Package org.apache.lucene.search

Class FloatVectorSimilarityQuery.Adaptive

java.lang.Object
org.apache.lucene.search.Query
org.apache.lucene.search.FloatVectorSimilarityQuery
org.apache.lucene.search.FloatVectorSimilarityQuery.Adaptive
Enclosing class:
FloatVectorSimilarityQuery
public static non-sealed class FloatVectorSimilarityQuery.Adaptive extends FloatVectorSimilarityQuery
A FloatVectorSimilarityQuery with an adaptive threshold for graph traversal.

  • Field Details

    • field

      protected final String field

    • resultSimilarity

      protected final float resultSimilarity

    • decay

      protected final float decay

    • filter

      protected final Query filter

    • searchStrategy

      protected final KnnSearchStrategy searchStrategy

  • Constructor Details

    • Adaptive

      public Adaptive(String field, float[] target, float resultSimilarity, float decay, Query filter, KnnSearchStrategy searchStrategy)
      Search for all (approximate) float vectors above a similarity threshold using VectorSimilarityCollector, with a caller-supplied KnnSearchStrategy. If a filter is applied, it traverses as many nodes as the cost of the filter, and then falls back to exact search if results are incomplete.
      Parameters:
      field - a field that has been indexed as a KnnFloatVectorField.
      target - the target of the search.
      resultSimilarity - similarity score for result collection.
      decay - decay factor for graph traversal buffer.
      filter - a filter applied before the vector search.
      searchStrategy - the KnnSearchStrategy to use during graph search. If null, this query's own default is used: an KnnSearchStrategy.Hnsw with filteredSearchThreshold == 0, which preserves this query's filter handling. Note this differs from KnnSearchStrategy.Hnsw.DEFAULT, which uses a threshold of 60. The underlying format may not support all strategies and is free to ignore the requested strategy.

    • Adaptive

      public Adaptive(String field, float[] target, float resultSimilarity, float decay, Query filter)
      Search for all (approximate) float vectors above a similarity threshold using VectorSimilarityCollector, with the default KnnSearchStrategy. If a filter is applied, it traverses as many nodes as the cost of the filter, and then falls back to exact search if results are incomplete.
      Parameters:
      field - a field that has been indexed as a KnnFloatVectorField.
      target - the target of the search.
      resultSimilarity - similarity score for result collection.
      decay - decay factor for graph traversal buffer.
      filter - a filter applied before the vector search.

    • Adaptive

      public Adaptive(String field, float[] target, float resultSimilarity, Query filter)
      Search for all (approximate) float vectors above a similarity threshold using VectorSimilarityCollector. If a filter is applied, it traverses as many nodes as the cost of the filter, and then falls back to exact search if results are incomplete.
      Parameters:
      field - a field that has been indexed as a KnnFloatVectorField.
      target - the target of the search.
      resultSimilarity - similarity score for result collection.
      filter - a filter applied before the vector search.

    • Adaptive

      public Adaptive(String field, float[] target, float resultSimilarity)
      Search for all (approximate) float vectors above a similarity threshold using VectorSimilarityCollector.
      Parameters:
      field - a field that has been indexed as a KnnFloatVectorField.
      target - the target of the search.
      resultSimilarity - similarity score for result collection.

  • Method Details

    • toString

      public String toString(String field)
      Description copied from class: Query
      Prints a query to a string, with field assumed to be the default field and omitted.
      Specified by:
      toString in class Query

    • getKnnCollectorManager

      protected KnnCollectorManager getKnnCollectorManager()
      Returns a KnnCollectorManager that always builds a VectorSimilarityCollector configured with this query's searchStrategy.

      The manager's strategy and LeafReaderContext parameters are intentionally ignored: top-k queries may override the strategy on a per-search basis, but similarity- threshold queries pin the strategy to the query's own value at construction time so that Query.equals(java.lang.Object) / getSearchStrategy() remain a faithful description of search behavior.

    • getSearchStrategy

      public KnnSearchStrategy getSearchStrategy()
      Returns:
      the KnnSearchStrategy used during graph search.

    • createWeight

      public Weight createWeight(IndexSearcher searcher, ScoreMode scoreMode, float boost) throws IOException
      Description copied from class: Query
      Expert: Constructs an appropriate Weight implementation for this query.

      Only implemented by primitive queries, which re-write to themselves.

      Overrides:
      createWeight in class Query
      Parameters:
      scoreMode - How the produced scorers will be consumed.
      boost - The boost that is propagated by the parent queries.
      Throws:
      IOException

    • visit

      public void visit(QueryVisitor visitor)
      Description copied from class: Query
      Recurse through the query tree, visiting any child queries.
      Specified by:
      visit in class Query
      Parameters:
      visitor - a QueryVisitor to be called by each query in the tree