Mercurial > hg > duomoOWLProject

changeset 9:4392a6adf85a default tip

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
new version der label mit language tag
author dwinter
date Thu, 16 Aug 2012 11:40:17 +0200
parents 919e9f3b5efd
children
files protege.properties src/de/mpiwg/dwinter/duomo/lexdump/LexOWLTransformer.java src/de/mpiwg/dwinter/duomo/lexdump/OWLImporter.java src/de/mpiwg/dwinter/duomo/stanford/AnalyseWithEvents.java
diffstat 4 files changed, 105 insertions(+), 8 deletions(-) [+]
line wrap: on
line diff
--- a/protege.properties	Thu Jun 21 17:08:22 2012 +0200
+++ b/protege.properties	Thu Aug 16 11:40:17 2012 +0200
@@ -1,5 +1,5 @@
 #Protege Properties
-#Thu Jun 21 17:04:13 CEST 2012
+#Wed Aug 15 12:39:38 CEST 2012
+history.projects.reopen=file\:/Users/dwinter/Documents/Projekte/Diss%20-%20data-mining/eclipseWorkspace/de.mpiwg.dwinter.duomo/examples/newspaper/newspaper.pprj,file\:/Users/dwinter/Documents/Projekte/Diss%20-%20data-mining/eclipseWorkspace/de.mpiwg.dwinter.duomo/examples/sqwrl/SQWRLExamples.pprj,file\:/Users/dwinter/Documents/Projekte/Diss%20-%20data-mining/eclipseWorkspace/de.mpiwg.dwinter.duomo/examples/pizza/pizza.owl.pprj
 SwitchableClassDefinitionType=edu.stanford.smi.protegex.owl.ui.cls.LogicClassDefinitionWidgetType
-history.projects.reopen=file\:/Users/dwinter/Documents/Projekte/Diss%20-%20data-mining/eclipseWorkspace/de.mpiwg.dwinter.duomo/examples/newspaper/newspaper.pprj,file\:/Users/dwinter/Documents/Projekte/Diss%20-%20data-mining/eclipseWorkspace/de.mpiwg.dwinter.duomo/examples/sqwrl/SQWRLExamples.pprj,file\:/Users/dwinter/Documents/Projekte/Diss%20-%20data-mining/eclipseWorkspace/de.mpiwg.dwinter.duomo/examples/pizza/pizza.owl.pprj
 OntURIBase=http\://www.owl-ontologies.com
--- a/src/de/mpiwg/dwinter/duomo/lexdump/LexOWLTransformer.java	Thu Jun 21 17:08:22 2012 +0200
+++ b/src/de/mpiwg/dwinter/duomo/lexdump/LexOWLTransformer.java	Thu Aug 16 11:40:17 2012 +0200
@@ -637,7 +637,15 @@
 		String segheader = segHeaderElement.getTextTrim();
 
 		OWLIndividual signatureInstance = owlDoc.createInstance("Signatur");
-		owlDoc.setProperty(signatureInstance, "rdfs:label", segheader);
+		
+		
+		OWLIndividual preferredId = owlDoc.createInstance("Identifier");
+		owlDoc.setProperty(preferredId, "rdfs:label", segheader);
+
+		owlDoc.setProperty(signatureInstance, "crm:P48_has_preferred_identifier",
+				preferredId);
+
+		
 		return signatureInstance;
 	}
 
--- a/src/de/mpiwg/dwinter/duomo/lexdump/OWLImporter.java	Thu Jun 21 17:08:22 2012 +0200
+++ b/src/de/mpiwg/dwinter/duomo/lexdump/OWLImporter.java	Thu Aug 16 11:40:17 2012 +0200
@@ -171,6 +171,42 @@
 	 * @param individual 
 	 * @param propertyName
 	 * @param value
+	 * @param lang
+	 */
+	public void setProperty(OWLIndividual individual, String propertyName,
+			Object value, String lang) {
+		RDFProperty prop = owlModel.getRDFProperty(propertyName);
+		individual.setPropertyValue(prop, value);
+		
+		
+		
+		//TODO: replace the following by  @see{org.openrdf.rio.trig.TriGWriter}
+		String valName="";
+		
+		String triple;
+		if (OWLIndividual.class.isInstance(value)){
+			valName="<"+((OWLIndividual)value).getName()+">";
+			triple = String.format("<%s> <%s> %s.\n", individual.getName(),prop.getName(),valName);}
+		else{
+			valName=escapeRDFLit((String)value);
+		
+		
+			triple = String.format("<%s> <%s> \"%s\"@%s.\n", individual.getName(),prop.getName(),valName,lang);
+		
+		}
+		try {
+			outRDF.write(triple);
+			outRDF.flush();
+		} catch (IOException e) {
+			// TODO Auto-generated catch block
+			e.printStackTrace();
+		}
+	}
+	/**
+	 * Erzeuge eine Prpoerty und schreibe die entsprechenden Informationen in das out-file.
+	 * @param individual 
+	 * @param propertyName
+	 * @param value
 	 */
 	public void setProperty(OWLIndividual individual, String propertyName,
 			Object value) {
@@ -491,6 +527,7 @@
 			ind = createInstance(classNameInstance);
 			OWLIndividual identifierInd = createInstance(classNameIdentifier);
 			setProperty(identifierInd, "rdfs:label", identifier);
+			setProperty(identifierInd, "rdfs:label", identifier,"en");
 			try {
 				fh.write(classNameInstance+" --" +classNameIdentifier+"---"+identifier+"\n");
 				fh.flush();
--- a/src/de/mpiwg/dwinter/duomo/stanford/AnalyseWithEvents.java	Thu Jun 21 17:08:22 2012 +0200
+++ b/src/de/mpiwg/dwinter/duomo/stanford/AnalyseWithEvents.java	Thu Aug 16 11:40:17 2012 +0200
@@ -12,6 +12,7 @@
 import java.io.IOException;
 import java.io.InputStreamReader;
 import java.io.Reader;
+import java.net.URLEncoder;
 import java.util.ArrayList;
 import java.util.Collection;
 import java.util.HashMap;
@@ -34,9 +35,11 @@
 import edu.stanford.nlp.trees.Tree;
 import edu.stanford.nlp.trees.TreebankLanguagePack;
 import edu.stanford.nlp.trees.TypedDependency;
-
 public class AnalyseWithEvents {
 
+	private int prepcount = 0;
+	private String prep_ent="http://entities.mpiwg-berlin.mpg.de/research/duomo/prep/";
+	private String prep_ont="http://ontologies.mpiwg-berlin.mpg.de/research/duomo/prep/";
 	public void analyse(String filename) throws IOException {
 
 		LexicalizedParser lp = LexicalizedParser.loadModel("edu/stanford/nlp/models/lexparser/englishPCFG.ser.gz");
@@ -158,8 +161,8 @@
 
 
 			}
-			//if (count > 5)
-			//	  break;
+//			if (count > 100)
+//				  break;
 		}
 		System.out.println(tuple);
 		System.out.println(tupleLong);
@@ -175,14 +178,21 @@
 
 
 		fw = new FileWriter("/tmp/tupleLong");
-
+		
+		FileWriter fw2 = new FileWriter("/tmp/tupleLong.nt3.rdf");
+		
 		for (String key : tupleLong.keySet()){
 			List<String> val = tupleLong.get(key);
 			
 			fw.write(key+"\t"+String.valueOf(val.size())+"\t"+val.toString()+"\n");
+			String res = writePrepAsTriple(fw2,key);
+			writeEventsToRes(fw2,res,val);
+			
+			
 		}
 		fw.close();
-
+		fw2.close();
+		
 		fw = new FileWriter("/tmp/words");
 
 		for (String key : words.keySet()){
@@ -193,6 +203,48 @@
 		fw.close();
 
 	}
+	private void writeEventsToRes(FileWriter fw2, String prepUri, List<String> val) throws IOException {
+		for (String res :val){
+			fw2.write("<"+res.replace("\"", "")+"><"+prep_ont+"contains> <"+prepUri+">.\n");
+		}
+		fw2.flush();
+		
+	}
+	private String writePrepAsTriple(FileWriter fw2, String prep) throws IOException {
+		
+		String[] splitted = prep.split("\t");
+		prepcount+=1;
+		String resUri=String.format(prep_ent+"prep_%s",prepcount);
+		fw2.write("<"+resUri+ "> rdf:type "+"<"+prep_ont+"Preposition>.\n");
+		
+		if (!splitted[2].equals("")){
+			String wd = URLEncoder.encode(splitted[2],"utf-8");
+		fw2.write("<"+resUri+ "> "+"<"+prep_ont+"main> <"+prep_ent+"Word_"+wd+">.\n");
+		fw2.write("<"+prep_ent+"Word_"+wd+"> rdfs:label \""+splitted[2]+"\"@en .\n");
+		fw2.write("<"+prep_ent+"Word_"+wd+"> rdf:type  "+"<"+prep_ont+"Word>.\n");
+		}
+		
+		if (!splitted[3].equals("")){
+			String wd = URLEncoder.encode(splitted[3],"utf-8");
+		fw2.write("<"+resUri+ "> "+"<"+prep_ont+"specification> <"+prep_ent+"Word_"+wd+">.\n");
+		fw2.write("<"+prep_ent+"Word_"+wd+"> rdfs:label \""+splitted[3]+"\"@en .\n");
+		fw2.write("<"+prep_ent+"Word_"+wd+"> rdf:type  "+"<"+prep_ont+"Word>.\n");
+		}
+		
+		if (!splitted[1].equals("")){
+			String wd = URLEncoder.encode(splitted[1],"utf-8");
+		fw2.write("<"+resUri+ "> "+"<"+prep_ont+"prepType> <"+prep_ent+"Type_"+wd+">.\n");
+		fw2.write("<"+prep_ent+"Type_"+wd+"> rdfs:label \""+splitted[1]+"\"@en .\n");
+		fw2.write("<"+prep_ent+"Word_"+wd+"> rdf:type  "+"<"+prep_ont+"Type>.\n");
+		}
+		
+		fw2.flush();
+		return resUri;
+		
+		
+		
+		
+	}
 	/**
 	 * @param args
 	 */