1
package org.apache.lucene.index;
4
* Licensed to the Apache Software Foundation (ASF) under one or more
5
* contributor license agreements. See the NOTICE file distributed with
6
* this work for additional information regarding copyright ownership.
7
* The ASF licenses this file to You under the Apache License, Version 2.0
8
* (the "License"); you may not use this file except in compliance with
9
* the License. You may obtain a copy of the License at
11
* http://www.apache.org/licenses/LICENSE-2.0
13
* Unless required by applicable law or agreed to in writing, software
14
* distributed under the License is distributed on an "AS IS" BASIS,
15
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
16
* See the License for the specific language governing permissions and
17
* limitations under the License.
20
import java.io.IOException;
21
import java.util.ArrayList;
22
import java.util.Random;
24
import java.util.concurrent.atomic.AtomicInteger;
25
import org.apache.lucene.analysis.Analyzer;
26
import org.apache.lucene.analysis.MockAnalyzer;
27
import org.apache.lucene.document.Document;
28
import org.apache.lucene.document.Field;
29
import org.apache.lucene.document.Field.Index;
30
import org.apache.lucene.document.Field.Store;
31
import org.apache.lucene.index.IndexWriterConfig.OpenMode;
32
import org.apache.lucene.index.SegmentNorms;
33
import org.apache.lucene.search.DefaultSimilarity;
34
import org.apache.lucene.search.Similarity;
35
import org.apache.lucene.store.Directory;
36
import org.apache.lucene.util.LuceneTestCase;
39
* Tests cloning IndexReader norms
41
public class TestIndexReaderCloneNorms extends LuceneTestCase {
43
private class SimilarityOne extends DefaultSimilarity {
45
public float computeNorm(String fieldName, FieldInvertState state) {
47
return state.getBoost();
51
private static final int NUM_FIELDS = 10;
53
private Similarity similarityOne;
55
private Analyzer anlzr;
57
private int numDocNorms;
59
private ArrayList<Float> norms;
61
private ArrayList<Float> modifiedNorms;
63
private float lastNorm = 0;
65
private float normDelta = (float) 0.001;
68
public void setUp() throws Exception {
70
similarityOne = new SimilarityOne();
71
anlzr = new MockAnalyzer(random);
75
* Test that norms values are preserved as the index is maintained. Including
76
* separate norms. Including merging indexes with seprate norms. Including
79
public void testNorms() throws IOException {
80
// test with a single index: index1
81
Directory dir1 = newDirectory();
82
IndexWriter.unlock(dir1);
84
norms = new ArrayList<Float>();
85
modifiedNorms = new ArrayList<Float>();
87
createIndex(random, dir1);
88
doTestNorms(random, dir1);
90
// test with a single index: index2
91
ArrayList<Float> norms1 = norms;
92
ArrayList<Float> modifiedNorms1 = modifiedNorms;
93
int numDocNorms1 = numDocNorms;
95
norms = new ArrayList<Float>();
96
modifiedNorms = new ArrayList<Float>();
99
Directory dir2 = newDirectory();
101
createIndex(random, dir2);
102
doTestNorms(random, dir2);
104
// add index1 and index2 to a third index: index3
105
Directory dir3 = newDirectory();
107
createIndex(random, dir3);
108
IndexWriter iw = new IndexWriter(dir3, newIndexWriterConfig(
109
TEST_VERSION_CURRENT, anlzr).setOpenMode(OpenMode.APPEND)
110
.setMaxBufferedDocs(5).setMergePolicy(newLogMergePolicy(3)));
111
iw.addIndexes(new Directory[] { dir1, dir2 });
115
norms1.addAll(norms);
117
modifiedNorms1.addAll(modifiedNorms);
118
modifiedNorms = modifiedNorms1;
119
numDocNorms += numDocNorms1;
123
doTestNorms(random, dir3);
126
iw = new IndexWriter(dir3, newIndexWriterConfig( TEST_VERSION_CURRENT,
127
anlzr).setOpenMode(OpenMode.APPEND).setMaxBufferedDocs(5).setMergePolicy(newLogMergePolicy(3)));
137
// try cloning and reopening the norms
138
private void doTestNorms(Random random, Directory dir) throws IOException {
139
addDocs(random, dir, 12, true);
140
IndexReader ir = IndexReader.open(dir, false);
142
modifyNormsForF1(ir);
143
IndexReader irc = (IndexReader) ir.clone();// IndexReader.open(dir, false);//ir.clone();
146
modifyNormsForF1(irc);
148
IndexReader irc3 = (IndexReader) irc.clone();
150
modifyNormsForF1(irc3);
159
public void testNormsClose() throws IOException {
160
Directory dir1 = newDirectory();
161
TestIndexReaderReopen.createIndex(random, dir1, false);
162
SegmentReader reader1 = SegmentReader.getOnlySegmentReader(dir1);
163
reader1.norms("field1");
164
SegmentNorms r1norm = reader1.norms.get("field1");
165
AtomicInteger r1BytesRef = r1norm.bytesRef();
166
SegmentReader reader2 = (SegmentReader)reader1.clone();
167
assertEquals(2, r1norm.bytesRef().get());
169
assertEquals(1, r1BytesRef.get());
170
reader2.norms("field1");
175
public void testNormsRefCounting() throws IOException {
176
Directory dir1 = newDirectory();
177
TestIndexReaderReopen.createIndex(random, dir1, false);
178
IndexReader reader1 = IndexReader.open(dir1, false);
180
IndexReader reader2C = (IndexReader) reader1.clone();
181
SegmentReader segmentReader2C = SegmentReader.getOnlySegmentReader(reader2C);
182
segmentReader2C.norms("field1"); // load the norms for the field
183
SegmentNorms reader2CNorm = segmentReader2C.norms.get("field1");
184
assertTrue("reader2CNorm.bytesRef()=" + reader2CNorm.bytesRef(), reader2CNorm.bytesRef().get() == 2);
188
IndexReader reader3C = (IndexReader) reader2C.clone();
189
SegmentReader segmentReader3C = SegmentReader.getOnlySegmentReader(reader3C);
190
SegmentNorms reader3CCNorm = segmentReader3C.norms.get("field1");
191
assertEquals(3, reader3CCNorm.bytesRef().get());
193
// edit a norm and the refcount should be 1
194
IndexReader reader4C = (IndexReader) reader3C.clone();
195
SegmentReader segmentReader4C = SegmentReader.getOnlySegmentReader(reader4C);
196
assertEquals(4, reader3CCNorm.bytesRef().get());
197
reader4C.setNorm(5, "field1", 0.33f);
199
// generate a cannot update exception in reader1
201
reader3C.setNorm(1, "field1", 0.99f);
202
fail("did not hit expected exception");
203
} catch (Exception ex) {
207
// norm values should be different
208
assertTrue(Similarity.getDefault().decodeNormValue(segmentReader3C.norms("field1")[5])
209
!= Similarity.getDefault().decodeNormValue(segmentReader4C.norms("field1")[5]));
210
SegmentNorms reader4CCNorm = segmentReader4C.norms.get("field1");
211
assertEquals(3, reader3CCNorm.bytesRef().get());
212
assertEquals(1, reader4CCNorm.bytesRef().get());
214
IndexReader reader5C = (IndexReader) reader4C.clone();
215
SegmentReader segmentReader5C = SegmentReader.getOnlySegmentReader(reader5C);
216
SegmentNorms reader5CCNorm = segmentReader5C.norms.get("field1");
217
reader5C.setNorm(5, "field1", 0.7f);
218
assertEquals(1, reader5CCNorm.bytesRef().get());
228
private void createIndex(Random random, Directory dir) throws IOException {
229
IndexWriter iw = new IndexWriter(dir, newIndexWriterConfig(
230
TEST_VERSION_CURRENT, anlzr).setOpenMode(OpenMode.CREATE)
231
.setMaxBufferedDocs(5).setSimilarity(similarityOne).setMergePolicy(newLogMergePolicy()));
232
setUseCompoundFile(iw.getConfig().getMergePolicy(), true);
233
setMergeFactor(iw.getConfig().getMergePolicy(), 3);
237
private void modifyNormsForF1(IndexReader ir) throws IOException {
239
// System.out.println("modifyNormsForF1 maxDoc: "+n);
240
for (int i = 0; i < n; i += 3) { // modify for every third doc
241
int k = (i * 3) % modifiedNorms.size();
242
float origNorm = modifiedNorms.get(i).floatValue();
243
float newNorm = modifiedNorms.get(k).floatValue();
244
// System.out.println("Modifying: for "+i+" from "+origNorm+" to
246
// System.out.println(" and: for "+k+" from "+newNorm+" to "+origNorm);
247
modifiedNorms.set(i, Float.valueOf(newNorm));
248
modifiedNorms.set(k, Float.valueOf(origNorm));
249
ir.setNorm(i, "f" + 1, newNorm);
250
ir.setNorm(k, "f" + 1, origNorm);
251
// System.out.println("setNorm i: "+i);
257
private void verifyIndex(Directory dir) throws IOException {
258
IndexReader ir = IndexReader.open(dir, false);
263
private void verifyIndex(IndexReader ir) throws IOException {
264
for (int i = 0; i < NUM_FIELDS; i++) {
265
String field = "f" + i;
266
byte b[] = ir.norms(field);
267
assertEquals("number of norms mismatches", numDocNorms, b.length);
268
ArrayList<Float> storedNorms = (i == 1 ? modifiedNorms : norms);
269
for (int j = 0; j < b.length; j++) {
270
float norm = Similarity.getDefault().decodeNormValue(b[j]);
271
float norm1 = storedNorms.get(j).floatValue();
272
assertEquals("stored norm value of " + field + " for doc " + j + " is "
273
+ norm + " - a mismatch!", norm, norm1, 0.000001);
278
private void addDocs(Random random, Directory dir, int ndocs, boolean compound)
280
IndexWriterConfig conf = newIndexWriterConfig(
281
TEST_VERSION_CURRENT, anlzr).setOpenMode(OpenMode.APPEND)
282
.setMaxBufferedDocs(5).setSimilarity(similarityOne).setMergePolicy(newLogMergePolicy());
283
LogMergePolicy lmp = (LogMergePolicy) conf.getMergePolicy();
284
lmp.setMergeFactor(3);
285
lmp.setUseCompoundFile(compound);
286
IndexWriter iw = new IndexWriter(dir, conf);
287
for (int i = 0; i < ndocs; i++) {
288
iw.addDocument(newDoc());
293
// create the next document
294
private Document newDoc() {
295
Document d = new Document();
296
float boost = nextNorm();
297
for (int i = 0; i < 10; i++) {
298
Field f = newField("f" + i, "v" + i, Store.NO, Index.NOT_ANALYZED);
305
// return unique norm values that are unchanged by encoding/decoding
306
private float nextNorm() {
307
float norm = lastNorm + normDelta;
309
float norm1 = Similarity.getDefault().decodeNormValue(
310
Similarity.getDefault().encodeNormValue(norm));
311
if (norm1 > lastNorm) {
312
// System.out.println(norm1+" > "+lastNorm);
318
norms.add(numDocNorms, Float.valueOf(norm));
319
modifiedNorms.add(numDocNorms, Float.valueOf(norm));
320
// System.out.println("creating norm("+numDocNorms+"): "+norm);
322
lastNorm = (norm > 10 ? 0 : norm); // there's a limit to how many distinct
323
// values can be stored in a ingle byte