Don't lowercase linguistics annotations

Tokens are already lowercased by our bundled linguistics components.
Lowercasing again when annotating precludes plugging in a lingustics
component which preserves casing.

6 files changed, 24 insertions, 23 deletions

diff --git a/indexinglanguage/src/main/java/com/yahoo/vespa/indexinglanguage/expressions/NGramExpression.java b/indexinglanguage/src/main/java/com/yahoo/vespa/indexinglanguage/expressions/NGramExpression.java
index 26058eeb8f3..fdfadf65400 100644
--- a/indexinglanguage/src/main/java/com/yahoo/vespa/indexinglanguage/expressions/NGramExpression.java
+++ b/indexinglanguage/src/main/java/com/yahoo/vespa/indexinglanguage/expressions/NGramExpression.java
@@ -15,6 +15,8 @@ import com.yahoo.vespa.indexinglanguage.linguistics.LinguisticsAnnotator;
 
 import java.util.Iterator;
 
+import static com.yahoo.language.LinguisticsCase.toLowerCase;
+
 /**
  * A filter which splits incoming text into n-grams.
  *
@@ -68,8 +70,9 @@ public final class NGramExpression extends Expression {
 
             // annotate gram as a word term
             String gramString = gram.extractFrom(output.getString());
-            typedSpan(gram.getStart(), gram.getCodePointCount(), TokenType.ALPHABETIC, spanList).
-                    annotate(LinguisticsAnnotator.lowerCaseTermAnnotation(gramString, gramString));
+            typedSpan(gram.getStart(),
+                      gram.getCodePointCount(),
+                      TokenType.ALPHABETIC, spanList).annotate(LinguisticsAnnotator.termAnnotation(toLowerCase(gramString), gramString));
 
             lastPosition = gram.getStart() + gram.getCodePointCount();
         }
diff --git a/indexinglanguage/src/main/java/com/yahoo/vespa/indexinglanguage/linguistics/LinguisticsAnnotator.java b/indexinglanguage/src/main/java/com/yahoo/vespa/indexinglanguage/linguistics/LinguisticsAnnotator.java
index 191d067effe..04019800d59 100644
--- a/indexinglanguage/src/main/java/com/yahoo/vespa/indexinglanguage/linguistics/LinguisticsAnnotator.java
+++ b/indexinglanguage/src/main/java/com/yahoo/vespa/indexinglanguage/linguistics/LinguisticsAnnotator.java
@@ -86,24 +86,23 @@ public class LinguisticsAnnotator {
     }
 
     /**
-     * Creates a TERM annotation which has the lowercase value as annotation (only) if it is different from the
+     * Creates a TERM annotation which has the term as annotation (only) if it is different from the
      * original.
      *
-     * @param termToLowerCase the term to lower case
-     * @param origTerm        the original term
+     * @param term the term
+     * @param origTerm the original term
      * @return the created TERM annotation
      */
-    public static Annotation lowerCaseTermAnnotation(String termToLowerCase, String origTerm) {
-        String annotationValue = toLowerCase(termToLowerCase);
-        if (annotationValue.equals(origTerm)) {
+    public static Annotation termAnnotation(String term, String origTerm) {
+        if (term.equals(origTerm))
             return new Annotation(AnnotationTypes.TERM);
-        }
-        return new Annotation(AnnotationTypes.TERM, new StringFieldValue(annotationValue));
+        else
+            return new Annotation(AnnotationTypes.TERM, new StringFieldValue(term));
     }
 
     private static void addAnnotation(Span here, String term, String orig, TermOccurrences termOccurrences) {
         if (termOccurrences.termCountBelowLimit(term)) {
-            here.annotate(lowerCaseTermAnnotation(term, orig));
+            here.annotate(termAnnotation(term, orig));
         }
     }
 
@@ -127,21 +126,20 @@ public class LinguisticsAnnotator {
         }
         if (mode == StemMode.ALL) {
             Span where = parent.span((int)token.getOffset(), token.getOrig().length());
-            String lowercasedOrig = toLowerCase(token.getOrig());
             addAnnotation(where, token.getOrig(), token.getOrig(), termOccurrences);
 
-            String lowercasedTerm = lowercasedOrig;
+            String lowercasedOrig = toLowerCase(token.getOrig());
+            String termOrIfNullOrig = lowercasedOrig;
             String term = token.getTokenString();
             if (term != null) {
-                lowercasedTerm = toLowerCase(term);
+                termOrIfNullOrig = term;
             }
-            if (! lowercasedOrig.equals(lowercasedTerm)) {
+            if (! lowercasedOrig.equals(termOrIfNullOrig)) {
                 addAnnotation(where, term, token.getOrig(), termOccurrences);
             }
             for (int i = 0; i < token.getNumStems(); i++) {
                 String stem = token.getStem(i);
-                String lowercasedStem = toLowerCase(stem);
-                if (! (lowercasedOrig.equals(lowercasedStem) || lowercasedTerm.equals(lowercasedStem))) {
+                if (! (lowercasedOrig.equals(stem) || termOrIfNullOrig.equals(stem))) {
                     addAnnotation(where, stem, token.getOrig(), termOccurrences);
                 }
             }
@@ -149,7 +147,7 @@ public class LinguisticsAnnotator {
             String term = token.getTokenString();
             if (term == null || term.trim().isEmpty()) return;
             if (termOccurrences.termCountBelowLimit(term))  {
-                parent.span((int)token.getOffset(), token.getOrig().length()).annotate(lowerCaseTermAnnotation(term, token.getOrig()));
+                parent.span((int)token.getOffset(), token.getOrig().length()).annotate(termAnnotation(term, token.getOrig()));
             }
         }
     }
diff --git a/indexinglanguage/src/test/java/com/yahoo/vespa/indexinglanguage/expressions/NGramTestCase.java b/indexinglanguage/src/test/java/com/yahoo/vespa/indexinglanguage/expressions/NGramTestCase.java
index bcde8751de8..b4e266ab3eb 100644
--- a/indexinglanguage/src/test/java/com/yahoo/vespa/indexinglanguage/expressions/NGramTestCase.java
+++ b/indexinglanguage/src/test/java/com/yahoo/vespa/indexinglanguage/expressions/NGramTestCase.java
@@ -57,8 +57,8 @@ public class NGramTestCase {
         new NGramExpression(new SimpleLinguistics(), 3).execute(context);
 
         StringFieldValue value = (StringFieldValue)context.getValue();
-        assertEquals("Grams are pure annotations - field value is unchanged", "en gul Bille sang... ",
-                     value.getString());
+        assertEquals("Grams are pure annotations - field value is unchanged",
+                     "en gul Bille sang... ", value.getString());
         SpanTree gramTree = value.getSpanTree(SpanTrees.LINGUISTICS);
         assertNotNull(gramTree);
         SpanList grams = (SpanList)gramTree.getRoot();
diff --git a/indexinglanguage/src/test/java/com/yahoo/vespa/indexinglanguage/linguistics/LinguisticsAnnotatorTestCase.java b/indexinglanguage/src/test/java/com/yahoo/vespa/indexinglanguage/linguistics/LinguisticsAnnotatorTestCase.java
index 67bff3843ee..7f7e039cb9b 100644
--- a/indexinglanguage/src/test/java/com/yahoo/vespa/indexinglanguage/linguistics/LinguisticsAnnotatorTestCase.java
+++ b/indexinglanguage/src/test/java/com/yahoo/vespa/indexinglanguage/linguistics/LinguisticsAnnotatorTestCase.java
@@ -90,7 +90,7 @@ public class LinguisticsAnnotatorTestCase {
                 if (!specialToken && !type.isIndexable()) {
                     continue;
                 }
-                assertAnnotations(expected, "foo", newToken("foo", "BAR", type, specialToken));
+                assertAnnotations(expected, "foo", newToken("foo", "bar", type, specialToken));
             }
         }
     }
diff --git a/linguistics/src/main/java/com/yahoo/language/LinguisticsCase.java b/linguistics/src/main/java/com/yahoo/language/LinguisticsCase.java
index 5ad6a382abd..f0439a21fec 100644
--- a/linguistics/src/main/java/com/yahoo/language/LinguisticsCase.java
+++ b/linguistics/src/main/java/com/yahoo/language/LinguisticsCase.java
@@ -26,6 +26,7 @@ public class LinguisticsCase {
     public static String toLowerCase(String in) {
         // def is picked from http://docs.oracle.com/javase/6/docs/api/java/lang/String.html#toLowerCase%28%29
         // Also, at the time of writing, English is the default language for queries
+        if (in == null) return null;
         return Lowercase.toLowerCase(in);
     }
 
diff --git a/linguistics/src/main/java/com/yahoo/language/process/GramSplitter.java b/linguistics/src/main/java/com/yahoo/language/process/GramSplitter.java
index 33f5ee7e4bb..9178c2d7e09 100644
--- a/linguistics/src/main/java/com/yahoo/language/process/GramSplitter.java
+++ b/linguistics/src/main/java/com/yahoo/language/process/GramSplitter.java
@@ -189,9 +189,8 @@ public class GramSplitter {
         @Override
         public boolean equals(Object o) {
             if (this == o) return true;
-            if ( ! (o instanceof Gram)) return false;
+            if ( ! (o instanceof Gram gram)) return false;
 
-            Gram gram = (Gram)o;
             if (codePointCount != gram.codePointCount) return false;
             if (start != gram.start) return false;
             return true;