Hybrid SearchInstall

Summary

Search for documents using a hybrid search that combines multiple retrievers and optionally a reranker to return matching documents based on a textual query.

Inputs

• Query (required) — The text string to perform the search for.
• Text Field Mapping — A comma-separated list of metadata fields whose values should be used as the text content of each document for searching.

Output shape

A list of document objects, each containing 'pageContent' (the matched text content) and 'metadata' fields.

Returns all matching documents found by the hybrid search. The outputs are mapped to the node's main output as an array of JSON objects. Optional reranker input can refine results if connected.

Examples

Example 1: Retrieve relevant documents related to a user query

Provide the query text and optionally customize which metadata fields to consider as the document content for search.