain(llm=llm, prompt=PARSING_PUT_PROMPT)
tools: List[BaseTool] = [
    RequestsGetToolWithParsing(
        requests_wrapper=requests_wrapper, llm_chain=get_llm_chain
    ),
    RequestsPostToolWithParsing(
        requests_wrapper=requests_wrapper, llm_chain=post_llm_chain
    ),
    RequestsPutToolWithParsing(
        requests_wrapper=requests_wrapper, llm_chain=put_llm_chain
    ),
]
This modification creates an instance of the "requests_put" tool and adds it to the list of allowed tools for the agent. The Hierarchical Planner should then recognize "requests_put" as a valid tool.

Please note that the "requests_put" tool is implemented in the LangChain codebase. It is a class named "RequestsPutTool" that inherits from the "BaseRequestsTool" and "BaseTool" classes. This tool is used for making a PUT request to an API endpoint. The tool's functionality is tested in the unit test "test_requests_put_tool" in the file "libs/langchain/tests/unit_tests/tools/requests/test_tool.py". The test ensures that the tool correctly makes PUT requests and handles input data.

I hope this helps! If you have any other questions or need further clarification, please don't hesitate to ask.

Best, Dosu

Sources
* [libs/langchain/langchain/agents/agent_toolkits/openapi/planner.py](https://github.com/langchain-ai/langchain/blob/master/libs/langchain/langchain/agents/agent_toolkits/openapi/planner.py)

* [libs/langchain/langchain/tools/requests/tool.py](https://github.com/langchain-ai/langchain/blob/master/libs/langchain/langchain/tools/requests/tool.py)

* [libs/langchain/tests/unit_tests/tools/requests/test_tool.py](https://github.com/langchain-ai/langchain/blob/master/libs/langchain/tests/unit_tests/tools/requests/test_tool.py)
This response is meant to be useful and save you time. It is not meant to be a precise solution, but rather a starting point for your own research.

Help me be more useful! Please leave a 👍 if this is helpful and 👎 if it is irrelevant.

If you want to continue the conversation, start your reply with @dosu-bot.

Why is it not included in the official repository though?!

dosubot[bot] commented 1 year ago

🤖

Hello,

Yes, your approach to include the "requests_put" tool in the _create_api_controller_agent function by creating an instance of it and adding it to the list of allowed tools is correct.

In the LangChain framework, the RequestsToolkit class has a get_tools method that returns a list of tools, which includes RequestsPutTool. This list of tools is then used in the OpenAPIToolkit class, where the get_tools method returns a list of tools that includes all the tools from RequestsToolkit (which includes RequestsPutTool) and json_agent_tool.

Here is the relevant code:

class RequestsToolkit(BaseToolkit):
    """Toolkit for making REST requests."""

    def get_tools(self) -> List[BaseTool]:
        """Return a list of tools."""
        return [
            ...
            RequestsPutTool(requests_wrapper=self.requests_wrapper),
            ...
        ]

class OpenAPIToolkit(BaseToolkit):
    """Toolkit for interacting with an OpenAPI API."""

    def get_tools(self) -> List[BaseTool]:
        """Get the tools in the toolkit."""
        request_toolkit = RequestsToolkit(requests_wrapper=self.requests_wrapper)
        return [*request_toolkit.get_tools(), json_agent_tool]

This code shows that RequestsPutTool is included in the list of tools returned by get_tools in OpenAPIToolkit, which means it will be available for use in the _create_api_controller_agent function.

As for your question about the RequestsPutToolWithParsing class, I wasn't able to find specific information about this class in the repository. It's possible that this class is a subclass of RequestsPutTool that includes additional functionality for parsing the response from a PUT request, but without more information, I can't confirm this.

If you have any more questions or need further clarification, feel free to ask.

Best, Dosu

Sources

libs/langchain/langchain/agents/agent_toolkits/openapi/toolkit.py

This response is meant to be useful and save you time. It is not meant to be a precise solution, but rather a starting point for your own research.

Help me be more useful! Please leave a 👍 if this is helpful and 👎 if it is irrelevant.

If you want to continue the conversation, start your reply with @dosu-bot.

dosubot[bot] commented 10 months ago

Hi, @codebanesr,

I'm helping the LangChain team manage their backlog and am marking this issue as stale. The issue was raised by you regarding the use of the requests_put tool in the hierarchical planner, which was not valid. I provided guidance on modifying the function to include the "requests_put" tool, and you confirmed making the necessary changes. Additionally, you questioned why the "requests_put" tool was not included in the official repository, to which I explained its inclusion in the LangChain framework.

Could you please confirm if this issue is still relevant to the latest version of the LangChain repository? If it is, please let the LangChain team know by commenting on the issue. Otherwise, feel free to close the issue yourself, or it will be automatically closed in 7 days. Thank you!

langchain-ai / langchain