Add tests for min_hash configuration and fix settings names (#84753)

Fixes #84578
elastic · Mar 9, 2022 · 6242118 · 6242118
1 parent 987a83b
commit 6242118
Show file tree

Hide file tree

Showing 3 changed files with 47 additions and 3 deletions.
diff --git a/docs/changelog/84753.yaml b/docs/changelog/84753.yaml
@@ -0,0 +1,6 @@
+pr: 84753
+summary: Add tests for `min_hash` configuration and fix settings names
+area: Analysis
+type: bug
+issues:
+ - 84578
diff --git a/...sis-common/src/main/java/org/elasticsearch/analysis/common/MinHashTokenFilterFactory.java b/...sis-common/src/main/java/org/elasticsearch/analysis/common/MinHashTokenFilterFactory.java
@@ -41,10 +41,10 @@ private Map<String, String> convertSettings(Settings settings) {
         if (settings.hasValue("hash_count")) {
             settingMap.put("hashCount", settings.get("hash_count"));
         }
-        if (settings.hasValue("bucketCount")) {
+        if (settings.hasValue("bucket_count")) {
             settingMap.put("bucketCount", settings.get("bucket_count"));
         }
-        if (settings.hasValue("hashSetSize")) {
+        if (settings.hasValue("hash_set_size")) {
             settingMap.put("hashSetSize", settings.get("hash_set_size"));
         }
         if (settings.hasValue("with_rotation")) {

diff --git a/...sis-common/src/test/java/org/elasticsearch/analysis/common/MinHashFilterFactoryTests.java b/...sis-common/src/test/java/org/elasticsearch/analysis/common/MinHashFilterFactoryTests.java
@@ -37,7 +37,7 @@ public void testDefault() throws IOException {
         assertStreamHasNumberOfTokens(tokenFilter.create(tokenizer), default_hash_count * default_bucket_size * default_hash_set_size);
     }
 
-    public void testSettings() throws IOException {
+    public void testRotationSettings() throws IOException {
         Settings settings = Settings.builder()
             .put("index.analysis.filter.test_min_hash.type", "min_hash")
             .put("index.analysis.filter.test_min_hash.hash_count", "1")
@@ -56,4 +56,42 @@ public void testSettings() throws IOException {
         // because with_rotation is false, we only expect 1 token here.
         assertStreamHasNumberOfTokens(tokenFilter.create(tokenizer), 1);
     }
+
+    public void testHashCountSettings() throws IOException {
+        Settings settings = Settings.builder()
+            .put("index.analysis.filter.test_min_hash.type", "min_hash")
+            .put("index.analysis.filter.test_min_hash.hash_count", "1")
+            .put("index.analysis.filter.test_min_hash.bucket_count", "1")
+            .put("index.analysis.filter.test_min_hash.hash_set_size", "1")
+            .put("index.analysis.filter.test_min_hash.with_rotation", true)
+            .put(Environment.PATH_HOME_SETTING.getKey(), createTempDir().toString())
+            .build();
+        ESTestCase.TestAnalysis analysis = AnalysisTestsHelper.createTestAnalysisFromSettings(settings, new CommonAnalysisPlugin());
+        TokenFilterFactory tokenFilter = analysis.tokenFilter.get("test_min_hash");
+        String source = "sushi";
+        Tokenizer tokenizer = new WhitespaceTokenizer();
+        tokenizer.setReader(new StringReader(source));
+
+        // Single bucket should result in a single output token
+        assertStreamHasNumberOfTokens(tokenFilter.create(tokenizer), 1);
+    }
+
+    public void testHashSetSize() throws IOException {
+        Settings settings = Settings.builder()
+            .put("index.analysis.filter.test_min_hash.type", "min_hash")
+            .put("index.analysis.filter.test_min_hash.hash_count", "1")
+            .put("index.analysis.filter.test_min_hash.bucket_count", "1")
+            .put("index.analysis.filter.test_min_hash.hash_set_size", "2")
+            .put("index.analysis.filter.test_min_hash.with_rotation", false)
+            .put(Environment.PATH_HOME_SETTING.getKey(), createTempDir().toString())
+            .build();
+        ESTestCase.TestAnalysis analysis = AnalysisTestsHelper.createTestAnalysisFromSettings(settings, new CommonAnalysisPlugin());
+        TokenFilterFactory tokenFilter = analysis.tokenFilter.get("test_min_hash");
+        String source = "another, longer test";
+        Tokenizer tokenizer = new WhitespaceTokenizer();
+        tokenizer.setReader(new StringReader(source));
+
+        // hash_set_size = 2 should give us two buckets
+        assertStreamHasNumberOfTokens(tokenFilter.create(tokenizer), 2);
+    }
 }