Speed up TermQuery (#14709)

// nightly-benchmarks-results-changed //

Author: gf2121

lucene/CHANGES.txt

@@ -122,6 +122,8 @@ Optimizations
 * GITHUB#14701: Optimize top-n bulk scorers by evaluating scoring windows in a
   term-at-a-time fashion instead of doc-at-a-time. (Adrien Grand)
 
+* GITHUB#14709: Speed up TermQuery by Scorer#nextDocsAndScores. (Guo Feng)
+
 Bug Fixes
 ---------------------
 * GITHUB#14654: ValueSource.fromDoubleValuesSource(dvs).getSortField() would throw errors when

lucene/core/src/java/org/apache/lucene/search/BatchScoreBulkScorer.java

@@ -0,0 +1,68 @@
+/*
+ * 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.search;
+
+import java.io.IOException;
+import org.apache.lucene.util.Bits;
+
+/**
+ * A bulk scorer used when {@link ScoreMode#needsScores()} is true and {@link
+ * Scorer#nextDocsAndScores} has optimizations to run faster than one-by-one iteration.
+ */
+class BatchScoreBulkScorer extends BulkScorer {
+
+  private final SimpleScorable scorable = new SimpleScorable();
+  private final DocAndScoreBuffer buffer = new DocAndScoreBuffer();
+  private final Scorer scorer;
+
+  BatchScoreBulkScorer(Scorer scorer) {
+    this.scorer = scorer;
+  }
+
+  @Override
+  public int score(LeafCollector collector, Bits acceptDocs, int min, int max) throws IOException {
+    if (collector.competitiveIterator() != null) {
+      return new Weight.DefaultBulkScorer(scorer).score(collector, acceptDocs, min, max);
+    }
+
+    collector.setScorer(scorable);
+    scorer.setMinCompetitiveScore(scorable.minCompetitiveScore);
+
+    if (scorer.docID() < min) {
+      scorer.iterator().advance(min);
+    }
+
+    for (scorer.nextDocsAndScores(max, acceptDocs, buffer);
+        buffer.size > 0;
+        scorer.nextDocsAndScores(max, acceptDocs, buffer)) {
+      for (int i = 0, size = buffer.size; i < size; i++) {
+        float score = scorable.score = buffer.scores[i];
+        if (score >= scorable.minCompetitiveScore) {
+          collector.collect(buffer.docs[i]);
+        }
+      }
+      scorer.setMinCompetitiveScore(scorable.minCompetitiveScore);
+    }
+
+    return scorer.docID();
+  }
+
+  @Override
+  public long cost() {
+    return scorer.iterator().cost();
+  }
+}

lucene/core/src/java/org/apache/lucene/search/ImpactsDISI.java

@@ -98,6 +98,15 @@ private int advanceTarget(int target) throws IOException {
     }
   }
 
+  /** If current doc is not competitive, move to a competitive one. */
+  void ensureCompetitive() throws IOException {
+    int doc = docID();
+    int advanceTarget = advanceTarget(doc);
+    if (advanceTarget != doc) {
+      in.advance(advanceTarget);
+    }
+  }
+
   @Override
   public int advance(int target) throws IOException {
     return in.advance(advanceTarget(target));

lucene/core/src/java/org/apache/lucene/search/TermQuery.java

@@ -173,7 +173,7 @@ public BulkScorer bulkScorer() throws IOException {
             return ConstantScoreScorerSupplier.fromIterator(iterator, 0f, scoreMode, maxDoc)
                 .bulkScorer();
           }
-          return super.bulkScorer();
+          return new BatchScoreBulkScorer(get(Long.MAX_VALUE));
         }
 
         @Override

lucene/core/src/java/org/apache/lucene/search/TermScorer.java

@@ -135,6 +135,10 @@ public void nextDocsAndScores(int upTo, Bits liveDocs, DocAndScoreBuffer buffer)
     }
 
     for (; ; ) {
+      if (impactsDisi != null) {
+        impactsDisi.ensureCompetitive();
+      }
+
       postingsEnum.nextPostings(upTo, docAndFreqBuffer);
       if (liveDocs != null && docAndFreqBuffer.size != 0) {
         // An empty return value indicates that there are no more docs before upTo. We may be

lucene/core/src/test/org/apache/lucene/search/TestBooleanScorer.java

@@ -201,7 +201,7 @@ public void testOptimizeTopLevelClauseOrNull() throws IOException {
     weight = searcher.createWeight(searcher.rewrite(query), ScoreMode.COMPLETE, 1);
     ss = weight.scorerSupplier(ctx);
     scorer = ((BooleanScorerSupplier) ss).booleanScorer();
-    assertThat(scorer, instanceOf(DefaultBulkScorer.class)); // term scorer
+    assertThat(scorer, instanceOf(BatchScoreBulkScorer.class)); // term scorer
 
     w.close();
     reader.close();