Expose discard_compound_token option to kuromoji_tokenizer (#57421)

This commit exposes the new Lucene option `discard_compound_token` to the Elasticsearch Kuromoji plugin.

Author: johtani

docs/plugins/analysis-kuromoji.asciidoc

@@ -70,7 +70,7 @@ unknown words.  It can be set to:
 
     Extended mode outputs unigrams for unknown words. Example output:
 
-    関西, 国際, 空港
+    関西, 関西国際空港, 国際, 空港
     ア, ブ, ラ, カ, ダ, ブ, ラ
 --
 
@@ -208,6 +208,12 @@ The above `analyze` request returns the following:
 }
 --------------------------------------------------
 
+`discard_compound_token`::
+    Whether original compound tokens should be discarded from the output with `search` mode. Defaults to `false`.
+    Example output with `search` or `extended` mode and this option `true`:
+
+    関西, 国際, 空港
+
 
 [[analysis-kuromoji-baseform]]
 ==== `kuromoji_baseform` token filter

plugins/analysis-kuromoji/src/main/java/org/elasticsearch/index/analysis/KuromojiTokenizerFactory.java

@@ -41,21 +41,24 @@ public class KuromojiTokenizerFactory extends AbstractTokenizerFactory {
     private static final String USER_DICT_RULES_OPTION = "user_dictionary_rules";
     private static final String NBEST_COST = "nbest_cost";
     private static final String NBEST_EXAMPLES = "nbest_examples";
+    private static final String DISCARD_COMPOUND_TOKEN = "discard_compound_token";
 
     private final UserDictionary userDictionary;
     private final Mode mode;
     private final String nBestExamples;
     private final int nBestCost;
 
-    private boolean discartPunctuation;
+    private boolean discardPunctuation;
+    private boolean discardCompoundToken;
 
     public KuromojiTokenizerFactory(IndexSettings indexSettings, Environment env, String name, Settings settings) {
         super(indexSettings, settings, name);
         mode = getMode(settings);
         userDictionary = getUserDictionary(env, settings);
-        discartPunctuation = settings.getAsBoolean("discard_punctuation", true);
+        discardPunctuation = settings.getAsBoolean("discard_punctuation", true);
         nBestCost = settings.getAsInt(NBEST_COST, -1);
         nBestExamples = settings.get(NBEST_EXAMPLES);
+        discardCompoundToken = settings.getAsBoolean(DISCARD_COMPOUND_TOKEN, false);
     }
 
     public static UserDictionary getUserDictionary(Environment env, Settings settings) {
@@ -108,7 +111,7 @@ public static JapaneseTokenizer.Mode getMode(Settings settings) {
 
     @Override
     public Tokenizer create() {
-        JapaneseTokenizer t = new JapaneseTokenizer(userDictionary, discartPunctuation, mode);
+        JapaneseTokenizer t = new JapaneseTokenizer(userDictionary, discardPunctuation, discardCompoundToken, mode);
         int nBestCost = this.nBestCost;
         if (nBestExamples != null) {
             nBestCost = Math.max(nBestCost, t.calcNBestCost(nBestExamples));

plugins/analysis-kuromoji/src/test/java/org/elasticsearch/index/analysis/KuromojiAnalysisTests.java

@@ -348,6 +348,17 @@ public void testKuromojiAnalyzerDuplicateUserDictRule() throws Exception {
         assertThat(exc.getMessage(), containsString("[制限スピード] in user dictionary at line [3]"));
     }
 
+    public void testDiscardCompoundToken() throws Exception {
+        TestAnalysis analysis = createTestAnalysis();
+        TokenizerFactory tokenizerFactory = analysis.tokenizer.get("kuromoji_discard_compound_token");
+        String source = "株式会社";
+        String[] expected = new String[] {"株式", "会社"};
+
+        Tokenizer tokenizer = tokenizerFactory.create();
+        tokenizer.setReader(new StringReader(source));
+        assertSimpleTSOutput(tokenizer, expected);
+    }
+
     private TestAnalysis createTestAnalysis(Settings analysisSettings) throws IOException {
         InputStream dict = KuromojiAnalysisTests.class.getResourceAsStream("user_dict.txt");
         Path home = createTempDir();

plugins/analysis-kuromoji/src/test/resources/org/elasticsearch/index/analysis/kuromoji_analysis.json

@@ -60,6 +60,10 @@
                     "type": "kuromoji_tokenizer",
                     "nbest_examples" : "/鳩山積み-鳩山/鳩山積み-鳩/",
                     "nbest_cost" : "1000"
+                },
+                "kuromoji_discard_compound_token": {
+                    "type": "kuromoji_tokenizer",
+                    "discard_compound_token": true
                 }
             },
             "analyzer" : {
@@ -68,7 +72,7 @@
                     "tokenizer" : "kuromoji_tokenizer"
                 }
             }
-            
+
         }
     }
 }