indexers.VectorStore.reverse_search

indexers.VectorStore.reverse_search(
    query,
    max_n_results=100,
    partial_match=False,
)

Looks up documents in vectors by label.

Performs a label-based (non-semantic) lookup against the stored documents using a VectorStoreReverseSearchInput. For each query entry, matching documents are found by comparing the query’s doc_label against the label column of self.vectors. When partial_match is enabled, a document is considered a match if its label starts with the query label.

Any preprocessing hooks set on the instance are applied to the input before searching, and any postprocessing hooks are applied to the output before it is returned.

Parameters

Name	Type	Description	Default
query	VectorStoreReverseSearchInput	Input object containing the doc labels to look up in self.vectors and their corresponding ids.	required
max_n_results	int	Maximum number of matching documents to return per query entry. Pass -1 to return all matches. Defaults to 100.	100
partial_match	bool	If True, matches documents whose label starts with the query label (prefix matching). If False, only exact label matches are returned. Defaults to False.	False

Returns

Name	Type	Description
VectorStoreReverseSearchOutput	VectorStoreReverseSearchOutput	Output object containing the matched documents, with columns for id, searched_doc_label, doc_label, doc_text, and any metadata columns configured on the VectorStore.

Raises

Name	Type	Description
	DataValidationError	Raised if invalid arguments are passed.
	HookError	Raised if a preprocessing or postprocessing hook raises an exception.
	ClassifaiError	Raised if the reverse search operation fails.