diff options
author | Jon Bratseth <bratseth@gmail.com> | 2021-05-05 12:12:25 +0200 |
---|---|---|
committer | Jon Bratseth <bratseth@gmail.com> | 2021-05-05 12:12:25 +0200 |
commit | 4b3e615954fa5cbaf761a55cddba786f51ab26f3 (patch) | |
tree | 31cc973d490bdc7b8257c664b31018ac8915efc4 /linguistics/src/main/java/com/yahoo/language/simple/SimpleLinguistics.java | |
parent | f8c8b6ab958f930b318eadc92b760dd26f414e58 (diff) |
Revert "Merge pull request #17746 from vespa-engine/revert-17738-revert-17737-revert-17736-bratseth/special-tokens"
This reverts commit 491856b396d003885e159345fe3f533f0fa35933, reversing
changes made to 3720186303f4aef1d185525eaf61092097a64ec9.
Diffstat (limited to 'linguistics/src/main/java/com/yahoo/language/simple/SimpleLinguistics.java')
-rw-r--r-- | linguistics/src/main/java/com/yahoo/language/simple/SimpleLinguistics.java | 7 |
1 files changed, 6 insertions, 1 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; } |