Menu
▾
▴
[DL-Learner SVN] SF.net SVN: dl-learner:[1156] trunk/src/dl-learner/org/dllearner
|
From: <ku...@us...> - 2008-08-31 15:04:23
|
Revision: 1156
http://dl-learner.svn.sourceforge.net/dl-learner/?rev=1156&view=rev
Author: kurzum
Date: 2008-08-31 15:04:13 +0000 (Sun, 31 Aug 2008)
Log Message:
-----------
Modified Paths:
--------------
trunk/src/dl-learner/org/dllearner/kb/sparql/SparqlKnowledgeSource.java
trunk/src/dl-learner/org/dllearner/scripts/SemanticBible2.java
Added Paths:
-----------
trunk/src/dl-learner/org/dllearner/utilities/StringFormatter.java
Modified: trunk/src/dl-learner/org/dllearner/kb/sparql/SparqlKnowledgeSource.java
===================================================================
--- trunk/src/dl-learner/org/dllearner/kb/sparql/SparqlKnowledgeSource.java 2008-08-30 16:43:11 UTC (rev 1155)
+++ trunk/src/dl-learner/org/dllearner/kb/sparql/SparqlKnowledgeSource.java 2008-08-31 15:04:13 UTC (rev 1156)
@@ -83,8 +83,8 @@
static final boolean debug = false;
static final boolean debugUseImprovedTupleAquisitor = debug && false; //switches tupleaquisitor
static final boolean debugExitAfterExtraction = debug && false; //switches sysex und rdf generation
- static final boolean debugAdditionallyGenerateRDF = debug && true;
+
private boolean useCache=true;
// ConfigOptions
public URL url;
@@ -100,6 +100,7 @@
// private Set<String> classList;
private String format = "N-TRIPLES";
private boolean dumpToFile = true;
+ private boolean convertNT2RDF = true ;
private boolean useLits = false;
private boolean getAllSuperClasses = true;
private boolean closeAfterRecursion = true;
@@ -172,6 +173,11 @@
"dumpToFile",
"Specifies whether the extracted ontology is written to a file or not.",
true));
+ options
+ .add(new BooleanConfigOption(
+ "convertNT2RDF",
+ "Specifies whether the extracted NTriples are converted to RDF and deleted.",
+ true));
options.add(new BooleanConfigOption("useLits",
"use Literals in SPARQL query"));
options
@@ -250,6 +256,8 @@
format = (String) entry.getValue();
} else if (option.equals("dumpToFile")) {
dumpToFile = (Boolean) entry.getValue();
+ } else if (option.equals("convertNT2RDF")) {
+ convertNT2RDF = (Boolean) entry.getValue();
} else if (option.equals("useLits")) {
useLits = (Boolean) entry.getValue();
} else if (option.equals("useCache")) {
@@ -329,16 +337,21 @@
new File(basedir).mkdir();
}
+ File dump = new File(basedir + filename);
+
FileWriter fw = new FileWriter(
- new File(basedir + filename), true);
+ dump , true);
fw.write(ont);
fw.flush();
fw.close();
- dumpFile = (new File(basedir + filename)).toURI().toURL();
- if(debugAdditionallyGenerateRDF){
- NT2RDF.convertNT2RDF(basedir + filename);
- //System.exit(0);
+
+ dumpFile = (dump).toURI().toURL();
+
+
+ if(convertNT2RDF){
+ NT2RDF.convertNT2RDF(dump.getAbsolutePath());
+
}
} catch (Exception e) {
e.printStackTrace();
Modified: trunk/src/dl-learner/org/dllearner/scripts/SemanticBible2.java
===================================================================
--- trunk/src/dl-learner/org/dllearner/scripts/SemanticBible2.java 2008-08-30 16:43:11 UTC (rev 1155)
+++ trunk/src/dl-learner/org/dllearner/scripts/SemanticBible2.java 2008-08-31 15:04:13 UTC (rev 1156)
@@ -20,35 +20,34 @@
package org.dllearner.scripts;
import java.io.File;
-import java.util.ArrayList;
import java.util.List;
+import java.util.SortedSet;
+import java.util.StringTokenizer;
+import java.util.TreeSet;
import org.apache.log4j.ConsoleAppender;
import org.apache.log4j.FileAppender;
import org.apache.log4j.Level;
import org.apache.log4j.Logger;
import org.apache.log4j.SimpleLayout;
-import org.dllearner.algorithms.refexamples.ExampleBasedROLComponent;
import org.dllearner.cli.Start;
import org.dllearner.core.Component;
import org.dllearner.core.ComponentManager;
-import org.dllearner.core.KnowledgeSource;
+import org.dllearner.core.EvaluatedDescription;
import org.dllearner.core.LearningAlgorithm;
import org.dllearner.core.ReasoningService;
-import org.dllearner.kb.extraction.ExtractionAlgorithm;
-import org.dllearner.kb.extraction.Manager;
-import org.dllearner.kb.sparql.Cache;
-import org.dllearner.kb.sparql.SparqlKnowledgeSource;
-import org.dllearner.kb.sparql.SparqlQuery;
-import org.dllearner.reasoning.FastInstanceChecker;
+import org.dllearner.core.owl.Description;
+import org.dllearner.core.owl.Individual;
import org.dllearner.utilities.Files;
-import org.dllearner.utilities.examples.AutomaticNegativeExampleFinderSPARQL;
-import org.dllearner.utilities.examples.AutomaticPositiveExampleFinderOWL;
-import org.dllearner.utilities.examples.AutomaticPositiveExampleFinderSPARQL;
+import org.dllearner.utilities.JamonMonitorLogger;
+import org.dllearner.utilities.StringFormatter;
+import org.dllearner.utilities.datastructures.SetManipulation;
import org.dllearner.utilities.owl.ReasoningServiceFactory;
import org.dllearner.utilities.owl.ReasoningServiceFactory.AvailableReasoners;
import org.dllearner.utilities.statistics.SimpleClock;
+import com.jamonapi.Monitor;
+
public class SemanticBible2 {
private static ReasoningService reasoningService;
@@ -62,7 +61,9 @@
public static String normaldir = dir+"normal/";
public static String tmpFilename = dir + "tmp.conf";
+ static File log = new File(dir+"results.txt");
+
//private static Class usedReasoner = FastInstanceChecker.class;
private static boolean useSPARQL = true;
@@ -71,90 +72,191 @@
* @param args
*/
public static void main(String[] args) {
- SimpleClock sc = new SimpleClock();
+
initLogger();
logger.info("Start");
File tmpFile = new File(tmpFilename);
+ String del="\t";
+ String cr="\n";
+ int max = 100;
+ SortedSet<String> confs = getFilesContaining(useSPARQL,"ten","all", "99+");
+ analyzeFiles(confs);
+ Files.createFile(log, "accOnFragment"+del+"accOnOnt"+del+"timeFragme"+del+"timeWhole"+cr);
- List<File> confs = getFilesContaining(useSPARQL,"ten","all", "99+");
- System.out.println(confs);
- //reasoningService = ReasoningServiceFactory.getReasoningService(ontologyPath, AvailableReasoners.OWLAPIREASONERPELLET);
+ reasoningService = ReasoningServiceFactory.getReasoningService(ontologyPath, AvailableReasoners.OWLAPIREASONERPELLET);
+ ComponentManager cm =ComponentManager.getInstance();
try{
- for (File file : confs) {
- System.out.println(file.getAbsolutePath());
- StringBuffer sbuf = new StringBuffer(Files.readFile( file));
+ int i=0;
+ for (String fileContent : confs) {
+ if(i>= max) {break;}i++;
+ String logLine ="";
+ SortedSet<Individual> posEx = SetManipulation.stringToInd(getIndividuals(fileContent, true));
+ SortedSet<Individual> negEx = SetManipulation.stringToInd(getIndividuals(fileContent, false));
+
+
+ StringBuffer sbuf = new StringBuffer(fileContent);
sbuf.insert(0, (useSPARQL)?sparqlOptions():normalOptions());
Files.createFile(tmpFile, sbuf.toString());
//System.out.println(tmpFile.getCanonicalPath());
+ Monitor m = JamonMonitorLogger.getTimeMonitor(SemanticBible2.class, "learn on fragment").start();
Start.main(new String[] { tmpFilename });
- ComponentManager cm =ComponentManager.getInstance();
- List<Component> comp = cm.getLiveComponents();
- for (Component component : comp) {
- System.out.println(component.getClass().getCanonicalName());
- if(component instanceof LearningAlgorithm){
- System.out.println("yyyy");
- System.exit(0);
- }
-
- }
+ m.stop();
+ LearningAlgorithm la = getLearningAlgorithm();
-
- Cache.getDefaultCache().clearCache();
+ EvaluatedDescription onFragment =(la.getCurrentlyBestEvaluatedDescription());
+ logLine += StringFormatter.doubleToPercent(onFragment.getAccuracy())+del;
+ SortedSet<Individual> retrieved = reasoningService.retrieval(onFragment.getDescription());
+ EvaluatedDescription onOnto = reEvaluateDescription(
+ onFragment.getDescription(), retrieved, posEx, negEx);
+ logLine += StringFormatter.doubleToPercent(onOnto.getAccuracy())+del;
+ logLine += m.getTotal()+del+"missing instead size of retrieve: "+retrieved.size()+cr;
+ Files.appendFile(log, logLine);
+ //Cache.getDefaultCache().clearCache();
cm.freeAllComponents();
System.exit(0);
}//end for
}catch (Exception e) {
e.printStackTrace();
}
- logger.info("finished");
+ logger.info("Finished");
}
- public static List<File> getFilesContaining(boolean sparql, String numExamples, String allOrEx, String acc) {
- List<File> ret = new ArrayList<File>();
- try{
+ public static EvaluatedDescription reEvaluateDescription(Description d, SortedSet<Individual> retrieved ,SortedSet<Individual> posEx ,SortedSet<Individual> negEx ){
+ SortedSet<Individual> PosAsPos = new TreeSet<Individual>();
+ SortedSet<Individual> PosAsNeg = new TreeSet<Individual>();
+ SortedSet<Individual> NegAsPos = new TreeSet<Individual>();
+ SortedSet<Individual> NegAsNeg = new TreeSet<Individual>();
+
+ // PosAsPos
+ PosAsPos.addAll(posEx);
+ PosAsPos.retainAll(retrieved);
+
+ // PosAsNeg
+ PosAsNeg.addAll(posEx);
+ PosAsNeg.removeAll(retrieved);
+
+ // NegAsPos
+ NegAsPos.addAll(negEx);
+ NegAsPos.retainAll(retrieved);
+
+ // PosAsNeg
+ NegAsNeg.addAll(negEx);
+ NegAsNeg.removeAll(retrieved);
+
+ return new EvaluatedDescription(d, PosAsPos, PosAsNeg, NegAsPos,NegAsNeg);
+
+ }
+
+ public static LearningAlgorithm getLearningAlgorithm(){
+ ComponentManager cm =ComponentManager.getInstance();
+
+ List<Component> comp = cm.getLiveComponents();
+ for (Component component : comp) {
+ if(component instanceof LearningAlgorithm){
+ return (LearningAlgorithm) component;
+ }
+
+ }
+ return null;
+ }
+
+ public static SortedSet<String> getFilesContaining(boolean sparql, String numExamples, String allOrEx, String acc) {
+ //List<File> ret = new ArrayList<File>();
+ SortedSet<String> ret = new TreeSet<String>();
+
String actualDir = (sparql)?sparqldir:normaldir;
- System.out.println(actualDir);
+ logger.info(actualDir);
File f = new File(actualDir);
String[] files = f.list();
-
+ int consistent = 0;
+ try{
for (int i = 0; i < files.length; i++) {
- System.out.println(files[i]);
- if(
- files[i].contains(numExamples)
+
+ if( files[i].contains(numExamples)
&& files[i].contains(allOrEx)
&& files[i].contains(acc)
){
- ret.add(new File(actualDir+files[i]));
+ consistent++;
+ ret.add(Files.readFile(new File(actualDir+files[i])));
+ if(ret.size() != consistent){
+ logger.info("double file: "+files[i]);
+ }
}
}
}catch (Exception e) {
- logger.warn("deleting cache failed");
+
e.printStackTrace();
}
+ if(consistent != ret.size()){
+ logger.info("double files"+consistent+"::"+ret.size());
+ System.exit(0);
+ }else{
+ logger.info("all files different");
+ }
return ret;
}
+ public static void analyzeFiles(SortedSet<String> l){
+ SortedSet<String> differentIndividuals = new TreeSet<String>();
+ for (String content : l) {
+ differentIndividuals.addAll(getIndividuals(content, true));
+ differentIndividuals.addAll(getIndividuals(content, false));
+
+ }
+ System.out.println("found diff inds "+differentIndividuals.size());
+
+ }
+ public static SortedSet<String> getIndividuals(String target, boolean posOrNeg){
+ if(posOrNeg){
+ return getAllStringsBetween(target, "+\"", "\"");
+ }else{
+ return getAllStringsBetween(target, "-\"", "\"");
+ }
+
+ }
+
+ public static SortedSet<String> getAllStringsBetween(String target, String start, String end){
+ SortedSet<String> ret = new TreeSet<String>();
+ StringTokenizer st = new StringTokenizer(target,"\n");
+ while(st.hasMoreElements()){
+ String line = st.nextToken();
+ if(line.contains(start)){
+ line = line.substring(line.indexOf(start)+start.length());
+ String current = line.substring(0,line.indexOf(end));
+ ret.add(current);
+ }
+ }
+
+ return ret;
+ }
+
+ public static String getCombinedOptions(){
+ String s="\n"+
+ "algorithm = refexamples;\n"+
+ "refexamples.useAllConstructor = true;\n"+
+ "refexamples.useNegation = true;\n"+
+ "refexamples.useCardinalityRestrictions = true;\n"+
+ "refexamples.guaranteeXgoodDescriptions = 1;\n"+
+ "refexamples.maxExecutionTimeInSeconds = 100;\n"+
+ "\n"+
+ "reasoner = owlAPI;\n"+
+ //"reasoner = fastInstanceChecker;\n"+
+ "owlAPIReasoner.reasonerType = pellet;\n\n"+
+ "";
+ return s;
+ }
+
public static String sparqlOptions (){
String s="// SPARQL options\n"+
"sparql.recursionDepth = 3;\n"+
"sparql.useLits = true;\n"+
"sparql.predefinedEndpoint = \"LOCALJOSEKIBIBLE\";\n"+
"import(\"lalala\",\"SPARQL\");\n"+
-
- "algorithm = refexamples;\n"+
- "refexamples.useAllConstructor = true;\n"+
- "refexamples.useNegation = true;\n"+
- "refexamples.useCardinalityRestrictions = true;\n"+
- "refexamples.guaranteeXgoodDescriptions = 1;\n"+
- "refexamples.maxExecutionTimeInSeconds = 1;\n"+
- "\n"+
- "reasoner = owlAPI;\n"+
- //"reasoner = fastInstanceChecker;\n"+
- //"owlAPIReasoner.reasonerType = pellet;\n\n";
+ getCombinedOptions()+
"";
return s;
}
@@ -162,15 +264,7 @@
public static String normalOptions (){
String s="\n"+
"import(\"NTNcombined.owl\");\n"+
- "algorithm = refexamples;\n"+
- "refexamples.useAllConstructor = true;\n"+
- "refexamples.useNegation = true;\n"+
- "refexamples.useCardinalityRestrictions = true;\n"+
- "refexamples.guaranteeXgoodDescriptions = 1;\n"+
- "\n"+
- "reasoner = owlAPI;\n"+
- //"reasoner = fastInstanceChecker;\n"+
- //"owlAPIReasoner.reasonerType = pellet;\n\n";
+ getCombinedOptions()+
"";
return s;
}
@@ -191,9 +285,9 @@
ConsoleAppender consoleAppender = new ConsoleAppender(layout);
logger.removeAllAppenders();
- logger.addAppender(consoleAppender);
+ //logger.addAppender(consoleAppender);
logger.addAppender(fileAppender);
- logger.setLevel(Level.DEBUG);
+ logger.setLevel(Level.INFO);
}
Added: trunk/src/dl-learner/org/dllearner/utilities/StringFormatter.java
===================================================================
--- trunk/src/dl-learner/org/dllearner/utilities/StringFormatter.java (rev 0)
+++ trunk/src/dl-learner/org/dllearner/utilities/StringFormatter.java 2008-08-31 15:04:13 UTC (rev 1156)
@@ -0,0 +1,26 @@
+package org.dllearner.utilities;
+
+
+public class StringFormatter {
+
+
+ /**
+ * formats a double value between 0 and 100 to a percentage
+ * ex: 0.7854684 will be return 78.5%
+ * @param d
+ * @return
+ */
+ public static String doubleToPercent(double d){
+ if(d>1.0 || d<0.0)return "bad format: "+d;
+ else if(d == 1.0){
+ return "100.0%";
+ }else if(d == 0.0 ){
+ return "0.0%";
+ }else {
+ String acc = (d*100)+"";
+ acc = acc.substring(0,"55.5".length());
+ return acc+"%";
+ }
+
+ }
+}
This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.
|
