Menu
▾
▴
[DL-Learner SVN] SF.net SVN: dl-learner: [329] trunk
|
From: <ku...@us...> - 2007-12-05 16:16:06
|
Revision: 329
http://dl-learner.svn.sourceforge.net/dl-learner/?rev=329&view=rev
Author: kurzum
Date: 2007-12-05 08:15:59 -0800 (Wed, 05 Dec 2007)
Log Message:
-----------
some minor changes, role domain learning doesn't really work, because the instances have to be chosen in a smarter way.
Modified Paths:
--------------
trunk/src/dl-learner/org/dllearner/kb/SparqlEndpointRestructured.java
trunk/src/dl-learner/org/dllearner/kb/sparql/ClassNode.java
trunk/src/dl-learner/org/dllearner/kb/sparql/InstanceNode.java
trunk/src/dl-learner/org/dllearner/kb/sparql/Manager.java
trunk/src/dl-learner/org/dllearner/kb/sparql/PredefinedEndpoint.java
trunk/src/dl-learner/org/dllearner/kb/sparql/PropertyNode.java
Added Paths:
-----------
trunk/examples/sparql/role_domain_placeOfBirth_yago.conf
trunk/examples/sparql/role_range_placeOfBirth_yago.conf
Added: trunk/examples/sparql/role_domain_placeOfBirth_yago.conf
===================================================================
--- trunk/examples/sparql/role_domain_placeOfBirth_yago.conf (rev 0)
+++ trunk/examples/sparql/role_domain_placeOfBirth_yago.conf 2007-12-05 16:15:59 UTC (rev 329)
@@ -0,0 +1,58 @@
+/**
+ *
+ * solutions:
+ *
+ */
+
+
+// recursion depth of extraction algorithm
+sparql2.recursionDepth = 1;
+
+// list of ignored roles
+sparql2.predList = {
+ "http://www.w3.org/2004/02/skos/core",
+ "http://www.w3.org/2002/07/owl#sameAs",
+ "http://xmlns.com/foaf/0.1/",
+ "http://dbpedia.org/property/reference",
+ "http://dbpedia.org/property/website",
+ "http://dbpedia.org/property/wikipage"
+};
+
+// list of ignored objects
+sparql2.objList = {
+ "http://dbpedia.org/resource/Category:Wikipedia_",
+ "http://dbpedia.org/resource/Category:Articles_",
+ "http://xmlns.com/foaf/0.1/",
+ "http://upload.wikimedia.org/wikipedia/commons",
+ "http://upload.wikimedia.org/wikipedia",
+ "http://www.geonames.org",
+ "http://www.w3.org/2006/03/wn/wn20/instances/synset",
+ "http://www4.wiwiss.fu-berlin.de/flickrwrappr",
+ "http://www.w3.org/2004/02/skos/core"
+};
+
+// use DBpedia endpoint
+import("http://dbpedia.openlinksw.com:8890/sparql","SPARQL2");
+
+// the set of objects as starting point for fragment selection
+// (should be identical to the set of examples)
+sparql2.instances = {
+ "http://dbpedia.org/resource/Angela_Merkel",
+ "http://dbpedia.org/resource/Gerhard_Schr%C3%B6der"
+};
+sparql2.role = "http://dbpedia.org/property/placeOfBirth";
+sparql2.learnDomain=true;
+sparql2.learnRange=false;
+
+
+/** examples **/
++"http://dbpedia.org/resource/Richard_Burton"
++"http://dbpedia.org/resource/Gregory_Bateson"
++"http://dbpedia.org/resource/Gennaro_Gattuso"
++"http://dbpedia.org/resource/Thomas_Bayes"
++"http://dbpedia.org/resource/W%C5%82adys%C5%82aw_Anders"
+-"http://dbpedia.org/resource/Roman_Dacia"
+-"http://dbpedia.org/resource/Petrovichi"
+-"http://dbpedia.org/resource/Scotland"
+-"http://dbpedia.org/resource/Lanuvium"
+-"http://dbpedia.org/resource/Riga"
\ No newline at end of file
Added: trunk/examples/sparql/role_range_placeOfBirth_yago.conf
===================================================================
--- trunk/examples/sparql/role_range_placeOfBirth_yago.conf (rev 0)
+++ trunk/examples/sparql/role_range_placeOfBirth_yago.conf 2007-12-05 16:15:59 UTC (rev 329)
@@ -0,0 +1,59 @@
+/**
+ *
+ * solutions:
+ *(http://dbpedia.org/class/yago/Isle109319456 OR http://dbpedia.org/class/yago/Location100027167)
+ */
+
+
+// recursion depth of extraction algorithm
+sparql2.recursionDepth = 1;
+
+// list of ignored roles
+sparql2.predList = {
+ "http://www.w3.org/2004/02/skos/core",
+ "http://www.w3.org/2002/07/owl#sameAs",
+ "http://xmlns.com/foaf/0.1/",
+ "http://dbpedia.org/property/reference",
+ "http://dbpedia.org/property/website",
+ "http://dbpedia.org/property/wikipage"
+};
+
+// list of ignored objects
+sparql2.objList = {
+ "http://dbpedia.org/resource/Category:Wikipedia_",
+ "http://dbpedia.org/resource/Category:Articles_",
+ "http://xmlns.com/foaf/0.1/",
+ "http://upload.wikimedia.org/wikipedia/commons",
+ "http://upload.wikimedia.org/wikipedia",
+ "http://www.geonames.org",
+ "http://www.w3.org/2006/03/wn/wn20/instances/synset",
+ "http://www4.wiwiss.fu-berlin.de/flickrwrappr",
+ "http://www.w3.org/2004/02/skos/core"
+};
+
+// use DBpedia endpoint
+import("http://dbpedia.openlinksw.com:8890/sparql","SPARQL2");
+
+// the set of objects as starting point for fragment selection
+// (should be identical to the set of examples)
+sparql2.instances = {
+ "http://dbpedia.org/resource/Angela_Merkel",
+ "http://dbpedia.org/resource/Gerhard_Schr%C3%B6der"
+};
+
+sparql2.role = "http://dbpedia.org/property/placeOfBirth";
+sparql2.learnDomain=false;
+sparql2.learnRange=true;
+
+
+/** examples **/
++"http://dbpedia.org/resource/Roman_Dacia"
+//+"http://dbpedia.org/resource/Petrovichi"
++"http://dbpedia.org/resource/Scotland"
++"http://dbpedia.org/resource/Lanuvium"
++"http://dbpedia.org/resource/Riga"
+-"http://dbpedia.org/resource/Richard_Burton"
+-"http://dbpedia.org/resource/Gregory_Bateson"
+-"http://dbpedia.org/resource/Gennaro_Gattuso"
+-"http://dbpedia.org/resource/Thomas_Bayes"
+-"http://dbpedia.org/resource/W%C5%82adys%C5%82aw_Anders"
\ No newline at end of file
Modified: trunk/src/dl-learner/org/dllearner/kb/SparqlEndpointRestructured.java
===================================================================
--- trunk/src/dl-learner/org/dllearner/kb/SparqlEndpointRestructured.java 2007-12-05 13:09:58 UTC (rev 328)
+++ trunk/src/dl-learner/org/dllearner/kb/SparqlEndpointRestructured.java 2007-12-05 16:15:59 UTC (rev 329)
@@ -26,11 +26,11 @@
import java.net.MalformedURLException;
import java.net.URI;
import java.net.URL;
-import java.util.ArrayList;
import java.util.Collection;
import java.util.HashMap;
import java.util.HashSet;
import java.util.LinkedList;
+import java.util.Random;
import java.util.Set;
import org.dllearner.core.KnowledgeSource;
@@ -83,6 +83,7 @@
private int breakSuperClassRetrievalAfter = 500;
private boolean learnDomain = false;
+ private boolean learnRange = false;
private String role;
private String blankNodeIdentifier = "bnode";
@@ -160,7 +161,8 @@
options.add(new BooleanConfigOption("getAllSuperClasses", "If true then all superclasses are retrieved until the most general class (owl:Thing) is reached.", true));
options.add(new BooleanConfigOption("learnDomain", "learns the Domain for a Role"));
- options.add(new StringConfigOption("role", "role to learn Domain from"));
+ options.add(new BooleanConfigOption("learnRange", "learns the Range for a Role"));
+ options.add(new StringConfigOption("role", "role to learn Domain/Range from"));
options.add(new StringConfigOption("blankNodeIdentifier",
"used to identify blanknodes in Tripels"));
@@ -214,6 +216,8 @@
getAllSuperClasses = (Boolean) entry.getValue();
} else if (option.equals("learnDomain")) {
learnDomain = (Boolean) entry.getValue();
+ }else if (option.equals("learnRange")) {
+ learnRange = (Boolean) entry.getValue();
} else if (option.equals("role")) {
role = (String) entry.getValue();
} else if (option.equals("blankNodeIdentifier")) {
@@ -260,32 +264,76 @@
}
// get Options for Filters
- System.out.println("aaa"+predefinedFilter);
+
if (predefinedFilter >= 1) {
sqt = PredefinedFilter.getFilter(predefinedFilter);
} else {
sqt = new SparqlQueryType("forbid", objList, predList, useLits + "");
- System.out.println(sqt);
+
}
// give everything to the manager
m.useConfiguration(sqt, sse, man, recursionDepth, getAllSuperClasses);
try {
String ont = "";
// used to learn a domain of a role
- if (learnDomain) {
- instances = m.getDomainInstancesForRole(role);
+ if (learnDomain || learnRange) {
+ Set<String> pos=new HashSet<String>();
+ Set<String> neg=new HashSet<String>();
+ if(learnDomain){
+ pos = m.getDomainInstancesForRole(role);
+ neg = m.getRangeInstancesForRole(role);
+ }else if(learnRange){
+ neg = m.getDomainInstancesForRole(role);
+ pos = m.getRangeInstancesForRole(role);
+ }
+ //choose 30
+
+
+ Set<String> tmp=new HashSet<String>();
+ for(String one:pos){
+ tmp.add(one);
+ if(tmp.size()>=5)break;
+ }
+ pos=tmp;
+ System.out.println(pos.size());
+
+ tmp=new HashSet<String>();
+ for(String one:neg){
+ tmp.add(one);
+ if(tmp.size()>=5)break;
+ }
+ neg=tmp;
+
+ instances=new HashSet<String>();
+ instances.addAll( pos);
+
+ instances.addAll(neg);
+
+ for(String one:pos){
+ System.out.println("+\""+one+"\"");
+ }
+ for(String one:neg){
+ System.out.println("-\""+one+"\"");
+ }
+
+ /*Random r= new Random();
+
+
+ Object[] arr=instances.toArray();
+ while(instances.size()>=30){
+
+ }*/
// add the role to the filter(a solution is always EXISTS
// role.TOP)
m.addPredicateFilter(role);
//System.out.println(instances);
// THIS is a workaround
- for(String one:instances){
- System.out.println("+\""+one+"\"");
- }
+
}
// the actual extraction is started here
ont = m.extract(instances);
+ System.out.println("Finished collecting Fragment");
if (dumpToFile) {
String filename = System.currentTimeMillis() + ".nt";
Modified: trunk/src/dl-learner/org/dllearner/kb/sparql/ClassNode.java
===================================================================
--- trunk/src/dl-learner/org/dllearner/kb/sparql/ClassNode.java 2007-12-05 13:09:58 UTC (rev 328)
+++ trunk/src/dl-learner/org/dllearner/kb/sparql/ClassNode.java 2007-12-05 16:15:59 UTC (rev 329)
@@ -23,6 +23,8 @@
import java.util.HashSet;
import java.util.Iterator;
import java.util.Set;
+import java.util.SortedSet;
+import java.util.TreeSet;
import java.util.Vector;
import org.dllearner.utilities.StringTuple;
Modified: trunk/src/dl-learner/org/dllearner/kb/sparql/InstanceNode.java
===================================================================
--- trunk/src/dl-learner/org/dllearner/kb/sparql/InstanceNode.java 2007-12-05 13:09:58 UTC (rev 328)
+++ trunk/src/dl-learner/org/dllearner/kb/sparql/InstanceNode.java 2007-12-05 16:15:59 UTC (rev 329)
@@ -23,6 +23,8 @@
import java.util.HashSet;
import java.util.Iterator;
import java.util.Set;
+import java.util.SortedSet;
+import java.util.TreeSet;
import java.util.Vector;
import org.dllearner.utilities.StringTuple;
Modified: trunk/src/dl-learner/org/dllearner/kb/sparql/Manager.java
===================================================================
--- trunk/src/dl-learner/org/dllearner/kb/sparql/Manager.java 2007-12-05 13:09:58 UTC (rev 328)
+++ trunk/src/dl-learner/org/dllearner/kb/sparql/Manager.java 2007-12-05 16:15:59 UTC (rev 329)
@@ -22,6 +22,8 @@
import java.net.URI;
import java.util.HashSet;
import java.util.Set;
+import java.util.SortedSet;
+import java.util.TreeSet;
import org.dllearner.utilities.StringTuple;
@@ -53,6 +55,7 @@
Set<StringTuple> t = this.typedSparqlQuery.getTupelsForRole(u);
Set<String> ret = new HashSet<String>();
for (StringTuple one : t) {
+
ret.add(one.a);
}
return ret;
@@ -68,6 +71,7 @@
Set<StringTuple> t = this.typedSparqlQuery.getTupelsForRole(u);
Set<String> ret = new HashSet<String>();
for (StringTuple one : t) {
+
ret.add(one.b);
}
return ret;
@@ -91,7 +95,7 @@
// this.TypedSparqlQuery.query(uri);
// System.out.println(ExtractionAlgorithm.getFirstNode(uri));
System.out.println("Start extracting");
- Set<String> ret = new HashSet<String>();
+ SortedSet<String> ret = new TreeSet<String>();
for (String one : instances) {
try {
@@ -101,12 +105,26 @@
e.printStackTrace();
}
}
-
- String nt = "";
- for (String str : ret) {
+ System.out.println("Finished extracting, start conversion");
+ StringBuffer nt = new StringBuffer();
+ Object[] arr=ret.toArray();
+ for (int i = 0; i < arr.length; i++) {
+ nt.append((String) arr[i]+"\n");
+ if(i%1000==0)System.out.println(i+" of "+arr.length+" triples done");
+ }
+ /*
+ String tmp="";
+ while ( ret.size() > 0) {
+ tmp=ret.first();
+ nt+=tmp;
+ ret.remove(tmp);
+ System.out.println(ret.size());
+
+ }
+ /*for (String str : ret) {
nt += str + "\n";
- }
- return nt;
+ }*/
+ return nt.toString();
}
public void addPredicateFilter(String str) {
Modified: trunk/src/dl-learner/org/dllearner/kb/sparql/PredefinedEndpoint.java
===================================================================
--- trunk/src/dl-learner/org/dllearner/kb/sparql/PredefinedEndpoint.java 2007-12-05 13:09:58 UTC (rev 328)
+++ trunk/src/dl-learner/org/dllearner/kb/sparql/PredefinedEndpoint.java 2007-12-05 16:15:59 UTC (rev 329)
@@ -76,6 +76,22 @@
}
return new SpecificSparqlEndpoint(u, "www4.wiwiss.fu-berlin.de", m);
}
+
+ /*
+ * it only has 4 classes
+ public static SpecificSparqlEndpoint dblp() {
+ URL u = null;
+ HashMap<String, String> m = new HashMap<String, String>();
+ // m.put("default-graph-uri", "http://dbpedia.org");
+ // m.put("format", "application/sparql-results.xml");
+ try {
+ u = new URL("http://www4.wiwiss.fu-berlin.de/dblp/sparql");
+ } catch (Exception e) {
+ e.printStackTrace();
+ }
+ return new SpecificSparqlEndpoint(u, "www4.wiwiss.fu-berlin.de", m);
+ }
+ */
public static SpecificSparqlEndpoint govTrack() {
URL u = null;
HashMap<String, String> m = new HashMap<String, String>();
@@ -88,5 +104,55 @@
}
return new SpecificSparqlEndpoint(u, "www.rdfabout.com", m);
}
+ public static SpecificSparqlEndpoint revyu() {
+ URL u = null;
+ HashMap<String, String> m = new HashMap<String, String>();
+ // m.put("default-graph-uri", "http://dbpedia.org");
+ // m.put("format", "application/sparql-results.xml");
+ //http://revyu.com/sparql?query=SELECT DISTINCT * WHERE {[] a ?c}
+ try {
+ u = new URL("http://revyu.com/sparql");
+ } catch (Exception e) {
+ e.printStackTrace();
+ }
+ return new SpecificSparqlEndpoint(u, "revyu.com", m);
+ }
+ // returns strange xml
+ /*public static SpecificSparqlEndpoint dbtune() {
+ URL u = null;
+ HashMap<String, String> m = new HashMap<String, String>();
+ // m.put("default-graph-uri", "http://dbpedia.org");
+ // m.put("format", "application/sparql-results.xml");
+ //http://dbtune.org:2020/sparql/?query=SELECT DISTINCT * WHERE {[] a ?c}Limit 10
+ http://dbtune.org:2020/evaluateQuery?repository=default&serialization=rdfxml&queryLanguage=SPARQL&query=SELECT+DISTINCT+*+WHERE+%7B%5B%5D+a+%3Fc%7D
+ &resultFormat=xml
+ &resourceFormat=ns&entailment=none
+ http://dbtune.org:2020/evaluateQuery
+ ?repository=default&serialization=rdfxml&queryLanguage=SPARQL
+ &query=SELECT+DISTINCT+*+WHERE+%7B%5B%5D+a+%3Fc%7D
+ &resultFormat=xml
+ &resourceFormat=ns&entailment=none
+ try {
+ u = new URL("http://dbtune.org:2020/sparql/");
+ } catch (Exception e) {
+ e.printStackTrace();
+ }
+ return new SpecificSparqlEndpoint(u, "dbtune.org", m);
+ }*/
+
+ public static SpecificSparqlEndpoint myopenlink() {
+ URL u = null;
+ HashMap<String, String> m = new HashMap<String, String>();
+ m.put("default-graph-uri", "http://myopenlink.net/dataspace");
+ m.put("format", "application/sparql-results.xml");
+ //http://myopenlink.net:8890/sparql/?query=select+distinct+%3FConcept+where+%7B%5B%5D+a+%3FConcept%7D
+ try {
+ u = new URL("http://myopenlink.net:8890/sparql/");
+ } catch (Exception e) {
+ e.printStackTrace();
+ }
+ return new SpecificSparqlEndpoint(u, "myopenlink.net", m);
+ }
+
}
Modified: trunk/src/dl-learner/org/dllearner/kb/sparql/PropertyNode.java
===================================================================
--- trunk/src/dl-learner/org/dllearner/kb/sparql/PropertyNode.java 2007-12-05 13:09:58 UTC (rev 328)
+++ trunk/src/dl-learner/org/dllearner/kb/sparql/PropertyNode.java 2007-12-05 16:15:59 UTC (rev 329)
@@ -23,6 +23,8 @@
import java.util.HashSet;
import java.util.Iterator;
import java.util.Set;
+import java.util.SortedSet;
+import java.util.TreeSet;
import java.util.Vector;
import org.dllearner.utilities.StringTuple;
This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.
|
