Fix sentence detection initialisation

lfoppiano · Jun 19, 2023 · edf8bff · edf8bff
1 parent cd16548
commit edf8bff
Show file tree

Hide file tree

Showing 2 changed files with 8 additions and 9 deletions.
diff --git a/src/main/java/org/grobid/core/engines/QuantityParser.java b/src/main/java/org/grobid/core/engines/QuantityParser.java
@@ -1,6 +1,7 @@
 package org.grobid.core.engines;
 
 import com.google.common.collect.Iterables;
+import jakarta.inject.Inject;
 import org.apache.commons.collections4.CollectionUtils;
 import org.apache.commons.lang3.StringUtils;
 import org.apache.commons.lang3.tuple.Pair;
@@ -17,8 +18,6 @@
 import org.grobid.core.engines.label.TaggingLabel;
 import org.grobid.core.exceptions.GrobidException;
 import org.grobid.core.features.FeaturesVectorQuantities;
-import org.grobid.core.lang.SentenceDetector;
-import org.grobid.core.lang.impl.OpenNLPSentenceDetector;
 import org.grobid.core.layout.BoundingBox;
 import org.grobid.core.layout.LayoutToken;
 import org.grobid.core.lexicon.QuantityLexicon;
@@ -28,7 +27,6 @@
 import org.slf4j.Logger;
 import org.slf4j.LoggerFactory;
 
-import jakarta.inject.Inject;
 import java.math.BigDecimal;
 import java.util.ArrayList;
 import java.util.Arrays;
@@ -181,9 +179,10 @@ public List<Measurement> process(List<LayoutToken> layoutTokens) {
     }
 
     protected List<OffsetPosition> getSentencesOffsets(List<LayoutToken> tokens) {
-        SentenceDetector segmenter = new OpenNLPSentenceDetector();
+        SentenceUtilities segmenter = SentenceUtilities.getInstance();
+
         String text = LayoutTokensUtil.toText(tokens);
-        List<OffsetPosition> results = segmenter.detect(text);
+        List<OffsetPosition> results = segmenter.runSentenceDetection(text);
 
         if (CollectionUtils.isEmpty(results)) {
             results = Arrays.asList(new OffsetPosition(0, text.length()));

diff --git a/src/main/java/org/grobid/core/utilities/TextParser.java b/src/main/java/org/grobid/core/utilities/TextParser.java
@@ -57,7 +57,7 @@ public class TextParser {
     private AbstractPredIdentifier predicater = null;
     private AbstractSRLabeler labeler = null;
     private DEPReader depReader = null;
-    private SentenceDetector segmenter;
+    private SentenceUtilities segmenter;
 
     // this is for version 1.3.0 of ClearNLP
     private CRolesetClassifier roleClassifier = null;
@@ -120,7 +120,7 @@ private void init(String dictionaryFile, String posModelFile, String depModelFil
 
         depReader = new DEPReader(0, 1, 2, 3, 4, 5, 6);
 
-        segmenter = new OpenNLPSentenceDetector();
+        segmenter = SentenceUtilities.getInstance();
     }
 
     /**
@@ -177,7 +177,7 @@ public synchronized List<Sentence> parseText(String text) throws GrobidException
         }
 
         List<Sentence> results = new ArrayList<>();
-        List<OffsetPosition> sentences = this.segmenter.detect(text);
+        List<OffsetPosition> sentences = this.segmenter.runSentenceDetection(text);
 
         if (CollectionUtils.isEmpty(sentences)) {
             // there is some text but not in a state so that a sentence at least can be
@@ -219,7 +219,7 @@ public List<Sentence> parse(BufferedReader reader) throws GrobidException {
 
         String text = reader.lines().collect(Collectors.joining());
 
-        List<OffsetPosition> sentences = segmenter.detect(text);
+        List<OffsetPosition> sentences = segmenter.runSentenceDetection(text);
 
         for (OffsetPosition sentencePosition : sentences) {
             String sentence = text.substring(sentencePosition.start, sentencePosition.end);