This is an automated email from the ASF dual-hosted git repository.
rzo1 pushed a commit to branch OPENNLP-1589
in repository https://gitbox.apache.org/repos/asf/opennlp.git
commit 43e6d393fe615a5d56b46a755a5ccec9914b8a74
Author: Richard Zowalla <r...@apache.org>
AuthorDate: Thu Jul 4 14:00:23 2024 +0200
OPENNLP-1589 - Use Arrays.equals(...)
---
.../tools/namefind/TokenNameFinderFactory.java | 4 +--
.../util/featuregen/CachedFeatureGenerator.java | 13 ++++-----
.../featuregen/CachedFeatureGeneratorTest.java | 34 ++++++++++++++++++----
3 files changed, 36 insertions(+), 15 deletions(-)
diff --git
a/opennlp-tools/src/main/java/opennlp/tools/namefind/TokenNameFinderFactory.java
b/opennlp-tools/src/main/java/opennlp/tools/namefind/TokenNameFinderFactory.java
index 6f31f0fe..da7277a4 100644
---
a/opennlp-tools/src/main/java/opennlp/tools/namefind/TokenNameFinderFactory.java
+++
b/opennlp-tools/src/main/java/opennlp/tools/namefind/TokenNameFinderFactory.java
@@ -212,13 +212,13 @@ public class TokenNameFinderFactory extends
BaseToolFactory {
AdaptiveFeatureGenerator featureGenerator = createFeatureGenerators();
if (featureGenerator == null) {
- featureGenerator = new CachedFeatureGenerator(
+ featureGenerator = new CachedFeatureGenerator(new
AggregatedFeatureGenerator(
new WindowFeatureGenerator(new TokenFeatureGenerator(), 2, 2),
new WindowFeatureGenerator(new TokenClassFeatureGenerator(true), 2,
2),
new OutcomePriorFeatureGenerator(),
new PreviousMapFeatureGenerator(),
new BigramNameFeatureGenerator(),
- new SentenceFeatureGenerator(true, false));
+ new SentenceFeatureGenerator(true, false)));
}
return new DefaultNameContextGenerator(featureGenerator);
diff --git
a/opennlp-tools/src/main/java/opennlp/tools/util/featuregen/CachedFeatureGenerator.java
b/opennlp-tools/src/main/java/opennlp/tools/util/featuregen/CachedFeatureGenerator.java
index e26ffbdd..560c1e4a 100644
---
a/opennlp-tools/src/main/java/opennlp/tools/util/featuregen/CachedFeatureGenerator.java
+++
b/opennlp-tools/src/main/java/opennlp/tools/util/featuregen/CachedFeatureGenerator.java
@@ -19,6 +19,7 @@
package opennlp.tools.util.featuregen;
import java.util.ArrayList;
+import java.util.Arrays;
import java.util.List;
import opennlp.tools.util.Cache;
@@ -37,15 +38,13 @@ public class CachedFeatureGenerator implements
AdaptiveFeatureGenerator {
private long numberOfCacheHits;
private long numberOfCacheMisses;
- @Deprecated
- public CachedFeatureGenerator(AdaptiveFeatureGenerator... generators) {
- this.generator = new AggregatedFeatureGenerator(generators);
- contextsCache = new Cache<>(100);
+ public CachedFeatureGenerator(AdaptiveFeatureGenerator generator, int
cacheSize) {
+ this.generator = generator;
+ contextsCache = new Cache<>(cacheSize);
}
public CachedFeatureGenerator(AdaptiveFeatureGenerator generator) {
- this.generator = generator;
- contextsCache = new Cache<>(100);
+ this(generator, 100);
}
@Override
@@ -54,7 +53,7 @@ public class CachedFeatureGenerator implements
AdaptiveFeatureGenerator {
List<String> cacheFeatures;
- if (tokens == prevTokens) {
+ if (Arrays.equals(prevTokens, tokens)) {
cacheFeatures = contextsCache.get(index);
if (cacheFeatures != null) {
diff --git
a/opennlp-tools/src/test/java/opennlp/tools/util/featuregen/CachedFeatureGeneratorTest.java
b/opennlp-tools/src/test/java/opennlp/tools/util/featuregen/CachedFeatureGeneratorTest.java
index 0dcb10a8..c17549c3 100644
---
a/opennlp-tools/src/test/java/opennlp/tools/util/featuregen/CachedFeatureGeneratorTest.java
+++
b/opennlp-tools/src/test/java/opennlp/tools/util/featuregen/CachedFeatureGeneratorTest.java
@@ -48,6 +48,31 @@ public class CachedFeatureGeneratorTest {
features = new ArrayList<>();
}
+ @Test
+ void testCachingOfRealWorldSentence() {
+ CachedFeatureGenerator generator = new
CachedFeatureGenerator(identityGenerator);
+ final String[] sentence = "He belongs to Apache \n Software Foundation
.".split(" ");
+ int testIndex = 0;
+
+ // after this call features are cached for testIndex
+ generator.createFeatures(features, sentence, testIndex, null);
+ Assertions.assertEquals(1, generator.getNumberOfCacheMisses());
+ Assertions.assertEquals(0, generator.getNumberOfCacheHits());
+
+ generator.createFeatures(features, sentence, testIndex, null);
+ Assertions.assertEquals(1, generator.getNumberOfCacheMisses());
+ Assertions.assertEquals(1, generator.getNumberOfCacheHits());
+
+ generator.createFeatures(features, sentence, testIndex + 1, null);
+ Assertions.assertEquals(2, generator.getNumberOfCacheMisses());
+ Assertions.assertEquals(1, generator.getNumberOfCacheHits());
+
+ generator.createFeatures(features, sentence, testIndex + 1, null);
+ Assertions.assertEquals(2, generator.getNumberOfCacheMisses());
+ Assertions.assertEquals(2, generator.getNumberOfCacheHits());
+
+ }
+
/**
* Tests if cache works for one sentence and two different token indexes.
*/
@@ -85,7 +110,7 @@ public class CachedFeatureGeneratorTest {
int testIndex2 = testIndex + 1;
- generator.createFeatures(features, Arrays.copyOf(testSentence1,
testSentence1.length), testIndex2, null);
+ generator.createFeatures(features, Arrays.copyOf(testSentence1,
testSentence1.length), testIndex2, null);
Assertions.assertEquals(2, generator.getNumberOfCacheMisses());
Assertions.assertEquals(1, generator.getNumberOfCacheHits());
@@ -115,7 +140,7 @@ public class CachedFeatureGeneratorTest {
features.clear();
// use another sentence but same index
- generator.createFeatures(features, testSentence2, testIndex, null);
+ generator.createFeatures(features, Arrays.copyOf(testSentence2,
testSentence2.length), testIndex, null);
Assertions.assertEquals(2, generator.getNumberOfCacheMisses());
Assertions.assertEquals(0, generator.getNumberOfCacheHits());
@@ -127,10 +152,7 @@ public class CachedFeatureGeneratorTest {
// check if features are really cached
final String expectedToken = testSentence2[testIndex];
-
- testSentence2[testIndex] = null;
-
- generator.createFeatures(features, testSentence2, testIndex, null);
+ generator.createFeatures(features, Arrays.copyOf(testSentence2,
testSentence2.length), testIndex, null);
Assertions.assertTrue(features.contains(expectedToken));
Assertions.assertEquals(1, features.size());
