+ TopFieldCollector tfc = TopFieldCollector.create(sort, numHits, fillFields,
+ true /* trackDocScores */);
+ searcher.search(weight, filter, tfc);
+
+
+ Also, TimeLimitedCollector still extends HitCollector (for
+ back-compat), however in 3.0 it will move to extend
+ Collector. Meanwhile, if you want to use it in the new
+ Collector-based search methods, you can wrap it with
+ HitCollectorWrapper. (Shai Erera via Mike McCandless)
+
Changes in runtime behavior
1. LUCENE-1424: QueryParser now by default uses constant score query
@@ -10,7 +32,25 @@
already does so for RangeQuery, as well). Call
setConstantScoreRewrite(false) to revert to BooleanQuery rewriting
method. (Mark Miller via Mike McCandless)
+
+ 2. LUCENE-1575: As of 2.9, the core collectors as well as
+ IndexSearcher's search methods that return top N results, no
+ longer filter out zero scoring documents. You can wrap them (as
+ well as any other Collector) with PositiveScoresOnlyCollector to
+ achieve the same effect:
+
+ TopDocsCollector tdc = new TopScoreDocCollector(10);
+ Collector c = new PositiveScoresOnlyCollector(tdc);
+ searcher.search(query, c);
+ TopDocs td = tdc.topDocs();
+ ...
+
+
+ In addition, note that the IndexSearcher methods which take Sort
+ as parameter will be changed in 3.0 to not track document
+ scores. (Shai Erera via Mike McCandless)
+
API Changes
1. LUCENE-1419: Add expert API to set custom indexing chain. This API is
@@ -69,6 +109,13 @@
12. LUCENE-1500: Added new InvalidTokenOffsetsException to Highlighter methods
to denote issues when offsets in TokenStream tokens exceed the length of the
provided text. (Mark Harwood)
+
+13. LUCENE-1575: HitCollector is now deprecated in favor of a new
+ Collector abstract class. For easy migration, people can use
+ HitCollectorWrapper which translates (wraps) HitCollector into
+ Collector. Note that this class is also deprecated and will be
+ removed when HitCollector is removed. (Shai Erera via Mike
+ McCandless)
Bug fixes
@@ -237,6 +284,12 @@
those segments that did not change, and also speeds up searches
that sort by relevance or by field values. (Mark Miller, Mike
McCandless)
+
+ 7. LUCENE-1575: The new Collector class decouples collect() from
+ score computation. Instead it offers a setScorer method which
+ implementors can use to save the Scorer and use it in collect()
+ when the document's score is required. (Shai Erera via Mike
+ McCandless)
Documentation
Index: src/test/org/apache/lucene/search/TestSort.java
===================================================================
--- src/test/org/apache/lucene/search/TestSort.java (revision 761300)
+++ src/test/org/apache/lucene/search/TestSort.java (working copy)
@@ -413,7 +413,7 @@
slotValues = new int[numHits];
}
- public void copy(int slot, int doc, float score) {
+ public void copy(int slot, int doc) {
slotValues[slot] = docValues[doc];
}
@@ -421,7 +421,7 @@
return slotValues[slot1] - slotValues[slot2];
}
- public int compareBottom(int doc, float score) {
+ public int compareBottom(int doc) {
return bottomValue - docValues[doc];
}
@@ -447,7 +447,7 @@
}
static class MyFieldComparatorSource extends FieldComparatorSource {
- public FieldComparator newComparator(String fieldname, IndexReader[] subReaders, int numHits, int sortPos, boolean reversed) {
+ public FieldComparator newComparator(String fieldname, int numHits, int sortPos, boolean reversed) {
return new MyFieldComparator(numHits);
}
}
@@ -803,7 +803,27 @@
assertEquals(docs1.scoreDocs[0].score, docs2.scoreDocs[0].score, 1e-6);
}
+
+ public void testSortWithoutFillFields() throws Exception {
+
+ // There was previously a bug in TopFieldCollector when fillFields was set
+ // to false - the same doc and score was set in ScoreDoc[] array. This test
+ // asserts that if fillFields is false, the documents are set properly. It
+ // does not use Searcher's default search methods (with Sort) since all set
+ // fillFields to true.
+ Sort sort = new Sort();
+ int nDocs=10;
+
+ TopDocsCollector tdc = TopFieldCollector.create(sort, nDocs, false, false);
+
+ full.search(new MatchAllDocsQuery(), tdc);
+ ScoreDoc[] sd = tdc.topDocs().scoreDocs;
+ for (int i = 1; i < sd.length; i++) {
+ assertTrue(sd[i].doc != sd[i - 1].doc);
+ }
+ }
+
// runs a variety of sorts useful for multisearchers
private void runMultiSorts (Searcher multi) throws Exception {
sort.setSort (SortField.FIELD_DOC);
Index: src/test/org/apache/lucene/search/TestMultiTermConstantScore.java
===================================================================
--- src/test/org/apache/lucene/search/TestMultiTermConstantScore.java (revision 761300)
+++ src/test/org/apache/lucene/search/TestMultiTermConstantScore.java (working copy)
@@ -168,11 +168,15 @@
// must use a non score normalizing method for this.
Query q = csrq("data", "1", "6", T, T);
q.setBoost(100);
- search.search(q, null, new MultiReaderHitCollector() {
- private int base = -1;
- public void collect(int doc, float score) {
- assertEquals("score for doc " + (doc + base) + " was not correct", 1.0f, score);
+ search.search(q, null, new Collector() {
+ private int base = 0;
+ private Scorer scorer;
+ public void setScorer(Scorer scorer) throws IOException {
+ this.scorer = scorer;
}
+ public void collect(int doc) throws IOException {
+ assertEquals("score for doc " + (doc + base) + " was not correct", 1.0f, scorer.score());
+ }
public void setNextReader(IndexReader reader, int docBase) {
base = docBase;
}
Index: src/test/org/apache/lucene/search/CheckHits.java
===================================================================
--- src/test/org/apache/lucene/search/CheckHits.java (revision 761300)
+++ src/test/org/apache/lucene/search/CheckHits.java (working copy)
@@ -89,9 +89,10 @@
}
final Set actual = new TreeSet();
- searcher.search(query, new MultiReaderHitCollector() {
- private int base = -1;
- public void collect(int doc, float score) {
+ searcher.search(query, new Collector() {
+ private int base = 0;
+ public void setScorer(Scorer scorer) throws IOException {}
+ public void collect(int doc) {
actual.add(new Integer(doc + base));
}
@@ -390,14 +391,22 @@
checkExplanations(query);
return super.search(query,filter,n,sort);
}
+ /** @deprecated use {@link #search(Query, Collector)} instead. */
public void search(Query query, HitCollector results) throws IOException {
+ search(query, new HitCollectorWrapper(results));
+ }
+ public void search(Query query, Collector results) throws IOException {
checkExplanations(query);
- super.search(query,results);
+ super.search(query, results);
}
+ /** @deprecated use {@link #search(Query, Filter, Collector)} instead. */
public void search(Query query, Filter filter,
HitCollector results) throws IOException {
+ search(query, filter, new HitCollectorWrapper(results));
+ }
+ public void search(Query query, Filter filter, Collector results) throws IOException {
checkExplanations(query);
- super.search(query,filter, results);
+ super.search(query, filter, results);
}
public TopDocs search(Query query, Filter filter,
int n) throws IOException {
@@ -416,7 +425,7 @@
*
* @see CheckHits#verifyExplanation
*/
- public static class ExplanationAsserter extends MultiReaderHitCollector {
+ public static class ExplanationAsserter extends Collector {
/**
* @deprecated
@@ -428,9 +437,10 @@
Searcher s;
String d;
boolean deep;
+
+ Scorer scorer;
+ private int base = 0;
- private int base = -1;
-
/** Constructs an instance which does shallow tests on the Explanation */
public ExplanationAsserter(Query q, String defaultFieldName, Searcher s) {
this(q,defaultFieldName,s,false);
@@ -441,8 +451,12 @@
this.d = q.toString(defaultFieldName);
this.deep=deep;
}
-
- public void collect(int doc, float score) {
+
+ public void setScorer(Scorer scorer) throws IOException {
+ this.scorer = scorer;
+ }
+
+ public void collect(int doc) throws IOException {
Explanation exp = null;
doc = doc + base;
try {
@@ -454,7 +468,7 @@
TestCase.assertNotNull("Explanation of [["+d+"]] for #"+doc+" is null",
exp);
- verifyExplanation(d,doc,score,deep,exp);
+ verifyExplanation(d,doc,scorer.score(),deep,exp);
}
public void setNextReader(IndexReader reader, int docBase) {
base = docBase;
Index: src/test/org/apache/lucene/search/TestTimeLimitedCollector.java
===================================================================
--- src/test/org/apache/lucene/search/TestTimeLimitedCollector.java (revision 761300)
+++ src/test/org/apache/lucene/search/TestTimeLimitedCollector.java (working copy)
@@ -17,10 +17,12 @@
* limitations under the License.
*/
+import java.io.IOException;
+import java.util.BitSet;
+
import org.apache.lucene.analysis.WhitespaceAnalyzer;
import org.apache.lucene.document.Document;
import org.apache.lucene.document.Field;
-import org.apache.lucene.index.IndexReader;
import org.apache.lucene.index.IndexWriter;
import org.apache.lucene.index.IndexWriter.MaxFieldLength;
import org.apache.lucene.queryParser.QueryParser;
@@ -28,9 +30,6 @@
import org.apache.lucene.store.RAMDirectory;
import org.apache.lucene.util.LuceneTestCase;
-import java.io.IOException;
-import java.util.BitSet;
-
/**
* Tests the TimeLimitedCollector. This test checks (1) search
* correctness (regardless of timeout), (2) expected timeout behavior,
@@ -101,7 +100,7 @@
iw.addDocument(d);
}
- private void search(HitCollector collector) throws Exception {
+ private void search(Collector collector) throws Exception {
searcher.search(query, collector);
}
@@ -117,12 +116,12 @@
int totalTLCResults = 0;
try {
MyHitCollector myHc = new MyHitCollector();
- search(myHc);
+ search(new HitCollectorWrapper(myHc));
totalResults = myHc.hitCount();
myHc = new MyHitCollector();
long oneHour = 3600000;
- HitCollector tlCollector = createTimedCollector(myHc, oneHour, false);
+ Collector tlCollector = createTimedCollector(myHc, oneHour, false);
search(tlCollector);
totalTLCResults = myHc.hitCount();
} catch (Exception e) {
@@ -132,10 +131,10 @@
assertEquals( "Wrong number of results!", totalResults, totalTLCResults );
}
- private HitCollector createTimedCollector(MyHitCollector hc, long timeAllowed, boolean greedy) {
+ private Collector createTimedCollector(MyHitCollector hc, long timeAllowed, boolean greedy) {
TimeLimitedCollector res = new TimeLimitedCollector(hc, timeAllowed);
res.setGreedy(greedy); // set to true to make sure at least one doc is collected.
- return res;
+ return new HitCollectorWrapper(res);
}
/**
@@ -156,7 +155,7 @@
// setup
MyHitCollector myHc = new MyHitCollector();
myHc.setSlowDown(SLOW_DOWN);
- HitCollector tlCollector = createTimedCollector(myHc, TIME_ALLOWED, greedy);
+ Collector tlCollector = createTimedCollector(myHc, TIME_ALLOWED, greedy);
// search
TimeLimitedCollector.TimeExceededException timoutException = null;
@@ -286,23 +285,16 @@
assertEquals("some threads failed!", N_THREADS,success.cardinality());
}
- // counting hit collector that can slow down at collect().
- private class MyHitCollector extends MultiReaderHitCollector
- {
+ // counting collector that can slow down at collect().
+ // TODO: re-extend Collector when we move to 3.0
+ private class MyHitCollector extends HitCollector {
private final BitSet bits = new BitSet();
private int slowdown = 0;
private int lastDocCollected = -1;
- private int docBase = -1;
+// private int docBase = 0; // TODO: reenstate when we move to 3.0
- /**
- * amount of time to wait on each collect to simulate a long iteration
- */
- public void setSlowDown( int milliseconds ) {
- slowdown = milliseconds;
- }
-
- public void collect( final int doc, final float score ) {
- int docId = doc + docBase;
+ public void collect(final int doc, final float score) {
+ int docId = doc;
if( slowdown > 0 ) {
try {
Thread.sleep(slowdown);
@@ -311,10 +303,17 @@
throw new RuntimeException(ie);
}
}
- assert docId >= 0: " base=" + docBase + " doc=" + doc;
- bits.set( docId );
+ assert docId >= 0: " doc=" + doc;
+ bits.set(docId);
lastDocCollected = docId;
}
+
+ /**
+ * amount of time to wait on each collect to simulate a long iteration
+ */
+ public void setSlowDown( int milliseconds ) {
+ slowdown = milliseconds;
+ }
public int hitCount() {
return bits.cardinality();
@@ -324,9 +323,29 @@
return lastDocCollected;
}
- public void setNextReader(IndexReader reader, int base) {
- docBase = base;
- }
+ // TODO: reenstate when we move to 3.0
+// public void setScorer(Scorer scorer) throws IOException {
+// // scorer is not needed
+// }
+//
+// public void collect(final int doc) throws IOException {
+// int docId = doc + docBase;
+// if( slowdown > 0 ) {
+// try {
+// Thread.sleep(slowdown);
+// } catch (InterruptedException ie) {
+// Thread.currentThread().interrupt();
+// throw new RuntimeException(ie);
+// }
+// }
+// assert docId >= 0: " base=" + docBase + " doc=" + doc;
+// bits.set( docId );
+// lastDocCollected = docId;
+// }
+//
+// public void setNextReader(IndexReader reader, int base) {
+// docBase = base;
+// }
}
Index: src/test/org/apache/lucene/search/TestScoreCachingWrappingScorer.java
===================================================================
--- src/test/org/apache/lucene/search/TestScoreCachingWrappingScorer.java (revision 0)
+++ src/test/org/apache/lucene/search/TestScoreCachingWrappingScorer.java (revision 0)
@@ -0,0 +1,111 @@
+package org.apache.lucene.search;
+
+/**
+ * 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.io.IOException;
+
+import org.apache.lucene.index.IndexReader;
+import org.apache.lucene.util.LuceneTestCase;
+
+public class TestScoreCachingWrappingScorer extends LuceneTestCase {
+
+ private static final class SimpleScorer extends Scorer {
+ private int idx = 0;
+ private int doc = -1;
+
+ public SimpleScorer() {
+ super(null);
+ }
+
+ public Explanation explain(int doc) throws IOException { return null; }
+
+ public float score() throws IOException {
+ // advance idx on purpose, so that consecutive calls to score will get
+ // different results. This is to emulate computation of a score. If
+ // ScoreCachingWrappingScorer is used, this should not be called more than
+ // once per document.
+ return idx == scores.length ? Float.NaN : scores[idx++];
+ }
+
+ public int doc() { return doc; }
+
+ public boolean next() throws IOException {
+ return ++doc == scores.length;
+ }
+
+ public boolean skipTo(int target) throws IOException {
+ doc = target;
+ return doc >= scores.length;
+ }
+ }
+
+ private static final class ScoreCachingCollector extends Collector {
+
+ private int idx = 0;
+ private Scorer scorer;
+ float[] mscores;
+
+ public ScoreCachingCollector(int numToCollect) {
+ mscores = new float[numToCollect];
+ }
+
+ public void collect(int doc) throws IOException {
+ // just a sanity check to avoid IOOB.
+ if (idx == mscores.length) {
+ return;
+ }
+
+ // just call score() a couple of times and record the score.
+ mscores[idx] = scorer.score();
+ mscores[idx] = scorer.score();
+ mscores[idx] = scorer.score();
+ ++idx;
+ }
+
+ public void setNextReader(IndexReader reader, int docBase)
+ throws IOException {
+ }
+
+ public void setScorer(Scorer scorer) throws IOException {
+ this.scorer = new ScoreCachingWrappingScorer(scorer);
+ }
+
+ }
+
+ private static final float[] scores = new float[] { 0.7767749f, 1.7839992f,
+ 8.9925785f, 7.9608946f, 0.07948637f, 2.6356435f, 7.4950366f, 7.1490803f,
+ 8.108544f, 4.961808f, 2.2423935f, 7.285586f, 4.6699767f };
+
+ public void testGetScores() throws Exception {
+
+ Scorer s = new SimpleScorer();
+ ScoreCachingCollector scc = new ScoreCachingCollector(scores.length);
+ scc.setScorer(s);
+
+ // We need to iterate on the scorer so that its doc() advances.
+ while (!s.next()) {
+ scc.collect(s.doc());
+ }
+
+ for (int i = 0; i < scores.length; i++) {
+ assertEquals(scores[i], scc.mscores[i], 0f);
+ }
+
+ }
+
+}
Index: src/test/org/apache/lucene/search/TestTopDocsCollector.java
===================================================================
--- src/test/org/apache/lucene/search/TestTopDocsCollector.java (revision 0)
+++ src/test/org/apache/lucene/search/TestTopDocsCollector.java (revision 0)
@@ -0,0 +1,198 @@
+package org.apache.lucene.search;
+
+/**
+ * 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.io.IOException;
+
+import org.apache.lucene.analysis.KeywordAnalyzer;
+import org.apache.lucene.document.Document;
+import org.apache.lucene.index.IndexReader;
+import org.apache.lucene.index.IndexWriter;
+import org.apache.lucene.index.IndexWriter.MaxFieldLength;
+import org.apache.lucene.store.Directory;
+import org.apache.lucene.store.RAMDirectory;
+import org.apache.lucene.util.LuceneTestCase;
+
+public class TestTopDocsCollector extends LuceneTestCase {
+
+ private static final class MyTopsDocCollector extends TopDocsCollector {
+
+ private int idx = 0;
+ private int base = 0;
+
+ public MyTopsDocCollector(int size) {
+ super(new HitQueue(size));
+ }
+
+ protected TopDocs newTopDocs(ScoreDoc[] results, int start) {
+ if (results == null) {
+ return EMPTY_TOPDOCS;
+ }
+
+ float maxScore = Float.NaN;
+ if (start == 0) {
+ maxScore = results[0].score;
+ } else {
+ for (int i = pq.size(); i > 1; i--) { pq.pop(); }
+ maxScore = ((ScoreDoc) pq.pop()).score;
+ }
+
+ return new TopDocs(totalHits, results, maxScore);
+ }
+
+ public void collect(int doc) throws IOException {
+ pq.insert(new ScoreDoc(doc + base, scores[idx++]));
+ }
+
+ public void setNextReader(IndexReader reader, int docBase)
+ throws IOException {
+ base = docBase;
+ }
+
+ public void setScorer(Scorer scorer) throws IOException {
+ // Don't do anything. Assign scores in random
+ }
+
+ }
+
+ // Scores array to be used by MyTopDocsCollector. If it is changed, MAX_SCORE
+ // must also change.
+ private static final float[] scores = new float[] {
+ 0.7767749f, 1.7839992f, 8.9925785f, 7.9608946f, 0.07948637f, 2.6356435f,
+ 7.4950366f, 7.1490803f, 8.108544f, 4.961808f, 2.2423935f, 7.285586f, 4.6699767f,
+ 2.9655676f, 6.953706f, 5.383931f, 6.9916306f, 8.365894f, 7.888485f, 8.723962f,
+ 3.1796896f, 0.39971232f, 1.3077754f, 6.8489285f, 9.17561f, 5.060466f, 7.9793315f,
+ 8.601509f, 4.1858315f, 0.28146625f
+ };
+
+ private static final float MAX_SCORE = 9.17561f;
+
+ private Directory dir = new RAMDirectory();
+
+ private TopDocsCollector doSearch(int numResults) throws IOException {
+ Query q = new MatchAllDocsQuery();
+ IndexSearcher searcher = new IndexSearcher(dir);
+ TopDocsCollector tdc = new MyTopsDocCollector(numResults);
+ searcher.search(q, tdc);
+ searcher.close();
+ return tdc;
+ }
+
+ protected void setUp() throws Exception {
+ super.setUp();
+
+ // populate an index with 30 documents, this should be enough for the test.
+ // The documents have no content - the test uses MatchAllDocsQuery().
+ IndexWriter writer = new IndexWriter(dir, new KeywordAnalyzer(), MaxFieldLength.UNLIMITED);
+ for (int i = 0; i < 30; i++) {
+ writer.addDocument(new Document());
+ }
+ writer.close();
+ }
+
+ protected void tearDown() throws Exception {
+ dir.close();
+ dir = null;
+ super.tearDown();
+ }
+
+ public void testInvalidArguments() throws Exception {
+ int numResults = 5;
+ TopDocsCollector tdc = doSearch(numResults);
+
+ // start < 0
+ assertEquals(0, tdc.topDocs(-1).scoreDocs.length);
+
+ // start > pq.size()
+ assertEquals(0, tdc.topDocs(numResults + 1).scoreDocs.length);
+
+ // start == pq.size()
+ assertEquals(0, tdc.topDocs(numResults).scoreDocs.length);
+
+ // howMany < 0
+ assertEquals(0, tdc.topDocs(0, -1).scoreDocs.length);
+
+ // howMany == 0
+ assertEquals(0, tdc.topDocs(0, 0).scoreDocs.length);
+
+ }
+
+ public void testZeroResults() throws Exception {
+ TopDocsCollector tdc = new MyTopsDocCollector(5);
+ assertEquals(0, tdc.topDocs(0, 1).scoreDocs.length);
+ }
+
+ public void testFirstResultsPage() throws Exception {
+ TopDocsCollector tdc = doSearch(15);
+ assertEquals(10, tdc.topDocs(0, 10).scoreDocs.length);
+ }
+
+ public void testSecondResultsPages() throws Exception {
+ TopDocsCollector tdc = doSearch(15);
+ // ask for more results than are available
+ assertEquals(5, tdc.topDocs(10, 10).scoreDocs.length);
+
+ // ask for 5 results (exactly what there should be
+ tdc = doSearch(15);
+ assertEquals(5, tdc.topDocs(10, 5).scoreDocs.length);
+
+ // ask for less results than there are
+ tdc = doSearch(15);
+ assertEquals(4, tdc.topDocs(10, 4).scoreDocs.length);
+ }
+
+ public void testGetAllResults() throws Exception {
+ TopDocsCollector tdc = doSearch(15);
+ assertEquals(15, tdc.topDocs().scoreDocs.length);
+ }
+
+ public void testGetResultsFromStart() throws Exception {
+ TopDocsCollector tdc = doSearch(15);
+ // should bring all results
+ assertEquals(15, tdc.topDocs(0).scoreDocs.length);
+
+ tdc = doSearch(15);
+ // get the last 5 only.
+ assertEquals(5, tdc.topDocs(10).scoreDocs.length);
+ }
+
+ public void testMaxScore() throws Exception {
+ // ask for all results
+ TopDocsCollector tdc = doSearch(15);
+ TopDocs td = tdc.topDocs();
+ assertEquals(MAX_SCORE, td.getMaxScore(), 0f);
+
+ // ask for 5 last results
+ tdc = doSearch(15);
+ td = tdc.topDocs(10);
+ assertEquals(MAX_SCORE, td.getMaxScore(), 0f);
+ }
+
+ // This does not test the PQ's correctness, but whether topDocs()
+ // implementations return the results in decreasing score order.
+ public void testResultsOrder() throws Exception {
+ TopDocsCollector tdc = doSearch(15);
+ ScoreDoc[] sd = tdc.topDocs().scoreDocs;
+
+ assertEquals(MAX_SCORE, sd[0].score, 0f);
+ for (int i = 1; i < sd.length; i++) {
+ assertTrue(sd[i - 1].score >= sd[i].score);
+ }
+ }
+
+}
Index: src/test/org/apache/lucene/search/TestTermScorer.java
===================================================================
--- src/test/org/apache/lucene/search/TestTermScorer.java (revision 761300)
+++ src/test/org/apache/lucene/search/TestTermScorer.java (working copy)
@@ -65,8 +65,7 @@
}
- public void test() throws IOException
- {
+ public void test() throws IOException {
Term allTerm = new Term(FIELD, "all");
TermQuery termQuery = new TermQuery(allTerm);
@@ -76,22 +75,26 @@
TermScorer ts = new TermScorer(weight,
indexReader.termDocs(allTerm), indexSearcher.getSimilarity(),
indexReader.norms(FIELD));
- assertTrue("ts is null and it shouldn't be", ts != null);
//we have 2 documents with the term all in them, one document for all the other values
final List docs = new ArrayList();
//must call next first
- ts.score(new MultiReaderHitCollector()
- {
- private int base = -1;
- public void collect(int doc, float score)
- {
- docs.add(new TestHit(doc + base, score));
- assertTrue("score " + score + " is not greater than 0", score > 0);
- assertTrue("Doc: " + doc + " does not equal: " + 0 +
- " or doc does not equaal: " + 5, doc == 0 || doc == 5);
+ ts.score(new Collector() {
+ private int base = 0;
+ private Scorer scorer;
+ public void setScorer(Scorer scorer) throws IOException {
+ this.scorer = scorer;
}
+
+ public void collect(int doc) throws IOException {
+ float score = scorer.score();
+ doc = doc + base;
+ docs.add(new TestHit(doc, score));
+ assertTrue("score " + score + " is not greater than 0", score > 0);
+ assertTrue("Doc: " + doc + " does not equal 0 or doc does not equal 5",
+ doc == 0 || doc == 5);
+ }
public void setNextReader(IndexReader reader, int docBase) {
base = docBase;
}
@@ -121,8 +124,7 @@
assertTrue(doc0.score + " does not equal: " + 1.6931472f, doc0.score == 1.6931472f);
}
- public void testNext() throws Exception
- {
+ public void testNext() throws Exception {
Term allTerm = new Term(FIELD, "all");
TermQuery termQuery = new TermQuery(allTerm);
@@ -132,7 +134,6 @@
TermScorer ts = new TermScorer(weight,
indexReader.termDocs(allTerm), indexSearcher.getSimilarity(),
indexReader.norms(FIELD));
- assertTrue("ts is null and it shouldn't be", ts != null);
assertTrue("next did not return a doc", ts.next() == true);
assertTrue("score is not correct", ts.score() == 1.6931472f);
assertTrue("next did not return a doc", ts.next() == true);
@@ -140,8 +141,7 @@
assertTrue("next returned a doc and it should not have", ts.next() == false);
}
- public void testSkipTo() throws Exception
- {
+ public void testSkipTo() throws Exception {
Term allTerm = new Term(FIELD, "all");
TermQuery termQuery = new TermQuery(allTerm);
@@ -151,7 +151,6 @@
TermScorer ts = new TermScorer(weight,
indexReader.termDocs(allTerm), indexSearcher.getSimilarity(),
indexReader.norms(FIELD));
- assertTrue("ts is null and it shouldn't be", ts != null);
assertTrue("Didn't skip", ts.skipTo(3) == true);
//The next doc should be doc 5
assertTrue("doc should be number 5", ts.doc() == 5);
@@ -167,7 +166,6 @@
TermScorer ts = new TermScorer(weight,
indexReader.termDocs(allTerm), indexSearcher.getSimilarity(),
indexReader.norms(FIELD));
- assertTrue("ts is null and it shouldn't be", ts != null);
Explanation explanation = ts.explain(0);
assertTrue("explanation is null and it shouldn't be", explanation != null);
//System.out.println("Explanation: " + explanation.toString());
@@ -185,7 +183,6 @@
ts = new TermScorer(weight, indexReader.termDocs(dogsTerm), indexSearcher.getSimilarity(),
indexReader.norms(FIELD));
- assertTrue("ts is null and it shouldn't be", ts != null);
explanation = ts.explain(1);
assertTrue("explanation is null and it shouldn't be", explanation != null);
//System.out.println("Explanation: " + explanation.toString());
@@ -202,23 +199,17 @@
}
- private class TestHit
- {
+ private class TestHit {
public int doc;
public float score;
- public TestHit(int doc, float score)
- {
+ public TestHit(int doc, float score) {
this.doc = doc;
this.score = score;
}
- public String toString()
- {
- return "TestHit{" +
- "doc=" + doc +
- ", score=" + score +
- "}";
+ public String toString() {
+ return "TestHit{" + "doc=" + doc + ", score=" + score + "}";
}
}
Index: src/test/org/apache/lucene/search/TestSetNorm.java
===================================================================
--- src/test/org/apache/lucene/search/TestSetNorm.java (revision 761300)
+++ src/test/org/apache/lucene/search/TestSetNorm.java (working copy)
@@ -17,6 +17,8 @@
* limitations under the License.
*/
+import java.io.IOException;
+
import org.apache.lucene.util.LuceneTestCase;
import org.apache.lucene.analysis.SimpleAnalyzer;
import org.apache.lucene.document.*;
@@ -62,11 +64,15 @@
new IndexSearcher(store).search
(new TermQuery(new Term("field", "word")),
- new MultiReaderHitCollector() {
- private int base = -1;
- public final void collect(int doc, float score) {
- scores[doc + base] = score;
+ new Collector() {
+ private int base = 0;
+ private Scorer scorer;
+ public void setScorer(Scorer scorer) throws IOException {
+ this.scorer = scorer;
}
+ public final void collect(int doc) throws IOException {
+ scores[doc + base] = scorer.score();
+ }
public void setNextReader(IndexReader reader, int docBase) {
base = docBase;
}
Index: src/test/org/apache/lucene/search/TestScorerPerf.java
===================================================================
--- src/test/org/apache/lucene/search/TestScorerPerf.java (revision 761300)
+++ src/test/org/apache/lucene/search/TestScorerPerf.java (working copy)
@@ -96,14 +96,16 @@
return sets;
}
- public static class CountingHitCollector extends MultiReaderHitCollector {
+ public static class CountingHitCollector extends Collector {
int count=0;
int sum=0;
- protected int docBase = -1;
+ protected int docBase = 0;
- public void collect(int doc, float score) {
+ public void setScorer(Scorer scorer) throws IOException {}
+
+ public void collect(int doc) {
count++;
- sum += docBase+doc; // use it to avoid any possibility of being optimized away
+ sum += docBase + doc; // use it to avoid any possibility of being optimized away
}
public int getCount() { return count; }
@@ -123,11 +125,12 @@
}
public void collect(int doc, float score) {
+
pos = answer.nextSetBit(pos+1);
if (pos != doc + docBase) {
throw new RuntimeException("Expected doc " + pos + " but got " + doc + docBase);
}
- super.collect(doc,score);
+ super.collect(doc);
}
}
Index: src/test/org/apache/lucene/search/TestSimilarity.java
===================================================================
--- src/test/org/apache/lucene/search/TestSimilarity.java (revision 761300)
+++ src/test/org/apache/lucene/search/TestSimilarity.java (working copy)
@@ -19,6 +19,7 @@
import org.apache.lucene.util.LuceneTestCase;
+import java.io.IOException;
import java.util.Collection;
import org.apache.lucene.index.IndexReader;
@@ -75,10 +76,14 @@
searcher.search
(new TermQuery(b),
- new MultiReaderHitCollector() {
- public final void collect(int doc, float score) {
- assertTrue(score == 1.0f);
+ new Collector() {
+ private Scorer scorer;
+ public void setScorer(Scorer scorer) throws IOException {
+ this.scorer = scorer;
}
+ public final void collect(int doc) throws IOException {
+ assertTrue(scorer.score() == 1.0f);
+ }
public void setNextReader(IndexReader reader, int docBase) {}
});
@@ -88,11 +93,15 @@
//System.out.println(bq.toString("field"));
searcher.search
(bq,
- new MultiReaderHitCollector() {
- private int base = -1;
- public final void collect(int doc, float score) {
+ new Collector() {
+ private int base = 0;
+ private Scorer scorer;
+ public void setScorer(Scorer scorer) throws IOException {
+ this.scorer = scorer;
+ }
+ public final void collect(int doc) throws IOException {
//System.out.println("Doc=" + doc + " score=" + score);
- assertTrue(score == (float)doc+base+1);
+ assertTrue(scorer.score() == (float)doc+base+1);
}
public void setNextReader(IndexReader reader, int docBase) {
base = docBase;
@@ -105,10 +114,14 @@
//System.out.println(pq.toString("field"));
searcher.search
(pq,
- new MultiReaderHitCollector() {
- public final void collect(int doc, float score) {
+ new Collector() {
+ private Scorer scorer;
+ public void setScorer(Scorer scorer) throws IOException {
+ this.scorer = scorer;
+ }
+ public final void collect(int doc) throws IOException {
//System.out.println("Doc=" + doc + " score=" + score);
- assertTrue(score == 1.0f);
+ assertTrue(scorer.score() == 1.0f);
}
public void setNextReader(IndexReader reader, int docBase) {}
});
@@ -117,10 +130,14 @@
//System.out.println(pq.toString("field"));
searcher.search
(pq,
- new MultiReaderHitCollector() {
- public final void collect(int doc, float score) {
+ new Collector() {
+ private Scorer scorer;
+ public void setScorer(Scorer scorer) throws IOException {
+ this.scorer = scorer;
+ }
+ public final void collect(int doc) throws IOException {
//System.out.println("Doc=" + doc + " score=" + score);
- assertTrue(score == 2.0f);
+ assertTrue(scorer.score() == 2.0f);
}
public void setNextReader(IndexReader reader, int docBase) {}
});
Index: src/test/org/apache/lucene/search/TestDocBoost.java
===================================================================
--- src/test/org/apache/lucene/search/TestDocBoost.java (revision 761300)
+++ src/test/org/apache/lucene/search/TestDocBoost.java (working copy)
@@ -17,6 +17,8 @@
* limitations under the License.
*/
+import java.io.IOException;
+
import org.apache.lucene.util.LuceneTestCase;
import org.apache.lucene.analysis.SimpleAnalyzer;
import org.apache.lucene.document.*;
@@ -66,11 +68,15 @@
new IndexSearcher(store).search
(new TermQuery(new Term("field", "word")),
- new MultiReaderHitCollector() {
- private int base = -1;
- public final void collect(int doc, float score) {
- scores[doc + base] = score;
+ new Collector() {
+ private int base = 0;
+ private Scorer scorer;
+ public void setScorer(Scorer scorer) throws IOException {
+ this.scorer = scorer;
}
+ public final void collect(int doc) throws IOException {
+ scores[doc + base] = scorer.score();
+ }
public void setNextReader(IndexReader reader, int docBase) {
base = docBase;
}
Index: src/test/org/apache/lucene/search/QueryUtils.java
===================================================================
--- src/test/org/apache/lucene/search/QueryUtils.java (revision 761300)
+++ src/test/org/apache/lucene/search/QueryUtils.java (working copy)
@@ -153,10 +153,15 @@
final int[] sdoc = new int[] {-1};
final float maxDiff = 1e-5f;
- s.search(q,new MultiReaderHitCollector() {
- private int base = -1;
- public void collect(int doc, float score) {
+ s.search(q,new Collector() {
+ private int base = 0;
+ private Scorer sc;
+ public void setScorer(Scorer scorer) throws IOException {
+ this.sc = scorer;
+ }
+ public void collect(int doc) throws IOException {
doc = doc + base;
+ float score = sc.score();
try {
int op = order[(opidx[0]++)%order.length];
//System.out.println(op==skip_op ? "skip("+(sdoc[0]+1)+")":"next()");
@@ -205,11 +210,16 @@
//System.out.println("checkFirstSkipTo: "+q);
final float maxDiff = 1e-5f;
final int lastDoc[] = {-1};
- s.search(q,new MultiReaderHitCollector() {
- private int base = -1;
- public void collect(int doc, float score) {
+ s.search(q,new Collector() {
+ private int base = 0;
+ private Scorer scorer;
+ public void setScorer(Scorer scorer) throws IOException {
+ this.scorer = scorer;
+ }
+ public void collect(int doc) throws IOException {
//System.out.println("doc="+doc);
doc = doc + base;
+ float score = scorer.score();
try {
for (int i=lastDoc[0]+1; i<=doc; i++) {
Weight w = q.weight(s);
Index: src/test/org/apache/lucene/search/TestPositiveScoresOnlyCollector.java
===================================================================
--- src/test/org/apache/lucene/search/TestPositiveScoresOnlyCollector.java (revision 0)
+++ src/test/org/apache/lucene/search/TestPositiveScoresOnlyCollector.java (revision 0)
@@ -0,0 +1,85 @@
+package org.apache.lucene.search;
+
+/**
+ * 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.io.IOException;
+
+import org.apache.lucene.util.LuceneTestCase;
+
+public class TestPositiveScoresOnlyCollector extends LuceneTestCase {
+
+ private static final class SimpleScorer extends Scorer {
+ private int idx = -1;
+
+ public SimpleScorer() {
+ super(null);
+ }
+
+ public Explanation explain(int doc) throws IOException { return null; }
+
+ public float score() throws IOException {
+ return idx == scores.length ? Float.NaN : scores[idx];
+ }
+
+ public int doc() { return idx; }
+
+ public boolean next() throws IOException {
+ return ++idx == scores.length;
+ }
+
+ public boolean skipTo(int target) throws IOException {
+ idx = target;
+ return idx >= scores.length;
+ }
+ }
+
+ // The scores must have positive as well as negative values
+ private static final float[] scores = new float[] { 0.7767749f, -1.7839992f,
+ 8.9925785f, 7.9608946f, -0.07948637f, 2.6356435f, 7.4950366f, 7.1490803f,
+ -8.108544f, 4.961808f, 2.2423935f, -7.285586f, 4.6699767f };
+
+ public void testNegativeScores() throws Exception {
+
+ // The Top*Collectors previously filtered out documents with <= scores. This
+ // behavior has changed. This test checks that if PositiveOnlyScoresFilter
+ // wraps one of these collectors, documents with <= 0 scores are indeed
+ // filtered.
+
+ int numPositiveScores = 0;
+ for (int i = 0; i < scores.length; i++) {
+ if (scores[i] > 0) {
+ ++numPositiveScores;
+ }
+ }
+
+ Scorer s = new SimpleScorer();
+ TopDocsCollector tdc = new TopScoreDocCollector(scores.length);
+ Collector c = new PositiveScoresOnlyCollector(tdc);
+ c.setScorer(s);
+ while (!s.next()) {
+ c.collect(0);
+ }
+ TopDocs td = tdc.topDocs();
+ ScoreDoc[] sd = td.scoreDocs;
+ assertEquals(numPositiveScores, td.totalHits);
+ for (int i = 0; i < sd.length; i++) {
+ assertTrue("only positive scores should return: " + sd[i].score, sd[i].score > 0);
+ }
+ }
+
+}
Index: src/test/org/apache/lucene/index/TestIndexReader.java
===================================================================
--- src/test/org/apache/lucene/index/TestIndexReader.java (revision 761300)
+++ src/test/org/apache/lucene/index/TestIndexReader.java (working copy)
@@ -40,8 +40,9 @@
import org.apache.lucene.document.SetBasedFieldSelector;
import org.apache.lucene.index.IndexReader.FieldOption;
import org.apache.lucene.search.IndexSearcher;
-import org.apache.lucene.search.MultiReaderHitCollector;
+import org.apache.lucene.search.Collector;
import org.apache.lucene.search.ScoreDoc;
+import org.apache.lucene.search.Scorer;
import org.apache.lucene.search.TermQuery;
import org.apache.lucene.search.Query;
import org.apache.lucene.search.FieldCache;
@@ -1651,7 +1652,7 @@
Query q = new TermQuery(new Term("id", "a"));
IndexSearcher s = new IndexSearcher(dir);
- s.search(q, new MultiReaderHitCollector() {
+ s.search(q, new Collector() {
int lastDocBase = -1;
public void setNextReader(IndexReader reader, int docBase) {
if (lastDocBase == -1) {
@@ -1663,13 +1664,14 @@
}
lastDocBase = docBase;
}
- public void collect(int doc, float score) {}
+ public void collect(int doc) {}
+ public void setScorer(Scorer scorer) {}
});
s.close();
IndexReader r = IndexReader.open(dir);
s = new IndexSearcher(r, true);
- s.search(q, new MultiReaderHitCollector() {
+ s.search(q, new Collector() {
int lastDocBase = -1;
public void setNextReader(IndexReader reader, int docBase) {
if (lastDocBase == -1) {
@@ -1681,7 +1683,8 @@
}
lastDocBase = docBase;
}
- public void collect(int doc, float score) {}
+ public void collect(int doc) {}
+ public void setScorer(Scorer scorer) {}
});
s.close();
r.close();
Index: src/test/org/apache/lucene/index/TestOmitTf.java
===================================================================
--- src/test/org/apache/lucene/index/TestOmitTf.java (revision 761300)
+++ src/test/org/apache/lucene/index/TestOmitTf.java (working copy)
@@ -17,6 +17,7 @@
* limitations under the License.
*/
+import java.io.IOException;
import java.util.Collection;
import org.apache.lucene.util.LuceneTestCase;
@@ -27,7 +28,8 @@
import org.apache.lucene.document.Field;
import org.apache.lucene.search.BooleanQuery;
import org.apache.lucene.search.IndexSearcher;
-import org.apache.lucene.search.MultiReaderHitCollector;
+import org.apache.lucene.search.Collector;
+import org.apache.lucene.search.Scorer;
import org.apache.lucene.search.Searcher;
import org.apache.lucene.search.Similarity;
import org.apache.lucene.search.TermQuery;
@@ -283,10 +285,15 @@
searcher.search(q1,
new CountingHitCollector() {
- public final void collect(int doc, float score) {
+ private Scorer scorer;
+ public final void setScorer(Scorer scorer) {
+ this.scorer = scorer;
+ }
+ public final void collect(int doc) throws IOException {
//System.out.println("Q1: Doc=" + doc + " score=" + score);
+ float score = scorer.score();
assertTrue(score==1.0f);
- super.collect(doc, score);
+ super.collect(doc);
}
});
//System.out.println(CountingHitCollector.getCount());
@@ -294,10 +301,15 @@
searcher.search(q2,
new CountingHitCollector() {
- public final void collect(int doc, float score) {
- //System.out.println("Q2: Doc=" + doc + " score=" + score);
+ private Scorer scorer;
+ public final void setScorer(Scorer scorer) {
+ this.scorer = scorer;
+ }
+ public final void collect(int doc) throws IOException {
+ //System.out.println("Q2: Doc=" + doc + " score=" + score);
+ float score = scorer.score();
assertTrue(score==1.0f+doc);
- super.collect(doc, score);
+ super.collect(doc);
}
});
//System.out.println(CountingHitCollector.getCount());
@@ -308,11 +320,16 @@
searcher.search(q3,
new CountingHitCollector() {
- public final void collect(int doc, float score) {
+ private Scorer scorer;
+ public final void setScorer(Scorer scorer) {
+ this.scorer = scorer;
+ }
+ public final void collect(int doc) throws IOException {
//System.out.println("Q1: Doc=" + doc + " score=" + score);
+ float score = scorer.score();
assertTrue(score==1.0f);
assertFalse(doc%2==0);
- super.collect(doc, score);
+ super.collect(doc);
}
});
//System.out.println(CountingHitCollector.getCount());
@@ -320,11 +337,16 @@
searcher.search(q4,
new CountingHitCollector() {
- public final void collect(int doc, float score) {
+ private Scorer scorer;
+ public final void setScorer(Scorer scorer) {
+ this.scorer = scorer;
+ }
+ public final void collect(int doc) throws IOException {
+ float score = scorer.score();
//System.out.println("Q1: Doc=" + doc + " score=" + score);
assertTrue(score==1.0f);
assertTrue(doc%2==0);
- super.collect(doc, score);
+ super.collect(doc);
}
});
//System.out.println(CountingHitCollector.getCount());
@@ -337,9 +359,9 @@
searcher.search(bq,
new CountingHitCollector() {
- public final void collect(int doc, float score) {
+ public final void collect(int doc) throws IOException {
//System.out.println("BQ: Doc=" + doc + " score=" + score);
- super.collect(doc, score);
+ super.collect(doc);
}
});
assertTrue(15 == CountingHitCollector.getCount());
@@ -348,12 +370,13 @@
dir.close();
}
- public static class CountingHitCollector extends MultiReaderHitCollector {
+ public static class CountingHitCollector extends Collector {
static int count=0;
static int sum=0;
private int docBase = -1;
CountingHitCollector(){count=0;sum=0;}
- public void collect(int doc, float score) {
+ public void setScorer(Scorer scorer) throws IOException {}
+ public void collect(int doc) throws IOException {
count++;
sum += doc + docBase; // use it to avoid any possibility of being optimized away
}
Index: src/java/org/apache/lucene/search/TopFieldDocCollector.java
===================================================================
--- src/java/org/apache/lucene/search/TopFieldDocCollector.java (revision 761300)
+++ src/java/org/apache/lucene/search/TopFieldDocCollector.java (working copy)
@@ -30,8 +30,8 @@
* documents are collected.
*
* @deprecated Please use {@link TopFieldCollector} instead.
- **/
-public class TopFieldDocCollector extends TopScoreDocCollector {
+ */
+public class TopFieldDocCollector extends TopDocCollector {
private FieldDoc reusableFD;
@@ -50,7 +50,7 @@
if (score > 0.0f) {
totalHits++;
if (reusableFD == null)
- reusableFD = new FieldDoc(doc + docBase, score);
+ reusableFD = new FieldDoc(doc, score);
else {
// Whereas TopScoreDocCollector can skip this if the
// score is not competitive, we cannot because the
@@ -58,7 +58,7 @@
// aren't in general congruent with "higher score
// wins"
reusableFD.score = score;
- reusableFD.doc = doc + docBase;
+ reusableFD.doc = doc;
}
reusableFD = (FieldDoc) hq.insertWithOverflow(reusableFD);
}
Index: src/java/org/apache/lucene/search/RemoteSearchable.java
===================================================================
--- src/java/org/apache/lucene/search/RemoteSearchable.java (revision 761300)
+++ src/java/org/apache/lucene/search/RemoteSearchable.java (working copy)
@@ -45,12 +45,17 @@
this.local = local;
}
-
+ /** @deprecated use {@link #search(Weight, Filter, Collector)} instead. */
public void search(Weight weight, Filter filter, HitCollector results)
throws IOException {
local.search(weight, filter, results);
}
+ public void search(Weight weight, Filter filter, Collector results)
+ throws IOException {
+ local.search(weight, filter, results);
+ }
+
public void close() throws IOException {
local.close();
}
Index: src/java/org/apache/lucene/search/BooleanScorer2.java
===================================================================
--- src/java/org/apache/lucene/search/BooleanScorer2.java (revision 761300)
+++ src/java/org/apache/lucene/search/BooleanScorer2.java (working copy)
@@ -300,8 +300,17 @@
* @param hc The collector to which all matching documents are passed through
* {@link HitCollector#collect(int, float)}.
* FieldCache.DEFAULT for maintaining
@@ -61,11 +61,9 @@
* priority order (highest priority first); cannot be null or empty
* @param size The number of hits to retain. Must be
* greater than zero.
- * @param subReaders Array of IndexReaders we will search,
- * in order that they will be searched
* @throws IOException
*/
- public FieldValueHitQueue(SortField[] fields, int size, IndexReader[] subReaders) throws IOException {
+ public FieldValueHitQueue(SortField[] fields, int size) throws IOException {
numComparators = fields.length;
comparators = new FieldComparator[numComparators];
reverseMul = new int[numComparators];
@@ -82,7 +80,7 @@
assert field.getType() != SortField.AUTO;
reverseMul[i] = field.reverse ? -1 : 1;
- comparators[i] = field.getComparator(subReaders, size, i, field.reverse);
+ comparators[i] = field.getComparator(size, i, field.reverse);
}
if (numComparators == 1) {
@@ -180,7 +178,7 @@
* Attempts to detect the given field type for an IndexReader.
*/
static int detectFieldType(IndexReader reader, String fieldKey) throws IOException {
- String field = ((String)fieldKey).intern();
+ String field = fieldKey.intern();
TermEnum enumerator = reader.terms (new Term (field));
try {
Term term = enumerator.term();
Index: src/java/org/apache/lucene/search/TopScoreDocCollector.java
===================================================================
--- src/java/org/apache/lucene/search/TopScoreDocCollector.java (revision 761300)
+++ src/java/org/apache/lucene/search/TopScoreDocCollector.java (working copy)
@@ -17,85 +17,71 @@
* limitations under the License.
*/
+import java.io.IOException;
+
import org.apache.lucene.index.IndexReader;
-import org.apache.lucene.util.PriorityQueue;
-/** A {@link MultiReaderHitCollector} implementation that
- * collects the top-scoring documents, returning them as a
- * {@link TopDocs}. This is used by {@link IndexSearcher}
- * to implement {@link TopDocs}-based search.
- *
- * This may be extended, overriding the {@link
- * MultiReaderHitCollector#collect} method to, e.g.,
- * conditionally invoke super() in order to
- * filter which documents are collected, but sure you
- * either take docBase into account, or also override
- * {@link MultiReaderHitCollector#setNextReader} method. */
-public class TopScoreDocCollector extends MultiReaderHitCollector {
+/**
+ * A {@link Collector} implementation that collects the top-scoring documents,
+ * returning them as a {@link TopDocs}. This is used by {@link IndexSearcher} to
+ * implement {@link TopDocs}-based search.
+ */
+public final class TopScoreDocCollector extends TopDocsCollector {
private ScoreDoc reusableSD;
-
- /** The total number of hits the collector encountered. */
- protected int totalHits;
-
- /** The priority queue which holds the top-scoring documents. */
- protected PriorityQueue hq;
-
- protected int docBase = 0;
+ private int docBase = 0;
+ private Scorer scorer;
/** Construct to collect a given number of hits.
* @param numHits the maximum number of hits to collect
*/
public TopScoreDocCollector(int numHits) {
- this(new HitQueue(numHits));
+ super(new HitQueue(numHits));
}
- /** Constructor to collect the top-scoring documents by using the given PQ.
- * @param hq the PQ to use by this instance.
- */
- protected TopScoreDocCollector(PriorityQueue hq) {
- this.hq = hq;
- }
-
- // javadoc inherited
- public void collect(int doc, float score) {
- if (score > 0.0f) {
- totalHits++;
- if (reusableSD == null) {
- reusableSD = new ScoreDoc(doc + docBase, score);
- } else if (score >= reusableSD.score) {
- // reusableSD holds the last "rejected" entry, so, if
- // this new score is not better than that, there's no
- // need to try inserting it
- reusableSD.doc = doc + docBase;
- reusableSD.score = score;
- } else {
- return;
- }
- reusableSD = (ScoreDoc) hq.insertWithOverflow(reusableSD);
+ protected TopDocs newTopDocs(ScoreDoc[] results, int start) {
+ if (results == null) {
+ return EMPTY_TOPDOCS;
}
- }
-
- /** The total number of documents that matched this query. */
- public int getTotalHits() {
- return totalHits;
- }
-
- /** The top-scoring hits. */
- public TopDocs topDocs() {
- ScoreDoc[] scoreDocs = new ScoreDoc[hq.size()];
- for (int i = hq.size()-1; i >= 0; i--) { // put docs in array
- scoreDocs[i] = (ScoreDoc) hq.pop();
+
+ // We need to compute maxScore in order to set it in TopDocs. If start == 0,
+ // it means the largest element is already in results, use its score as
+ // maxScore. Otherwise pop everything else, until the largest element is
+ // extracted and use its score as maxScore.
+ float maxScore = Float.NaN;
+ if (start == 0) {
+ maxScore = results[0].score;
+ } else {
+ for (int i = pq.size(); i > 1; i--) { pq.pop(); }
+ maxScore = ((ScoreDoc) pq.pop()).score;
}
-
- float maxScore = (totalHits==0)
- ? Float.NEGATIVE_INFINITY
- : scoreDocs[0].score;
- return new TopDocs(totalHits, scoreDocs, maxScore);
+ return new TopDocs(totalHits, results, maxScore);
}
+ // javadoc inherited
+ public void collect(int doc) throws IOException {
+ float score = scorer.score();
+ totalHits++;
+ if (reusableSD == null) {
+ reusableSD = new ScoreDoc(doc + docBase, score);
+ } else if (score >= reusableSD.score) {
+ // reusableSD holds the last "rejected" entry, so, if
+ // this new score is not better than that, there's no
+ // need to try inserting it
+ reusableSD.doc = doc + docBase;
+ reusableSD.score = score;
+ } else {
+ return;
+ }
+ reusableSD = (ScoreDoc) pq.insertWithOverflow(reusableSD);
+ }
+
public void setNextReader(IndexReader reader, int base) {
docBase = base;
}
+
+ public void setScorer(Scorer scorer) throws IOException {
+ this.scorer = scorer;
+ }
}
Index: src/java/org/apache/lucene/search/Searcher.java
===================================================================
--- src/java/org/apache/lucene/search/Searcher.java (revision 761300)
+++ src/java/org/apache/lucene/search/Searcher.java (working copy)
@@ -76,9 +76,13 @@
* the top n hits for query, applying
* filter if non-null, and sorting the hits by the criteria in
* sort.
+ *
+ * NOTE: currently, this method tracks document scores and sets them in
+ * the returned {@link FieldDoc}, however in 3.0 it will move to not track
+ * document scores. If document scores tracking is still needed, you can use
+ * {@link #search(Weight, Filter, Collector)} and pass in a
+ * {@link TopFieldCollector} instance.
*
- *
Applications should usually call {@link - * Searcher#search(Query,Filter,Sort)} instead. * @throws BooleanQuery.TooManyClauses */ public TopFieldDocs search(Query query, Filter filter, int n, @@ -99,6 +103,7 @@ * In other words, the score will not necessarily be a float whose value is * between 0 and 1. * @throws BooleanQuery.TooManyClauses + * @deprecated use {@link #search(Query, Collector)} instead. */ public void search(Query query, HitCollector results) throws IOException { @@ -106,6 +111,24 @@ } /** Lower-level search API. + * + *
{@link Collector#collect(int)} is called for every matching document. + * + *
Applications should only use this if they need all of the + * matching documents. The high-level search API ({@link + * Searcher#search(Query)}) is usually more efficient, as it skips + * non-high-scoring hits. + *
Note: The score passed to this method is a raw score.
+ * In other words, the score will not necessarily be a float whose value is
+ * between 0 and 1.
+ * @throws BooleanQuery.TooManyClauses
+ */
+ public void search(Query query, Collector results)
+ throws IOException {
+ search(query, (Filter)null, results);
+ }
+
+ /** Lower-level search API.
*
*
{@link HitCollector#collect(int,float)} is called for every matching * document. @@ -120,11 +143,33 @@ * @param filter if non-null, used to permit documents to be collected. * @param results to receive hits * @throws BooleanQuery.TooManyClauses + * @deprecated use {@link #search(Query, Filter, Collector)} instead. */ public void search(Query query, Filter filter, HitCollector results) throws IOException { search(createWeight(query), filter, results); } + + /** Lower-level search API. + * + *
{@link Collector#collect(int)} is called for every matching
+ * document.
+ *
Collector-based access to remote indexes is discouraged.
+ *
+ *
Applications should only use this if they need all of the
+ * matching documents. The high-level search API ({@link
+ * Searcher#search(Query, Filter, int)}) is usually more efficient, as it skips
+ * non-high-scoring hits.
+ *
+ * @param query to match documents
+ * @param filter if non-null, used to permit documents to be collected.
+ * @param results to receive hits
+ * @throws BooleanQuery.TooManyClauses
+ */
+ public void search(Query query, Filter filter, Collector results)
+ throws IOException {
+ search(createWeight(query), filter, results);
+ }
/** Finds the top n
* hits for query, applying filter if non-null.
@@ -197,7 +242,11 @@
/* The following abstract methods were added as a workaround for GCJ bug #15411.
* http://gcc.gnu.org/bugzilla/show_bug.cgi?id=15411
*/
+ /**
+ * @deprecated use {@link #search(Weight, Filter, Collector)} instead.
+ */
abstract public void search(Weight weight, Filter filter, HitCollector results) throws IOException;
+ abstract public void search(Weight weight, Filter filter, Collector results) throws IOException;
abstract public void close() throws IOException;
abstract public int docFreq(Term term) throws IOException;
abstract public int maxDoc() throws IOException;
Index: src/java/org/apache/lucene/search/Collector.java
===================================================================
--- src/java/org/apache/lucene/search/Collector.java (revision 0)
+++ src/java/org/apache/lucene/search/Collector.java (revision 0)
@@ -0,0 +1,160 @@
+package org.apache.lucene.search;
+
+/**
+ * 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.io.IOException;
+
+import org.apache.lucene.index.IndexReader;
+
+/**
+ *
Expert: Collectors are primarily meant to be used to + * gather raw results from a search, and implement sorting + * or custom result filtering, collation, etc.
+ * + *As of 2.9, this class replaces the deprecated + * HitCollector, and offers an API for efficient collection + * of hits across sequential {@link IndexReader}s. {@link + * IndexSearcher} advances the collector through each of the + * sub readers, in an arbitrary order. This results in a + * higher performance means of collection.
+ * + *Lucene's core collectors are derived from Collector: + * + *
Collector decouples the score from the collected doc: + * the score computation is skipped entirely if it's not + * needed. Collectors that do need the score should + * implement the {@link #setScorer} method, to hold onto the + * passed {@link Scorer} instance, and call {@link + * Scorer#score()} within collect to compute the current + * hit's score. If your collector may request the score + * for a single hit multiple times, you should use {@link + * ScoreCachingWrappingScorer}.
+ * + *NOTE: The doc that is passed to the collect + * method is relative to the current reader. If your + * collector needs to resolve this to the docID space of the + * Multi*Reader, you must re-base it by recording the + * docBase from the most recent setNextReader call. Here's + * a simple example showing how to collect docIDs into a + * BitSet:
+ * + *
+ * Searcher searcher = new IndexSearcher(indexReader);
+ * final BitSet bits = new BitSet(indexReader.maxDoc());
+ * searcher.search(query, new Collector() {
+ * private int docBase = 0;
+ *
+ * // ignore scorer
+ * public void setScorer(Scorer scorer) {
+ * }
+ *
+ * public void collect(int doc) {
+ * bits.set(doc + docBase);
+ * }
+ *
+ * public void setNextReader(IndexReader reader, int docBase) {
+ * this.docBase = docBase;
+ * }
+ * });
+ *
+ *
+ * Not all collectors will need to rebase the docID. For + * example, a collector that simply counts the total number + * of hits would skip it.
+ * + *NOTE: Prior Lucene releases silently filtered + * out hits with score <= 0, but the core Collectors no + * longer do that. It's very unusual to have such hits (a + * negative query boost, or function query, could cause it + * to happen). If you need that behavior, use {@link + * PositiveScoresOnlyCollector}.
+ * + *NOTE: This API is experimental and might change + * in incompatible ways in the next release.
+ */ +public abstract class Collector { + + /** + * Called before successive calls to {@link #collect(int)}. Implementations + * that need the score of the current document (passed-in to + * {@link #collect(int)}), should save the passed-in Scorer and call + * scorer.score() when needed. + */ + public abstract void setScorer(Scorer scorer) throws IOException; + + /** + * Called once for every document matching a query, with the unbased document + * number. + * + *+ * Note: This is called in an inner search loop. For good search performance, + * implementations of this method should not call {@link Searcher#doc(int)} or + * {@link org.apache.lucene.index.IndexReader#document(int)} on every document + * number encountered. Doing so can slow searches by an order of magnitude or + * more. + */ + public abstract void collect(int doc) throws IOException; + + /** + * Called before collecting from each IndexReader. All doc ids in + * {@link #collect(int)} will correspond to reader. + * + * Add docBase to the current IndexReaders internal document id to re-base ids + * in {@link #collect(int)}. + * + * @param reader + * next IndexReader + * @param docBase + */ + public abstract void setNextReader(IndexReader reader, int docBase) throws IOException; + +} Index: src/java/org/apache/lucene/search/MultiReaderHitCollector.java =================================================================== --- src/java/org/apache/lucene/search/MultiReaderHitCollector.java (revision 761300) +++ src/java/org/apache/lucene/search/MultiReaderHitCollector.java (working copy) @@ -1,53 +0,0 @@ -package org.apache.lucene.search; - -/** - * 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.io.IOException; - -import org.apache.lucene.index.IndexReader; - -/** - * Expert: A HitCollector that can be used to collect hits - * across sequential IndexReaders. For a Multi*Reader, this - * collector advances through each of the sub readers, in an - * arbitrary order. This results in a higher performance - * means of collection. - * - * NOTE: The doc that is passed to the collect method - * is relative to the current reader. You must re-base the - * doc, by recording the docBase from the last setNextReader - * call, to map it to the docID space of the - * Multi*Reader. - * - * NOTE: This API is experimental and might change in - * incompatible ways in the next release. - */ -public abstract class MultiReaderHitCollector extends HitCollector { - /** - * Called before collecting from each IndexReader. All doc - * ids in {@link #collect(int, float)} will correspond to reader. - * - * Add docBase to the current IndexReaders internal document id to - * re-base ids in {@link #collect(int, float)}. - * - * @param reader next IndexReader - * @param docBase - * @throws IOException - */ - public abstract void setNextReader(IndexReader reader, int docBase) throws IOException; -} Index: src/java/org/apache/lucene/search/Searchable.java =================================================================== --- src/java/org/apache/lucene/search/Searchable.java (revision 761300) +++ src/java/org/apache/lucene/search/Searchable.java (working copy) @@ -19,7 +19,7 @@ import org.apache.lucene.document.Document; import org.apache.lucene.document.FieldSelector; -import org.apache.lucene.index.IndexReader; +import org.apache.lucene.index.IndexReader; // for javadoc import org.apache.lucene.index.Term; import org.apache.lucene.index.CorruptIndexException; @@ -51,10 +51,32 @@ * @param filter if non-null, used to permit documents to be collected. * @param results to receive hits * @throws BooleanQuery.TooManyClauses + * @deprecated use {@link #search(Weight, Filter, Collector)} instead. */ void search(Weight weight, Filter filter, HitCollector results) throws IOException; + /** + * Lower-level search API. + * + *
+ * {@link Collector#collect(int)} is called for every document.
+ * Collector-based access to remote indexes is discouraged.
+ *
+ *
+ * Applications should only use this if they need all of the matching + * documents. The high-level search API ({@link Searcher#search(Query)}) is + * usually more efficient, as it skips non-high-scoring hits. + * + * @param weight + * to match documents + * @param filter + * if non-null, used to permit documents to be collected. + * @param collector + * to receive hits + * @throws BooleanQuery.TooManyClauses + */ + void search(Weight weight, Filter filter, Collector collector) throws IOException; /** Frees resources associated with this Searcher. * Be careful not to call this method while you are still using objects @@ -147,6 +169,13 @@ * *
Applications should usually call {@link
* Searcher#search(Query,Filter,Sort)} instead.
+ *
+ * NOTE: currently, this method tracks document scores and sets them in
+ * the returned {@link FieldDoc}, however in 3.0 it will move to not track
+ * document scores. If document scores tracking is still needed, you can use
+ * {@link #search(Weight, Filter, Collector)} and pass in a
+ * {@link TopFieldCollector} instance.
+ *
* @throws BooleanQuery.TooManyClauses
*/
TopFieldDocs search(Weight weight, Filter filter, int n, Sort sort)
Index: src/java/org/apache/lucene/search/PositiveScoresOnlyCollector.java
===================================================================
--- src/java/org/apache/lucene/search/PositiveScoresOnlyCollector.java (revision 0)
+++ src/java/org/apache/lucene/search/PositiveScoresOnlyCollector.java (revision 0)
@@ -0,0 +1,56 @@
+package org.apache.lucene.search;
+
+/**
+ * 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.io.IOException;
+
+import org.apache.lucene.index.IndexReader;
+
+/**
+ * A {@link Collector} implementation which wraps another
+ * {@link Collector} and makes sure only documents with
+ * scores > 0 are collected.
+ */
+
+public class PositiveScoresOnlyCollector extends Collector {
+
+ final private Collector c;
+ private Scorer scorer;
+
+ public PositiveScoresOnlyCollector(Collector c) {
+ this.c = c;
+ }
+
+ public void collect(int doc) throws IOException {
+ if (scorer.score() > 0) {
+ c.collect(doc);
+ }
+ }
+
+ public void setNextReader(IndexReader reader, int docBase) throws IOException {
+ c.setNextReader(reader, docBase);
+ }
+
+ public void setScorer(Scorer scorer) throws IOException {
+ // Set a ScoreCachingWrappingScorer in case the wrapped Collector will call
+ // score() also.
+ this.scorer = new ScoreCachingWrappingScorer(scorer);
+ c.setScorer(this.scorer);
+ }
+
+}
Index: src/java/org/apache/lucene/search/HitCollector.java
===================================================================
--- src/java/org/apache/lucene/search/HitCollector.java (revision 761300)
+++ src/java/org/apache/lucene/search/HitCollector.java (working copy)
@@ -17,14 +17,13 @@
* limitations under the License.
*/
-/** Lower-level search API.
- *
HitCollectors are primarily meant to be used to implement queries,
- * sorting and filtering. See {@link
- * MultiReaderHitCollector} for a lower level and
- * higher performance (on a multi-segment index) API.
- * @see Searcher#search(Query,HitCollector)
- * @version $Id$
- */
+/**
++ * Lower-level search API.
++ * HitCollectors are primarily meant to be used to implement queries, sorting
++ * and filtering.
++ *
++ * @deprecated use {@link Collector} instead.
+ */
public abstract class HitCollector {
/** Called once for every document matching a query, with the document
* number and its raw score.
@@ -51,4 +50,5 @@
* between 0 and 1.
*/
public abstract void collect(int doc, float score);
+
}
Index: src/java/org/apache/lucene/search/TopDocs.java
===================================================================
--- src/java/org/apache/lucene/search/TopDocs.java (revision 761300)
+++ src/java/org/apache/lucene/search/TopDocs.java (working copy)
@@ -29,7 +29,10 @@
/** Expert: Stores the maximum score value encountered, needed for normalizing. */
private float maxScore;
- /** Expert: Returns the maximum score value encountered. */
+ /**
+ * Expert: Returns the maximum score value encountered. Note that in case
+ * scores are not tracked, this returns {@link Float#NaN}.
+ */
public float getMaxScore() {
return maxScore;
}
@@ -38,7 +41,12 @@
public void setMaxScore(float maxScore) {
this.maxScore=maxScore;
}
-
+
+ /** Expert: Constructs a TopDocs with a default maxScore=Float.NaN. */
+ TopDocs(int totalHits, ScoreDoc[] scoreDocs) {
+ this(totalHits, scoreDocs, Float.NaN);
+ }
+
/** Expert: Constructs a TopDocs.*/
public TopDocs(int totalHits, ScoreDoc[] scoreDocs, float maxScore) {
this.totalHits = totalHits;
Index: src/java/org/apache/lucene/search/FieldComparator.java
===================================================================
--- src/java/org/apache/lucene/search/FieldComparator.java (revision 761300)
+++ src/java/org/apache/lucene/search/FieldComparator.java (working copy)
@@ -62,11 +62,11 @@
return values[slot1] - values[slot2];
}
- public int compareBottom(int doc, float score) {
+ public int compareBottom(int doc) {
return bottom - currentReaderValues[doc];
}
- public void copy(int slot, int doc, float score) {
+ public void copy(int slot, int doc) {
values[slot] = currentReaderValues[doc];
}
@@ -87,7 +87,7 @@
public Comparable value(int slot) {
return new Byte(values[slot]);
}
- };
+ }
/** Sorts by ascending docID */
public static final class DocComparator extends FieldComparator {
@@ -104,12 +104,12 @@
return docIDs[slot1] - docIDs[slot2];
}
- public int compareBottom(int doc, float score) {
+ public int compareBottom(int doc) {
// No overflow risk because docIDs are non-negative
return bottom - (docBase + doc);
}
- public void copy(int slot, int doc, float score) {
+ public void copy(int slot, int doc) {
docIDs[slot] = docBase + doc;
}
@@ -131,7 +131,7 @@
public Comparable value(int slot) {
return new Integer(docIDs[slot]);
}
- };
+ }
/** Parses field's values as double (using {@link
* ExtendedFieldCache#getDoubles} and sorts by ascending value */
@@ -160,7 +160,7 @@
}
}
- public int compareBottom(int doc, float score) {
+ public int compareBottom(int doc) {
final double v2 = currentReaderValues[doc];
if (bottom > v2) {
return 1;
@@ -171,7 +171,7 @@
}
}
- public void copy(int slot, int doc, float score) {
+ public void copy(int slot, int doc) {
values[slot] = currentReaderValues[doc];
}
@@ -192,7 +192,7 @@
public Comparable value(int slot) {
return new Double(values[slot]);
}
- };
+ }
/** Parses field's values as float (using {@link
* FieldCache#getFloats} and sorts by ascending value */
@@ -223,7 +223,7 @@
}
}
- public int compareBottom(int doc, float score) {
+ public int compareBottom(int doc) {
// TODO: are there sneaky non-branch ways to compute
// sign of float?
final float v2 = currentReaderValues[doc];
@@ -236,7 +236,7 @@
}
}
- public void copy(int slot, int doc, float score) {
+ public void copy(int slot, int doc) {
values[slot] = currentReaderValues[doc];
}
@@ -256,7 +256,7 @@
public Comparable value(int slot) {
return new Float(values[slot]);
}
- };
+ }
/** Parses field's values as int (using {@link
* FieldCache#getInts} and sorts by ascending value */
@@ -289,7 +289,7 @@
}
}
- public int compareBottom(int doc, float score) {
+ public int compareBottom(int doc) {
// TODO: there are sneaky non-branch ways to compute
// -1/+1/0 sign
// Cannot return bottom - values[slot2] because that
@@ -304,7 +304,7 @@
}
}
- public void copy(int slot, int doc, float score) {
+ public void copy(int slot, int doc) {
values[slot] = currentReaderValues[doc];
}
@@ -324,7 +324,7 @@
public Comparable value(int slot) {
return new Integer(values[slot]);
}
- };
+ }
/** Parses field's values as long (using {@link
* ExtendedFieldCache#getLongs} and sorts by ascending value */
@@ -355,7 +355,7 @@
}
}
- public int compareBottom(int doc, float score) {
+ public int compareBottom(int doc) {
// TODO: there are sneaky non-branch ways to compute
// -1/+1/0 sign
final long v2 = currentReaderValues[doc];
@@ -368,7 +368,7 @@
}
}
- public void copy(int slot, int doc, float score) {
+ public void copy(int slot, int doc) {
values[slot] = currentReaderValues[doc];
}
@@ -389,7 +389,7 @@
public Comparable value(int slot) {
return new Long(values[slot]);
}
- };
+ }
/** Sorts by descending relevance. NOTE: if you are
* sorting only by descending relevance and then
@@ -400,7 +400,8 @@
public static final class RelevanceComparator extends FieldComparator {
private final float[] scores;
private float bottom;
-
+ private Scorer scorer;
+
RelevanceComparator(int numHits) {
scores = new float[numHits];
}
@@ -408,27 +409,16 @@
public int compare(int slot1, int slot2) {
final float score1 = scores[slot1];
final float score2 = scores[slot2];
- if (score1 > score2) {
- return -1;
- } else if (score1 < score2) {
- return 1;
- } else {
- return 0;
- }
+ return score1 > score2 ? -1 : (score1 < score2 ? 1 : 0);
}
- public int compareBottom(int doc, float score) {
- if (bottom > score) {
- return -1;
- } else if (bottom < score) {
- return 1;
- } else {
- return 0;
- }
+ public int compareBottom(int doc) throws IOException {
+ float score = scorer.score();
+ return bottom > score ? -1 : (bottom < score ? 1 : 0);
}
- public void copy(int slot, int doc, float score) {
- scores[slot] = score;
+ public void copy(int slot, int doc) throws IOException {
+ scores[slot] = scorer.score();
}
public void setNextReader(IndexReader reader, int docBase, int numSlotsFull) {
@@ -438,6 +428,12 @@
this.bottom = scores[bottom];
}
+ public void setScorer(Scorer scorer) {
+ // wrap with a ScoreCachingWrappingScorer so that successive calls to
+ // score() will not incur score computation over and over again.
+ this.scorer = new ScoreCachingWrappingScorer(scorer);
+ }
+
public int sortType() {
return SortField.SCORE;
}
@@ -445,7 +441,7 @@
public Comparable value(int slot) {
return new Float(scores[slot]);
}
- };
+ }
/** Parses field's values as short (using {@link
* FieldCache#getShorts} and sorts by ascending value */
@@ -466,11 +462,11 @@
return values[slot1] - values[slot2];
}
- public int compareBottom(int doc, float score) {
+ public int compareBottom(int doc) {
return bottom - currentReaderValues[doc];
}
- public void copy(int slot, int doc, float score) {
+ public void copy(int slot, int doc) {
values[slot] = currentReaderValues[doc];
}
@@ -491,7 +487,7 @@
public Comparable value(int slot) {
return new Short(values[slot]);
}
- };
+ }
/** Sorts by a field's value using the Collator for a
* given Locale.*/
@@ -523,7 +519,7 @@
return collator.compare(val1, val2);
}
- public int compareBottom(int doc, float score) {
+ public int compareBottom(int doc) {
final String val2 = currentReaderValues[doc];
if (bottom == null) {
if (val2 == null) {
@@ -536,7 +532,7 @@
return collator.compare(bottom, val2);
}
- public void copy(int slot, int doc, float score) {
+ public void copy(int slot, int doc) {
values[slot] = currentReaderValues[doc];
}
@@ -556,7 +552,7 @@
public Comparable value(int slot) {
return values[slot];
}
- };
+ }
// NOTE: there were a number of other interesting String
// comparators explored, but this one seemed to perform
@@ -608,7 +604,7 @@
return val1.compareTo(val2);
}
- public int compareBottom(int doc, float score) {
+ public int compareBottom(int doc) {
assert bottomSlot != -1;
int order = this.order[doc];
final int cmp = bottomOrd - order;
@@ -659,7 +655,7 @@
ords[slot] = index;
}
- public void copy(int slot, int doc, float score) {
+ public void copy(int slot, int doc) {
final int ord = order[doc];
ords[slot] = ord;
assert ord >= 0;
@@ -709,7 +705,7 @@
public String getField() {
return field;
}
- };
+ }
/** Sorts by field's natural String sort order. All
* comparisons are done using String.compareTo, which is
@@ -742,7 +738,7 @@
return val1.compareTo(val2);
}
- public int compareBottom(int doc, float score) {
+ public int compareBottom(int doc) {
final String val2 = currentReaderValues[doc];
if (bottom == null) {
if (val2 == null) {
@@ -755,7 +751,7 @@
return bottom.compareTo(val2);
}
- public void copy(int slot, int doc, float score) {
+ public void copy(int slot, int doc) {
values[slot] = currentReaderValues[doc];
}
@@ -775,11 +771,11 @@
public Comparable value(int slot) {
return values[slot];
}
- };
+ }
final protected static int binarySearch(String[] a, String key) {
return binarySearch(a, key, 0, a.length-1);
- };
+ }
final protected static int binarySearch(String[] a, String key, int low, int high) {
@@ -801,7 +797,7 @@
return mid;
}
return -(low + 1);
- };
+ }
/**
* Compare hit at slot1 with hit at slot2. Return
@@ -827,22 +823,20 @@
* only invoked after setBottom has been called.
*
* @param doc that was hit
- * @param score of the hit
* @return any N < 0 if the doc's value is sorted after
* the bottom entry (not competitive), any N > 0 if the
* doc's value is sorted before the bottom entry and 0 if
* they are equal.
*/
- public abstract int compareBottom(int doc, float score);
+ public abstract int compareBottom(int doc) throws IOException;
/**
* Copy hit (doc,score) to hit slot.
*
* @param slot which slot to copy the hit to
* @param doc docID relative to current reader
- * @param score hit score
*/
- public abstract void copy(int slot, int doc, float score);
+ public abstract void copy(int slot, int doc) throws IOException;
/**
* Set a new Reader. All doc correspond to the current Reader.
@@ -854,6 +848,12 @@
*/
public abstract void setNextReader(IndexReader reader, int docBase, int numSlotsFull) throws IOException;
+ /** Sets the Scorer to use in case a document's score is needed. */
+ public void setScorer(Scorer scorer) {
+ // Empty implementation since most comparators don't need the score. This
+ // can be overridden by those that need it.
+ }
+
/**
* @return SortField.TYPE
*/
Index: src/java/org/apache/lucene/search/SortField.java
===================================================================
--- src/java/org/apache/lucene/search/SortField.java (revision 761300)
+++ src/java/org/apache/lucene/search/SortField.java (working copy)
@@ -21,8 +21,6 @@
import java.io.Serializable;
import java.util.Locale;
-import org.apache.lucene.index.IndexReader;
-
/**
* Stores information about how to sort documents by terms in an individual
* field. Fields must be indexed in order to sort by them.
@@ -434,8 +432,6 @@
/** Returns the {@link FieldComparator} to use for sorting.
- * @param subReaders array of {@link IndexReader} search
- * will step through
* @param numHits number of top hits the queue will store
* @param sortPos position of this SortField within {@link
* Sort}. The comparator is primary if sortPos==0,
@@ -444,7 +440,7 @@
* @param reversed True if the SortField is reversed
* @return {@link FieldComparator} to use when sorting
*/
- protected FieldComparator getComparator(final IndexReader[] subReaders, final int numHits, final int sortPos, final boolean reversed) throws IOException {
+ protected FieldComparator getComparator(final int numHits, final int sortPos, final boolean reversed) throws IOException {
if (locale != null) {
// TODO: it'd be nice to allow FieldCache.getStringIndex
@@ -480,7 +476,7 @@
case SortField.CUSTOM:
assert factory == null && comparatorSource != null;
- return comparatorSource.newComparator(field, subReaders, numHits, sortPos, reversed);
+ return comparatorSource.newComparator(field, numHits, sortPos, reversed);
case SortField.STRING:
return new FieldComparator.StringOrdValComparator(numHits, field, sortPos, reversed);
Index: src/java/org/apache/lucene/search/DisjunctionSumScorer.java
===================================================================
--- src/java/org/apache/lucene/search/DisjunctionSumScorer.java (revision 761300)
+++ src/java/org/apache/lucene/search/DisjunctionSumScorer.java (working copy)
@@ -112,10 +112,20 @@
* @param hc The collector to which all matching documents are passed through
* {@link HitCollector#collect(int, float)}.
*
When this method is used the {@link #explain(int)} method should not be used.
+ * @deprecated use {@link #score(Collector)} instead.
*/
public void score(HitCollector hc) throws IOException {
+ score(new HitCollectorWrapper(hc));
+ }
+
+ /** Scores and collects all matching documents.
+ * @param collector The collector to which all matching documents are passed through.
+ *
When this method is used the {@link #explain(int)} method should not be used.
+ */
+ public void score(Collector collector) throws IOException {
+ collector.setScorer(this);
while (next()) {
- hc.collect(currentDoc, currentScore);
+ collector.collect(currentDoc);
}
}
@@ -126,10 +136,23 @@
* {@link HitCollector#collect(int, float)}.
* @param max Do not score documents past this.
* @return true if more matching documents may remain.
+ * @deprecated use {@link #score(Collector, int)} instead.
*/
protected boolean score(HitCollector hc, int max) throws IOException {
+ return score(new HitCollectorWrapper(hc), max);
+ }
+
+ /** Expert: Collects matching documents in a range. Hook for optimization.
+ * Note that {@link #next()} must be called once before this method is called
+ * for the first time.
+ * @param collector The collector to which all matching documents are passed through.
+ * @param max Do not score documents past this.
+ * @return true if more matching documents may remain.
+ */
+ protected boolean score(Collector collector, int max) throws IOException {
+ collector.setScorer(this);
while (currentDoc < max) {
- hc.collect(currentDoc, currentScore);
+ collector.collect(currentDoc);
if (!next()) {
return false;
}
Index: src/java/org/apache/lucene/search/HitCollectorWrapper.java
===================================================================
--- src/java/org/apache/lucene/search/HitCollectorWrapper.java (revision 0)
+++ src/java/org/apache/lucene/search/HitCollectorWrapper.java (revision 0)
@@ -0,0 +1,49 @@
+package org.apache.lucene.search;
+
+/**
+ * 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.io.IOException;
+
+import org.apache.lucene.index.IndexReader;
+
+/**
+ * Wrapper for non expert ({@link HitCollector}) implementations, which simply
+ * re-bases the incoming docID before calling {@link HitCollector#collect}.
+ * @deprecated this class should be removed when {@link HitCollector} is removed.
+ */
+class HitCollectorWrapper extends Collector {
+ private HitCollector collector;
+ private int base = 0;
+ private Scorer scorer = null;
+
+ public HitCollectorWrapper(HitCollector collector) {
+ this.collector = collector;
+ }
+
+ public void setNextReader(IndexReader reader, int docBase) {
+ base = docBase;
+ }
+
+ public void collect(int doc) throws IOException {
+ collector.collect(doc + base, scorer.score());
+ }
+
+ public void setScorer(Scorer scorer) throws IOException {
+ this.scorer = scorer;
+ }
+}
Index: src/java/org/apache/lucene/search/BooleanScorer.java
===================================================================
--- src/java/org/apache/lucene/search/BooleanScorer.java (revision 761300)
+++ src/java/org/apache/lucene/search/BooleanScorer.java (working copy)
@@ -80,11 +80,11 @@
public boolean done;
public boolean required = false;
public boolean prohibited = false;
- public MultiReaderHitCollector collector;
+ public Collector collector;
public SubScorer next;
public SubScorer(Scorer scorer, boolean required, boolean prohibited,
- MultiReaderHitCollector collector, SubScorer next)
+ Collector collector, SubScorer next)
throws IOException {
this.scorer = scorer;
this.done = !scorer.next();
@@ -128,18 +128,32 @@
private int end;
private Bucket current;
+ /** @deprecated use {@link #score(Collector)} instead. */
public void score(HitCollector hc) throws IOException {
next();
score(hc, Integer.MAX_VALUE);
}
+
+ public void score(Collector collector) throws IOException {
+ next();
+ score(collector, Integer.MAX_VALUE);
+ }
+ /** @deprecated use {@link #score(Collector, int)} instead. */
protected boolean score(HitCollector hc, int max) throws IOException {
+ return score(new HitCollectorWrapper(hc), max);
+ }
+
+ protected boolean score(Collector collector, int max) throws IOException {
if (coordFactors == null)
computeCoordFactors();
boolean more;
Bucket tmp;
+ BucketScorer bs = new BucketScorer();
+ // The internal loop will set the score and doc before calling collect.
+ collector.setScorer(bs);
do {
bucketTable.first = null;
@@ -158,7 +172,9 @@
}
if (current.coord >= minNrShouldMatch) {
- hc.collect(current.doc, current.score * coordFactors[current.coord]);
+ bs.score = current.score * coordFactors[current.coord];
+ bs.doc = current.doc;
+ collector.collect(current.doc);
}
}
@@ -210,8 +226,9 @@
end += BucketTable.SIZE;
for (SubScorer sub = scorers; sub != null; sub = sub.next) {
Scorer scorer = sub.scorer;
+ sub.collector.setScorer(scorer);
while (!sub.done && scorer.doc() < end) {
- sub.collector.collect(scorer.doc(), scorer.score());
+ sub.collector.collect(scorer.doc());
sub.done = !scorer.next();
}
if (!sub.done) {
@@ -237,6 +254,42 @@
Bucket next; // next valid bucket
}
+ // An internal class which is used in score(Collector, int) for setting the
+ // current score. This is required since Collector exposes a setScorer method
+ // and implementations that need the score will call scorer.score().
+ // Therefore the only methods that are implemented are score() and doc().
+ private static final class BucketScorer extends Scorer {
+
+ float score;
+ int doc;
+
+ public BucketScorer() {
+ super(null);
+ }
+
+
+ public Explanation explain(int doc) throws IOException {
+ return null;
+ }
+
+ public float score() throws IOException {
+ return score;
+ }
+
+ public int doc() {
+ return doc;
+ }
+
+ public boolean next() throws IOException {
+ return false;
+ }
+
+ public boolean skipTo(int target) throws IOException {
+ return false;
+ }
+
+ }
+
/** A simple hash table of document scores within a range. */
static final class BucketTable {
public static final int SIZE = 1 << 11;
@@ -249,19 +302,25 @@
public final int size() { return SIZE; }
- public MultiReaderHitCollector newCollector(int mask) {
- return new Collector(mask, this);
+ public Collector newCollector(int mask) {
+ return new BolleanScorerCollector(mask, this);
}
}
- static final class Collector extends MultiReaderHitCollector {
+ private static final class BolleanScorerCollector extends Collector {
private BucketTable bucketTable;
private int mask;
- public Collector(int mask, BucketTable bucketTable) {
+ private Scorer scorer;
+
+ public BolleanScorerCollector(int mask, BucketTable bucketTable) {
this.mask = mask;
this.bucketTable = bucketTable;
}
- public final void collect(final int doc, final float score) {
+ public void setScorer(Scorer scorer) throws IOException {
+ this.scorer = scorer;
+ }
+
+ public final void collect(final int doc) throws IOException {
final BucketTable table = bucketTable;
final int i = doc & BucketTable.MASK;
Bucket bucket = table.buckets[i];
@@ -270,14 +329,14 @@
if (bucket.doc != doc) { // invalid bucket
bucket.doc = doc; // set doc
- bucket.score = score; // initialize score
+ bucket.score = scorer.score(); // initialize score
bucket.bits = mask; // initialize mask
bucket.coord = 1; // initialize coord
bucket.next = table.first; // push onto valid list
table.first = bucket;
} else { // valid bucket
- bucket.score += score; // increment score
+ bucket.score += scorer.score(); // increment score
bucket.bits |= mask; // add bits in mask
bucket.coord++; // increment coord
}
Index: src/java/org/apache/lucene/search/Scorer.java
===================================================================
--- src/java/org/apache/lucene/search/Scorer.java (revision 761300)
+++ src/java/org/apache/lucene/search/Scorer.java (working copy)
@@ -52,10 +52,20 @@
* @param hc The collector to which all matching documents are passed through
* {@link HitCollector#collect(int, float)}.
*
When this method is used the {@link #explain(int)} method should not be used.
+ * @deprecated use {@link #score(Collector)} instead.
*/
public void score(HitCollector hc) throws IOException {
+ score(new HitCollectorWrapper(hc));
+ }
+
+ /** Scores and collects all matching documents.
+ * @param collector The collector to which all matching documents are passed.
+ *
When this method is used the {@link #explain(int)} method should not be used.
+ */
+ public void score(Collector collector) throws IOException {
+ collector.setScorer(this);
while (next()) {
- hc.collect(doc(), score());
+ collector.collect(doc());
}
}
@@ -66,10 +76,23 @@
* {@link HitCollector#collect(int, float)}.
* @param max Do not score documents past this.
* @return true if more matching documents may remain.
+ * @deprecated use {@link #score(Collector, int)} instead.
*/
protected boolean score(HitCollector hc, int max) throws IOException {
+ return score(new HitCollectorWrapper(hc), max);
+ }
+
+ /** Expert: Collects matching documents in a range. Hook for optimization.
+ * Note that {@link #next()} must be called once before this method is called
+ * for the first time.
+ * @param collector The collector to which all matching documents are passed.
+ * @param max Do not score documents past this.
+ * @return true if more matching documents may remain.
+ */
+ protected boolean score(Collector collector, int max) throws IOException {
+ collector.setScorer(this);
while (doc() < max) {
- hc.collect(doc(), score());
+ collector.collect(doc());
if (!next())
return false;
}
Index: src/java/org/apache/lucene/search/MultiSearcher.java
===================================================================
--- src/java/org/apache/lucene/search/MultiSearcher.java (revision 761300)
+++ src/java/org/apache/lucene/search/MultiSearcher.java (working copy)
@@ -97,9 +97,14 @@
throw new UnsupportedOperationException();
}
+ /** @deprecated use {@link #search(Weight, Filter, Collector)} instead. */
public void search(Weight weight, Filter filter, HitCollector results) {
throw new UnsupportedOperationException();
}
+
+ public void search(Weight weight, Filter filter, Collector collector) {
+ throw new UnsupportedOperationException();
+ }
public TopDocs search(Weight weight,Filter filter,int n) {
throw new UnsupportedOperationException();
@@ -251,41 +256,32 @@
return new TopFieldDocs (totalHits, scoreDocs, hq.getFields(), maxScore);
}
-
// inherit javadoc
+ /** @deprecated use {@link #search(Weight, Filter, Collector)} instead. */
public void search(Weight weight, Filter filter, final HitCollector results)
throws IOException {
+ search(weight, filter, new HitCollectorWrapper(results));
+ }
+
+ // inherit javadoc
+ public void search(Weight weight, Filter filter, final Collector collector)
+ throws IOException {
for (int i = 0; i < searchables.length; i++) {
-
+
final int start = starts[i];
-
- final MultiReaderHitCollector hc;
- if (results instanceof MultiReaderHitCollector) {
- // results can shift
- final MultiReaderHitCollector resultsMulti = (MultiReaderHitCollector) results;
- hc = new MultiReaderHitCollector() {
- public void collect(int doc, float score) {
- resultsMulti.collect(doc, score);
- }
-
- public void setNextReader(IndexReader reader, int docBase) throws IOException {
- resultsMulti.setNextReader(reader, start+docBase);
- }
- };
- } else {
- // We must shift the docIDs
- hc = new MultiReaderHitCollector() {
- private int docBase;
- public void collect(int doc, float score) {
- results.collect(doc + docBase + start, score);
- }
-
- public void setNextReader(IndexReader reader, int docBase) {
- this.docBase = docBase;
- }
- };
- }
+ final Collector hc = new Collector() {
+ public void setScorer(Scorer scorer) throws IOException {
+ collector.setScorer(scorer);
+ }
+ public void collect(int doc) throws IOException {
+ collector.collect(doc);
+ }
+ public void setNextReader(IndexReader reader, int docBase) throws IOException {
+ collector.setNextReader(reader, start + docBase);
+ }
+ };
+
searchables[i].search(weight, filter, hc);
}
}
Index: src/java/org/apache/lucene/search/QueryWrapperFilter.java
===================================================================
--- src/java/org/apache/lucene/search/QueryWrapperFilter.java (revision 761300)
+++ src/java/org/apache/lucene/search/QueryWrapperFilter.java (working copy)
@@ -50,9 +50,12 @@
public BitSet bits(IndexReader reader) throws IOException {
final BitSet bits = new BitSet(reader.maxDoc());
- new IndexSearcher(reader).search(query, new MultiReaderHitCollector() {
- private int base = -1;
- public final void collect(int doc, float score) {
+ new IndexSearcher(reader).search(query, new Collector() {
+ private int base = 0;
+ public void setScorer(Scorer scorer) throws IOException {
+ // score is not needed by this collector
+ }
+ public final void collect(int doc) {
bits.set(doc + base); // set bit for hit
}
public void setNextReader(IndexReader reader, int docBase) {
Index: src/java/org/apache/lucene/search/TimeLimitedCollector.java
===================================================================
--- src/java/org/apache/lucene/search/TimeLimitedCollector.java (revision 761300)
+++ src/java/org/apache/lucene/search/TimeLimitedCollector.java (working copy)
@@ -1,7 +1,5 @@
package org.apache.lucene.search;
-import org.apache.lucene.index.IndexReader;
-
/**
* Licensed to the Apache Software Foundation (ASF) under one or more
* contributor license agreements. See the NOTICE file distributed with
@@ -19,16 +17,20 @@
* limitations under the License.
*/
-import java.io.IOException;
-
/**
- *
The TimeLimitedCollector is used to timeout search requests that - * take longer than the maximum allowed search time limit. After this - * time is exceeded, the search thread is stopped by throwing a - * TimeExceeded Exception.
+ *+ * The TimeLimitedCollector is used to timeout search requests that take longer + * than the maximum allowed search time limit. After this time is exceeded, the + * search thread is stopped by throwing a TimeExceeded Exception. + *
* + * NOTE: For backwards compatibility, this class still extends + * {@link HitCollector}. In 3.0 this class will be changed to extend the new + * {@link Collector} class and its {@link HitCollector} related methods will be + * removed. Meanwhile, you can wrap it with {@link HitCollectorWrapper} in + * order to use the {@link Collector}-based search methods. */ -public class TimeLimitedCollector extends MultiReaderHitCollector { +public class TimeLimitedCollector extends HitCollector { /** * Default timer resolution. @@ -72,7 +74,7 @@ } public void run() { - while( true ) { + while (true) { // TODO: Use System.nanoTime() when Lucene moves to Java SE 5. time += resolution; try { @@ -92,9 +94,7 @@ } } - /** - * Thrown when elapsed search time exceeds allowed search time. - */ + /** Thrown when elapsed search time exceeds allowed search time. */ public static class TimeExceededException extends RuntimeException { private long timeAllowed; private long timeElapsed; @@ -105,21 +105,15 @@ this.timeElapsed = timeElapsed; this.lastDocCollected = lastDocCollected; } - /** - * Returns allowed time (milliseconds). - */ + /** Returns allowed time (milliseconds). */ public long getTimeAllowed() { return timeAllowed; } - /** - * Returns elapsed time (milliseconds). - */ + /** Returns elapsed time (milliseconds). */ public long getTimeElapsed() { return timeElapsed; } - /** - * Returns last doc that was collected when the search time exceeded. - */ + /** Returns last doc that was collected when the search time exceeded. */ public int getLastDocCollected() { return lastDocCollected; } @@ -136,40 +130,56 @@ private final long t0; private final long timeout; - private final MultiReaderHitCollector hc; + + private final HitCollector collector; + + // TODO: reenstate when we move to 3.0 and rebase this class on Collector. +// private final Collector collector; /** - * Create a TimeLimitedCollector wrapper over another HitCollector with a specified timeout. - * @param hc the wrapped HitCollector + * Create a TimeLimitedCollector wrapper over another {@link HitCollector} with a specified timeout. + * @param collector the wrapped {@link HitCollector} * @param timeAllowed max time allowed for collecting hits after which {@link TimeExceededException} is thrown + * @deprecated use {@link #TimeLimitedCollector(Collector, long)} instead. */ - public TimeLimitedCollector( final HitCollector hc, final long timeAllowed ) { - if (hc instanceof MultiReaderHitCollector) { - this.hc = (MultiReaderHitCollector) hc; - } else { - this.hc = new IndexSearcher.MultiReaderCollectorWrapper(hc); - } + public TimeLimitedCollector(final HitCollector collector, final long timeAllowed) { + this.collector = collector; t0 = TIMER_THREAD.getMilliseconds(); this.timeout = t0 + timeAllowed; } + + // TODO: reenstate when we move to 3.0 and rebase this class on Collector. +// /** +// * Create a TimeLimitedCollector wrapper over another {@link Collector} with a specified timeout. +// * @param collector the wrapped {@link Collector} +// * @param timeAllowed max time allowed for collecting hits after which {@link TimeExceededException} is thrown +// */ +// public TimeLimitedCollector(final Collector collector, final long timeAllowed ) { +// this.collector = collector; +// t0 = TIMER_THREAD.getMilliseconds(); +// this.timeout = t0 + timeAllowed; +// } /** - * Calls collect() on the decorated HitCollector. + * Calls {@link HitCollector#collect(int, float)} on the decorated {@link HitCollector} + * unless the allowed time has passed, in which case it throws an exception. * - * @throws TimeExceededException if the time allowed has been exceeded. + * @throws TimeExceededException + * if the time allowed has exceeded. + * @deprecated this will be removed in 3.0 */ - public void collect( final int doc, final float score ) { + public void collect(final int doc, final float score) { long time = TIMER_THREAD.getMilliseconds(); - if( timeout < time) { + if (timeout < time) { if (greedy) { //System.out.println(this+" greedy: before failing, collecting doc: "+doc+" "+(time-t0)); - hc.collect( doc, score ); + collector.collect(doc, score); } //System.out.println(this+" failing on: "+doc+" "+(time-t0)); throw new TimeExceededException( timeout-t0, time-t0, doc ); } //System.out.println(this+" collecting: "+doc+" "+(time-t0)); - hc.collect( doc, score ); + collector.collect(doc, score); } /** @@ -219,7 +229,34 @@ this.greedy = greedy; } - public void setNextReader(IndexReader reader, int base) throws IOException { - hc.setNextReader(reader, base); - } + // TODO: reenstate when we move to 3.0 and rebase this class on Collector. +// /** +// * Calls {@link Collector#collect(int)} on the decorated {@link Collector} +// * unless the allowed time has passed, in which case it throws an exception. +// * +// * @throws TimeExceededException +// * if the time allowed has exceeded. +// */ +// public void collect(final int doc) throws IOException { +// long time = TIMER_THREAD.getMilliseconds(); +// if (timeout < time) { +// if (greedy) { +// //System.out.println(this+" greedy: before failing, collecting doc: "+doc+" "+(time-t0)); +// collector.collect(doc); +// } +// //System.out.println(this+" failing on: "+doc+" "+(time-t0)); +// throw new TimeExceededException( timeout-t0, time-t0, doc ); +// } +// //System.out.println(this+" collecting: "+doc+" "+(time-t0)); +// collector.collect(doc); +// } +// +// public void setNextReader(IndexReader reader, int base) throws IOException { +// collector.setNextReader(reader, base); +// } +// +// public void setScorer(Scorer scorer) throws IOException { +// collector.setScorer(scorer); +// } + } Index: src/java/org/apache/lucene/search/TopFieldCollector.java =================================================================== --- src/java/org/apache/lucene/search/TopFieldCollector.java (revision 761300) +++ src/java/org/apache/lucene/search/TopFieldCollector.java (working copy) @@ -21,201 +21,319 @@ import org.apache.lucene.index.IndexReader; import org.apache.lucene.search.FieldValueHitQueue.Entry; +import org.apache.lucene.util.PriorityQueue; /** - * A {@link HitCollector} that sorts by {@link SortField} using + * A {@link Collector} that sorts by {@link SortField} using * {@link FieldComparator}s. - * + * * NOTE: This API is experimental and might change in * incompatible ways in the next release. */ -public final class TopFieldCollector extends MultiReaderHitCollector { +public abstract class TopFieldCollector extends TopDocsCollector { + // TODO: one optimization we could do is to pre-fill + // the queue with sentinel value that guaranteed to + // always compare lower than a real hit; this would + // save having to check queueFull on each insert - private final FieldValueHitQueue queue; + /* + * Implements a TopFieldCollector over one SortField criteria, without + * tracking document scores. + */ + private static class OneComparatorNonScoringCollector extends TopFieldCollector { - private final FieldComparator[] comparators; - private FieldComparator comparator1; - private final int numComparators; - private int[] reverseMul; - private int reverseMul1 = 0; - - private final int numHits; - private int totalHits; - private FieldValueHitQueue.Entry bottom = null; - - /** Stores the maximum score value encountered, needed for normalizing. */ - private float maxScore = Float.NEGATIVE_INFINITY; - - private boolean queueFull; - - private boolean fillFields; - - public TopFieldCollector(Sort sort, int numHits, IndexReader[] subReaders, boolean fillFields) - throws IOException { - - if (sort.fields.length == 0) { - throw new IllegalArgumentException("Sort must contain at least one field"); + final private FieldComparator comparator; + final private int reverseMul; + + public OneComparatorNonScoringCollector(FieldValueHitQueue queue, + int numHits, boolean fillFields) throws IOException { + super(queue, numHits, fillFields); + comparator = queue.getComparators()[0]; + reverseMul = queue.getReverseMul()[0]; } - - queue = new FieldValueHitQueue(sort.fields, numHits, subReaders); - comparators = queue.getComparators(); - reverseMul = queue.getReverseMul(); - numComparators = comparators.length; - - if (numComparators == 1) { - comparator1 = comparators[0]; - reverseMul1 = reverseMul[0]; - } else { - comparator1 = null; - reverseMul1 = 0; + + public void collect(int doc) throws IOException { + ++totalHits; + if (queueFull) { + // Fastmatch: return if this hit is not competitive + final int cmp = reverseMul * comparator.compareBottom(doc); + if (cmp < 0 || (cmp == 0 && doc + docBase > bottom.docID)) { + return; + } + + // This hit is competitive - replace bottom element in queue & adjustTop + comparator.copy(bottom.slot, doc); + updateBottom(doc, docScore); + comparator.setBottom(bottom.slot); + } else { + // Startup transient: queue hasn't gathered numHits yet + final int slot = totalHits - 1; + // Copy hit into queue + comparator.copy(slot, doc); + add(slot, doc, docScore); + if (queueFull) { + comparator.setBottom(bottom.slot); + } + } } - this.numHits = numHits; - this.fillFields = fillFields; + + public void setNextReader(IndexReader reader, int docBase) throws IOException { + final int numSlotsFull = queueFull ? pq.size() : totalHits; + this.docBase = docBase; + comparator.setNextReader(reader, docBase, numSlotsFull); + } + + public void setScorer(Scorer scorer) throws IOException { + comparator.setScorer(scorer); + } + } - int currentDocBase; + /* + * Implements a TopFieldCollector over one SortField criteria, with tracking + * document scores. + */ + private static class OneComparatorScoringCollector extends OneComparatorNonScoringCollector { - // javadoc inherited - public void setNextReader(IndexReader reader, int docBase) throws IOException { - final int numSlotsFull; - if (queueFull) - numSlotsFull = numHits; - else - numSlotsFull = totalHits; - - currentDocBase = docBase; - - for (int i = 0; i < numComparators; i++) { - comparators[i].setNextReader(reader, docBase, numSlotsFull); + private Scorer scorer; + + public OneComparatorScoringCollector(FieldValueHitQueue queue, + int numHits, boolean fillFields) throws IOException { + super(queue, numHits, fillFields); } + + public void collect(int doc) throws IOException { + docScore = scorer.score(); + maxScore = Math.max(maxScore, docScore); + super.collect(doc); + } + + public void setScorer(Scorer scorer) throws IOException { + this.scorer = scorer; + super.setScorer(scorer); + } } - private final void updateBottom(int doc, float score) { - bottom.docID = currentDocBase + doc; - bottom.score = score; - queue.adjustTop(); - bottom = (FieldValueHitQueue.Entry) queue.top(); - } - - private final void add(int slot, int doc, float score) { - queue.put(new FieldValueHitQueue.Entry(slot, currentDocBase+doc, score)); - bottom = (FieldValueHitQueue.Entry) queue.top(); - queueFull = totalHits == numHits; - } - - // javadoc inherited - public void collect(int doc, float score) { - if (score > 0.0f) { - - maxScore = Math.max(maxScore, score); - totalHits++; - - // TODO: one optimization we could do is to pre-fill - // the queue with sentinel value that guaranteed to - // always compare lower than a real hit; this would - // save having to check queueFull on each insert - + /* + * Implements a TopFieldCollector over multiple SortField criteria, without + * tracking document scores. + */ + private static class MultiComparatorNonScoringCollector extends TopFieldCollector { + + final private FieldComparator[] comparators; + final private int[] reverseMul; + + public MultiComparatorNonScoringCollector(FieldValueHitQueue queue, + int numHits, boolean fillFields) throws IOException { + super(queue, numHits, fillFields); + comparators = queue.getComparators(); + reverseMul = queue.getReverseMul(); + } + + public void collect(int doc) throws IOException { + ++totalHits; if (queueFull) { - - if (numComparators == 1) { - // Common case - - // Fastmatch: return if this hit is not competitive - final int cmp = reverseMul1 * comparator1.compareBottom(doc, score); - if (cmp < 0) { + // Fastmatch: return if this hit is not competitive + for (int i = 0;; i++) { + final int c = reverseMul[i] * comparators[i].compareBottom(doc); + if (c < 0) { // Definitely not competitive return; - } else if (cmp == 0 && doc + currentDocBase > bottom.docID) { - // Definitely not competitive - return; - } - - // This hit is competitive -- replace bottom - // element in queue & adjustTop - comparator1.copy(bottom.slot, doc, score); - - updateBottom(doc, score); - - comparator1.setBottom(bottom.slot); - - } else { - - // Fastmatch: return if this hit is not competitive - for(int i=0;;i++) { - final int c = reverseMul[i] * comparators[i].compareBottom(doc, score); - if (c < 0) { + } else if (c > 0) { + // Definitely competitive + break; + } else if (i == comparators.length - 1) { + // This is the equals case. + if (doc + docBase > bottom.docID) { // Definitely not competitive return; - } else if (c > 0) { - // Definitely competitive - break; - } else if (i == numComparators-1) { - // This is the equals case. - if (doc + currentDocBase > bottom.docID) { - // Definitely not competitive - return; - } else { - break; - } } + break; } + } - // This hit is competitive -- replace bottom - // element in queue & adjustTop - for (int i = 0; i < numComparators; i++) { - comparators[i].copy(bottom.slot, doc, score); - } + // This hit is competitive - replace bottom element in queue & adjustTop + for (int i = 0; i < comparators.length; i++) { + comparators[i].copy(bottom.slot, doc); + } - updateBottom(doc, score); + updateBottom(doc, docScore); - for(int i=0;iresults is null it means there are no results to return,
+ * either because there were 0 calls to collect() or because the arguments to
+ * topDocs were invalid.
+ */
+ protected TopDocs newTopDocs(ScoreDoc[] results, int start) {
+ return results == null ? EMPTY_TOPDOCS : new TopDocs(totalHits, results);
+ }
+
+ /** The total number of documents that matched this query. */
+ public int getTotalHits() {
+ return totalHits;
+ }
+
+ /** Returns the top docs that were collected by this collector. */
+ public final TopDocs topDocs() {
+ return topDocs(0, pq.size());
+ }
+
+ /**
+ * Returns the documents in the rage [start .. pq.size()) that were collected
+ * by this collector. Note that if start >= pq.size(), an empty TopDocs is
+ * returned.start, you should call {@link #topDocs()} and work
+ * with the returned {@link TopDocs} object, which will contain all the
+ * results this search execution collected.
+ */
+ public final TopDocs topDocs(int start) {
+ return topDocs(start, pq.size());
+ }
+
+ /**
+ * Returns the documents in the rage [start .. start+howMany) that were
+ * collected by this collector. Note that if start >= pq.size(), an empty
+ * TopDocs is returned, and if pq.size() - start < howMany, then only the
+ * available documents in [start .. pq.size()) are returned.{@link Collector#collect(int)} is called for every matching document. + * + *
Applications should only use this if they need all of the
+ * matching documents. The high-level search API ({@link
+ * Searcher#search(Query)}) is usually more efficient, as it skips
+ * non-high-scoring hits.
+ *
+ * @param weight to match documents
+ * @param filter if non-null, a bitset used to eliminate some documents
+ * @param collector to receive hits
+ *
+ * @todo parallelize this one too
+ */
+ public void search(Weight weight, Filter filter, final Collector collector)
+ throws IOException {
+ for (int i = 0; i < searchables.length; i++) {
- final MultiReaderHitCollector hc;
- if (results instanceof MultiReaderHitCollector) {
- // results can shift
- final MultiReaderHitCollector resultsMulti = (MultiReaderHitCollector) results;
- hc = new MultiReaderHitCollector() {
- public void collect(int doc, float score) {
- resultsMulti.collect(doc, score);
- }
+ final int start = starts[i];
- public void setNextReader(IndexReader reader, int docBase) throws IOException {
- resultsMulti.setNextReader(reader, start+docBase);
- }
- };
- } else {
- // We must shift the docIDs
- hc = new MultiReaderHitCollector() {
- private int docBase;
- public void collect(int doc, float score) {
- results.collect(doc + docBase + start, score);
- }
+ final Collector hc = new Collector() {
+ public void setScorer(Scorer scorer) throws IOException {
+ collector.setScorer(scorer);
+ }
+ public void collect(int doc) throws IOException {
+ collector.collect(doc);
+ }
+
+ public void setNextReader(IndexReader reader, int docBase) throws IOException {
+ collector.setNextReader(reader, start + docBase);
+ }
+ };
+
+ searchables[i].search(weight, filter, hc);
+ }
+ }
- public void setNextReader(IndexReader reader, int docBase) {
- this.docBase = docBase;
- }
- };
- }
-
- searchables[i].search(weight, filter, hc);
- }
- }
-
/*
* TODO: this one could be parallelized too
* @see org.apache.lucene.search.Searchable#rewrite(org.apache.lucene.search.Query)
Index: src/java/org/apache/lucene/search/IndexSearcher.java
===================================================================
--- src/java/org/apache/lucene/search/IndexSearcher.java (revision 761300)
+++ src/java/org/apache/lucene/search/IndexSearcher.java (working copy)
@@ -188,12 +188,16 @@
throws IOException {
return search(weight, filter, nDocs, sort, true);
}
-
- /**
- * Just like {@link #search(Weight, Filter, int, Sort)},
- * but you choose whether or not the fields in the
- * returned {@link FieldDoc} instances should be set by
- * specifying fillFields.
+
+ /**
+ * Just like {@link #search(Weight, Filter, int, Sort)}, but you choose
+ * whether or not the fields in the returned {@link FieldDoc} instances should
+ * be set by specifying fillFields.
+ * NOTE: currently, this method tracks document scores and sets them in
+ * the returned {@link FieldDoc}, however in 3.0 it will move to not track
+ * document scores. If document scores tracking is still needed, you can use
+ * {@link #search(Weight, Filter, Collector)} and pass in a
+ * {@link TopFieldCollector} instance.
*/
public TopFieldDocs search(Weight weight, Filter filter, final int nDocs,
Sort sort, boolean fillFields)
@@ -222,29 +226,32 @@
if (legacy) {
// Search the single top-level reader
- TopScoreDocCollector collector = new TopFieldDocCollector(reader, sort, nDocs);
- collector.setNextReader(reader, 0);
- doSearch(reader, weight, filter, collector);
+ TopDocCollector collector = new TopFieldDocCollector(reader, sort, nDocs);
+ HitCollectorWrapper hcw = new HitCollectorWrapper(collector);
+ hcw.setNextReader(reader, 0);
+ doSearch(reader, weight, filter, hcw);
return (TopFieldDocs) collector.topDocs();
- } else {
- // Search each sub-reader
- TopFieldCollector collector = new TopFieldCollector(sort, nDocs, sortedSubReaders, fillFields);
- search(weight, filter, collector);
- return (TopFieldDocs) collector.topDocs();
}
+ // Search each sub-reader
+ // TODO: by default we should create a TopFieldCollector which does not
+ // track document scores. Currently the default is set to true, however it
+ // will change in 3.0.
+ TopFieldCollector collector = TopFieldCollector.create(sort, nDocs, fillFields, true);
+ search(weight, filter, collector);
+ return (TopFieldDocs) collector.topDocs();
}
// inherit javadoc
+ /** @deprecated use {@link #search(Weight, Filter, Collector)} instead. */
public void search(Weight weight, Filter filter, HitCollector results)
throws IOException {
-
- final MultiReaderHitCollector collector;
- if (results instanceof MultiReaderHitCollector) {
- collector = (MultiReaderHitCollector) results;
- } else {
- collector = new MultiReaderCollectorWrapper(results);
- }
-
+ search(weight, filter, new HitCollectorWrapper(results));
+ }
+
+ // inherit javadoc
+ public void search(Weight weight, Filter filter, Collector collector)
+ throws IOException {
+
for (int i = 0; i < sortedSubReaders.length; i++) { // search each subreader
collector.setNextReader(sortedSubReaders[i], sortedStarts[i]);
doSearch(sortedSubReaders[i], weight, filter, collector);
@@ -252,14 +259,14 @@
}
private void doSearch(IndexReader reader, Weight weight, Filter filter,
- final HitCollector results) throws IOException {
+ final Collector collector) throws IOException {
Scorer scorer = weight.scorer(reader);
if (scorer == null)
return;
if (filter == null) {
- scorer.score(results);
+ scorer.score(collector);
return;
}
@@ -267,6 +274,7 @@
boolean more = filterDocIdIterator.next() && scorer.skipTo(filterDocIdIterator.doc());
+ collector.setScorer(scorer);
while (more) {
int filterDocId = filterDocIdIterator.doc();
if (filterDocId > scorer.doc() && !scorer.skipTo(filterDocId)) {
@@ -274,7 +282,7 @@
} else {
int scorerDocId = scorer.doc();
if (scorerDocId == filterDocId) { // permitted by filter
- results.collect(scorerDocId, scorer.score());
+ collector.collect(scorerDocId);
more = filterDocIdIterator.next();
} else {
more = filterDocIdIterator.skipTo(scorerDocId);
@@ -295,26 +303,4 @@
public Explanation explain(Weight weight, int doc) throws IOException {
return weight.explain(reader, doc);
}
-
- /**
- * Wrapper for non expert ({@link HitCollector})
- * implementations, which simply re-bases the incoming
- * docID before calling {@link HitCollector#collect}.
- */
- static class MultiReaderCollectorWrapper extends MultiReaderHitCollector {
- private HitCollector collector;
- private int base = -1;
-
- public MultiReaderCollectorWrapper(HitCollector collector) {
- this.collector = collector;
- }
-
- public void collect(int doc, float score) {
- collector.collect(doc + base, score);
- }
-
- public void setNextReader(IndexReader reader, int docBase) {
- base = docBase;
- }
- }
}
Index: src/java/org/apache/lucene/search/TermScorer.java
===================================================================
--- src/java/org/apache/lucene/search/TermScorer.java (revision 761300)
+++ src/java/org/apache/lucene/search/TermScorer.java (working copy)
@@ -24,6 +24,23 @@
/** Expert: A Scorer for documents matching a Term.
*/
final class TermScorer extends Scorer {
+
+ // Used internally in score(Collector, int) for setting the score of the current document.
+ private static final class InternalScorer extends Scorer {
+ private int d;
+ private float score;
+
+ public InternalScorer() {
+ super(null);
+ }
+
+ public Explanation explain(int doc) throws IOException { return null; }
+ public float score() throws IOException { return score; }
+ public int doc() { return d; }
+ public boolean next() throws IOException { return false; }
+ public boolean skipTo(int target) throws IOException { return false; }
+ }
+
private Weight weight;
private TermDocs termDocs;
private byte[] norms;
@@ -56,36 +73,52 @@
scoreCache[i] = getSimilarity().tf(i) * weightValue;
}
+ /** @deprecated use {@link #score(Collector)} instead. */
public void score(HitCollector hc) throws IOException {
+ score(new HitCollectorWrapper(hc));
+ }
+
+ public void score(Collector c) throws IOException {
next();
- score(hc, Integer.MAX_VALUE);
+ score(c, Integer.MAX_VALUE);
}
+ /** @deprecated use {@link #score(Collector, int)} instead. */
protected boolean score(HitCollector c, int end) throws IOException {
+ return score(new HitCollectorWrapper(c), end);
+ }
+
+ protected boolean score(Collector c, int end) throws IOException {
Similarity similarity = getSimilarity(); // cache sim in local
float[] normDecoder = Similarity.getNormDecoder();
+ InternalScorer s = new InternalScorer();
+ c.setScorer(s);
while (doc < end) { // for docs in window
int f = freqs[pointer];
float score = // compute tf(f)*weight
f < SCORE_CACHE_SIZE // check cache
- ? scoreCache[f] // cache hit
- : similarity.tf(f)*weightValue; // cache miss
-
- score *= normDecoder[norms[doc] & 0xFF]; // normalize for field
-
- c.collect(doc, score); // collect score
-
- if (++pointer >= pointerMax) {
- pointerMax = termDocs.read(docs, freqs); // refill buffers
- if (pointerMax != 0) {
- pointer = 0;
- } else {
- termDocs.close(); // close stream
- doc = Integer.MAX_VALUE; // set to sentinel value
- return false;
- }
- }
- doc = docs[pointer];
+ ? scoreCache[f] // cache hit
+ : similarity.tf(f)*weightValue; // cache miss
+
+ score *= normDecoder[norms[doc] & 0xFF]; // normalize for field
+
+ // Set the Scorer doc and score before calling collect in case it will be
+ // used in collect()
+ s.d = doc;
+ s.score = score;
+ c.collect(doc); // collect score
+
+ if (++pointer >= pointerMax) {
+ pointerMax = termDocs.read(docs, freqs); // refill buffers
+ if (pointerMax != 0) {
+ pointer = 0;
+ } else {
+ termDocs.close(); // close stream
+ doc = Integer.MAX_VALUE; // set to sentinel value
+ return false;
+ }
+ }
+ doc = docs[pointer];
}
return true;
}
Index: src/java/org/apache/lucene/search/FieldComparatorSource.java
===================================================================
--- src/java/org/apache/lucene/search/FieldComparatorSource.java (revision 761300)
+++ src/java/org/apache/lucene/search/FieldComparatorSource.java (working copy)
@@ -18,7 +18,6 @@
*/
import java.io.IOException;
-import org.apache.lucene.index.IndexReader;
/**
* Provides a {@link FieldComparator} for custom field sorting.
@@ -38,6 +37,6 @@
* @throws IOException
* If an error occurs reading the index.
*/
- public abstract FieldComparator newComparator(String fieldname, IndexReader[] subReaders, int numHits, int sortPos, boolean reversed)
+ public abstract FieldComparator newComparator(String fieldname, int numHits, int sortPos, boolean reversed)
throws IOException;
}
Index: contrib/miscellaneous/src/test/org/apache/lucene/index/TestFieldNormModifier.java
===================================================================
--- contrib/miscellaneous/src/test/org/apache/lucene/index/TestFieldNormModifier.java (revision 761300)
+++ contrib/miscellaneous/src/test/org/apache/lucene/index/TestFieldNormModifier.java (working copy)
@@ -22,16 +22,18 @@
import junit.framework.TestCase;
+import org.apache.lucene.analysis.SimpleAnalyzer;
+import org.apache.lucene.document.Document;
+import org.apache.lucene.document.Field;
+import org.apache.lucene.index.IndexWriter.MaxFieldLength;
+import org.apache.lucene.search.Collector;
+import org.apache.lucene.search.DefaultSimilarity;
import org.apache.lucene.search.IndexSearcher;
-import org.apache.lucene.search.MultiReaderHitCollector;
+import org.apache.lucene.search.Scorer;
import org.apache.lucene.search.Similarity;
-import org.apache.lucene.search.DefaultSimilarity;
import org.apache.lucene.search.TermQuery;
+import org.apache.lucene.store.Directory;
import org.apache.lucene.store.RAMDirectory;
-import org.apache.lucene.store.Directory;
-import org.apache.lucene.analysis.SimpleAnalyzer;
-import org.apache.lucene.document.Document;
-import org.apache.lucene.document.Field;
/**
* Tests changing of field norms with a custom similarity and with fake norms.
@@ -52,12 +54,12 @@
/** inverts the normal notion of lengthNorm */
public static Similarity s = new DefaultSimilarity() {
public float lengthNorm(String fieldName, int numTokens) {
- return (float)numTokens;
+ return numTokens;
}
};
public void setUp() throws Exception {
- IndexWriter writer = new IndexWriter(store, new SimpleAnalyzer(), true);
+ IndexWriter writer = new IndexWriter(store, new SimpleAnalyzer(), true, MaxFieldLength.UNLIMITED);
for (int i = 0; i < NUM_DOCS; i++) {
Document d = new Document();
@@ -123,14 +125,19 @@
float lastScore = 0.0f;
// default similarity should put docs with shorter length first
- searcher.search(new TermQuery(new Term("field", "word")), new MultiReaderHitCollector() {
- private int docBase = -1;
- public final void collect(int doc, float score) {
- scores[doc + docBase] = score;
+ searcher.search(new TermQuery(new Term("field", "word")), new Collector() {
+ private int docBase = 0;
+ private Scorer scorer;
+
+ public final void collect(int doc) throws IOException {
+ scores[doc + docBase] = scorer.score();
}
public void setNextReader(IndexReader reader, int docBase) {
this.docBase = docBase;
}
+ public void setScorer(Scorer scorer) throws IOException {
+ this.scorer = scorer;
+ }
});
searcher.close();
@@ -147,14 +154,18 @@
// new norm (with default similarity) should put longer docs first
searcher = new IndexSearcher(store);
- searcher.search(new TermQuery(new Term("field", "word")), new MultiReaderHitCollector() {
- private int docBase = -1;
- public final void collect(int doc, float score) {
- scores[doc + docBase] = score;
+ searcher.search(new TermQuery(new Term("field", "word")), new Collector() {
+ private int docBase = 0;
+ private Scorer scorer;
+ public final void collect(int doc) throws IOException {
+ scores[doc + docBase] = scorer.score();
}
public void setNextReader(IndexReader reader, int docBase) {
this.docBase = docBase;
}
+ public void setScorer(Scorer scorer) throws IOException {
+ this.scorer = scorer;
+ }
});
searcher.close();
@@ -188,15 +199,18 @@
float lastScore = 0.0f;
// default similarity should return the same score for all documents for this query
- searcher.search(new TermQuery(new Term("untokfield", "20061212")), new MultiReaderHitCollector() {
- private int docBase = -1;
- private int lastMax;
- public final void collect(int doc, float score) {
- scores[doc + docBase] = score;
+ searcher.search(new TermQuery(new Term("untokfield", "20061212")), new Collector() {
+ private int docBase = 0;
+ private Scorer scorer;
+ public final void collect(int doc) throws IOException {
+ scores[doc + docBase] = scorer.score();
}
public void setNextReader(IndexReader reader, int docBase) {
this.docBase = docBase;
}
+ public void setScorer(Scorer scorer) throws IOException {
+ this.scorer = scorer;
+ }
});
searcher.close();
Index: contrib/miscellaneous/src/test/org/apache/lucene/misc/TestLengthNormModifier.java
===================================================================
--- contrib/miscellaneous/src/test/org/apache/lucene/misc/TestLengthNormModifier.java (revision 761300)
+++ contrib/miscellaneous/src/test/org/apache/lucene/misc/TestLengthNormModifier.java (working copy)
@@ -17,21 +17,26 @@
* limitations under the License.
*/
+import java.io.IOException;
+
import junit.framework.TestCase;
+import org.apache.lucene.analysis.SimpleAnalyzer;
+import org.apache.lucene.document.Document;
+import org.apache.lucene.document.Field;
+import org.apache.lucene.index.FieldNormModifier;
+import org.apache.lucene.index.IndexReader;
+import org.apache.lucene.index.IndexWriter;
import org.apache.lucene.index.Term;
-import org.apache.lucene.index.IndexWriter;
-import org.apache.lucene.index.IndexReader;
+import org.apache.lucene.index.IndexWriter.MaxFieldLength;
+import org.apache.lucene.search.Collector;
+import org.apache.lucene.search.DefaultSimilarity;
import org.apache.lucene.search.IndexSearcher;
-import org.apache.lucene.search.MultiReaderHitCollector;
+import org.apache.lucene.search.Scorer;
import org.apache.lucene.search.Similarity;
-import org.apache.lucene.search.DefaultSimilarity;
import org.apache.lucene.search.TermQuery;
+import org.apache.lucene.store.Directory;
import org.apache.lucene.store.RAMDirectory;
-import org.apache.lucene.store.Directory;
-import org.apache.lucene.analysis.SimpleAnalyzer;
-import org.apache.lucene.document.Document;
-import org.apache.lucene.document.Field;
/**
* Tests changing the norms after changing the simularity
@@ -52,13 +57,12 @@
/** inverts the normal notion of lengthNorm */
public static Similarity s = new DefaultSimilarity() {
public float lengthNorm(String fieldName, int numTokens) {
- return (float)numTokens;
+ return numTokens;
}
};
public void setUp() throws Exception {
- IndexWriter writer = new
- IndexWriter(store, new SimpleAnalyzer(), true);
+ IndexWriter writer = new IndexWriter(store, new SimpleAnalyzer(), true, MaxFieldLength.UNLIMITED);
for (int i = 0; i < NUM_DOCS; i++) {
Document d = new Document();
@@ -79,9 +83,9 @@
}
public void testMissingField() {
- LengthNormModifier lnm = new LengthNormModifier(store, s);
+ FieldNormModifier fnm = new FieldNormModifier(store, s);
try {
- lnm.reSetNorms("nobodyherebutuschickens");
+ fnm.reSetNorms("nobodyherebutuschickens");
} catch (Exception e) {
assertNull("caught something", e);
}
@@ -100,9 +104,9 @@
r.close();
- LengthNormModifier lnm = new LengthNormModifier(store, s);
+ FieldNormModifier fnm = new FieldNormModifier(store, s);
try {
- lnm.reSetNorms("nonorm");
+ fnm.reSetNorms("nonorm");
} catch (Exception e) {
assertNull("caught something", e);
}
@@ -129,14 +133,18 @@
// default similarity should put docs with shorter length first
searcher = new IndexSearcher(store);
- searcher.search(new TermQuery(new Term("field", "word")), new MultiReaderHitCollector() {
- private int docBase = -1;
- public final void collect(int doc, float score) {
- scores[doc + docBase] = score;
+ searcher.search(new TermQuery(new Term("field", "word")), new Collector() {
+ private int docBase = 0;
+ private Scorer scorer;
+ public final void collect(int doc) throws IOException {
+ scores[doc + docBase] = scorer.score();
}
public void setNextReader(IndexReader reader, int docBase) {
this.docBase = docBase;
}
+ public void setScorer(Scorer scorer) throws IOException {
+ this.scorer = scorer;
+ }
});
searcher.close();
@@ -151,22 +159,26 @@
// override the norms to be inverted
Similarity s = new DefaultSimilarity() {
public float lengthNorm(String fieldName, int numTokens) {
- return (float)numTokens;
+ return numTokens;
}
};
- LengthNormModifier lnm = new LengthNormModifier(store, s);
- lnm.reSetNorms("field");
+ FieldNormModifier fnm = new FieldNormModifier(store, s);
+ fnm.reSetNorms("field");
// new norm (with default similarity) should put longer docs first
searcher = new IndexSearcher(store);
- searcher.search(new TermQuery(new Term("field", "word")), new MultiReaderHitCollector() {
- private int docBase = -1;
- public final void collect(int doc, float score) {
- scores[doc + docBase] = score;
+ searcher.search(new TermQuery(new Term("field", "word")), new Collector() {
+ private int docBase = 0;
+ private Scorer scorer;
+ public final void collect(int doc) throws IOException {
+ scores[doc + docBase] = scorer.score();
}
public void setNextReader(IndexReader reader, int docBase) {
this.docBase = docBase;
}
+ public void setScorer(Scorer scorer) throws IOException {
+ this.scorer = scorer;
+ }
});
searcher.close();