Merge pull request #17736 from vespa-engine/bratseth/special-tokens

Bratseth/special tokens

2 files changed, 23 insertions, 11 deletions

diff --git a/linguistics/src/main/java/com/yahoo/language/simple/SimpleLinguistics.java b/linguistics/src/main/java/com/yahoo/language/simple/SimpleLinguistics.java
index e1a04b2985d..4ffe2a866d8 100644
--- a/linguistics/src/main/java/com/yahoo/language/simple/SimpleLinguistics.java
+++ b/linguistics/src/main/java/com/yahoo/language/simple/SimpleLinguistics.java
@@ -11,10 +11,14 @@ import com.yahoo.language.process.GramSplitter;
 import com.yahoo.language.process.Normalizer;
 import com.yahoo.language.process.Segmenter;
 import com.yahoo.language.process.SegmenterImpl;
+import com.yahoo.language.process.SpecialTokenRegistry;
 import com.yahoo.language.process.Stemmer;
 import com.yahoo.language.process.StemmerImpl;
 import com.yahoo.language.process.Tokenizer;
 import com.yahoo.language.process.Transformer;
+import com.yahoo.vespa.configdefinition.SpecialtokensConfig;
+
+import java.util.List;
 
 /**
  * Factory of simple linguistic processor implementations.
@@ -31,6 +35,7 @@ public class SimpleLinguistics implements Linguistics {
     private final Detector detector;
     private final CharacterClasses characterClasses;
     private final GramSplitter gramSplitter;
+    private final SpecialTokenRegistry specialTokenRegistry = new SpecialTokenRegistry(List.of());
 
     @Inject
     public SimpleLinguistics() {
@@ -45,7 +50,7 @@ public class SimpleLinguistics implements Linguistics {
     public Stemmer getStemmer() { return new StemmerImpl(getTokenizer()); }
 
     @Override
-    public Tokenizer getTokenizer() { return new SimpleTokenizer(normalizer, transformer); }
+    public Tokenizer getTokenizer() { return new SimpleTokenizer(normalizer, transformer, specialTokenRegistry); }
 
     @Override
     public Normalizer getNormalizer() { return normalizer; }
diff --git a/linguistics/src/main/java/com/yahoo/language/simple/SimpleTokenizer.java b/linguistics/src/main/java/com/yahoo/language/simple/SimpleTokenizer.java
index 7df432f496d..740307c0cca 100644
--- a/linguistics/src/main/java/com/yahoo/language/simple/SimpleTokenizer.java
+++ b/linguistics/src/main/java/com/yahoo/language/simple/SimpleTokenizer.java
@@ -23,11 +23,13 @@ import java.util.logging.Level;
  */
 public class SimpleTokenizer implements Tokenizer {
 
+    private static final Logger log = Logger.getLogger(SimpleTokenizer.class.getName());
     private final static int SPACE_CODE = 32;
+
     private final Normalizer normalizer;
     private final Transformer transformer;
     private final KStemmer stemmer = new KStemmer();
-    private static final Logger log = Logger.getLogger(SimpleTokenizer.class.getName());
+    private final SpecialTokenRegistry specialTokenRegistry;
 
     public SimpleTokenizer() {
         this(new SimpleNormalizer(), new SimpleTransformer());
@@ -38,8 +40,13 @@ public class SimpleTokenizer implements Tokenizer {
     }
 
     public SimpleTokenizer(Normalizer normalizer, Transformer transformer) {
+        this(normalizer, transformer, new SpecialTokenRegistry(List.of()));
+    }
+
+    public SimpleTokenizer(Normalizer normalizer, Transformer transformer, SpecialTokenRegistry specialTokenRegistry) {
         this.normalizer = normalizer;
         this.transformer = transformer;
+        this.specialTokenRegistry = specialTokenRegistry;
     }
 
     @Override
@@ -56,8 +63,8 @@ public class SimpleTokenizer implements Tokenizer {
                 String original = input.substring(prev, next);
                 String token = processToken(original, language, stemMode, removeAccents);
                 tokens.add(new SimpleToken(original).setOffset(prev)
-                                                .setType(prevType)
-                                                .setTokenString(token));
+                                                    .setType(prevType)
+                                                    .setTokenString(token));
                 prev = next;
                 prevType = nextType;
             }
@@ -67,20 +74,20 @@ public class SimpleTokenizer implements Tokenizer {
     }
 
     private String processToken(String token, Language language, StemMode stemMode, boolean removeAccents) {
-        final String original = token;
-        log.log(Level.FINEST, () -> "processToken '"+original+"'");
+        String original = token;
+        log.log(Level.FINEST, () -> "processToken '" + original + "'");
         token = normalizer.normalize(token);
         token = LinguisticsCase.toLowerCase(token);
         if (removeAccents)
             token = transformer.accentDrop(token, language);
         if (stemMode != StemMode.NONE) {
-            final String oldToken = token;
+            String oldToken = token;
             token = stemmer.stem(token);
-            final String newToken = token;
-            log.log(Level.FINEST, () -> "stem '"+oldToken+"' to '"+newToken+"'");
+            String newToken = token;
+            log.log(Level.FINEST, () -> "stem '" + oldToken+"' to '" + newToken+"'");
         }
-        final String result = token;
-        log.log(Level.FINEST, () -> "processed token is: "+result);
+        String result = token;
+        log.log(Level.FINEST, () -> "processed token is: " + result);
         return result;
     }