[DL-Learner SVN] SF.net SVN: dl-learner:[4104] trunk/components-core/src

[<dfl...@us...>]

Revision: 4104
          http://sourceforge.net/p/dl-learner/code/4104
Author:   dfleischhacker
Date:     2013-09-09 11:21:51 +0000 (Mon, 09 Sep 2013)
Log Message:
-----------
Add test for entity linking and make use of lemmatizing optional in SimpleSemanticIndex

Modified Paths:
--------------
    trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/semantic/simple/SimpleSemanticIndex.java
    trunk/components-core/src/test/java/org/dllearner/algorithms/isle/ISLETest.java

Modified: trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/semantic/simple/SimpleSemanticIndex.java
===================================================================
--- trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/semantic/simple/SimpleSemanticIndex.java	2013-09-09 11:03:04 UTC (rev 4103)
+++ trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/semantic/simple/SimpleSemanticIndex.java	2013-09-09 11:21:51 UTC (rev 4104)
@@ -23,20 +23,42 @@
 
     /**
      * Initializes the semantic index to use {@code ontology} for finding all labels of an entity and
+     * {@code syntacticIndex} to query for documents containing these labels. This consutrctor initializes with
+     * full lemmatizing enabled.
+     *
+     * @param ontology       ontology to retrieve entity labels from
+     * @param syntacticIndex index to query for documents containing the labels
+     */
+    public SimpleSemanticIndex(OWLOntology ontology, SyntacticIndex syntacticIndex) {
+        this(ontology, syntacticIndex, true);
+    }
+
+    /**
+     * Initializes the semantic index to use {@code ontology} for finding all labels of an entity and
      * {@code syntacticIndex} to query for documents containing these labels.
      *
      * @param ontology       ontology to retrieve entity labels from
      * @param syntacticIndex index to query for documents containing the labels
+     * @param useWordNormalization    whether word normalization should be used or not
      */
-    public SimpleSemanticIndex(OWLOntology ontology, SyntacticIndex syntacticIndex) {
+    public SimpleSemanticIndex(OWLOntology ontology, SyntacticIndex syntacticIndex, boolean useWordNormalization) {
         super(ontology);
-        SimpleEntityCandidatesTrie trie = new SimpleEntityCandidatesTrie(new RDFSLabelEntityTextRetriever(ontology),
-                ontology, new SimpleEntityCandidatesTrie.LemmatizingWordNetNameGenerator(5));
+        SimpleEntityCandidatesTrie trie;
+        if (useWordNormalization) {
+            trie = new SimpleEntityCandidatesTrie(new RDFSLabelEntityTextRetriever(ontology),
+                    ontology, new SimpleEntityCandidatesTrie.LemmatizingWordNetNameGenerator(5));
+        }
+        else {
+            trie = new SimpleEntityCandidatesTrie(new RDFSLabelEntityTextRetriever(ontology),
+                    ontology, new SimpleEntityCandidatesTrie.DummyNameGenerator());
+        }
 //        trie.printTrie();
+        TrieLinguisticAnnotator linguisticAnnotator = new TrieLinguisticAnnotator(trie);
+        linguisticAnnotator.setNormalizeWords(useWordNormalization);
         setSemanticAnnotator(new SemanticAnnotator(
                 new SimpleWordSenseDisambiguation(ontology),
                 new TrieEntityCandidateGenerator(ontology, trie),
-                new TrieLinguisticAnnotator(trie)));
+                linguisticAnnotator));
 
     }
 

Modified: trunk/components-core/src/test/java/org/dllearner/algorithms/isle/ISLETest.java
===================================================================
--- trunk/components-core/src/test/java/org/dllearner/algorithms/isle/ISLETest.java	2013-09-09 11:03:04 UTC (rev 4103)
+++ trunk/components-core/src/test/java/org/dllearner/algorithms/isle/ISLETest.java	2013-09-09 11:21:51 UTC (rev 4104)
@@ -3,16 +3,11 @@
  */
 package org.dllearner.algorithms.isle;
 
-import java.io.File;
-import java.io.IOException;
-import java.text.DecimalFormat;
-import java.util.HashSet;
-import java.util.Map;
-import java.util.Set;
-
+import com.google.common.base.Charsets;
+import com.google.common.base.Joiner;
+import com.google.common.io.Files;
 import org.dllearner.algorithms.celoe.CELOE;
-import org.dllearner.algorithms.isle.index.AnnotatedDocument;
-import org.dllearner.algorithms.isle.index.TextDocument;
+import org.dllearner.algorithms.isle.index.*;
 import org.dllearner.algorithms.isle.index.semantic.SemanticIndex;
 import org.dllearner.algorithms.isle.index.semantic.simple.SimpleSemanticIndex;
 import org.dllearner.algorithms.isle.index.syntactic.OWLOntologyLuceneSyntacticIndexCreator;
@@ -36,12 +31,14 @@
 import org.semanticweb.owlapi.model.OWLDataFactory;
 import org.semanticweb.owlapi.model.OWLOntology;
 import org.semanticweb.owlapi.model.OWLOntologyManager;
-
 import uk.ac.manchester.cs.owl.owlapi.OWLDataFactoryImpl;
 
-import com.google.common.base.Charsets;
-import com.google.common.base.Joiner;
-import com.google.common.io.Files;
+import java.io.File;
+import java.io.IOException;
+import java.text.DecimalFormat;
+import java.util.HashSet;
+import java.util.Map;
+import java.util.Set;
 
 /**
  * Some tests for the ISLE algorithm.
@@ -160,6 +157,39 @@
 		isle.start();
 	}
 
+    @Test
+    public void testEntityLinkingWithLemmatizing() throws Exception {
+        EntityCandidatesTrie ect = new SimpleEntityCandidatesTrie(new RDFSLabelEntityTextRetriever(ontology), ontology,
+                new SimpleEntityCandidatesTrie.LemmatizingWordNetNameGenerator(5));
+        LinguisticAnnotator linguisticAnnotator = new TrieLinguisticAnnotator(ect);
+        WordSenseDisambiguation wsd = new SimpleWordSenseDisambiguation(ontology);
+        EntityCandidateGenerator ecg = new TrieEntityCandidateGenerator(ontology, ect);
+        SemanticAnnotator semanticAnnotator = new SemanticAnnotator(wsd, ecg, linguisticAnnotator);
+
+        Set<TextDocument> docs = createDocuments();
+        for (TextDocument doc : docs) {
+            AnnotatedDocument annotated = semanticAnnotator.processDocument(doc);
+            System.out.println(annotated);
+        }
+    }
+
+    @Test
+    public void testEntityLinkingWithSimpleStringMatching() throws Exception {
+        EntityCandidatesTrie ect = new SimpleEntityCandidatesTrie(new RDFSLabelEntityTextRetriever(ontology), ontology,
+                new SimpleEntityCandidatesTrie.DummyNameGenerator());
+        TrieLinguisticAnnotator linguisticAnnotator = new TrieLinguisticAnnotator(ect);
+        linguisticAnnotator.setNormalizeWords(false);
+        WordSenseDisambiguation wsd = new SimpleWordSenseDisambiguation(ontology);
+        EntityCandidateGenerator ecg = new TrieEntityCandidateGenerator(ontology, ect);
+        SemanticAnnotator semanticAnnotator = new SemanticAnnotator(wsd, ecg, linguisticAnnotator);
+
+        Set<TextDocument> docs = createDocuments();
+        for (TextDocument doc : docs) {
+            AnnotatedDocument annotated = semanticAnnotator.processDocument(doc);
+            System.out.println(annotated);
+        }
+    }
+
 	@Test
 	public void compareISLE() throws Exception {
 		KnowledgeSource ks = new OWLAPIOntology(ontology);
@@ -170,7 +200,7 @@
 		lp.setClassToDescribe(cls);
 		lp.init();
 		
-		semanticIndex = new SimpleSemanticIndex(ontology, syntacticIndex);
+		semanticIndex = new SimpleSemanticIndex(ontology, syntacticIndex, false);
 		semanticIndex.buildIndex(createDocuments());
 		
 		relevance = new PMIRelevanceMetric(semanticIndex);

This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.