zhaih commented on code in PR #12657: URL: https://github.com/apache/lucene/pull/12657#discussion_r1357779437
########## lucene/core/src/java/org/apache/lucene/codecs/lucene95/IncrementalHnswGraphMerger.java: ########## @@ -0,0 +1,209 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ +package org.apache.lucene.codecs.lucene95; + +import static org.apache.lucene.search.DocIdSetIterator.NO_MORE_DOCS; + +import java.io.IOException; +import java.util.Collections; +import java.util.HashMap; +import java.util.Map; +import org.apache.lucene.codecs.HnswGraphProvider; +import org.apache.lucene.codecs.KnnVectorsReader; +import org.apache.lucene.codecs.KnnVectorsWriter; +import org.apache.lucene.codecs.perfield.PerFieldKnnVectorsFormat; +import org.apache.lucene.index.ByteVectorValues; +import org.apache.lucene.index.FieldInfo; +import org.apache.lucene.index.FloatVectorValues; +import org.apache.lucene.index.MergeState; +import org.apache.lucene.search.DocIdSetIterator; +import org.apache.lucene.util.Bits; +import org.apache.lucene.util.hnsw.HnswGraph; +import org.apache.lucene.util.hnsw.HnswGraphBuilder; +import org.apache.lucene.util.hnsw.InitializedHnswGraphBuilder; +import org.apache.lucene.util.hnsw.RandomVectorScorerSupplier; + +/** + * This selects the biggest Hnsw graph from the provided merge state and initializes a new + * HnswGraphBuilder with that graph as a starting point. + */ +public class IncrementalHnswGraphMerger { + + private KnnVectorsReader curReader; + private MergeState.DocMap curDocMap; + private int curGraphSize; Review Comment: maybe rename to `initGraphReader`, `initGraphDocMap` and `initGraphSize`? ########## lucene/core/src/java/org/apache/lucene/util/hnsw/InitializedHnswGraphBuilder.java: ########## @@ -0,0 +1,78 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package org.apache.lucene.util.hnsw; + +import static org.apache.lucene.search.DocIdSetIterator.NO_MORE_DOCS; + +import java.io.IOException; +import java.util.Map; +import java.util.Set; +import org.apache.lucene.util.CollectionUtil; + +/** + * This creates a graph builder that is initialized with the provided HnswGraph. This is useful for + * merging HnswGraphs from multiple segments. + * + * @lucene.internal + */ +public final class InitializedHnswGraphBuilder extends HnswGraphBuilder { Review Comment: I wonder whether it'll be better to allow `HnswGraphBuilder` to accept a `OnHeapHnswGraph` as a constructor parameter and then when we init from graph we will use this `InitializedHnswGraphBuilder` to build a graph and then pass the built graph and a node filter (to avoid reindex the same nodes) to the normal `HnswGraphBuilder`? Then for example if we want to have a multi-thread `ConcurrentHnswGraphBuilder` we can still use this `InitializedHnswGraphBuilder` to build the init graph and pass it to the `ConcurentHnswGraphBuilder`? I mentioned this because in my draft concurrent HNSW merge PR #12660 I do need to pass the HNSW graph to a builder per thread, altho it can be done in various ways. But I still feel using parent/child class to separate this can make things a little bit hard later? Like, if I want a concurrent builder will the concurrent builder extend from this class? If so we need to be quite careful not to inherit a wrong behavior from the original `HnswGraphBuilder` and things can become quite complex? -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: issues-unsubscr...@lucene.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: issues-unsubscr...@lucene.apache.org For additional commands, e-mail: issues-h...@lucene.apache.org
