Remove abbreviations & Cache

Author: dfuchss

framework/common/pom.xml

@@ -40,16 +40,6 @@
       <groupId>org.apache.commons</groupId>
       <artifactId>commons-text</artifactId>
     </dependency>
-    <dependency>
-      <groupId>org.apache.jena</groupId>
-      <artifactId>jena-arq</artifactId>
-      <version>4.10.0</version>
-    </dependency>
-    <dependency>
-      <groupId>org.apache.jena</groupId>
-      <artifactId>jena-core</artifactId>
-      <version>4.10.0</version>
-    </dependency>
     <dependency>
       <groupId>org.apache.opennlp</groupId>
       <artifactId>opennlp-tools</artifactId>
@@ -72,11 +62,6 @@
       <artifactId>jgrapht-core</artifactId>
       <version>${jgrapht.version}</version>
     </dependency>
-    <dependency>
-      <groupId>org.jsoup</groupId>
-      <artifactId>jsoup</artifactId>
-      <version>1.17.2</version>
-    </dependency>
     <dependency>
       <groupId>org.xerial</groupId>
       <artifactId>sqlite-jdbc</artifactId>

framework/common/src/main/java/edu/kit/kastel/mcse/ardoco/core/api/stage/textextraction/PhraseAbbreviation.java

@@ -1,10 +1,7 @@
 /* Licensed under MIT 2021-2024. */
 package edu.kit.kastel.mcse.ardoco.core.api.stage.textextraction;
 
-import java.util.Optional;
-
 import org.eclipse.collections.api.factory.Lists;
-import org.eclipse.collections.api.factory.SortedSets;
 import org.eclipse.collections.api.list.ImmutableList;
 import org.eclipse.collections.api.list.MutableList;
 import org.eclipse.collections.api.map.sorted.ImmutableSortedMap;
@@ -15,7 +12,6 @@
 import edu.kit.kastel.mcse.ardoco.core.api.text.Phrase;
 import edu.kit.kastel.mcse.ardoco.core.api.text.Word;
 import edu.kit.kastel.mcse.ardoco.core.common.tuple.Pair;
-import edu.kit.kastel.mcse.ardoco.core.common.util.AbbreviationDisambiguationHelper;
 import edu.kit.kastel.mcse.ardoco.core.configuration.IConfigurable;
 import edu.kit.kastel.mcse.ardoco.core.data.Confidence;
 import edu.kit.kastel.mcse.ardoco.core.data.PipelineStepData;
@@ -42,7 +38,7 @@ public interface TextState extends IConfigurable, PipelineStepData {
      * @param probability probability to be a name mapping
      */
     default NounMapping addNounMapping(Word word, MappingKind kind, Claimant claimant, double probability) {
-        return getTextStateStrategy().addOrExtendNounMapping(word, kind, claimant, probability, Lists.immutable.with(word.getText()));
+        return this.getTextStateStrategy().addOrExtendNounMapping(word, kind, claimant, probability, Lists.immutable.with(word.getText()));
     }
 
     /**
@@ -56,7 +52,7 @@ default NounMapping addNounMapping(Word word, MappingKind kind, Claimant claiman
      * @param surfaceForms list of the appearances of the mapping
      */
     default NounMapping addNounMapping(Word word, MappingKind kind, Claimant claimant, double probability, ImmutableList<String> surfaceForms) {
-        return getTextStateStrategy().addOrExtendNounMapping(word, kind, claimant, probability, surfaceForms);
+        return this.getTextStateStrategy().addOrExtendNounMapping(word, kind, claimant, probability, surfaceForms);
     }
 
     /**
@@ -74,7 +70,7 @@ default NounMapping addNounMapping(Word word, MappingKind kind, Claimant claiman
      */
     default NounMapping addNounMapping(ImmutableSortedSet<Word> words, MappingKind kind, Claimant claimant, double probability,
             ImmutableList<Word> referenceWords, ImmutableList<String> surfaceForms, String reference) {
-        return getTextStateStrategy().addNounMapping(words, kind, claimant, probability, referenceWords, surfaceForms, reference);
+        return this.getTextStateStrategy().addNounMapping(words, kind, claimant, probability, referenceWords, surfaceForms, reference);
     }
 
     /**
@@ -90,7 +86,7 @@ default NounMapping addNounMapping(ImmutableSortedSet<Word> words, MappingKind k
      */
     default NounMapping addNounMapping(ImmutableSortedSet<Word> words, ImmutableSortedMap<MappingKind, Confidence> distribution,
             ImmutableList<Word> referenceWords, ImmutableList<String> surfaceForms, String reference) {
-        return getTextStateStrategy().addNounMapping(words, distribution, referenceWords, surfaceForms, reference);
+        return this.getTextStateStrategy().addNounMapping(words, distribution, referenceWords, surfaceForms, reference);
     }
 
     /**
@@ -137,7 +133,7 @@ default NounMapping addNounMapping(ImmutableSortedSet<Word> words, ImmutableSort
      * @param phrase the phrase
      */
     default ImmutableList<PhraseMapping> getPhraseMappings(Phrase phrase) {
-        return Lists.immutable.fromStream(getPhraseMappings().stream().filter(pm -> pm.getPhrases().contains(phrase)));
+        return Lists.immutable.fromStream(this.getPhraseMappings().stream().filter(pm -> pm.getPhrases().contains(phrase)));
     }
 
     ImmutableList<NounMapping> getNounMappingsOfKind(MappingKind mappingKind);
@@ -166,78 +162,4 @@ void mergePhraseMappingsAndNounMappings(PhraseMapping phraseMapping, PhraseMappi
     boolean isWordContainedByMappingKind(Word word, MappingKind kind);
 
     ImmutableList<NounMapping> getNounMappingsWithSimilarReference(String reference);
-
-    /**
-     * Adds a {@link WordAbbreviation} for the abbreviation with the specified word meaning to the state using the state's {@link TextStateStrategy}.
-     *
-     * @param abbreviation the abbreviation
-     * @param word         the meaning
-     * @return the resulting word abbreviation in the state
-     */
-    default WordAbbreviation addWordAbbreviation(String abbreviation, Word word) {
-        logger.debug("Added word abbreviation for {} with meaning {}", abbreviation, word.getText());
-        var wordAbbreviation = getTextStateStrategy().addOrExtendWordAbbreviation(abbreviation, word);
-        AbbreviationDisambiguationHelper.addTransient(wordAbbreviation);
-        return wordAbbreviation;
-    }
-
-    /**
-     * Adds a {@link PhraseAbbreviation} for the abbreviation with the specified phrase meaning to the state using the state's {@link TextStateStrategy}.
-     *
-     * @param abbreviation the abbrevation
-     * @param phrase       the meaning
-     * @return the resulting phrase abbreviation in the state
-     */
-    default PhraseAbbreviation addPhraseAbbreviation(String abbreviation, Phrase phrase) {
-        logger.debug("Added phrase abbreviation for {} with meaning {}", abbreviation, phrase.getText());
-        var phraseAbbreviation = getTextStateStrategy().addOrExtendPhraseAbbreviation(abbreviation, phrase);
-        AbbreviationDisambiguationHelper.addTransient(phraseAbbreviation);
-        return phraseAbbreviation;
-    }
-
-    /**
-     * {@return all word abbreviations from the state}
-     */
-    ImmutableSortedSet<WordAbbreviation> getWordAbbreviations();
-
-    /**
-     * {@return all word abbreviations from the state that have the specified meaning}
-     *
-     * @param word the meaning to search for
-     */
-    default ImmutableSortedSet<WordAbbreviation> getWordAbbreviations(Word word) {
-        return SortedSets.immutable.ofAll(getWordAbbreviations().stream().filter(w -> w.getWords().contains(word)).toList());
-    }
-
-    /**
-     * {@return the optional word abbreviation from the state that has the specified abbreviation}
-     *
-     * @param abbreviation the abbreviation
-     */
-    default Optional<WordAbbreviation> getWordAbbreviation(String abbreviation) {
-        return getWordAbbreviations().stream().filter(w -> w.getAbbreviation().equals(abbreviation)).findFirst();
-    }
-
-    /**
-     * {@return all phrase abbreviations from the state}
-     */
-    ImmutableSortedSet<PhraseAbbreviation> getPhraseAbbreviations();
-
-    /**
-     * {@return all phrase abbreviations from the state that have the specified meaning}
-     *
-     * @param phrase the meaning to search for
-     */
-    default ImmutableSortedSet<PhraseAbbreviation> getPhraseAbbreviations(Phrase phrase) {
-        return SortedSets.immutable.ofAll(getPhraseAbbreviations().stream().filter(p -> p.getPhrases().contains(phrase)).toList());
-    }
-
-    /**
-     * {@return the optional phrase abbreviation from the state that has the specified abbreviation}
-     *
-     * @param abbreviation the abbreviation
-     */
-    default Optional<PhraseAbbreviation> getPhraseAbbreviation(String abbreviation) {
-        return getPhraseAbbreviations().stream().filter(p -> p.getAbbreviation().equals(abbreviation)).findFirst();
-    }
 }

framework/common/src/main/java/edu/kit/kastel/mcse/ardoco/core/api/stage/textextraction/TextState.java

@@ -9,7 +9,6 @@
 import org.eclipse.collections.api.map.sorted.ImmutableSortedMap;
 import org.eclipse.collections.api.set.sorted.ImmutableSortedSet;
 
-import edu.kit.kastel.mcse.ardoco.core.api.text.Phrase;
 import edu.kit.kastel.mcse.ardoco.core.api.text.Word;
 import edu.kit.kastel.mcse.ardoco.core.common.AggregationFunctions;
 import edu.kit.kastel.mcse.ardoco.core.data.Confidence;
@@ -120,25 +119,5 @@ default String calculateNounMappingReference(ImmutableList<Word> referenceWords)
         return refBuilder.toString();
     }
 
-    /**
-     * Tries to add a word abbreviation to the state. If the abbreviation already exists, it is extended.
-     *
-     * @param abbreviation the abbreviation
-     * @param word         the word
-     * @return the resulting {@link edu.kit.kastel.mcse.ardoco.core.api.Disambiguation} in the stage
-     */
-
-    WordAbbreviation addOrExtendWordAbbreviation(String abbreviation, Word word);
-
-    /**
-     * Tries to add a phrase abbreviation to the state. If the abbreviation already exists, it is extended.
-     *
-     * @param abbreviation the abbreviation
-     * @param phrase       the phrase
-     * @return the resulting {@link edu.kit.kastel.mcse.ardoco.core.api.Disambiguation} in the stage
-     */
-
-    PhraseAbbreviation addOrExtendPhraseAbbreviation(String abbreviation, Phrase phrase);
-
     ImmutableList<NounMapping> getNounMappingsWithSimilarReference(String reference);
 }

framework/common/src/main/java/edu/kit/kastel/mcse/ardoco/core/api/stage/textextraction/TextStateStrategy.java

@@ -2,20 +2,19 @@
 package edu.kit.kastel.mcse.ardoco.core.common.tuple;
 
 import java.io.Serializable;
-
-import com.github.jsonldjava.shaded.com.google.common.base.Objects;
+import java.util.Objects;
 
 public record Pair<T extends Serializable, U extends Serializable>(T first, U second) implements Serializable {
 
     public boolean hasElement(Serializable element) {
-        return Objects.equal(this.first, element) || Objects.equal(this.second, element);
+        return Objects.equals(this.first, element) || Objects.equals(this.second, element);
     }
 
     public Serializable getOtherElement(Serializable element) {
-        if (Objects.equal(this.first, element)) {
+        if (Objects.equals(this.first, element)) {
             return this.second;
         }
-        if (Objects.equal(this.second, element)) {
+        if (Objects.equals(this.second, element)) {
             return this.first;
         }
         throw new IllegalArgumentException("Unknown element: " + element);