Module org.elasticsearch.server
Package org.elasticsearch.search.vectors
package org.elasticsearch.search.vectors
-
ClassDescriptionA query wrapper that ensures the inner query is always eligible for query caching.Exact knn query.A Lucene query that selects the highest-scoring child document for each parent block.An extension of
AcceptDocsthat provides additional methods to get an approximate cost and a BitSet representation of the accepted documents.An AcceptDocs that wraps a Bits instance.An AcceptDocs that accepts all documents.An AcceptDocs that wraps a ScorerSupplier.Exact knn query builder.AIVFKnnFloatVectorQuerythat uses the IVF search strategy.A query that matches the provided docs with their scores.A query that matches the provided docs with their scores.Defines a kNN search to run in the search request.A builder used inRestKnnSearchActionto convert the kNN REST request into aSearchRequestBuilder.A query that performs kNN search using Lucene'sKnnFloatVectorQueryorKnnByteVectorQuery.This interface includes the declaration of an abstract method, profile().Provides a mechanism for building a KNN query vector in an asynchronous manner during the rewrite phaseA LuceneQuerythat applies vector-based rescoring to an inner query's results.A wrapper class for the Lucene query generated by SparseVectorQueryBuilder#toQuery(SearchExecutionContext) (found in x-pack/core/ml/search).This query provides a simple post-filter for the provided Query to limit the results of the inner query to those that have a similarity above a certain threshold