load_query_constructor_chain#

langchain.chains.query_constructor.base.load_query_constructor_chain(
llm: BaseLanguageModel,
document_contents: str,
attribute_info: Sequence[AttributeInfo | dict],
examples: list | None = None,
allowed_comparators: Sequence[Comparator] = (Comparator.EQ, Comparator.NE, Comparator.GT, Comparator.GTE, Comparator.LT, Comparator.LTE, Comparator.CONTAIN, Comparator.LIKE, Comparator.IN, Comparator.NIN),
allowed_operators: Sequence[Operator] = (Operator.AND, Operator.OR, Operator.NOT),
enable_limit: bool = False,
schema_prompt: BasePromptTemplate | None = None,
**kwargs: Any,
) → LLMChain[source]#

Deprecated since version 0.2.13: Use load_query_constructor_runnable() instead. It will not be removed until langchain==1.0.

Load a query constructor chain.

Parameters:

  • llm (BaseLanguageModel) – BaseLanguageModel to use for the chain.

  • document_contents (str) – The contents of the document to be queried.

  • attribute_info (Sequence[AttributeInfo | dict]) – Sequence of attributes in the document.

  • examples (list | None) – Optional list of examples to use for the chain.

  • allowed_comparators (Sequence[Comparator]) – Sequence of allowed comparators. Defaults to all Comparators.

  • allowed_operators (Sequence[Operator]) – Sequence of allowed operators. Defaults to all Operators.

  • enable_limit (bool) – Whether to enable the limit operator. Defaults to False.

  • schema_prompt (BasePromptTemplate | None) – Prompt for describing query schema. Should have string input variables allowed_comparators and allowed_operators.

  • **kwargs (Any) – Arbitrary named params to pass to LLMChain.

Returns:

A LLMChain that can be used to construct queries.

Return type:

LLMChain