LUCENE-10311: remove complex cost estimation and abstraction leakage around it

lucene/core/src/java/org/apache/lucene/document/LatLonPointDistanceQuery.java

@@ -148,7 +148,7 @@ public ScorerSupplier scorerSupplier(LeafReaderContext context) throws IOExcepti
         LatLonPoint.checkCompatible(fieldInfo);
 
         // matching docids
-        DocIdSetBuilder result = new DocIdSetBuilder(reader.maxDoc(), values, field);
+        DocIdSetBuilder result = new DocIdSetBuilder(reader.maxDoc());
         final IntersectVisitor visitor = getIntersectVisitor(result);
 
         final Weight weight = this;

lucene/core/src/java/org/apache/lucene/document/RangeFieldQuery.java

@@ -538,7 +538,7 @@ public long cost() {
         } else {
           return new ScorerSupplier() {
 
-            final DocIdSetBuilder result = new DocIdSetBuilder(reader.maxDoc(), values, field);
+            final DocIdSetBuilder result = new DocIdSetBuilder(reader.maxDoc());
             final IntersectVisitor visitor = getIntersectVisitor(result);
             long cost = -1;
 

lucene/core/src/java/org/apache/lucene/document/SpatialQuery.java

@@ -187,7 +187,7 @@ && hasAnyHits(spatialVisitor, queryRelation, values) == false) {
             return null;
           }
           // walk the tree to get matching documents
-          return new RelationScorerSupplier(values, spatialVisitor, queryRelation, field) {
+          return new RelationScorerSupplier(values, spatialVisitor, queryRelation) {
             @Override
             public Scorer get(long leadCost) throws IOException {
               return getScorer(reader, weight, score(), scoreMode);
@@ -252,18 +252,15 @@ private abstract static class RelationScorerSupplier extends ScorerSupplier {
     private final PointValues values;
     private final SpatialVisitor spatialVisitor;
     private final QueryRelation queryRelation;
-    private final String field;
     private long cost = -1;
 
     RelationScorerSupplier(
         final PointValues values,
         SpatialVisitor spatialVisitor,
-        final QueryRelation queryRelation,
-        final String field) {
+        final QueryRelation queryRelation) {
       this.values = values;
       this.spatialVisitor = spatialVisitor;
       this.queryRelation = queryRelation;
-      this.field = field;
     }
 
     protected Scorer getScorer(
@@ -311,7 +308,7 @@ && cost() > reader.maxDoc() / 2) {
             cost[0] == 0 ? DocIdSetIterator.empty() : new BitSetIterator(result, cost[0]);
         return new ConstantScoreScorer(weight, boost, scoreMode, iterator);
       } else {
-        final DocIdSetBuilder docIdSetBuilder = new DocIdSetBuilder(reader.maxDoc(), values, field);
+        final DocIdSetBuilder docIdSetBuilder = new DocIdSetBuilder(reader.maxDoc());
         values.intersect(getSparseVisitor(spatialVisitor, queryRelation, docIdSetBuilder));
         final DocIdSetIterator iterator = docIdSetBuilder.build().iterator();
         return new ConstantScoreScorer(weight, boost, scoreMode, iterator);

lucene/core/src/java/org/apache/lucene/document/XYPointInGeometryQuery.java

@@ -149,7 +149,7 @@ public ScorerSupplier scorerSupplier(LeafReaderContext context) throws IOExcepti
         return new ScorerSupplier() {
 
           long cost = -1;
-          DocIdSetBuilder result = new DocIdSetBuilder(reader.maxDoc(), values, field);
+          DocIdSetBuilder result = new DocIdSetBuilder(reader.maxDoc());
           final IntersectVisitor visitor = getIntersectVisitor(result, tree);
 
           @Override

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

@@ -180,7 +180,7 @@ private WeightOrDocIdSet rewrite(LeafReaderContext context) throws IOException {
         }
 
         // Too many terms: go back to the terms we already collected and start building the bit set
-        DocIdSetBuilder builder = new DocIdSetBuilder(context.reader().maxDoc(), terms);
+        DocIdSetBuilder builder = new DocIdSetBuilder(context.reader().maxDoc());
         if (collectedTerms.isEmpty() == false) {
           TermsEnum termsEnum2 = terms.iterator();
           for (TermAndState t : collectedTerms) {

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

@@ -173,7 +173,7 @@ public Scorer scorer(LeafReaderContext context) throws IOException {
                   + bytesPerDim);
         }
 
-        DocIdSetBuilder result = new DocIdSetBuilder(reader.maxDoc(), values, field);
+        DocIdSetBuilder result = new DocIdSetBuilder(reader.maxDoc());
 
         if (numDims == 1) {
 

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

@@ -331,7 +331,7 @@ public long cost() {
         } else {
           return new ScorerSupplier() {
 
-            final DocIdSetBuilder result = new DocIdSetBuilder(reader.maxDoc(), values, field);
+            final DocIdSetBuilder result = new DocIdSetBuilder(reader.maxDoc());
             final IntersectVisitor visitor = getIntersectVisitor(result);
             long cost = -1;
 

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

@@ -288,7 +288,7 @@ private WeightOrDocIdSet rewrite(LeafReaderContext context) throws IOException {
               matchingTerms.add(new TermAndState(field, termsEnum));
             } else {
               assert matchingTerms.size() == threshold;
-              builder = new DocIdSetBuilder(reader.maxDoc(), terms);
+              builder = new DocIdSetBuilder(reader.maxDoc());
               docs = termsEnum.postings(docs, PostingsEnum.NONE);
               builder.add(docs);
               for (TermAndState t : matchingTerms) {

lucene/core/src/java/org/apache/lucene/util/DocIdSetBuilder.java

@@ -19,8 +19,6 @@
 import java.io.IOException;
 import java.util.ArrayList;
 import java.util.List;
-import org.apache.lucene.index.PointValues;
-import org.apache.lucene.index.Terms;
 import org.apache.lucene.search.DocIdSet;
 import org.apache.lucene.search.DocIdSetIterator;
 import org.apache.lucene.util.packed.PackedInts;
@@ -100,52 +98,17 @@ public void add(int doc) {
 
   private final int maxDoc;
   private final int threshold;
-  // pkg-private for testing
-  final boolean multivalued;
-  final double numValuesPerDoc;
 
   private List<Buffer> buffers = new ArrayList<>();
   private int totalAllocated; // accumulated size of the allocated buffers
 
   private FixedBitSet bitSet;
 
-  private long counter = -1;
   private BulkAdder adder;
 
   /** Create a builder that can contain doc IDs between {@code 0} and {@code maxDoc}. */
   public DocIdSetBuilder(int maxDoc) {
-    this(maxDoc, -1, -1);
-  }
-
-  /**
-   * Create a {@link DocIdSetBuilder} instance that is optimized for accumulating docs that match
-   * the given {@link Terms}.
-   */
-  public DocIdSetBuilder(int maxDoc, Terms terms) throws IOException {
-    this(maxDoc, terms.getDocCount(), terms.getSumDocFreq());
-  }
-
-  /**
-   * Create a {@link DocIdSetBuilder} instance that is optimized for accumulating docs that match
-   * the given {@link PointValues}.
-   */
-  public DocIdSetBuilder(int maxDoc, PointValues values, String field) throws IOException {
-    this(maxDoc, values.getDocCount(), values.size());
-  }
-
-  DocIdSetBuilder(int maxDoc, int docCount, long valueCount) {
     this.maxDoc = maxDoc;
-    this.multivalued = docCount < 0 || docCount != valueCount;
-    if (docCount <= 0 || valueCount < 0) {
-      // assume one value per doc, this means the cost will be overestimated
-      // if the docs are actually multi-valued
-      this.numValuesPerDoc = 1;
-    } else {
-      // otherwise compute from index stats
-      this.numValuesPerDoc = (double) valueCount / docCount;
-    }
-
-    assert numValuesPerDoc >= 1 : "valueCount=" + valueCount + " docCount=" + docCount;
 
     // For ridiculously small sets, we'll just use a sorted int[]
     // maxDoc >>> 7 is a good value if you want to save memory, lower values
@@ -190,10 +153,8 @@ public BulkAdder grow(int numDocs) {
         ensureBufferCapacity(numDocs);
       } else {
         upgradeToBitSet();
-        counter += numDocs;
       }
     } else {
-      counter += numDocs;
     }
     return adder;
   }
@@ -247,17 +208,14 @@ private void growBuffer(Buffer buffer, int additionalCapacity) {
   private void upgradeToBitSet() {
     assert bitSet == null;
     FixedBitSet bitSet = new FixedBitSet(maxDoc);
-    long counter = 0;
     for (Buffer buffer : buffers) {
       int[] array = buffer.array;
       int length = buffer.length;
-      counter += length;
       for (int i = 0; i < length; ++i) {
         bitSet.set(array[i]);
       }
     }
     this.bitSet = bitSet;
-    this.counter = counter;
     this.buffers = null;
     this.adder = new FixedBitSetAdder(bitSet);
   }
@@ -266,20 +224,12 @@ private void upgradeToBitSet() {
   public DocIdSet build() {
     try {
       if (bitSet != null) {
-        assert counter >= 0;
-        final long cost = Math.round(counter / numValuesPerDoc);
-        return new BitDocIdSet(bitSet, cost);
+        return new BitDocIdSet(bitSet);
       } else {
         Buffer concatenated = concat(buffers);
         LSBRadixSorter sorter = new LSBRadixSorter();
         sorter.sort(PackedInts.bitsRequired(maxDoc - 1), concatenated.array, concatenated.length);
-        final int l;
-        if (multivalued) {
-          l = dedup(concatenated.array, concatenated.length);
-        } else {
-          assert noDups(concatenated.array, concatenated.length);
-          l = concatenated.length;
-        }
+        final int l = dedup(concatenated.array, concatenated.length);
         assert l <= concatenated.length;
         concatenated.array[l] = DocIdSetIterator.NO_MORE_DOCS;
         return new IntArrayDocIdSet(concatenated.array, l);
@@ -336,11 +286,4 @@ private static int dedup(int[] arr, int length) {
     }
     return l;
   }
-
-  private static boolean noDups(int[] a, int len) {
-    for (int i = 1; i < len; ++i) {
-      assert a[i - 1] < a[i];
-    }
-    return true;
-  }
 }