ConversationalChatAgent Class — langchain Architecture

Architecture documentation for the ConversationalChatAgent class in base.py from the langchain codebase.

Class python

Entity Profile

Dependency Diagram

graph TD
  37dfdb3a_fc75_8e43_a082_e000915175c2["ConversationalChatAgent"]
  37fcc3f3_2798_3648_915c_2bffdd19bff7["Agent"]
  37dfdb3a_fc75_8e43_a082_e000915175c2 -->|extends| 37fcc3f3_2798_3648_915c_2bffdd19bff7
  13cdba7d_c4a4_edb6_521f_09c39f95cf58["base.py"]
  37dfdb3a_fc75_8e43_a082_e000915175c2 -->|defined in| 13cdba7d_c4a4_edb6_521f_09c39f95cf58
  5e0d227f_9765_818c_3c02_743fc1bc0c86["_get_default_output_parser()"]
  37dfdb3a_fc75_8e43_a082_e000915175c2 -->|method| 5e0d227f_9765_818c_3c02_743fc1bc0c86
  00ca2fb7_cdfd_af44_4ed6_3ce8b4278290["_agent_type()"]
  37dfdb3a_fc75_8e43_a082_e000915175c2 -->|method| 00ca2fb7_cdfd_af44_4ed6_3ce8b4278290
  feb1728c_6b31_023b_2cde_af5cdf93de3c["observation_prefix()"]
  37dfdb3a_fc75_8e43_a082_e000915175c2 -->|method| feb1728c_6b31_023b_2cde_af5cdf93de3c
  750ee5bd_8767_0a1d_24d7_62ec922493ff["llm_prefix()"]
  37dfdb3a_fc75_8e43_a082_e000915175c2 -->|method| 750ee5bd_8767_0a1d_24d7_62ec922493ff
  37ebc42f_11ae_0dcf_a66a_08e26fb475df["_validate_tools()"]
  37dfdb3a_fc75_8e43_a082_e000915175c2 -->|method| 37ebc42f_11ae_0dcf_a66a_08e26fb475df
  a6206bc9_89e6_ac37_3164_590e3974b4f6["create_prompt()"]
  37dfdb3a_fc75_8e43_a082_e000915175c2 -->|method| a6206bc9_89e6_ac37_3164_590e3974b4f6
  3a5094fa_f1ec_4d12_ae46_7af52ecae3e1["_construct_scratchpad()"]
  37dfdb3a_fc75_8e43_a082_e000915175c2 -->|method| 3a5094fa_f1ec_4d12_ae46_7af52ecae3e1
  792d39b0_6cc6_56c1_0b6c_cc11d6f11d45["from_llm_and_tools()"]
  37dfdb3a_fc75_8e43_a082_e000915175c2 -->|method| 792d39b0_6cc6_56c1_0b6c_cc11d6f11d45

Relationship Graph

Source Code

libs/langchain/langchain_classic/agents/conversational_chat/base.py lines 37–181

class ConversationalChatAgent(Agent):
    """An agent designed to hold a conversation in addition to using tools."""

    output_parser: AgentOutputParser = Field(default_factory=ConvoOutputParser)
    """Output parser for the agent."""
    template_tool_response: str = TEMPLATE_TOOL_RESPONSE
    """Template for the tool response."""

    @classmethod
    @override
    def _get_default_output_parser(cls, **kwargs: Any) -> AgentOutputParser:
        return ConvoOutputParser()

    @property
    def _agent_type(self) -> str:
        raise NotImplementedError

    @property
    def observation_prefix(self) -> str:
        """Prefix to append the observation with.

        Returns:
            "Observation: "
        """
        return "Observation: "

    @property
    def llm_prefix(self) -> str:
        """Prefix to append the llm call with.

        Returns:
            "Thought: "
        """
        return "Thought:"

    @classmethod
    def _validate_tools(cls, tools: Sequence[BaseTool]) -> None:
        super()._validate_tools(tools)
        validate_tools_single_input(cls.__name__, tools)

    @classmethod
    def create_prompt(
        cls,
        tools: Sequence[BaseTool],
        system_message: str = PREFIX,
        human_message: str = SUFFIX,
        input_variables: list[str] | None = None,
        output_parser: BaseOutputParser | None = None,
    ) -> BasePromptTemplate:
        """Create a prompt for the agent.

        Args:
            tools: The tools to use.
            system_message: The `SystemMessage` to use.
            human_message: The `HumanMessage` to use.
            input_variables: The input variables to use.
            output_parser: The output parser to use.

        Returns:
            A `PromptTemplate`.
        """
        tool_strings = "\n".join(
            [f"> {tool.name}: {tool.description}" for tool in tools],
        )
        tool_names = ", ".join([tool.name for tool in tools])
        _output_parser = output_parser or cls._get_default_output_parser()
        format_instructions = human_message.format(
            format_instructions=_output_parser.get_format_instructions(),
        )
        final_prompt = format_instructions.format(
            tool_names=tool_names,
            tools=tool_strings,
        )
        if input_variables is None:
            input_variables = ["input", "chat_history", "agent_scratchpad"]
        messages = [
            SystemMessagePromptTemplate.from_template(system_message),
            MessagesPlaceholder(variable_name="chat_history"),
            HumanMessagePromptTemplate.from_template(final_prompt),
            MessagesPlaceholder(variable_name="agent_scratchpad"),
        ]