Menu
▾
▴
[DL-Learner SVN] SF.net SVN: dl-learner:[4208] trunk/components-core/src/main/java/org/ dllearner/algorithms/isle
|
From: <dfl...@us...> - 2013-12-10 15:41:39
|
Revision: 4208
http://sourceforge.net/p/dl-learner/code/4208
Author: dfleischhacker
Date: 2013-12-10 15:41:36 +0000 (Tue, 10 Dec 2013)
Log Message:
-----------
Adapt WSD interfaces to scored candidates
Modified Paths:
--------------
trunk/components-core/src/main/java/org/dllearner/algorithms/isle/EntityCandidateGenerator.java
trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/EntityCandidatesTrie.java
trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/SemanticAnnotator.java
trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/SimpleEntityCandidateGenerator.java
trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/SimpleEntityCandidatesTrie.java
trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/TrieEntityCandidateGenerator.java
trunk/components-core/src/main/java/org/dllearner/algorithms/isle/wsd/RandomWordSenseDisambiguation.java
trunk/components-core/src/main/java/org/dllearner/algorithms/isle/wsd/SimpleWordSenseDisambiguation.java
trunk/components-core/src/main/java/org/dllearner/algorithms/isle/wsd/StructureBasedWordSenseDisambiguation.java
trunk/components-core/src/main/java/org/dllearner/algorithms/isle/wsd/WordSenseDisambiguation.java
Modified: trunk/components-core/src/main/java/org/dllearner/algorithms/isle/EntityCandidateGenerator.java
===================================================================
--- trunk/components-core/src/main/java/org/dllearner/algorithms/isle/EntityCandidateGenerator.java 2013-12-10 15:25:13 UTC (rev 4207)
+++ trunk/components-core/src/main/java/org/dllearner/algorithms/isle/EntityCandidateGenerator.java 2013-12-10 15:41:36 UTC (rev 4208)
@@ -3,13 +3,14 @@
*/
package org.dllearner.algorithms.isle;
-import java.util.HashMap;
-import java.util.Set;
-
import org.dllearner.algorithms.isle.index.Annotation;
+import org.dllearner.algorithms.isle.index.EntityScorePair;
import org.dllearner.core.owl.Entity;
import org.semanticweb.owlapi.model.OWLOntology;
+import java.util.HashMap;
+import java.util.Set;
+
/**
* @author Lorenz Buehmann
*
@@ -22,8 +23,8 @@
this.ontology = ontology;
}
- public abstract Set<Entity> getCandidates(Annotation annotation);
+ public abstract Set<EntityScorePair> getCandidates(Annotation annotation);
- public abstract HashMap<Annotation,Set<Entity>> getCandidatesMap(Set<Annotation> annotations);
+ public abstract HashMap<Annotation,Set<EntityScorePair>> getCandidatesMap(Set<Annotation> annotations);
}
Modified: trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/EntityCandidatesTrie.java
===================================================================
--- trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/EntityCandidatesTrie.java 2013-12-10 15:25:13 UTC (rev 4207)
+++ trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/EntityCandidatesTrie.java 2013-12-10 15:41:36 UTC (rev 4208)
@@ -17,10 +17,9 @@
/**
* Gets set of candidate entities for a list of tokens
- * @param s
* @return
*/
- public Set<Entity> getCandidateEntities(List<Token> tokens);
+ public Set<EntityScorePair> getCandidateEntities(List<Token> tokens);
/**
@@ -28,14 +27,12 @@
* ontology string when the parameter string has been added to the trie after generation by using
* WordNet or other additional methods.
*
- * @param s the string to search in the trie
* @return string generating the path of the longest match in the trie
*/
public List<Token> getGeneratingStringForLongestMatch(List<Token> tokens);
/**
* Gets the longest matching string
- * @param s
* @return
*/
public List<Token> getLongestMatchingText(List<Token> tokens);
Modified: trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/SemanticAnnotator.java
===================================================================
--- trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/SemanticAnnotator.java 2013-12-10 15:25:13 UTC (rev 4207)
+++ trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/SemanticAnnotator.java 2013-12-10 15:41:36 UTC (rev 4208)
@@ -1,13 +1,12 @@
package org.dllearner.algorithms.isle.index;
+import org.dllearner.algorithms.isle.EntityCandidateGenerator;
+import org.dllearner.algorithms.isle.wsd.WordSenseDisambiguation;
+
import java.util.HashMap;
import java.util.HashSet;
import java.util.Set;
-import org.dllearner.algorithms.isle.EntityCandidateGenerator;
-import org.dllearner.algorithms.isle.wsd.WordSenseDisambiguation;
-import org.dllearner.core.owl.Entity;
-
/**
* Provides methods to annotate documents.
*
@@ -23,7 +22,6 @@
/**
* Initialize this semantic annotator to use the entities from the provided ontology.
*
- * @param ontology the ontology to use entities from
*/
public SemanticAnnotator(WordSenseDisambiguation wordSenseDisambiguation,
EntityCandidateGenerator entityCandidateGenerator, LinguisticAnnotator linguisticAnnotator) {
@@ -41,9 +39,9 @@
public AnnotatedDocument processDocument(TextDocument document){
Set<Annotation> annotations = linguisticAnnotator.annotate(document);
Set<SemanticAnnotation> semanticAnnotations = new HashSet<SemanticAnnotation>();
- HashMap<Annotation,Set<Entity>> candidatesMap = entityCandidateGenerator.getCandidatesMap(annotations);
+ HashMap<Annotation, Set<EntityScorePair>> candidatesMap = entityCandidateGenerator.getCandidatesMap(annotations);
for (Annotation annotation : candidatesMap.keySet()) {
- Set<Entity> candidateEntities = candidatesMap.get(annotation);
+ Set<EntityScorePair> candidateEntities = candidatesMap.get(annotation);
if (candidateEntities == null || candidateEntities.size() == 0) {
continue;
}
Modified: trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/SimpleEntityCandidateGenerator.java
===================================================================
--- trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/SimpleEntityCandidateGenerator.java 2013-12-10 15:25:13 UTC (rev 4207)
+++ trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/SimpleEntityCandidateGenerator.java 2013-12-10 15:41:36 UTC (rev 4208)
@@ -3,16 +3,16 @@
*/
package org.dllearner.algorithms.isle.index;
-import java.util.HashMap;
-import java.util.HashSet;
-import java.util.Set;
-
import org.dllearner.algorithms.isle.EntityCandidateGenerator;
import org.dllearner.core.owl.Entity;
import org.dllearner.utilities.owl.OWLAPIConverter;
import org.semanticweb.owlapi.model.OWLEntity;
import org.semanticweb.owlapi.model.OWLOntology;
+import java.util.HashMap;
+import java.util.HashSet;
+import java.util.Set;
+
/**
* @author Lorenz Buehmann
*
@@ -36,13 +36,17 @@
* @see org.dllearner.algorithms.isle.EntityCandidateGenerator#getCandidates(org.dllearner.algorithms.isle.index.Annotation)
*/
@Override
- public Set<Entity> getCandidates(Annotation annotation) {
- return allEntities;
- }
+ public Set<EntityScorePair> getCandidates(Annotation annotation) {
+ HashSet<EntityScorePair> result = new HashSet<>();
+ for (Entity e : allEntities) {
+ result.add(new EntityScorePair(e, 1.0));
+ }
+ return result;
+ }
@Override
- public HashMap<Annotation, Set<Entity>> getCandidatesMap(Set<Annotation> annotations) {
- HashMap<Annotation, Set<Entity>> result = new HashMap<Annotation, Set<Entity>>();
+ public HashMap<Annotation, Set<EntityScorePair>> getCandidatesMap(Set<Annotation> annotations) {
+ HashMap<Annotation, Set<EntityScorePair>> result = new HashMap<Annotation, Set<EntityScorePair>>();
for (Annotation annotation: annotations)
result.put(annotation, getCandidates(annotation));
Modified: trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/SimpleEntityCandidatesTrie.java
===================================================================
--- trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/SimpleEntityCandidatesTrie.java 2013-12-10 15:25:13 UTC (rev 4207)
+++ trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/SimpleEntityCandidatesTrie.java 2013-12-10 15:41:36 UTC (rev 4208)
@@ -1,160 +1,156 @@
-package org.dllearner.algorithms.isle.index;
-
-import net.didion.jwnl.data.POS;
-import org.dllearner.algorithms.isle.WordNet;
-import org.dllearner.algorithms.isle.textretrieval.EntityTextRetriever;
-import org.dllearner.core.owl.Entity;
-import org.semanticweb.owlapi.model.OWLOntology;
-
-import java.util.*;
-import java.util.Map.Entry;
-
-public class SimpleEntityCandidatesTrie implements EntityCandidatesTrie {
- TokenTree tree;
- EntityTextRetriever entityTextRetriever;
-
-// /**
-// * Initialize the trie with strings from the provided ontology using a no-op name generator, i.e., only the
-// * actual ontology strings are added and no expansion is done.
-// *
-// * @param entityTextRetriever the text retriever to use
-// * @param ontology the ontology to get strings from
-// */
-// public SimpleEntityCandidatesTrie(EntityTextRetriever entityTextRetriever, OWLOntology ontology) {
-// this(entityTextRetriever, ontology, new DummyNameGenerator());
-// }
-
- /**
- * Initialize the trie with strings from the provided ontology and use the given entity name generator
- * for generating alternative words.
- *
- * @param entityTextRetriever the text retriever to use
- * @param ontology the ontology to get strings from
- */
- public SimpleEntityCandidatesTrie(EntityTextRetriever entityTextRetriever, OWLOntology ontology) {
- this.entityTextRetriever = entityTextRetriever;
- buildTrie(ontology);
- }
-
- public void buildTrie(OWLOntology ontology) {
- this.tree = new TokenTree();
- Map<Entity, Set<List<Token>>> entity2TokenSet = entityTextRetriever.getRelevantText(ontology);
-
-
- for (Entry<Entity, Set<List<Token>>> entry : entity2TokenSet.entrySet()) {
- Entity entity = entry.getKey();
- Set<List<Token>> tokenSet = entry.getValue();
- for (List<Token> tokens : tokenSet) {
- addAlternativeFormsFromWordNet(tokens);
- addEntry(tokens, entity);
- addSubsequences(entity, tokens);
- }
- }
- }
-
- /**
- * Adds the subsequences of a test
- * @param entity
- * @param tokens
- */
- private void addSubsequences(Entity entity, List<Token> tokens) {
- tree.add(tokens, entity);
- for (int size = 1; size < tokens.size(); size++) {
- for (int start = 0; start < tokens.size() - size + 1; start++) {
- ArrayList<Token> subsequence = new ArrayList<>();
- for (int i = 0; i < size; i++) {
- subsequence.add(tokens.get(start + i));
- }
- addEntry(subsequence, entity);
- }
- }
- }
-
- private void addAlternativeFormsFromWordNet(List<Token> tokens) {
- for (Token t : tokens) {
- POS wordnetPos = null;
- String posTag = t.getPOSTag();
- if (posTag.startsWith("N")) {//nouns
- wordnetPos = POS.NOUN;
- }
- else if (posTag.startsWith("V")) {//verbs
- wordnetPos = POS.VERB;
- }
- else if (posTag.startsWith("J")) {//adjectives
- wordnetPos = POS.ADJECTIVE;
- }
- else if (posTag.startsWith("R")) {//adverbs
- wordnetPos = POS.ADVERB;
- }
- if (wordnetPos == null) {
- continue;
- }
- //String[] synonyms = LinguisticUtil.getInstance().getSynonymsForWord(t.getRawForm(), wordnetPos);
- Set<WordNet.LemmaScorePair> alternativeFormPairs = LinguisticUtil.getInstance()
- .getScoredHyponyms(t.getRawForm(), wordnetPos);
-
- for (WordNet.LemmaScorePair synonym : alternativeFormPairs) {
- // ignore all multi word synonyms
- if (synonym.getLemma().contains("_")) {
- continue;
- }
- //t.addAlternativeForm(LinguisticUtil.getInstance().getNormalizedForm(synonym));
- t.addAlternativeForm(synonym.getLemma(), synonym.getScore());
- }
- }
- }
-
- @Override
- public void addEntry(List<Token> s, Entity e) {
- tree.add(s, e);
- }
-
- public void addEntry(List<Token> s, Entity e, List<Token> originalTokens) {
- tree.add(s, e, originalTokens);
- }
-
- @Override
- public Set<Entity> getCandidateEntities(List<Token> tokens) {
- Set<Entity> res = tree.getAllEntities(tokens);
- System.out.println("Unscored: " + res);
- Set<EntityScorePair> scored = tree.getAllEntitiesScored(tokens);
- System.out.println("Scored: " + scored);
-
- return res;
- }
-
- @Override
- public List<Token> getGeneratingStringForLongestMatch(List<Token> tokens) {
- return tree.getOriginalTokensForLongestMatch(tokens);
- }
-
- @Override
- public List<Token> getLongestMatchingText(List<Token> tokens) {
- return tree.getLongestMatch(tokens);
- }
-
- public String toString() {
- return tree.toString();
- }
-
- public static void main(String[] args) {
- String[] tokens = "this is a long and very complex text".split(" ");
-
- List<String>[] wordnetTokens = (ArrayList<String>[]) new ArrayList[tokens.length];
-
- // generate list of lemmatized wordnet synonyms for each token
- for (int i = 0; i < tokens.length; i++) {
- wordnetTokens[i] = new ArrayList<String>();
- wordnetTokens[i].add(LinguisticUtil.getInstance().getNormalizedForm(tokens[i]));
- for (String w : LinguisticUtil.getInstance().getTopSynonymsForWord(tokens[i], 5)) {
- System.out.println("Adding: " + LinguisticUtil.getInstance().getNormalizedForm(w));
- wordnetTokens[i].add(LinguisticUtil.getInstance().getNormalizedForm(w).replaceAll("_", " "));
- }
- }
- }
-
- public void printTrie() {
- System.out.println(this.toString());
-
- }
-}
+package org.dllearner.algorithms.isle.index;
+
+import net.didion.jwnl.data.POS;
+import org.dllearner.algorithms.isle.WordNet;
+import org.dllearner.algorithms.isle.textretrieval.EntityTextRetriever;
+import org.dllearner.core.owl.Entity;
+import org.semanticweb.owlapi.model.OWLOntology;
+
+import java.util.*;
+import java.util.Map.Entry;
+
+public class SimpleEntityCandidatesTrie implements EntityCandidatesTrie {
+ TokenTree tree;
+ EntityTextRetriever entityTextRetriever;
+
+// /**
+// * Initialize the trie with strings from the provided ontology using a no-op name generator, i.e., only the
+// * actual ontology strings are added and no expansion is done.
+// *
+// * @param entityTextRetriever the text retriever to use
+// * @param ontology the ontology to get strings from
+// */
+// public SimpleEntityCandidatesTrie(EntityTextRetriever entityTextRetriever, OWLOntology ontology) {
+// this(entityTextRetriever, ontology, new DummyNameGenerator());
+// }
+
+ /**
+ * Initialize the trie with strings from the provided ontology and use the given entity name generator
+ * for generating alternative words.
+ *
+ * @param entityTextRetriever the text retriever to use
+ * @param ontology the ontology to get strings from
+ */
+ public SimpleEntityCandidatesTrie(EntityTextRetriever entityTextRetriever, OWLOntology ontology) {
+ this.entityTextRetriever = entityTextRetriever;
+ buildTrie(ontology);
+ }
+
+ public void buildTrie(OWLOntology ontology) {
+ this.tree = new TokenTree();
+ Map<Entity, Set<List<Token>>> entity2TokenSet = entityTextRetriever.getRelevantText(ontology);
+
+
+ for (Entry<Entity, Set<List<Token>>> entry : entity2TokenSet.entrySet()) {
+ Entity entity = entry.getKey();
+ Set<List<Token>> tokenSet = entry.getValue();
+ for (List<Token> tokens : tokenSet) {
+ addAlternativeFormsFromWordNet(tokens);
+ addEntry(tokens, entity);
+ addSubsequences(entity, tokens);
+ }
+ }
+ }
+
+ /**
+ * Adds the subsequences of a test
+ * @param entity
+ * @param tokens
+ */
+ private void addSubsequences(Entity entity, List<Token> tokens) {
+ tree.add(tokens, entity);
+ for (int size = 1; size < tokens.size(); size++) {
+ for (int start = 0; start < tokens.size() - size + 1; start++) {
+ ArrayList<Token> subsequence = new ArrayList<>();
+ for (int i = 0; i < size; i++) {
+ subsequence.add(tokens.get(start + i));
+ }
+ addEntry(subsequence, entity);
+ }
+ }
+ }
+
+ private void addAlternativeFormsFromWordNet(List<Token> tokens) {
+ for (Token t : tokens) {
+ POS wordnetPos = null;
+ String posTag = t.getPOSTag();
+ if (posTag.startsWith("N")) {//nouns
+ wordnetPos = POS.NOUN;
+ }
+ else if (posTag.startsWith("V")) {//verbs
+ wordnetPos = POS.VERB;
+ }
+ else if (posTag.startsWith("J")) {//adjectives
+ wordnetPos = POS.ADJECTIVE;
+ }
+ else if (posTag.startsWith("R")) {//adverbs
+ wordnetPos = POS.ADVERB;
+ }
+ if (wordnetPos == null) {
+ continue;
+ }
+ //String[] synonyms = LinguisticUtil.getInstance().getSynonymsForWord(t.getRawForm(), wordnetPos);
+ Set<WordNet.LemmaScorePair> alternativeFormPairs = LinguisticUtil.getInstance()
+ .getScoredHyponyms(t.getRawForm(), wordnetPos);
+
+ for (WordNet.LemmaScorePair synonym : alternativeFormPairs) {
+ // ignore all multi word synonyms
+ if (synonym.getLemma().contains("_")) {
+ continue;
+ }
+ //t.addAlternativeForm(LinguisticUtil.getInstance().getNormalizedForm(synonym));
+ t.addAlternativeForm(synonym.getLemma(), synonym.getScore());
+ }
+ }
+ }
+
+ @Override
+ public void addEntry(List<Token> s, Entity e) {
+ tree.add(s, e);
+ }
+
+ public void addEntry(List<Token> s, Entity e, List<Token> originalTokens) {
+ tree.add(s, e, originalTokens);
+ }
+
+ @Override
+ public Set<EntityScorePair> getCandidateEntities(List<Token> tokens) {
+ Set<EntityScorePair> res = tree.getAllEntitiesScored(tokens);
+ return res;
+ }
+
+ @Override
+ public List<Token> getGeneratingStringForLongestMatch(List<Token> tokens) {
+ return tree.getOriginalTokensForLongestMatch(tokens);
+ }
+
+ @Override
+ public List<Token> getLongestMatchingText(List<Token> tokens) {
+ return tree.getLongestMatch(tokens);
+ }
+
+ public String toString() {
+ return tree.toString();
+ }
+
+ public static void main(String[] args) {
+ String[] tokens = "this is a long and very complex text".split(" ");
+
+ List<String>[] wordnetTokens = (ArrayList<String>[]) new ArrayList[tokens.length];
+
+ // generate list of lemmatized wordnet synonyms for each token
+ for (int i = 0; i < tokens.length; i++) {
+ wordnetTokens[i] = new ArrayList<String>();
+ wordnetTokens[i].add(LinguisticUtil.getInstance().getNormalizedForm(tokens[i]));
+ for (String w : LinguisticUtil.getInstance().getTopSynonymsForWord(tokens[i], 5)) {
+ System.out.println("Adding: " + LinguisticUtil.getInstance().getNormalizedForm(w));
+ wordnetTokens[i].add(LinguisticUtil.getInstance().getNormalizedForm(w).replaceAll("_", " "));
+ }
+ }
+ }
+
+ public void printTrie() {
+ System.out.println(this.toString());
+
+ }
+}
Modified: trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/TrieEntityCandidateGenerator.java
===================================================================
--- trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/TrieEntityCandidateGenerator.java 2013-12-10 15:25:13 UTC (rev 4207)
+++ trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/TrieEntityCandidateGenerator.java 2013-12-10 15:41:36 UTC (rev 4208)
@@ -3,7 +3,6 @@
import com.google.common.collect.Lists;
import org.dllearner.algorithms.isle.EntityCandidateGenerator;
import org.dllearner.algorithms.isle.StopWordFilter;
-import org.dllearner.core.owl.Entity;
import org.semanticweb.owlapi.model.OWLOntology;
import java.util.ArrayList;
@@ -27,8 +26,8 @@
this.candidatesTrie = candidatesTrie;
}
- public Set<Entity> getCandidates(Annotation annotation) {
- Set<Entity> candidateEntities = candidatesTrie.getCandidateEntities(annotation.getTokens());
+ public Set<EntityScorePair> getCandidates(Annotation annotation) {
+ Set<EntityScorePair> candidateEntities = candidatesTrie.getCandidateEntities(annotation.getTokens());
System.out.println(annotation + " --> " + candidateEntities);
return candidateEntities;
}
@@ -39,7 +38,7 @@
* @param window : maximum distance between the annotations
* @return
*/
- public void postProcess(HashMap<Annotation,Set<Entity>> candidatesMap, int window, StopWordFilter stopWordFilter) {
+ public void postProcess(HashMap<Annotation,Set<EntityScorePair>> candidatesMap, int window, StopWordFilter stopWordFilter) {
Set<Annotation> annotations = candidatesMap.keySet();
List<Annotation> sortedAnnotations = new ArrayList<Annotation>(annotations);
//TODO refactoring
@@ -119,8 +118,8 @@
}
@Override
- public HashMap<Annotation, Set<Entity>> getCandidatesMap(Set<Annotation> annotations) {
- HashMap<Annotation, Set<Entity>> candidatesMap = new HashMap<Annotation, Set<Entity>>();
+ public HashMap<Annotation, Set<EntityScorePair>> getCandidatesMap(Set<Annotation> annotations) {
+ HashMap<Annotation, Set<EntityScorePair>> candidatesMap = new HashMap<>();
for (Annotation annotation: annotations)
candidatesMap.put(annotation, getCandidates(annotation));
Modified: trunk/components-core/src/main/java/org/dllearner/algorithms/isle/wsd/RandomWordSenseDisambiguation.java
===================================================================
--- trunk/components-core/src/main/java/org/dllearner/algorithms/isle/wsd/RandomWordSenseDisambiguation.java 2013-12-10 15:25:13 UTC (rev 4207)
+++ trunk/components-core/src/main/java/org/dllearner/algorithms/isle/wsd/RandomWordSenseDisambiguation.java 2013-12-10 15:41:36 UTC (rev 4208)
@@ -18,14 +18,15 @@
*/
package org.dllearner.algorithms.isle.wsd;
-import java.util.Random;
-import java.util.Set;
-
import org.dllearner.algorithms.isle.index.Annotation;
+import org.dllearner.algorithms.isle.index.EntityScorePair;
import org.dllearner.algorithms.isle.index.SemanticAnnotation;
import org.dllearner.core.owl.Entity;
import org.semanticweb.owlapi.model.OWLOntology;
+import java.util.Random;
+import java.util.Set;
+
/**
* Disambiguation by randomly selecting one of the candidates (baseline method).
*
@@ -43,17 +44,17 @@
@Override
public SemanticAnnotation disambiguate(Annotation annotation,
- Set<Entity> candidateEntities) {
+ Set<EntityScorePair> candidateEntities) {
int pos = random.nextInt(candidateEntities.size());
int i = 0;
- for(Entity e : candidateEntities)
- {
- if (i == pos) {
- return new SemanticAnnotation(annotation, e);
- }
- i++;
- }
- return null;
+ for(EntityScorePair esp : candidateEntities) {
+ Entity e = esp.getEntity();
+ if (i == pos) {
+ return new SemanticAnnotation(annotation, e);
+ }
+ i++;
+ }
+ return null;
}
}
Modified: trunk/components-core/src/main/java/org/dllearner/algorithms/isle/wsd/SimpleWordSenseDisambiguation.java
===================================================================
--- trunk/components-core/src/main/java/org/dllearner/algorithms/isle/wsd/SimpleWordSenseDisambiguation.java 2013-12-10 15:25:13 UTC (rev 4207)
+++ trunk/components-core/src/main/java/org/dllearner/algorithms/isle/wsd/SimpleWordSenseDisambiguation.java 2013-12-10 15:41:36 UTC (rev 4208)
@@ -3,26 +3,20 @@
*/
package org.dllearner.algorithms.isle.wsd;
-import java.util.HashSet;
-import java.util.Set;
-
import org.apache.log4j.Logger;
import org.dllearner.algorithms.isle.index.Annotation;
+import org.dllearner.algorithms.isle.index.EntityScorePair;
import org.dllearner.algorithms.isle.index.SemanticAnnotation;
import org.dllearner.core.owl.Entity;
import org.dllearner.utilities.owl.OWLAPIConverter;
-import org.semanticweb.owlapi.model.IRI;
-import org.semanticweb.owlapi.model.OWLAnnotationAssertionAxiom;
-import org.semanticweb.owlapi.model.OWLAnnotationProperty;
-import org.semanticweb.owlapi.model.OWLDataFactory;
-import org.semanticweb.owlapi.model.OWLEntity;
-import org.semanticweb.owlapi.model.OWLLiteral;
-import org.semanticweb.owlapi.model.OWLOntology;
+import org.semanticweb.owlapi.model.*;
import org.semanticweb.owlapi.util.IRIShortFormProvider;
import org.semanticweb.owlapi.util.SimpleIRIShortFormProvider;
-
import uk.ac.manchester.cs.owl.owlapi.OWLDataFactoryImpl;
+import java.util.HashSet;
+import java.util.Set;
+
/**
* @author Lorenz Buehmann
*
@@ -47,26 +41,27 @@
* @see org.dllearner.algorithms.isle.WordSenseDisambiguation#disambiguate(org.dllearner.algorithms.isle.index.Annotation, java.util.Set)
*/
@Override
- public SemanticAnnotation disambiguate(Annotation annotation, Set<Entity> candidateEntities) {
+ public SemanticAnnotation disambiguate(Annotation annotation, Set<EntityScorePair> candidateEntities) {
logger.debug("Linguistic annotations:\n" + annotation);
logger.debug("Candidate entities:" + candidateEntities);
String token = annotation.getString().trim();
//check if annotation token matches label of entity or the part behind #(resp. /)
- for (Entity entity : candidateEntities) {
- Set<String> labels = getLabels(entity);
- for (String label : labels) {
- if(label.equals(token)){
- logger.debug("Disambiguated entity: " + entity);
- return new SemanticAnnotation(annotation, entity);
- }
- }
- String shortForm = sfp.getShortForm(IRI.create(entity.getURI()));
- if(annotation.equals(shortForm)){
- logger.debug("Disambiguated entity: " + entity);
- return new SemanticAnnotation(annotation, entity);
- }
- }
- return null;
+ for (EntityScorePair entityScorePair : candidateEntities) {
+ Entity entity = entityScorePair.getEntity();
+ Set<String> labels = getLabels(entity);
+ for (String label : labels) {
+ if (label.equals(token)) {
+ logger.debug("Disambiguated entity: " + entity);
+ return new SemanticAnnotation(annotation, entity);
+ }
+ }
+ String shortForm = sfp.getShortForm(IRI.create(entity.getURI()));
+ if (annotation.equals(shortForm)) {
+ logger.debug("Disambiguated entity: " + entity);
+ return new SemanticAnnotation(annotation, entity);
+ }
+ }
+ return null;
}
private Set<String> getLabels(Entity entity){
Modified: trunk/components-core/src/main/java/org/dllearner/algorithms/isle/wsd/StructureBasedWordSenseDisambiguation.java
===================================================================
--- trunk/components-core/src/main/java/org/dllearner/algorithms/isle/wsd/StructureBasedWordSenseDisambiguation.java 2013-12-10 15:25:13 UTC (rev 4207)
+++ trunk/components-core/src/main/java/org/dllearner/algorithms/isle/wsd/StructureBasedWordSenseDisambiguation.java 2013-12-10 15:41:36 UTC (rev 4208)
@@ -3,21 +3,21 @@
*/
package org.dllearner.algorithms.isle.wsd;
-import java.io.IOException;
-import java.util.Collection;
-import java.util.HashSet;
-import java.util.List;
-import java.util.Set;
-
+import com.google.common.base.Joiner;
+import com.google.common.collect.Sets;
import org.dllearner.algorithms.isle.StructuralEntityContext;
import org.dllearner.algorithms.isle.VSMCosineDocumentSimilarity;
import org.dllearner.algorithms.isle.index.Annotation;
+import org.dllearner.algorithms.isle.index.EntityScorePair;
import org.dllearner.algorithms.isle.index.SemanticAnnotation;
import org.dllearner.core.owl.Entity;
import org.semanticweb.owlapi.model.OWLOntology;
-import com.google.common.base.Joiner;
-import com.google.common.collect.Sets;
+import java.io.IOException;
+import java.util.Collection;
+import java.util.HashSet;
+import java.util.List;
+import java.util.Set;
/**
* @author Lorenz Buehmann
@@ -39,7 +39,7 @@
* @see org.dllearner.algorithms.isle.wsd.WordSenseDisambiguation#disambiguate(org.dllearner.algorithms.isle.index.Annotation, java.util.Set)
*/
@Override
- public SemanticAnnotation disambiguate(Annotation annotation, Set<Entity> candidateEntities) {
+ public SemanticAnnotation disambiguate(Annotation annotation, Set<EntityScorePair> candidateEntities) {
if(!candidateEntities.isEmpty()){
//get the context of the annotated token
List<String> tokenContext = contextExtractor.extractContext(annotation);
@@ -47,19 +47,20 @@
//compare this context with the context of each entity candidate
double maxScore = Double.NEGATIVE_INFINITY;
Entity bestEntity = null;
- for (Entity entity : candidateEntities) {
- //get the context of the entity by analyzing the structure of the ontology
- Set<String> entityContext = StructuralEntityContext.getContextInNaturalLanguage(ontology, entity);
- //compute the VSM Cosine Similarity
- double score = computeScore(tokenContext, entityContext);
- //set best entity
- if(score > maxScore){
- maxScore = score;
- bestEntity = entity;
- }
- }
-
- return new SemanticAnnotation(annotation, bestEntity);
+ for (EntityScorePair entityScorePair : candidateEntities) {
+ Entity entity = entityScorePair.getEntity();
+ //get the context of the entity by analyzing the structure of the ontology
+ Set<String> entityContext = StructuralEntityContext.getContextInNaturalLanguage(ontology, entity);
+ //compute the VSM Cosine Similarity
+ double score = computeScore(tokenContext, entityContext);
+ //set best entity
+ if (score > maxScore) {
+ maxScore = score;
+ bestEntity = entity;
+ }
+ }
+
+ return new SemanticAnnotation(annotation, bestEntity);
}
return null;
}
Modified: trunk/components-core/src/main/java/org/dllearner/algorithms/isle/wsd/WordSenseDisambiguation.java
===================================================================
--- trunk/components-core/src/main/java/org/dllearner/algorithms/isle/wsd/WordSenseDisambiguation.java 2013-12-10 15:25:13 UTC (rev 4207)
+++ trunk/components-core/src/main/java/org/dllearner/algorithms/isle/wsd/WordSenseDisambiguation.java 2013-12-10 15:41:36 UTC (rev 4208)
@@ -1,12 +1,12 @@
package org.dllearner.algorithms.isle.wsd;
-import java.util.Set;
-
import org.dllearner.algorithms.isle.index.Annotation;
+import org.dllearner.algorithms.isle.index.EntityScorePair;
import org.dllearner.algorithms.isle.index.SemanticAnnotation;
-import org.dllearner.core.owl.Entity;
import org.semanticweb.owlapi.model.OWLOntology;
+import java.util.Set;
+
/**
* Abstract class for the word sense disambiguation component.
*
@@ -27,9 +27,10 @@
/**
* Chooses the correct entity for the given annotation from a set of candidate entities.
*
+ *
* @param annotation the annotation to find entity for
* @param candidateEntities the set of candidate entities
* @return semantic annotation containing the given annotation and the chosen entity
*/
- public abstract SemanticAnnotation disambiguate(Annotation annotation, Set<Entity> candidateEntities);
+ public abstract SemanticAnnotation disambiguate(Annotation annotation, Set<EntityScorePair> candidateEntities);
}
This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.
|
