Menu
▾
▴
[DL-Learner SVN] SF.net SVN: dl-learner:[4026] trunk/components-core/src
|
From: <dfl...@us...> - 2013-08-19 09:54:10
|
Revision: 4026
http://sourceforge.net/p/dl-learner/code/4026
Author: dfleischhacker
Date: 2013-08-19 09:54:07 +0000 (Mon, 19 Aug 2013)
Log Message:
-----------
TR API: More restructuring of class structure
Modified Paths:
--------------
trunk/components-core/src/main/java/org/dllearner/algorithms/isle/metrics/AbstractRelevanceMetric.java
trunk/components-core/src/main/java/org/dllearner/algorithms/isle/metrics/PMIRelevanceMetric.java
trunk/components-core/src/test/java/org/dllearner/algorithms/isle/ISLETest.java
Added Paths:
-----------
trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/SemanticAnnotator.java
trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/semantic/
trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/semantic/SemanticIndex.java
trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/semantic/SemanticIndexFactory.java
trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/semantic/simple/
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/SimpleSemanticIndexFactory.java
trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/syntactic/
trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/syntactic/LuceneSyntacticIndex.java
trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/syntactic/OWLOntologyLuceneSyntacticIndexCreator.java
trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/syntactic/SyntacticIndex.java
Removed Paths:
-------------
trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/LuceneSyntacticIndex.java
trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/OWLOntologyLuceneSyntacticIndexCreator.java
trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/SemanticIndex.java
trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/SemanticIndexCreator.java
trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/SimpleSemanticIndex.java
trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/SyntacticIndex.java
Deleted: trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/LuceneSyntacticIndex.java
===================================================================
--- trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/LuceneSyntacticIndex.java 2013-08-19 09:53:14 UTC (rev 4025)
+++ trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/LuceneSyntacticIndex.java 2013-08-19 09:54:07 UTC (rev 4026)
@@ -1,99 +0,0 @@
-/**
- *
- */
-package org.dllearner.algorithms.isle.index;
-
-import org.apache.lucene.analysis.standard.StandardAnalyzer;
-import org.apache.lucene.document.Document;
-import org.apache.lucene.index.DirectoryReader;
-import org.apache.lucene.index.IndexReader;
-import org.apache.lucene.queryparser.classic.ParseException;
-import org.apache.lucene.queryparser.classic.QueryParser;
-import org.apache.lucene.search.IndexSearcher;
-import org.apache.lucene.search.Query;
-import org.apache.lucene.search.ScoreDoc;
-import org.apache.lucene.search.TotalHitCountCollector;
-import org.apache.lucene.store.Directory;
-import org.apache.lucene.store.FSDirectory;
-import org.apache.lucene.util.Version;
-
-import java.io.File;
-import java.io.IOException;
-import java.util.HashSet;
-import java.util.Set;
-
-/**
- * @author Lorenz Buehmann
- *
- */
-public class LuceneSyntacticIndex implements SyntacticIndex {
-
- private IndexSearcher searcher;
- private QueryParser parser;
- private IndexReader indexReader;
- private String searchField;
-
- public LuceneSyntacticIndex(IndexReader indexReader, String searchField) throws Exception {
- this.indexReader = indexReader;
- this.searchField = searchField;
- searcher = new IndexSearcher(indexReader);
- StandardAnalyzer analyzer = new StandardAnalyzer( Version.LUCENE_43);
- parser = new QueryParser( Version.LUCENE_43, searchField, analyzer );
- }
-
- public LuceneSyntacticIndex(Directory directory, String searchField) throws Exception {
- this(DirectoryReader.open(directory), searchField);
- }
-
- public LuceneSyntacticIndex(String indexDirectory, String searchField) throws Exception {
- this(DirectoryReader.open(FSDirectory.open(new File(indexDirectory))), searchField);
- }
-
- /* (non-Javadoc)
- * @see org.dllearner.algorithms.isle.SyntacticIndex#getDocuments(java.lang.String)
- */
- @Override
- public Set<org.dllearner.algorithms.isle.index.Document> getDocuments(String searchString) {
- Set<org.dllearner.algorithms.isle.index.Document> documents = new HashSet<org.dllearner.algorithms.isle.index.Document>();
- try {
- Query query = parser.parse(searchString);
- ScoreDoc[] result = searcher.search(query, getSize()).scoreDocs;
- for (int i = 0; i < result.length; i++) {
- Document doc = searcher.doc(result[i].doc);
- documents.add(new TextDocument(doc.get(searchField)));
- }
- } catch (ParseException e) {
- e.printStackTrace();
- } catch (IOException e) {
- e.printStackTrace();
- }
- return documents;
- }
-
- /* (non-Javadoc)
- * @see org.dllearner.algorithms.isle.SyntacticIndex#getSize()
- */
- @Override
- public int getSize() {
- return indexReader.numDocs();
- }
-
- /* (non-Javadoc)
- * @see org.dllearner.algorithms.isle.SyntacticIndex#count(java.lang.String)
- */
- @Override
- public int count(String searchString) {
- try {
- Query query = parser.parse(searchString);
- TotalHitCountCollector results = new TotalHitCountCollector();
- searcher.search(query, results);
- return results.getTotalHits();
- } catch (ParseException e) {
- e.printStackTrace();
- } catch (IOException e) {
- e.printStackTrace();
- }
- return -1;
- }
-
-}
Deleted: trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/OWLOntologyLuceneSyntacticIndexCreator.java
===================================================================
--- trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/OWLOntologyLuceneSyntacticIndexCreator.java 2013-08-19 09:53:14 UTC (rev 4025)
+++ trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/OWLOntologyLuceneSyntacticIndexCreator.java 2013-08-19 09:54:07 UTC (rev 4026)
@@ -1,94 +0,0 @@
-/**
- *
- */
-package org.dllearner.algorithms.isle.index;
-
-import org.apache.lucene.analysis.Analyzer;
-import org.apache.lucene.analysis.standard.StandardAnalyzer;
-import org.apache.lucene.document.Field;
-import org.apache.lucene.document.FieldType;
-import org.apache.lucene.document.StringField;
-import org.apache.lucene.document.TextField;
-import org.apache.lucene.index.IndexWriter;
-import org.apache.lucene.index.IndexWriterConfig;
-import org.apache.lucene.store.Directory;
-import org.apache.lucene.store.RAMDirectory;
-import org.apache.lucene.util.Version;
-import org.semanticweb.owlapi.model.*;
-import org.semanticweb.owlapi.vocab.OWLRDFVocabulary;
-import uk.ac.manchester.cs.owl.owlapi.OWLDataFactoryImpl;
-
-import java.io.IOException;
-import java.util.HashSet;
-import java.util.Set;
-
-/**
- * Creates a Lucene Index for the labels if classes and properties.
- * @author Lorenz Buehmann
- *
- */
-public class OWLOntologyLuceneSyntacticIndexCreator {
-
- private Directory directory = new RAMDirectory();
- private OWLOntology ontology;
- private Set<OWLEntity> schemaEntities;
-
- private OWLDataFactory df = new OWLDataFactoryImpl();
- private OWLAnnotationProperty annotationProperty = df.getOWLAnnotationProperty(OWLRDFVocabulary.RDFS_LABEL.getIRI());
- private String language = "en";
- private String searchField;
-
- public OWLOntologyLuceneSyntacticIndexCreator(OWLOntology ontology, OWLAnnotationProperty annotationProperty, String searchField) throws IOException {
- this.ontology = ontology;
- this.annotationProperty = annotationProperty;
- this.searchField = searchField;
-
- schemaEntities = new HashSet<OWLEntity>();
- schemaEntities.addAll(ontology.getClassesInSignature());
- schemaEntities.addAll(ontology.getObjectPropertiesInSignature());
- schemaEntities.addAll(ontology.getDataPropertiesInSignature());
- }
-
- public SyntacticIndex buildIndex() throws Exception{
- Analyzer analyzer = new StandardAnalyzer(Version.LUCENE_43);
- IndexWriterConfig indexWriterConfig = new IndexWriterConfig(Version.LUCENE_43, analyzer);
- IndexWriter writer = new IndexWriter(directory, indexWriterConfig);
- System.out.println( "Creating index ..." );
-
- Set<org.apache.lucene.document.Document> luceneDocuments = new HashSet<org.apache.lucene.document.Document>();
- FieldType stringType = new FieldType(StringField.TYPE_STORED);
- stringType.setStoreTermVectors(false);
- FieldType textType = new FieldType(TextField.TYPE_STORED);
- textType.setStoreTermVectors(false);
-
- for (OWLEntity entity : schemaEntities) {
- String label = null;
- Set<OWLAnnotation> annotations = entity.getAnnotations(ontology, annotationProperty);
- for (OWLAnnotation annotation : annotations) {
- if (annotation.getValue() instanceof OWLLiteral) {
- OWLLiteral val = (OWLLiteral) annotation.getValue();
- if (val.hasLang(language)) {
- label = val.getLiteral();
- }
- }
- }
-
- if(label != null){
- org.apache.lucene.document.Document luceneDocument = new org.apache.lucene.document.Document();
- luceneDocument.add(new Field("uri", entity.toStringID(), stringType));
- luceneDocument.add(new Field(searchField, label, textType));
- luceneDocuments.add(luceneDocument);
- }
-
- }
- writer.addDocuments(luceneDocuments);
-
- System.out.println("Done.");
- writer.close();
-
- return new LuceneSyntacticIndex(directory, searchField);
- }
-
-
-
-}
Added: 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 (rev 0)
+++ trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/SemanticAnnotator.java 2013-08-19 09:54:07 UTC (rev 4026)
@@ -0,0 +1,10 @@
+package org.dllearner.algorithms.isle.index;
+
+/**
+ *
+ */
+public interface SemanticAnnotator {
+ /**
+ * Processes the given document and return
+ */
+}
Deleted: trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/SemanticIndex.java
===================================================================
--- trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/SemanticIndex.java 2013-08-19 09:53:14 UTC (rev 4025)
+++ trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/SemanticIndex.java 2013-08-19 09:54:07 UTC (rev 4026)
@@ -1,37 +0,0 @@
-package org.dllearner.algorithms.isle.index;
-
-import org.dllearner.core.owl.Entity;
-
-import java.util.Set;
-
-/**
- * Interface for an index which is able to resolve a given entity's URI to the set of documents containing
- * this entity, i.e., documents which contain words disambiguated to the given entity.
- *
- * @author Lorenz Buehmann
- * @author Daniel Fleischhacker
- */
-public interface SemanticIndex {
- /**
- * Returns the set of documents which reference the given entity using one of its surface forms.
- *
- * @param entity entity to retrieve documents
- * @return documents referencing given entity
- */
- public Set<Document> getDocuments(Entity entity);
-
- /**
- * Returns the number of documents for the given entity.
- *
- * @param entity entity to return number of referencing documents for
- * @return number of documents for the given entity in this index
- */
- public int count(Entity entity);
-
- /**
- * Returns the total number of documents contained in the index.
- *
- * @return the total number of documents contained in the index
- */
- public int getSize();
-}
Deleted: trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/SemanticIndexCreator.java
===================================================================
--- trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/SemanticIndexCreator.java 2013-08-19 09:53:14 UTC (rev 4025)
+++ trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/SemanticIndexCreator.java 2013-08-19 09:54:07 UTC (rev 4026)
@@ -1,22 +0,0 @@
-/**
- *
- */
-package org.dllearner.algorithms.isle.index;
-
-/**
- * This gets a syntactic index and returns a semantic index by applying WSD etc.
- * @author Lorenz Buehmann
- *
- */
-public class SemanticIndexCreator {
-
- private SyntacticIndex syntacticIndex;
-
- public SemanticIndexCreator(SyntacticIndex syntacticIndex) {
- this.syntacticIndex = syntacticIndex;
- }
-
- public SemanticIndex createSemanticIndex(){
- return null;
- }
-}
Deleted: trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/SimpleSemanticIndex.java
===================================================================
--- trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/SimpleSemanticIndex.java 2013-08-19 09:53:14 UTC (rev 4025)
+++ trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/SimpleSemanticIndex.java 2013-08-19 09:54:07 UTC (rev 4026)
@@ -1,67 +0,0 @@
-/**
- *
- */
-package org.dllearner.algorithms.isle.index;
-
-import org.dllearner.algorithms.isle.textretrieval.RDFSLabelEntityTextRetriever;
-import org.dllearner.core.owl.Entity;
-import org.semanticweb.owlapi.model.OWLOntology;
-
-import java.util.HashSet;
-import java.util.Map;
-import java.util.Map.Entry;
-import java.util.Set;
-
-/**
- * @author Lorenz Buehmann
- *
- */
-public class SimpleSemanticIndex implements SemanticIndex {
-
- private SyntacticIndex syntacticIndex;
- private RDFSLabelEntityTextRetriever labelRetriever;
-
- /**
- *
- */
- public SimpleSemanticIndex(OWLOntology ontology, SyntacticIndex syntacticIndex) {
- this.syntacticIndex = syntacticIndex;
- labelRetriever = new RDFSLabelEntityTextRetriever(ontology);
- }
-
-
- /* (non-Javadoc)
- * @see org.dllearner.algorithms.isle.SemanticIndex#getDocuments(org.dllearner.core.owl.Entity)
- */
- @Override
- public Set<Document> getDocuments(Entity entity) {
- Set<Document> documents = new HashSet<Document>();
- Map<String, Double> relevantText = labelRetriever.getRelevantText(entity);
-
- for (Entry<String, Double> entry : relevantText.entrySet()) {
- String label = entry.getKey();
- documents.addAll(syntacticIndex.getDocuments(label));
- }
-
- return documents;
- }
-
- /* (non-Javadoc)
- * @see org.dllearner.algorithms.isle.SemanticIndex#count(java.lang.String)
- */
- @Override
- public int count(Entity entity) {
- return getDocuments(entity).size();
- }
-
- /* (non-Javadoc)
- * @see org.dllearner.algorithms.isle.SemanticIndex#getSize()
- */
- @Override
- public int getSize() {
- return syntacticIndex.getSize();
- }
-
-
-
-}
Deleted: trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/SyntacticIndex.java
===================================================================
--- trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/SyntacticIndex.java 2013-08-19 09:53:14 UTC (rev 4025)
+++ trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/SyntacticIndex.java 2013-08-19 09:54:07 UTC (rev 4026)
@@ -1,41 +0,0 @@
-/**
- *
- */
-package org.dllearner.algorithms.isle.index;
-
-import java.util.Set;
-
-/**
- * Interface for a syntactic index, e.g., a basic string-based inverted index.
- *
- * @author Lorenz Buehmann
- * @author Daniel Fleischhacker
- */
-public interface SyntacticIndex {
-
- /**
- * Returns a set of documents based on how the underlying index is processing the given
- * search string.
- *
- * @param searchString query specifying the documents to retrieve
- * @return set of documents retrieved based on the given query string
- */
- Set<Document> getDocuments(String searchString);
-
- /**
- * Returns the number of documents based on how the underlying index is processing the
- * given search string.
- *
- * @param searchString query specifying the documents to include in the number of documents
- * @return number of documents retrieved based on the given query string
- */
- int count(String searchString);
-
- /**
- * Returns the total number of documents contained in the index.
- *
- * @return the total number of documents contained in the index
- */
- int getSize();
-
-}
Copied: trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/semantic/SemanticIndex.java (from rev 4025, trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/SemanticIndex.java)
===================================================================
--- trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/semantic/SemanticIndex.java (rev 0)
+++ trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/semantic/SemanticIndex.java 2013-08-19 09:54:07 UTC (rev 4026)
@@ -0,0 +1,38 @@
+package org.dllearner.algorithms.isle.index.semantic;
+
+import org.dllearner.algorithms.isle.index.Document;
+import org.dllearner.core.owl.Entity;
+
+import java.util.Set;
+
+/**
+ * Interface for an index which is able to resolve a given entity's URI to the set of documents containing
+ * this entity, i.e., documents which contain words disambiguated to the given entity.
+ *
+ * @author Lorenz Buehmann
+ * @author Daniel Fleischhacker
+ */
+public interface SemanticIndex {
+ /**
+ * Returns the set of documents which reference the given entity using one of its surface forms.
+ *
+ * @param entity entity to retrieve documents
+ * @return documents referencing given entity
+ */
+ public Set<Document> getDocuments(Entity entity);
+
+ /**
+ * Returns the number of documents for the given entity.
+ *
+ * @param entity entity to return number of referencing documents for
+ * @return number of documents for the given entity in this index
+ */
+ public int count(Entity entity);
+
+ /**
+ * Returns the total number of documents contained in the index.
+ *
+ * @return the total number of documents contained in the index
+ */
+ public int getSize();
+}
Added: trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/semantic/SemanticIndexFactory.java
===================================================================
--- trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/semantic/SemanticIndexFactory.java (rev 0)
+++ trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/semantic/SemanticIndexFactory.java 2013-08-19 09:54:07 UTC (rev 4026)
@@ -0,0 +1,18 @@
+package org.dllearner.algorithms.isle.index.semantic;
+
+import java.io.File;
+
+/**
+ * Provides methods for creating semantic indexes.
+ *
+ * @author Daniel Fleischhacker
+ */
+public interface SemanticIndexFactory {
+ /**
+ * Returns a newly created semantic index for the collection of files contained in the given {@code directory}.
+ *
+ * @param inputDirectory directory containing files to create index from
+ * @return semantic index for the files in the given input directory
+ */
+ public SemanticIndex createIndex(File inputDirectory);
+}
Added: 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 (rev 0)
+++ trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/semantic/simple/SimpleSemanticIndex.java 2013-08-19 09:54:07 UTC (rev 4026)
@@ -0,0 +1,73 @@
+/**
+ *
+ */
+package org.dllearner.algorithms.isle.index.semantic.simple;
+
+import org.dllearner.algorithms.isle.index.Document;
+import org.dllearner.algorithms.isle.index.syntactic.SyntacticIndex;
+import org.dllearner.algorithms.isle.index.semantic.SemanticIndex;
+import org.dllearner.algorithms.isle.textretrieval.RDFSLabelEntityTextRetriever;
+import org.dllearner.core.owl.Entity;
+import org.semanticweb.owlapi.model.OWLOntology;
+
+import java.util.HashSet;
+import java.util.Map;
+import java.util.Map.Entry;
+import java.util.Set;
+
+/**
+ * A semantic index which returns all documents which contain at least one of the labels assigned to a specific
+ * entity in a provided ontology.
+ *
+ * @author Lorenz Buehmann
+ */
+public class SimpleSemanticIndex implements SemanticIndex {
+ private SyntacticIndex syntacticIndex;
+ private RDFSLabelEntityTextRetriever labelRetriever;
+
+ /**
+ * 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
+ */
+ public SimpleSemanticIndex(OWLOntology ontology, SyntacticIndex syntacticIndex) {
+ this.syntacticIndex = syntacticIndex;
+ labelRetriever = new RDFSLabelEntityTextRetriever(ontology);
+ }
+
+ /* (non-Javadoc)
+ * @see org.dllearner.algorithms.isle.SemanticIndex#getDocuments(org.dllearner.core.owl.Entity)
+ */
+ @Override
+ public Set<Document> getDocuments(Entity entity) {
+ Set<Document> documents = new HashSet<Document>();
+ Map<String, Double> relevantText = labelRetriever.getRelevantText(entity);
+
+ for (Entry<String, Double> entry : relevantText.entrySet()) {
+ String label = entry.getKey();
+ documents.addAll(syntacticIndex.getDocuments(label));
+ }
+
+ return documents;
+ }
+
+ /* (non-Javadoc)
+ * @see org.dllearner.algorithms.isle.SemanticIndex#count(java.lang.String)
+ */
+ @Override
+ public int count(Entity entity) {
+ return getDocuments(entity).size();
+ }
+
+ /* (non-Javadoc)
+ * @see org.dllearner.algorithms.isle.SemanticIndex#getSize()
+ */
+ @Override
+ public int getSize() {
+ return syntacticIndex.getSize();
+ }
+
+
+}
Added: trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/semantic/simple/SimpleSemanticIndexFactory.java
===================================================================
--- trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/semantic/simple/SimpleSemanticIndexFactory.java (rev 0)
+++ trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/semantic/simple/SimpleSemanticIndexFactory.java 2013-08-19 09:54:07 UTC (rev 4026)
@@ -0,0 +1,40 @@
+/**
+ *
+ */
+package org.dllearner.algorithms.isle.index.semantic.simple;
+
+import org.dllearner.algorithms.isle.index.syntactic.SyntacticIndex;
+import org.dllearner.algorithms.isle.index.semantic.SemanticIndex;
+import org.dllearner.algorithms.isle.index.semantic.SemanticIndexFactory;
+import org.semanticweb.owlapi.model.OWLOntology;
+
+import java.io.File;
+
+/**
+ * This gets a syntactic index and returns a semantic index by applying WSD etc.
+ *
+ * @author Lorenz Buehmann
+ * @author Daniel Fleischhacker
+ */
+public class SimpleSemanticIndexFactory implements SemanticIndexFactory {
+ private OWLOntology ontology;
+ private SyntacticIndex syntacticIndex;
+
+ /**
+ * Initializes a semantic index factory for creating simple semantic indexes. Simple semantic indexes use
+ * the labels assigned to an entity in {@code ontology} as its surface forms and return the all documents
+ * from the given syntactic index which contain at least one of these surface forms.
+ *
+ * @param syntacticIndex the syntactic index in which occurrences of the labels are searched
+ * @param ontology the ontology retrieve the entities' labels from
+ */
+ public SimpleSemanticIndexFactory(SyntacticIndex syntacticIndex, OWLOntology ontology) {
+ this.syntacticIndex = syntacticIndex;
+ this.ontology = ontology;
+ }
+
+ @Override
+ public SemanticIndex createIndex(File inputDirectory) {
+ return new SimpleSemanticIndex(ontology, syntacticIndex);
+ }
+}
Copied: trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/syntactic/LuceneSyntacticIndex.java (from rev 4025, trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/LuceneSyntacticIndex.java)
===================================================================
--- trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/syntactic/LuceneSyntacticIndex.java (rev 0)
+++ trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/syntactic/LuceneSyntacticIndex.java 2013-08-19 09:54:07 UTC (rev 4026)
@@ -0,0 +1,100 @@
+/**
+ *
+ */
+package org.dllearner.algorithms.isle.index.syntactic;
+
+import org.apache.lucene.analysis.standard.StandardAnalyzer;
+import org.apache.lucene.document.Document;
+import org.apache.lucene.index.DirectoryReader;
+import org.apache.lucene.index.IndexReader;
+import org.apache.lucene.queryparser.classic.ParseException;
+import org.apache.lucene.queryparser.classic.QueryParser;
+import org.apache.lucene.search.IndexSearcher;
+import org.apache.lucene.search.Query;
+import org.apache.lucene.search.ScoreDoc;
+import org.apache.lucene.search.TotalHitCountCollector;
+import org.apache.lucene.store.Directory;
+import org.apache.lucene.store.FSDirectory;
+import org.apache.lucene.util.Version;
+import org.dllearner.algorithms.isle.index.TextDocument;
+
+import java.io.File;
+import java.io.IOException;
+import java.util.HashSet;
+import java.util.Set;
+
+/**
+ * @author Lorenz Buehmann
+ *
+ */
+public class LuceneSyntacticIndex implements SyntacticIndex {
+
+ private IndexSearcher searcher;
+ private QueryParser parser;
+ private IndexReader indexReader;
+ private String searchField;
+
+ public LuceneSyntacticIndex(IndexReader indexReader, String searchField) throws Exception {
+ this.indexReader = indexReader;
+ this.searchField = searchField;
+ searcher = new IndexSearcher(indexReader);
+ StandardAnalyzer analyzer = new StandardAnalyzer( Version.LUCENE_43);
+ parser = new QueryParser( Version.LUCENE_43, searchField, analyzer );
+ }
+
+ public LuceneSyntacticIndex(Directory directory, String searchField) throws Exception {
+ this(DirectoryReader.open(directory), searchField);
+ }
+
+ public LuceneSyntacticIndex(String indexDirectory, String searchField) throws Exception {
+ this(DirectoryReader.open(FSDirectory.open(new File(indexDirectory))), searchField);
+ }
+
+ /* (non-Javadoc)
+ * @see org.dllearner.algorithms.isle.SyntacticIndex#getDocuments(java.lang.String)
+ */
+ @Override
+ public Set<org.dllearner.algorithms.isle.index.Document> getDocuments(String searchString) {
+ Set<org.dllearner.algorithms.isle.index.Document> documents = new HashSet<org.dllearner.algorithms.isle.index.Document>();
+ try {
+ Query query = parser.parse(searchString);
+ ScoreDoc[] result = searcher.search(query, getSize()).scoreDocs;
+ for (int i = 0; i < result.length; i++) {
+ Document doc = searcher.doc(result[i].doc);
+ documents.add(new TextDocument(doc.get(searchField)));
+ }
+ } catch (ParseException e) {
+ e.printStackTrace();
+ } catch (IOException e) {
+ e.printStackTrace();
+ }
+ return documents;
+ }
+
+ /* (non-Javadoc)
+ * @see org.dllearner.algorithms.isle.SyntacticIndex#getSize()
+ */
+ @Override
+ public int getSize() {
+ return indexReader.numDocs();
+ }
+
+ /* (non-Javadoc)
+ * @see org.dllearner.algorithms.isle.SyntacticIndex#count(java.lang.String)
+ */
+ @Override
+ public int count(String searchString) {
+ try {
+ Query query = parser.parse(searchString);
+ TotalHitCountCollector results = new TotalHitCountCollector();
+ searcher.search(query, results);
+ return results.getTotalHits();
+ } catch (ParseException e) {
+ e.printStackTrace();
+ } catch (IOException e) {
+ e.printStackTrace();
+ }
+ return -1;
+ }
+
+}
Copied: trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/syntactic/OWLOntologyLuceneSyntacticIndexCreator.java (from rev 4025, trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/OWLOntologyLuceneSyntacticIndexCreator.java)
===================================================================
--- trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/syntactic/OWLOntologyLuceneSyntacticIndexCreator.java (rev 0)
+++ trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/syntactic/OWLOntologyLuceneSyntacticIndexCreator.java 2013-08-19 09:54:07 UTC (rev 4026)
@@ -0,0 +1,94 @@
+/**
+ *
+ */
+package org.dllearner.algorithms.isle.index.syntactic;
+
+import org.apache.lucene.analysis.Analyzer;
+import org.apache.lucene.analysis.standard.StandardAnalyzer;
+import org.apache.lucene.document.Field;
+import org.apache.lucene.document.FieldType;
+import org.apache.lucene.document.StringField;
+import org.apache.lucene.document.TextField;
+import org.apache.lucene.index.IndexWriter;
+import org.apache.lucene.index.IndexWriterConfig;
+import org.apache.lucene.store.Directory;
+import org.apache.lucene.store.RAMDirectory;
+import org.apache.lucene.util.Version;
+import org.semanticweb.owlapi.model.*;
+import org.semanticweb.owlapi.vocab.OWLRDFVocabulary;
+import uk.ac.manchester.cs.owl.owlapi.OWLDataFactoryImpl;
+
+import java.io.IOException;
+import java.util.HashSet;
+import java.util.Set;
+
+/**
+ * Creates a Lucene Index for the labels if classes and properties.
+ * @author Lorenz Buehmann
+ *
+ */
+public class OWLOntologyLuceneSyntacticIndexCreator {
+
+ private Directory directory = new RAMDirectory();
+ private OWLOntology ontology;
+ private Set<OWLEntity> schemaEntities;
+
+ private OWLDataFactory df = new OWLDataFactoryImpl();
+ private OWLAnnotationProperty annotationProperty = df.getOWLAnnotationProperty(OWLRDFVocabulary.RDFS_LABEL.getIRI());
+ private String language = "en";
+ private String searchField;
+
+ public OWLOntologyLuceneSyntacticIndexCreator(OWLOntology ontology, OWLAnnotationProperty annotationProperty, String searchField) throws IOException {
+ this.ontology = ontology;
+ this.annotationProperty = annotationProperty;
+ this.searchField = searchField;
+
+ schemaEntities = new HashSet<OWLEntity>();
+ schemaEntities.addAll(ontology.getClassesInSignature());
+ schemaEntities.addAll(ontology.getObjectPropertiesInSignature());
+ schemaEntities.addAll(ontology.getDataPropertiesInSignature());
+ }
+
+ public SyntacticIndex buildIndex() throws Exception{
+ Analyzer analyzer = new StandardAnalyzer(Version.LUCENE_43);
+ IndexWriterConfig indexWriterConfig = new IndexWriterConfig(Version.LUCENE_43, analyzer);
+ IndexWriter writer = new IndexWriter(directory, indexWriterConfig);
+ System.out.println( "Creating index ..." );
+
+ Set<org.apache.lucene.document.Document> luceneDocuments = new HashSet<org.apache.lucene.document.Document>();
+ FieldType stringType = new FieldType(StringField.TYPE_STORED);
+ stringType.setStoreTermVectors(false);
+ FieldType textType = new FieldType(TextField.TYPE_STORED);
+ textType.setStoreTermVectors(false);
+
+ for (OWLEntity entity : schemaEntities) {
+ String label = null;
+ Set<OWLAnnotation> annotations = entity.getAnnotations(ontology, annotationProperty);
+ for (OWLAnnotation annotation : annotations) {
+ if (annotation.getValue() instanceof OWLLiteral) {
+ OWLLiteral val = (OWLLiteral) annotation.getValue();
+ if (val.hasLang(language)) {
+ label = val.getLiteral();
+ }
+ }
+ }
+
+ if(label != null){
+ org.apache.lucene.document.Document luceneDocument = new org.apache.lucene.document.Document();
+ luceneDocument.add(new Field("uri", entity.toStringID(), stringType));
+ luceneDocument.add(new Field(searchField, label, textType));
+ luceneDocuments.add(luceneDocument);
+ }
+
+ }
+ writer.addDocuments(luceneDocuments);
+
+ System.out.println("Done.");
+ writer.close();
+
+ return new LuceneSyntacticIndex(directory, searchField);
+ }
+
+
+
+}
Copied: trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/syntactic/SyntacticIndex.java (from rev 4025, trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/SyntacticIndex.java)
===================================================================
--- trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/syntactic/SyntacticIndex.java (rev 0)
+++ trunk/components-core/src/main/java/org/dllearner/algorithms/isle/index/syntactic/SyntacticIndex.java 2013-08-19 09:54:07 UTC (rev 4026)
@@ -0,0 +1,43 @@
+/**
+ *
+ */
+package org.dllearner.algorithms.isle.index.syntactic;
+
+import org.dllearner.algorithms.isle.index.Document;
+
+import java.util.Set;
+
+/**
+ * Interface for a syntactic index, e.g., a basic string-based inverted index.
+ *
+ * @author Lorenz Buehmann
+ * @author Daniel Fleischhacker
+ */
+public interface SyntacticIndex {
+
+ /**
+ * Returns a set of documents based on how the underlying index is processing the given
+ * search string.
+ *
+ * @param searchString query specifying the documents to retrieve
+ * @return set of documents retrieved based on the given query string
+ */
+ Set<Document> getDocuments(String searchString);
+
+ /**
+ * Returns the number of documents based on how the underlying index is processing the
+ * given search string.
+ *
+ * @param searchString query specifying the documents to include in the number of documents
+ * @return number of documents retrieved based on the given query string
+ */
+ int count(String searchString);
+
+ /**
+ * Returns the total number of documents contained in the index.
+ *
+ * @return the total number of documents contained in the index
+ */
+ int getSize();
+
+}
Modified: trunk/components-core/src/main/java/org/dllearner/algorithms/isle/metrics/AbstractRelevanceMetric.java
===================================================================
--- trunk/components-core/src/main/java/org/dllearner/algorithms/isle/metrics/AbstractRelevanceMetric.java 2013-08-19 09:53:14 UTC (rev 4025)
+++ trunk/components-core/src/main/java/org/dllearner/algorithms/isle/metrics/AbstractRelevanceMetric.java 2013-08-19 09:54:07 UTC (rev 4026)
@@ -3,12 +3,12 @@
*/
package org.dllearner.algorithms.isle.metrics;
+import org.dllearner.algorithms.isle.index.semantic.SemanticIndex;
+import org.semanticweb.owlapi.model.OWLEntity;
+
import java.util.HashMap;
import java.util.Map;
-import org.dllearner.algorithms.isle.index.SemanticIndex;
-import org.semanticweb.owlapi.model.OWLEntity;
-
/**
* @author Lorenz Buehmann
*
Modified: trunk/components-core/src/main/java/org/dllearner/algorithms/isle/metrics/PMIRelevanceMetric.java
===================================================================
--- trunk/components-core/src/main/java/org/dllearner/algorithms/isle/metrics/PMIRelevanceMetric.java 2013-08-19 09:53:14 UTC (rev 4025)
+++ trunk/components-core/src/main/java/org/dllearner/algorithms/isle/metrics/PMIRelevanceMetric.java 2013-08-19 09:54:07 UTC (rev 4026)
@@ -3,12 +3,12 @@
*/
package org.dllearner.algorithms.isle.metrics;
-import java.util.Set;
-
-import org.dllearner.algorithms.isle.index.SemanticIndex;
+import com.google.common.collect.Sets;
+import org.dllearner.algorithms.isle.index.Document;
+import org.dllearner.algorithms.isle.index.semantic.SemanticIndex;
import org.dllearner.core.owl.Entity;
-import com.google.common.collect.Sets;
+import java.util.Set;
/**
* @author Lorenz Buehmann
@@ -22,9 +22,9 @@
@Override
public double getRelevance(Entity entityA, Entity entityB){
- Set<String> documentsA = index.getDocuments(entityA);
- Set<String> documentsB = index.getDocuments(entityB);
- Set<String> documentsAB = Sets.intersection(documentsA, documentsB);
+ Set<Document> documentsA = index.getDocuments(entityA);
+ Set<Document> documentsB = index.getDocuments(entityB);
+ Set<Document> documentsAB = Sets.intersection(documentsA, documentsB);
int nrOfDocuments = index.getSize();
double dPClass = nrOfDocuments == 0 ? 0 : ((double) documentsA.size() / (double) nrOfDocuments);
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 2013-08-19 09:53:14 UTC (rev 4025)
+++ trunk/components-core/src/test/java/org/dllearner/algorithms/isle/ISLETest.java 2013-08-19 09:54:07 UTC (rev 4026)
@@ -3,13 +3,11 @@
*/
package org.dllearner.algorithms.isle;
-import java.io.File;
-import java.util.Map;
-
-import org.dllearner.algorithms.isle.index.OWLOntologyLuceneSyntacticIndexCreator;
-import org.dllearner.algorithms.isle.index.SemanticIndex;
-import org.dllearner.algorithms.isle.index.SimpleSemanticIndex;
-import org.dllearner.algorithms.isle.index.SyntacticIndex;
+import com.google.common.base.Joiner;
+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;
+import org.dllearner.algorithms.isle.index.syntactic.SyntacticIndex;
import org.dllearner.algorithms.isle.metrics.PMIRelevanceMetric;
import org.dllearner.algorithms.isle.metrics.RelevanceMetric;
import org.dllearner.algorithms.isle.metrics.RelevanceUtils;
@@ -28,10 +26,10 @@
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.Joiner;
+import java.io.File;
+import java.util.Map;
/**
* @author Lorenz Buehmann
This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.
|
