motleycrew.tools.autogen_chat_tool

Functions

create_autogen_chat_tool(name, description, ...)

get_last_message(chat_result)

Classes

AutoGenChatTool(name, description, prompt, ...)

A tool for incorporating AutoGen chats into MotleyCrew.
motleycrew.tools.autogen_chat_tool.get_last_message(chat_result: None) → str
class motleycrew.tools.autogen_chat_tool.AutoGenChatTool(name: str, description: str, prompt: str | ~langchain_core.prompts.base.BasePromptTemplate, initiator: None, recipient: None, result_extractor: ~typing.Callable[[None], ~typing.Any] = <function get_last_message>, input_schema: ~typing.Type[~pydantic.v1.main.BaseModel] | None = None, return_direct: bool = False, exceptions_to_reflect: ~typing.List[Exception] | None = None)

Bases: MotleyTool

A tool for incorporating AutoGen chats into MotleyCrew.

__init__(name: str, description: str, prompt: str | ~langchain_core.prompts.base.BasePromptTemplate, initiator: None, recipient: None, result_extractor: ~typing.Callable[[None], ~typing.Any] = <function get_last_message>, input_schema: ~typing.Type[~pydantic.v1.main.BaseModel] | None = None, return_direct: bool = False, exceptions_to_reflect: ~typing.List[Exception] | None = None)
Parameters:

  • name – Name of the tool.

  • description – Description of the tool.

  • prompt – Prompt to use for the tool. Can be a string or a PromptTemplate object.

  • initiator – The agent initiating the chat.

  • recipient – The first recipient agent. This is the agent that you would specify in initiate_chat arguments.

  • result_extractor – Function to extract the result from the chat result.

  • input_schema – Input schema for the tool. The input variables should match the variables in the prompt. If not provided, a schema will be generated based on the input variables in the prompt, if any, with string fields.

motleycrew.tools.autogen_chat_tool.create_autogen_chat_tool(name: str, description: str, prompt: str | BasePromptTemplate, initiator: None, recipient: None, result_extractor: Callable[[None], Any], input_schema: Type[BaseModel] | None = None)