SelfAskWithSearchChain Class — langchain Architecture

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

Class python

Entity Profile

Dependency Diagram

graph TD
  8b500caa_9954_b0fb_3155_01b166f2ddb7["SelfAskWithSearchChain"]
  493d4ce4_2303_08e0_2337_2bee34fe2662["AgentExecutor"]
  8b500caa_9954_b0fb_3155_01b166f2ddb7 -->|extends| 493d4ce4_2303_08e0_2337_2bee34fe2662
  1fab5269_943e_e7a5_4729_eee14e04448a["base.py"]
  8b500caa_9954_b0fb_3155_01b166f2ddb7 -->|defined in| 1fab5269_943e_e7a5_4729_eee14e04448a
  e8fc2157_5457_e275_2219_a88c3412bb53["__init__()"]
  8b500caa_9954_b0fb_3155_01b166f2ddb7 -->|method| e8fc2157_5457_e275_2219_a88c3412bb53

Relationship Graph

Source Code

libs/langchain/langchain_classic/agents/self_ask_with_search/base.py lines 77–94

class SelfAskWithSearchChain(AgentExecutor):
    """[Deprecated] Chain that does self-ask with search."""

    def __init__(
        self,
        llm: BaseLanguageModel,
        search_chain: GoogleSerperAPIWrapper | SearchApiAPIWrapper | SerpAPIWrapper,
        **kwargs: Any,
    ):
        """Initialize only with an LLM and a search chain."""
        search_tool = Tool(
            name="Intermediate Answer",
            func=search_chain.run,
            coroutine=search_chain.arun,
            description="Search",
        )
        agent = SelfAskWithSearchAgent.from_llm_and_tools(llm, [search_tool])
        super().__init__(agent=agent, tools=[search_tool], **kwargs)