diff --git a/lucene/CHANGES.txt b/lucene/CHANGES.txt index c81ed4fae99..a38c96aa61d 100644 --- a/lucene/CHANGES.txt +++ b/lucene/CHANGES.txt @@ -737,6 +737,10 @@ New features is no longer needed and discouraged for that use case. Directly wrapping Query improves performance, as out-of-order collection is now supported. (Uwe Schindler) + +* LUCENE-2864: Add getMaxTermFrequency (maximum within-document TF) to + FieldInvertState so that it can be used in Similarity.computeNorm. + (Robert Muir) Optimizations diff --git a/lucene/src/java/org/apache/lucene/index/FieldInvertState.java b/lucene/src/java/org/apache/lucene/index/FieldInvertState.java index 321c89211c2..8c4e92ad4ea 100644 --- a/lucene/src/java/org/apache/lucene/index/FieldInvertState.java +++ b/lucene/src/java/org/apache/lucene/index/FieldInvertState.java @@ -30,6 +30,7 @@ public final class FieldInvertState { int length; int numOverlap; int offset; + int maxTermFrequency; float boost; AttributeSource attributeSource; @@ -53,6 +54,7 @@ public final class FieldInvertState { length = 0; numOverlap = 0; offset = 0; + maxTermFrequency = 0; boost = docBoost; attributeSource = null; } @@ -110,6 +112,15 @@ public final class FieldInvertState { public void setBoost(float boost) { this.boost = boost; } + + /** + * Get the maximum term-frequency encountered for any term in the field. A + * field containing "the quick brown fox jumps over the lazy dog" would have + * a value of 2, because "the" appears twice. + */ + public int getMaxTermFrequency() { + return maxTermFrequency; + } public AttributeSource getAttributeSource() { return attributeSource; diff --git a/lucene/src/java/org/apache/lucene/index/FreqProxTermsWriterPerField.java b/lucene/src/java/org/apache/lucene/index/FreqProxTermsWriterPerField.java index f22c484d1a5..f47d2a71bc0 100644 --- a/lucene/src/java/org/apache/lucene/index/FreqProxTermsWriterPerField.java +++ b/lucene/src/java/org/apache/lucene/index/FreqProxTermsWriterPerField.java @@ -125,6 +125,7 @@ final class FreqProxTermsWriterPerField extends TermsHashConsumerPerField implem postings.docFreqs[termID] = 1; writeProx(termID, fieldState.position); } + fieldState.maxTermFrequency = Math.max(1, fieldState.maxTermFrequency); } @Override @@ -158,11 +159,12 @@ final class FreqProxTermsWriterPerField extends TermsHashConsumerPerField implem termsHashPerField.writeVInt(0, postings.docFreqs[termID]); } postings.docFreqs[termID] = 1; + fieldState.maxTermFrequency = Math.max(1, fieldState.maxTermFrequency); postings.lastDocCodes[termID] = (docState.docID - postings.lastDocIDs[termID]) << 1; postings.lastDocIDs[termID] = docState.docID; writeProx(termID, fieldState.position); } else { - postings.docFreqs[termID]++; + fieldState.maxTermFrequency = Math.max(fieldState.maxTermFrequency, ++postings.docFreqs[termID]); writeProx(termID, fieldState.position-postings.lastPositions[termID]); } } diff --git a/lucene/src/test/org/apache/lucene/index/TestMaxTermFrequency.java b/lucene/src/test/org/apache/lucene/index/TestMaxTermFrequency.java new file mode 100644 index 00000000000..15a1fefe823 --- /dev/null +++ b/lucene/src/test/org/apache/lucene/index/TestMaxTermFrequency.java @@ -0,0 +1,116 @@ +package org.apache.lucene.index; + +/** + * 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. + */ + +import java.util.ArrayList; +import java.util.Arrays; +import java.util.Collections; +import java.util.List; + +import org.apache.lucene.analysis.MockAnalyzer; +import org.apache.lucene.analysis.MockTokenizer; +import org.apache.lucene.document.Document; +import org.apache.lucene.document.Field; +import org.apache.lucene.search.DefaultSimilarity; +import org.apache.lucene.store.Directory; +import org.apache.lucene.util.LuceneTestCase; +import org.apache.lucene.util._TestUtil; + +/** + * Tests the maxTermFrequency statistic in FieldInvertState + */ +public class TestMaxTermFrequency extends LuceneTestCase { + Directory dir; + IndexReader reader; + /* expected maxTermFrequency values for our documents */ + ArrayList expected = new ArrayList(); + + @Override + public void setUp() throws Exception { + super.setUp(); + dir = newDirectory(); + IndexWriterConfig config = newIndexWriterConfig(TEST_VERSION_CURRENT, + new MockAnalyzer(MockTokenizer.SIMPLE, true)); + config.setSimilarity(new TestSimilarity()); + RandomIndexWriter writer = new RandomIndexWriter(random, dir, config); + Document doc = new Document(); + Field foo = newField("foo", "", Field.Store.NO, Field.Index.ANALYZED); + doc.add(foo); + for (int i = 0; i < 100; i++) { + foo.setValue(addValue()); + writer.addDocument(doc); + } + reader = writer.getReader(); + writer.close(); + } + + @Override + public void tearDown() throws Exception { + reader.close(); + dir.close(); + super.tearDown(); + } + + public void test() throws Exception { + byte fooNorms[] = MultiNorms.norms(reader, "foo"); + for (int i = 0; i < reader.maxDoc(); i++) + assertEquals(expected.get(i).intValue(), fooNorms[i] & 0xff); + } + + /** + * Makes a bunch of single-char tokens (the max freq will at most be 255). + * shuffles them around, and returns the whole list with Arrays.toString(). + * This works fine because we use lettertokenizer. + * puts the max-frequency term into expected, to be checked against the norm. + */ + private String addValue() { + List terms = new ArrayList(); + int maxCeiling = _TestUtil.nextInt(random, 0, 255); + int max = 0; + for (char ch = 'a'; ch <= 'z'; ch++) { + int num = _TestUtil.nextInt(random, 0, maxCeiling); + for (int i = 0; i < num; i++) + terms.add(Character.toString(ch)); + max = Math.max(max, num); + } + expected.add(max); + Collections.shuffle(terms, random); + return Arrays.toString(terms.toArray(new String[terms.size()])); + } + + /** + * Simple similarity that encodes maxTermFrequency directly as a byte + */ + class TestSimilarity extends DefaultSimilarity { + + @Override + public byte encodeNormValue(float f) { + return (byte) f; + } + + @Override + public float decodeNormValue(byte b) { + return (float) b; + } + + @Override + public float computeNorm(String field, FieldInvertState state) { + return (float) state.getMaxTermFrequency(); + } + } +}