Package org.apache.lucene.backward_codecs.lucene90

Class Lucene90OnHeapHnswGraph

java.lang.Object

org.apache.lucene.util.hnsw.HnswGraph

org.apache.lucene.backward_codecs.lucene90.Lucene90OnHeapHnswGraph

public final class Lucene90OnHeapHnswGraph
extends HnswGraph

An HnswGraph where all nodes and connections are held in memory. This class is used to construct the HNSW graph before it's written to the index.

Nested Class Summary

Nested classes/interfaces inherited from class org.apache.lucene.util.hnsw.HnswGraph

HnswGraph.NodesIterator

Field Summary

Fields inherited from class org.apache.lucene.util.hnsw.HnswGraph

EMPTY

Method Summary

Modifier and Type	Method	Description
int	entryNode()
Lucene90NeighborArray	getNeighbors(int node)	Returns the NeighborQueue connected to the given node.
HnswGraph.NodesIterator	getNodesOnLevel(int level)
int	nextNeighbor()
int	numLevels()
static NeighborQueue	search(float[] query, int topK, int numSeed, RandomAccessVectorValues vectors, VectorSimilarityFunction similarityFunction, HnswGraph graphValues, Bits acceptOrds, int visitedLimit, SplittableRandom random)	Searches for the nearest neighbors of a query vector.
void	seek(int level, int targetNode)
int	size()

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Method Detail

search

public static NeighborQueue search(float[] query,
                                   int topK,
                                   int numSeed,
                                   RandomAccessVectorValues vectors,
                                   VectorSimilarityFunction similarityFunction,
                                   HnswGraph graphValues,
                                   Bits acceptOrds,
                                   int visitedLimit,
                                   SplittableRandom random)
                            throws IOException

Searches for the nearest neighbors of a query vector.

Parameters:

query - search query vector

topK - the number of nodes to be returned

numSeed - the size of the queue maintained while searching, and controls the number of random entry points to sample

vectors - vector values

graphValues - the graph values. May represent the entire graph, or a level in a hierarchical graph.

acceptOrds - Bits that represents the allowed document ordinals to match, or null if they are all allowed to match.

random - a source of randomness, used for generating entry points to the graph

Returns:

a priority queue holding the closest neighbors found

Throws:

IOException

getNeighbors

public Lucene90NeighborArray getNeighbors(int node)

Returns the NeighborQueue connected to the given node.

Parameters:

node - the node whose neighbors are returned

size

public int size()

Specified by:

size in class HnswGraph

seek

public void seek(int level,
                 int targetNode)

Specified by:

seek in class HnswGraph

nextNeighbor

public int nextNeighbor()

Specified by:

nextNeighbor in class HnswGraph

numLevels

public int numLevels()

Specified by:

numLevels in class HnswGraph

entryNode

public int entryNode()

Specified by:

entryNode in class HnswGraph

getNodesOnLevel

public HnswGraph.NodesIterator getNodesOnLevel(int level)

Specified by:

getNodesOnLevel in class HnswGraph