MapReduceChain Class — langchain Architecture

Architecture documentation for the MapReduceChain class in mapreduce.py from the langchain codebase.

Class python

Entity Profile

Dependency Diagram

graph TD
  5201e8ad_1740_ce23_1ab6_cafdbc91f78c["MapReduceChain"]
  f3cef70e_11b0_61c9_7ec0_7308f4b45056["Chain"]
  5201e8ad_1740_ce23_1ab6_cafdbc91f78c -->|extends| f3cef70e_11b0_61c9_7ec0_7308f4b45056
  2b877f61_7216_3d99_402d_5c7078ad6783["mapreduce.py"]
  5201e8ad_1740_ce23_1ab6_cafdbc91f78c -->|defined in| 2b877f61_7216_3d99_402d_5c7078ad6783
  d4b15165_0193_5a67_e7ae_50a24fe2f0aa["from_params()"]
  5201e8ad_1740_ce23_1ab6_cafdbc91f78c -->|method| d4b15165_0193_5a67_e7ae_50a24fe2f0aa
  1f8d05e7_9a50_ba11_3f5e_938bf022930b["input_keys()"]
  5201e8ad_1740_ce23_1ab6_cafdbc91f78c -->|method| 1f8d05e7_9a50_ba11_3f5e_938bf022930b
  dbf5f27c_025c_c644_9a87_7598ca49d74b["output_keys()"]
  5201e8ad_1740_ce23_1ab6_cafdbc91f78c -->|method| dbf5f27c_025c_c644_9a87_7598ca49d74b
  d4377005_d592_4f81_ac05_8d26750ca867["_call()"]
  5201e8ad_1740_ce23_1ab6_cafdbc91f78c -->|method| d4377005_d592_4f81_ac05_8d26750ca867

Relationship Graph

Source Code

libs/langchain/langchain_classic/chains/mapreduce.py lines 39–116

class MapReduceChain(Chain):
    """Map-reduce chain."""

    combine_documents_chain: BaseCombineDocumentsChain
    """Chain to use to combine documents."""
    text_splitter: TextSplitter
    """Text splitter to use."""
    input_key: str = "input_text"
    output_key: str = "output_text"

    @classmethod
    def from_params(
        cls,
        llm: BaseLanguageModel,
        prompt: BasePromptTemplate,
        text_splitter: TextSplitter,
        callbacks: Callbacks = None,
        combine_chain_kwargs: Mapping[str, Any] | None = None,
        reduce_chain_kwargs: Mapping[str, Any] | None = None,
        **kwargs: Any,
    ) -> MapReduceChain:
        """Construct a map-reduce chain that uses the chain for map and reduce."""
        llm_chain = LLMChain(llm=llm, prompt=prompt, callbacks=callbacks)
        stuff_chain = StuffDocumentsChain(
            llm_chain=llm_chain,
            callbacks=callbacks,
            **(reduce_chain_kwargs or {}),
        )
        reduce_documents_chain = ReduceDocumentsChain(
            combine_documents_chain=stuff_chain,
        )
        combine_documents_chain = MapReduceDocumentsChain(
            llm_chain=llm_chain,
            reduce_documents_chain=reduce_documents_chain,
            callbacks=callbacks,
            **(combine_chain_kwargs or {}),
        )
        return cls(
            combine_documents_chain=combine_documents_chain,
            text_splitter=text_splitter,
            callbacks=callbacks,
            **kwargs,
        )

    model_config = ConfigDict(
        arbitrary_types_allowed=True,
        extra="forbid",
    )

    @property
    def input_keys(self) -> list[str]:
        """Expect input key."""
        return [self.input_key]

    @property
    def output_keys(self) -> list[str]:
        """Return output key."""
        return [self.output_key]

    def _call(
        self,
        inputs: dict[str, str],
        run_manager: CallbackManagerForChainRun | None = None,
    ) -> dict[str, str]:
        _run_manager = run_manager or CallbackManagerForChainRun.get_noop_manager()
        # Split the larger text into smaller chunks.
        doc_text = inputs.pop(self.input_key)
        texts = self.text_splitter.split_text(doc_text)
        docs = [Document(page_content=text) for text in texts]
        _inputs: dict[str, Any] = {
            **inputs,
            self.combine_documents_chain.input_key: docs,
        }
        outputs = self.combine_documents_chain.run(
            _inputs,
            callbacks=_run_manager.get_child(),
        )
        return {self.output_key: outputs}