Package org.apache.lucene.codecs.lucene90

Class Lucene90HnswVectorsFormat

  • All Implemented Interfaces:
    NamedSPILoader.NamedSPI
    public final class Lucene90HnswVectorsFormat
extends KnnVectorsFormat
    Lucene 9.0 vector format, which encodes numeric vector values and an optional associated graph connecting the documents having values. The graph is used to power HNSW search. The format consists of three files:

    .vec (vector data) file

    This file stores all the floating-point vector data ordered by field, document ordinal, and vector dimension. The floats are stored in little-endian byte order.

    .vex (vector index) file

    Stores graphs connecting the documents for each field. For each document having a vector for a given field, this is stored as:

    • [int32] the number of neighbor nodes
    • array[vint] the neighbor ordinals, delta-encoded (initially subtracting -1)

    .vem (vector metadata) file

    For each field:

    • [int32] field number
    • [int32] vector similarity function ordinal
    • [vlong] offset to this field's vectors in the .vec file
    • [vlong] length of this field's vectors, in bytes
    • [vlong] offset to this field's index in the .vex file
    • [vlong] length of this field's index data, in bytes
    • [int] dimension of this field's vectors
    • [int] the number of documents having values for this field
    • array[vint] the docids of documents having vectors, in order
    • array[vlong] for each document having a vector, the offset (delta-encoded relative to the previous document) of its entry in the .vex file
    WARNING: This API is experimental and might change in incompatible ways in the next release.