Add a new configuration setting synonym_analyzer for `synonym_graph…

…` and `synonym`. (#16488) * Add custom synonym_analyzer * synonym_analyzer configuration setting --------- Signed-off-by: Prudhvi Godithi <[email protected]>
opensearch-project · Nov 12, 2024 · 5068fad · 5068fad
1 parent 53d41d3
commit 5068fad
Show file tree

Hide file tree

Showing 10 changed files with 250 additions and 16 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -17,6 +17,7 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 - Support for keyword fields in star-tree index ([#16233](https://github.com/opensearch-project/OpenSearch/pull/16233))
 - Add a flag in QueryShardContext to differentiate inner hit query ([#16600](https://github.com/opensearch-project/OpenSearch/pull/16600))
 - Add vertical scaling and SoftReference for snapshot repository data cache ([#16489](https://github.com/opensearch-project/OpenSearch/pull/16489))
+- Add new configuration setting `synonym_analyzer`, to the `synonym` and `synonym_graph` filters, enabling the specification of a custom analyzer for reading the synonym file ([#16488](https://github.com/opensearch-project/OpenSearch/pull/16488)).
 
 ### Dependencies
 - Bump `com.azure:azure-storage-common` from 12.25.1 to 12.27.1 ([#16521](https://github.com/opensearch-project/OpenSearch/pull/16521))

diff --git a/...lysis-common/src/main/java/org/opensearch/analysis/common/CommonAnalysisModulePlugin.java b/...lysis-common/src/main/java/org/opensearch/analysis/common/CommonAnalysisModulePlugin.java
@@ -146,6 +146,7 @@
 import org.opensearch.index.analysis.PreConfiguredTokenizer;
 import org.opensearch.index.analysis.TokenFilterFactory;
 import org.opensearch.index.analysis.TokenizerFactory;
+import org.opensearch.indices.analysis.AnalysisModule;
 import org.opensearch.indices.analysis.AnalysisModule.AnalysisProvider;
 import org.opensearch.indices.analysis.PreBuiltCacheFactory.CachingStrategy;
 import org.opensearch.plugins.AnalysisPlugin;
@@ -247,7 +248,7 @@ public Map<String, AnalysisProvider<AnalyzerProvider<? extends Analyzer>>> getAn
     }
 
     @Override
-    public Map<String, AnalysisProvider<TokenFilterFactory>> getTokenFilters() {
+    public Map<String, AnalysisProvider<TokenFilterFactory>> getTokenFilters(AnalysisModule analysisModule) {
         Map<String, AnalysisProvider<TokenFilterFactory>> filters = new TreeMap<>();
         filters.put("apostrophe", ApostropheFilterFactory::new);
         filters.put("arabic_normalization", ArabicNormalizationFilterFactory::new);
@@ -332,14 +333,36 @@ public Map<String, AnalysisProvider<TokenFilterFactory>> getTokenFilters() {
         filters.put("sorani_normalization", SoraniNormalizationFilterFactory::new);
         filters.put("stemmer_override", requiresAnalysisSettings(StemmerOverrideTokenFilterFactory::new));
         filters.put("stemmer", StemmerTokenFilterFactory::new);
-        filters.put("synonym", requiresAnalysisSettings(SynonymTokenFilterFactory::new));
-        filters.put("synonym_graph", requiresAnalysisSettings(SynonymGraphTokenFilterFactory::new));
         filters.put("trim", TrimTokenFilterFactory::new);
         filters.put("truncate", requiresAnalysisSettings(TruncateTokenFilterFactory::new));
         filters.put("unique", UniqueTokenFilterFactory::new);
         filters.put("uppercase", UpperCaseTokenFilterFactory::new);
         filters.put("word_delimiter_graph", WordDelimiterGraphTokenFilterFactory::new);
         filters.put("word_delimiter", WordDelimiterTokenFilterFactory::new);
+        filters.put(
+            "synonym",
+            requiresAnalysisSettings(
+                (indexSettings, environment, name, settings) -> new SynonymTokenFilterFactory(
+                    indexSettings,
+                    environment,
+                    name,
+                    settings,
+                    analysisModule.getAnalysisRegistry()
+                )
+            )
+        );
+        filters.put(
+            "synonym_graph",
+            requiresAnalysisSettings(
+                (indexSettings, environment, name, settings) -> new SynonymGraphTokenFilterFactory(
+                    indexSettings,
+                    environment,
+                    name,
+                    settings,
+                    analysisModule.getAnalysisRegistry()
+                )
+            )
+        );
         return filters;
     }
 

diff --git a/...s-common/src/main/java/org/opensearch/analysis/common/SynonymGraphTokenFilterFactory.java b/...s-common/src/main/java/org/opensearch/analysis/common/SynonymGraphTokenFilterFactory.java
@@ -40,6 +40,7 @@
 import org.opensearch.env.Environment;
 import org.opensearch.index.IndexSettings;
 import org.opensearch.index.analysis.AnalysisMode;
+import org.opensearch.index.analysis.AnalysisRegistry;
 import org.opensearch.index.analysis.CharFilterFactory;
 import org.opensearch.index.analysis.TokenFilterFactory;
 import org.opensearch.index.analysis.TokenizerFactory;
@@ -49,8 +50,14 @@
 
 public class SynonymGraphTokenFilterFactory extends SynonymTokenFilterFactory {
 
-    SynonymGraphTokenFilterFactory(IndexSettings indexSettings, Environment env, String name, Settings settings) {
-        super(indexSettings, env, name, settings);
+    SynonymGraphTokenFilterFactory(
+        IndexSettings indexSettings,
+        Environment env,
+        String name,
+        Settings settings,
+        AnalysisRegistry analysisRegistry
+    ) {
+        super(indexSettings, env, name, settings, analysisRegistry);
     }
 
     @Override

diff --git a/...alysis-common/src/main/java/org/opensearch/analysis/common/SynonymTokenFilterFactory.java b/...alysis-common/src/main/java/org/opensearch/analysis/common/SynonymTokenFilterFactory.java
@@ -44,11 +44,13 @@
 import org.opensearch.index.analysis.AbstractTokenFilterFactory;
 import org.opensearch.index.analysis.Analysis;
 import org.opensearch.index.analysis.AnalysisMode;
+import org.opensearch.index.analysis.AnalysisRegistry;
 import org.opensearch.index.analysis.CharFilterFactory;
 import org.opensearch.index.analysis.CustomAnalyzer;
 import org.opensearch.index.analysis.TokenFilterFactory;
 import org.opensearch.index.analysis.TokenizerFactory;
 
+import java.io.IOException;
 import java.io.Reader;
 import java.io.StringReader;
 import java.util.List;
@@ -64,8 +66,16 @@ public class SynonymTokenFilterFactory extends AbstractTokenFilterFactory {
     protected final Settings settings;
     protected final Environment environment;
     protected final AnalysisMode analysisMode;
-
-    SynonymTokenFilterFactory(IndexSettings indexSettings, Environment env, String name, Settings settings) {
+    private final String synonymAnalyzerName;
+    private final AnalysisRegistry analysisRegistry;
+
+    SynonymTokenFilterFactory(
+        IndexSettings indexSettings,
+        Environment env,
+        String name,
+        Settings settings,
+        AnalysisRegistry analysisRegistry
+    ) {
         super(indexSettings, name, settings);
         this.settings = settings;
 
@@ -83,6 +93,8 @@ public class SynonymTokenFilterFactory extends AbstractTokenFilterFactory {
         boolean updateable = settings.getAsBoolean("updateable", false);
         this.analysisMode = updateable ? AnalysisMode.SEARCH_TIME : AnalysisMode.ALL;
         this.environment = env;
+        this.synonymAnalyzerName = settings.get("synonym_analyzer", null);
+        this.analysisRegistry = analysisRegistry;
     }
 
     @Override
@@ -137,6 +149,17 @@ Analyzer buildSynonymAnalyzer(
         List<TokenFilterFactory> tokenFilters,
         Function<String, TokenFilterFactory> allFilters
     ) {
+        if (synonymAnalyzerName != null) {
+            Analyzer customSynonymAnalyzer;
+            try {
+                customSynonymAnalyzer = analysisRegistry.getAnalyzer(synonymAnalyzerName);
+            } catch (IOException e) {
+                throw new RuntimeException(e);
+            }
+            if (customSynonymAnalyzer != null) {
+                return customSynonymAnalyzer;
+            }
+        }
         return new CustomAnalyzer(
             tokenizer,
             charFilters.toArray(new CharFilterFactory[0]),
@@ -177,5 +200,4 @@ Reader getRulesFromSettings(Environment env) {
         }
         return rulesReader;
     }
-
 }
diff --git a/...lysis-common/src/test/java/org/opensearch/analysis/common/CommonAnalysisFactoryTests.java b/...lysis-common/src/test/java/org/opensearch/analysis/common/CommonAnalysisFactoryTests.java
@@ -39,12 +39,16 @@
 import org.apache.lucene.analysis.snowball.SnowballPorterFilterFactory;
 import org.apache.lucene.analysis.te.TeluguNormalizationFilterFactory;
 import org.apache.lucene.analysis.te.TeluguStemFilterFactory;
+import org.opensearch.index.analysis.TokenFilterFactory;
 import org.opensearch.indices.analysis.AnalysisFactoryTestCase;
+import org.opensearch.indices.analysis.AnalysisModule;
 
 import java.util.List;
 import java.util.Map;
 import java.util.TreeMap;
 
+import org.mockito.Mock;
+
 import static java.util.Collections.emptyList;
 import static java.util.stream.Collectors.toList;
 
@@ -53,6 +57,9 @@ public CommonAnalysisFactoryTests() {
         super(new CommonAnalysisModulePlugin());
     }
 
+    @Mock
+    private AnalysisModule analysisModule;
+
     @Override
     protected Map<String, Class<?>> getTokenizers() {
         Map<String, Class<?>> tokenizers = new TreeMap<>(super.getTokenizers());
@@ -302,4 +309,19 @@ private void markedTestCase(String name, Map<String, Class<?>> map) {
             unmarked
         );
     }
+
+    /**
+     * Tests the getTokenFilters(AnalysisModule) method to verify:
+     * 1. All token filters are properly loaded
+     * 2. Basic filters remain available
+     * 3. Synonym filters remain available when AnalysisModule is provided
+     */
+    public void testGetTokenFiltersWithAnalysisModule() {
+        CommonAnalysisModulePlugin plugin = (CommonAnalysisModulePlugin) getAnalysisPlugin();
+        Map<String, AnalysisModule.AnalysisProvider<TokenFilterFactory>> filters = plugin.getTokenFilters(analysisModule);
+        assertNotNull("Token filters should not be null", filters);
+        assertTrue("Should contain basic filters", filters.containsKey("lowercase"));
+        assertTrue("Should contain synonym filter", filters.containsKey("synonym"));
+        assertTrue("Should contain synonym_graph filter", filters.containsKey("synonym_graph"));
+    }
 }
diff --git a/...s/analysis-common/src/test/java/org/opensearch/analysis/common/SynonymsAnalysisTests.java b/...s/analysis-common/src/test/java/org/opensearch/analysis/common/SynonymsAnalysisTests.java
@@ -41,11 +41,14 @@
 import org.opensearch.cluster.metadata.IndexMetadata;
 import org.opensearch.common.settings.Settings;
 import org.opensearch.env.Environment;
+import org.opensearch.env.TestEnvironment;
 import org.opensearch.index.IndexSettings;
+import org.opensearch.index.analysis.AnalysisRegistry;
 import org.opensearch.index.analysis.IndexAnalyzers;
 import org.opensearch.index.analysis.PreConfiguredTokenFilter;
 import org.opensearch.index.analysis.TokenFilterFactory;
 import org.opensearch.index.analysis.TokenizerFactory;
+import org.opensearch.indices.analysis.AnalysisModule;
 import org.opensearch.test.IndexSettingsModule;
 import org.opensearch.test.OpenSearchTestCase;
 import org.opensearch.test.VersionUtils;
@@ -63,6 +66,7 @@
 import static org.hamcrest.Matchers.equalTo;
 import static org.hamcrest.Matchers.instanceOf;
 import static org.hamcrest.Matchers.startsWith;
+import static org.apache.lucene.tests.analysis.BaseTokenStreamTestCase.assertTokenStreamContents;
 
 public class SynonymsAnalysisTests extends OpenSearchTestCase {
     private IndexAnalyzers indexAnalyzers;
@@ -255,14 +259,16 @@ public void testTokenFiltersBypassSynonymAnalysis() throws IOException {
             .put("hyphenation_patterns_path", "foo")
             .build();
         IndexSettings idxSettings = IndexSettingsModule.newIndexSettings("index", settings);
-
+        Environment environment = TestEnvironment.newEnvironment(settings);
+        AnalysisModule analysisModule = new AnalysisModule(environment, Collections.singletonList(new CommonAnalysisModulePlugin()));
+        AnalysisRegistry analysisRegistry = analysisModule.getAnalysisRegistry();
         String[] bypassingFactories = new String[] { "dictionary_decompounder" };
 
         CommonAnalysisModulePlugin plugin = new CommonAnalysisModulePlugin();
         for (String factory : bypassingFactories) {
-            TokenFilterFactory tff = plugin.getTokenFilters().get(factory).get(idxSettings, null, factory, settings);
-            TokenizerFactory tok = new KeywordTokenizerFactory(idxSettings, null, "keyword", settings);
-            SynonymTokenFilterFactory stff = new SynonymTokenFilterFactory(idxSettings, null, "synonym", settings);
+            TokenFilterFactory tff = plugin.getTokenFilters(analysisModule).get(factory).get(idxSettings, environment, factory, settings);
+            TokenizerFactory tok = new KeywordTokenizerFactory(idxSettings, environment, "keyword", settings);
+            SynonymTokenFilterFactory stff = new SynonymTokenFilterFactory(idxSettings, environment, "synonym", settings, analysisRegistry);
             Analyzer analyzer = stff.buildSynonymAnalyzer(tok, Collections.emptyList(), Collections.singletonList(tff), null);
 
             try (TokenStream ts = analyzer.tokenStream("field", "text")) {
@@ -319,7 +325,11 @@ public void testDisallowedTokenFilters() throws IOException {
             .putList("common_words", "a", "b")
             .put("output_unigrams", "true")
             .build();
+
+        Environment environment = TestEnvironment.newEnvironment(settings);
         IndexSettings idxSettings = IndexSettingsModule.newIndexSettings("index", settings);
+        AnalysisModule analysisModule = new AnalysisModule(environment, Collections.singletonList(new CommonAnalysisModulePlugin()));
+        AnalysisRegistry analysisRegistry = analysisModule.getAnalysisRegistry();
         CommonAnalysisModulePlugin plugin = new CommonAnalysisModulePlugin();
 
         String[] disallowedFactories = new String[] {
@@ -333,9 +343,9 @@ public void testDisallowedTokenFilters() throws IOException {
             "fingerprint" };
 
         for (String factory : disallowedFactories) {
-            TokenFilterFactory tff = plugin.getTokenFilters().get(factory).get(idxSettings, null, factory, settings);
-            TokenizerFactory tok = new KeywordTokenizerFactory(idxSettings, null, "keyword", settings);
-            SynonymTokenFilterFactory stff = new SynonymTokenFilterFactory(idxSettings, null, "synonym", settings);
+            TokenFilterFactory tff = plugin.getTokenFilters(analysisModule).get(factory).get(idxSettings, environment, factory, settings);
+            TokenizerFactory tok = new KeywordTokenizerFactory(idxSettings, environment, "keyword", settings);
+            SynonymTokenFilterFactory stff = new SynonymTokenFilterFactory(idxSettings, environment, "synonym", settings, analysisRegistry);
 
             IllegalArgumentException e = expectThrows(
                 IllegalArgumentException.class,
@@ -362,4 +372,75 @@ private void match(String analyzerName, String source, String target) throws IOE
         MatcherAssert.assertThat(target, equalTo(sb.toString().trim()));
     }
 
+    /**
+     * Tests the integration of word delimiter and synonym graph filters with synonym_analyzer based on issue #16263.
+     * This test verifies the correct handling of:
+     * 1. Hyphenated words with word delimiter (e.g., "note-book" → ["notebook", "note", "book"])
+     * 2. Multi-word synonyms (e.g., "mobile phone" → ["smartphone"])
+     * 3. Single word synonyms (e.g., "laptop" → ["notebook"])
+     *
+     * @see <a href="https://github.com/opensearch-project/OpenSearch/issues/16263">Issue #16263</a>
+     */
+    public void testSynonymAnalyzerWithWordDelimiter() throws IOException {
+        Settings settings = Settings.builder()
+            .put(IndexMetadata.SETTING_VERSION_CREATED, Version.CURRENT)
+            .put("path.home", createTempDir().toString())
+            .put("index.analysis.filter.custom_word_delimiter.type", "word_delimiter_graph")
+            .put("index.analysis.filter.custom_word_delimiter.generate_word_parts", true)
+            .put("index.analysis.filter.custom_word_delimiter.catenate_all", true)
+            .put("index.analysis.filter.custom_word_delimiter.split_on_numerics", false)
+            .put("index.analysis.filter.custom_word_delimiter.split_on_case_change", false)
+            .put("index.analysis.filter.custom_pattern_replace_filter.type", "pattern_replace")
+            .put("index.analysis.filter.custom_pattern_replace_filter.pattern", "(-)")
+            .put("index.analysis.filter.custom_pattern_replace_filter.replacement", " ")
+            .put("index.analysis.filter.custom_pattern_replace_filter.all", true)
+            .put("index.analysis.filter.custom_synonym_graph_filter.type", "synonym_graph")
+            .putList(
+                "index.analysis.filter.custom_synonym_graph_filter.synonyms",
+                "laptop => notebook",
+                "smartphone, mobile phone, cell phone => smartphone",
+                "tv, television => television"
+            )
+            .put("index.analysis.filter.custom_synonym_graph_filter.synonym_analyzer", "standard")
+            .put("index.analysis.analyzer.text_en_index.type", "custom")
+            .put("index.analysis.analyzer.text_en_index.tokenizer", "whitespace")
+            .putList(
+                "index.analysis.analyzer.text_en_index.filter",
+                "lowercase",
+                "custom_word_delimiter",
+                "custom_synonym_graph_filter",
+                "custom_pattern_replace_filter",
+                "flatten_graph"
+            )
+            .build();
+        Environment environment = TestEnvironment.newEnvironment(settings);
+        IndexSettings indexSettings = IndexSettingsModule.newIndexSettings("test", settings);
+        AnalysisModule module = new AnalysisModule(environment, Collections.singletonList(new CommonAnalysisModulePlugin()));
+        IndexAnalyzers analyzers = module.getAnalysisRegistry().build(indexSettings);
+        try (TokenStream ts = analyzers.get("text_en_index").tokenStream("", "note-book")) {
+            assertTokenStreamContents(
+                ts,
+                new String[] { "notebook", "note", "book" },
+                new int[] { 0, 0, 5 },
+                new int[] { 9, 4, 9 },
+                new String[] { "word", "word", "word" },
+                new int[] { 1, 0, 1 },
+                new int[] { 2, 1, 1 }
+            );
+        }
+        try (TokenStream ts = analyzers.get("text_en_index").tokenStream("", "mobile phone")) {
+            assertTokenStreamContents(
+                ts,
+                new String[] { "smartphone" },
+                new int[] { 0 },
+                new int[] { 12 },
+                new String[] { "SYNONYM" },
+                new int[] { 1 },
+                new int[] { 1 }
+            );
+        }
+        try (TokenStream ts = analyzers.get("text_en_index").tokenStream("", "laptop")) {
+            assertTokenStreamContents(ts, new String[] { "notebook" }, new int[] { 0 }, new int[] { 6 });
+        }
+    }
 }
diff --git a/server/src/main/java/org/opensearch/indices/analysis/AnalysisModule.java b/server/src/main/java/org/opensearch/indices/analysis/AnalysisModule.java
@@ -165,7 +165,12 @@ public boolean requiresAnalysisSettings() {
             )
         );
 
-        tokenFilters.extractAndRegister(plugins, AnalysisPlugin::getTokenFilters);
+        for (AnalysisPlugin plugin : plugins) {
+            Map<String, AnalysisProvider<TokenFilterFactory>> filters = plugin.getTokenFilters(this);
+            for (Map.Entry<String, AnalysisProvider<TokenFilterFactory>> entry : filters.entrySet()) {
+                tokenFilters.register(entry.getKey(), entry.getValue());
+            }
+        }
         return tokenFilters;
     }
 

diff --git a/server/src/main/java/org/opensearch/plugins/AnalysisPlugin.java b/server/src/main/java/org/opensearch/plugins/AnalysisPlugin.java
@@ -47,6 +47,7 @@
 import org.opensearch.index.analysis.PreConfiguredTokenizer;
 import org.opensearch.index.analysis.TokenFilterFactory;
 import org.opensearch.index.analysis.TokenizerFactory;
+import org.opensearch.indices.analysis.AnalysisModule;
 import org.opensearch.indices.analysis.AnalysisModule.AnalysisProvider;
 
 import java.io.IOException;
@@ -84,6 +85,14 @@ default Map<String, AnalysisProvider<CharFilterFactory>> getCharFilters() {
         return emptyMap();
     }
 
+    /**
+     * Override to add additional {@link TokenFilter}s that need access to the AnalysisModule.
+     * The default implementation for plugins that don't need AnalysisModule calls the existing getTokenFilters() method.
+     */
+    default Map<String, AnalysisProvider<TokenFilterFactory>> getTokenFilters(AnalysisModule analysisModule) {
+        return getTokenFilters();
+    }
+
     /**
      * Override to add additional {@link TokenFilter}s. See {@link #requiresAnalysisSettings(AnalysisProvider)}
      * how to on get the configuration from the index.