jainankitk commented on code in PR #14439: URL: https://github.com/apache/lucene/pull/14439#discussion_r2054906523
########## lucene/sandbox/src/java/org/apache/lucene/sandbox/facet/plain/histograms/PointTreeBulkCollector.java: ########## @@ -0,0 +1,227 @@ +/* + * 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.sandbox.facet.plain.histograms; + +import static org.apache.lucene.search.DocIdSetIterator.NO_MORE_DOCS; + +import java.io.IOException; +import java.util.function.Function; +import org.apache.lucene.index.PointValues; +import org.apache.lucene.internal.hppc.LongIntHashMap; +import org.apache.lucene.search.DocIdSetIterator; +import org.apache.lucene.util.NumericUtils; + +class PointTreeBulkCollector { + private static Function<byte[], Long> bytesToLong(int numBytes) { + if (numBytes == Long.BYTES) { + // Used by LongPoint, DoublePoint + return a -> NumericUtils.sortableBytesToLong(a, 0); + } else if (numBytes == Integer.BYTES) { + // Used by IntPoint, FloatPoint, LatLonPoint, LatLonShape + return a -> (long) NumericUtils.sortableBytesToInt(a, 0); + } + + return null; + } + + static boolean canCollectEfficiently(final PointValues pointValues, final long bucketWidth) + throws IOException { + // TODO: Do we really need pointValues.getDocCount() == pointValues.size() for this optimization + if (pointValues == null + || pointValues.getNumDimensions() != 1 + || pointValues.getDocCount() != pointValues.size()) { + return false; + } + + final Function<byte[], Long> byteToLong = bytesToLong(pointValues.getBytesPerDimension()); + if (byteToLong == null) { + return false; + } + + long leafMinBucket = + Math.floorDiv(byteToLong.apply(pointValues.getMinPackedValue()), bucketWidth); + long leafMaxBucket = + Math.floorDiv(byteToLong.apply(pointValues.getMaxPackedValue()), bucketWidth); + + // We want that # leaf nodes is more than # buckets so that we can completely skip over + // some of the leaf nodes. Higher this ratio, more efficient it is than naive approach! + if ((pointValues.size() / 512) < (leafMaxBucket - leafMinBucket)) { Review Comment: > I imagine that 512 is the size of leaves with the default codec. It's a bit annoying to rely on a codec parameter. I'm tempted to replace it with a simpler heuristic like the collector does, something like leafMaxBucket - leafMinBucket <= 1024. While I share your annoyance of relying on codec parameter, `leafMaxBucket - leafMinBucket <= 1024` works for prior case as it uses dense array that is just function of this. But here, if we collect billion documents, even ` leafMaxBucket - leafMinBucket <= 10^5` is fairly reasonable. Maybe we can use something more generic and reasonable like 1000 instead of sticking exactly to the codec parameter? -- 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
