Improved tree filter

SmartDataAnalytics · Aug 28, 2016 · 8443717 · 8443717
1 parent 34fb62f
commit 8443717
Show file tree

Hide file tree

Showing 4 changed files with 49 additions and 22 deletions.
diff --git a/...ore/src/main/java/org/dllearner/algorithms/qtl/util/filters/PredicateExistenceFilter.java b/...ore/src/main/java/org/dllearner/algorithms/qtl/util/filters/PredicateExistenceFilter.java
@@ -20,7 +20,10 @@
 
 import java.util.HashSet;
 import java.util.Set;
+import java.util.SortedSet;
 
+import com.google.common.collect.Sets;
+import org.apache.jena.vocabulary.RDF;
 import org.dllearner.algorithms.qtl.QueryTreeUtils;
 import org.dllearner.algorithms.qtl.datastructures.impl.RDFResourceTree;
 import org.dllearner.algorithms.qtl.util.Entailment;
@@ -30,6 +33,7 @@
 /**
  * A query tree filter that removes edges whose existence is supposed to be
  * semantically meaningless from user perspective.
+ *
  * @author Lorenz Buehmann
  *
  */
@@ -54,6 +58,7 @@ public void setExistentialMeaninglessProperties(Set<Node> existentialMeaningless
 
 	/**
 	 * Returns a new tree based on the input tree.
+	 *
 	 * @param tree the input tree
 	 * @return a filtered new tree
 	 */
@@ -67,10 +72,18 @@ public RDFResourceTree filter(RDFResourceTree tree) {
 
 		for(Node edge : tree.getEdges()) {
 			if(existentialMeaninglessProperties.contains(edge)) {
+				// if the edge is meaningless
+				// 1. process all children
 				for (RDFResourceTree child : tree.getChildren(edge)) {
-					if(child.isResourceNode() || child.isLiteralValueNode() || !child.isLeaf()) {
+					if(child.isResourceNode() || child.isLiteralValueNode()) {
 						RDFResourceTree newChild = filter(child);
 						newTree.addChild(newChild, edge);
+					} else {
+						RDFResourceTree newChild = filter(child);
+						SortedSet<Node> childEdges = newChild.getEdges();
+						if(!childEdges.isEmpty() && !(childEdges.size() == 1 && childEdges.contains(RDF.type.asNode()))) {
+							newTree.addChild(newChild, edge);
+						}
 					}
 				}
 			} else {

diff --git a/.../main/java/org/dllearner/algorithms/qtl/util/filters/PredicateExistenceFilterDBpedia.java b/.../main/java/org/dllearner/algorithms/qtl/util/filters/PredicateExistenceFilterDBpedia.java
@@ -18,15 +18,6 @@
  */
 package org.dllearner.algorithms.qtl.util.filters;
 
-import java.io.File;
-import java.io.IOException;
-import java.net.URISyntaxException;
-import java.util.List;
-import java.util.Set;
-import java.util.TreeSet;
-
-import org.dllearner.kb.SparqlEndpointKS;
-
 import com.google.common.base.Charsets;
 import com.google.common.io.Files;
 import org.apache.jena.graph.Node;
@@ -36,6 +27,14 @@
 import org.apache.jena.query.ResultSet;
 import org.apache.jena.rdf.model.Resource;
 import org.apache.jena.sparql.util.NodeComparator;
+import org.dllearner.kb.SparqlEndpointKS;
+
+import java.io.File;
+import java.io.IOException;
+import java.net.URISyntaxException;
+import java.util.List;
+import java.util.Set;
+import java.util.TreeSet;
 
 /**
  * @author Lorenz Buehmann

diff --git a/components-core/src/test/java/org/dllearner/algorithms/qtl/experiments/QTLEvaluation.java b/components-core/src/test/java/org/dllearner/algorithms/qtl/experiments/QTLEvaluation.java
@@ -488,6 +488,14 @@ public void run(int maxNrOfProcessedQueries, int maxTreeDepth, int[] exampleInte
 		logger.info("got {} empty queries.", emptyQueries.size());
 		queries.removeAll(emptyQueries);
 
+		// min. pos examples
+		Set<String> lowNrOfExamplesQueries = query2Examples.entrySet().stream()
+				.filter(e -> e.getValue().correctPosExampleCandidates.size() < 2)
+				.map(e -> e.getKey())
+				.collect(Collectors.toSet());
+		logger.info("got {} queries with < 2 pos. examples.", emptyQueries.size());
+		queries.removeAll(lowNrOfExamplesQueries);
+
 
 		final int totalNrOfQTLRuns = heuristics.length * this.measures.length * nrOfExamplesIntervals.length * noiseIntervals.length * queries.size();
 		logger.info("#QTL runs: " + totalNrOfQTLRuns);
@@ -569,7 +577,7 @@ public void run(int maxNrOfProcessedQueries, int maxTreeDepth, int[] exampleInte
 						// loop over SPARQL queries
 						for (final String sparqlQuery : queries) {
 
-							if(!(sparqlQuery.contains("Queen_Victoria")))continue;
+//							if(!(sparqlQuery.contains("Queen_Victoria")))continue;
 
 							tp.submit(() -> {
 
@@ -1439,11 +1447,11 @@ private List<RDFResourceTree> getQueryTrees(List<String> resources){
 
 	private RDFResourceTree getQueryTree(String resource){
 		// get CBD
-		logger.debug("loading data for {} ...", resource);
+		logger.info("loading data for {} ...", resource);
 		MonitorFactory.getTimeMonitor(TimeMonitors.CBD_RETRIEVAL.name()).start();
 		Model cbd = cbdGen.getConciseBoundedDescription(resource, maxTreeDepth);
 		MonitorFactory.getTimeMonitor(TimeMonitors.CBD_RETRIEVAL.name()).stop();
-		logger.debug("got {} triples in {}ms.", cbd.size(), MonitorFactory.getTimeMonitor(TimeMonitors.CBD_RETRIEVAL.name()).getLastValue());
+		logger.info("got {} triples in {}ms.", cbd.size(), MonitorFactory.getTimeMonitor(TimeMonitors.CBD_RETRIEVAL.name()).getLastValue());
 
 		// rewrite NAN to NaN to avoid parse exception
 		try(ByteArrayOutputStream baos = new ByteArrayOutputStream()){
@@ -1459,11 +1467,11 @@ private RDFResourceTree getQueryTree(String resource){
 
 
 		// generate tree
-		logger.debug("generating query tree for {} ...", resource);
+		logger.info("generating query tree for {} ...", resource);
 		MonitorFactory.getTimeMonitor(TimeMonitors.TREE_GENERATION.name()).start();
 		RDFResourceTree tree = queryTreeFactory.getQueryTree(resource, cbd, maxTreeDepth);
 		MonitorFactory.getTimeMonitor(TimeMonitors.TREE_GENERATION.name()).stop();
-		logger.debug("generating query tree for {} took {}ms.", resource, MonitorFactory.getTimeMonitor(TimeMonitors.TREE_GENERATION.name()).getLastValue());
+		logger.info("generating query tree for {} took {}ms.", resource, MonitorFactory.getTimeMonitor(TimeMonitors.TREE_GENERATION.name()).getLastValue());
 
 //		System.out.println(tree.getStringRepresentation());
 
@@ -1810,7 +1818,14 @@ private Score computeScore(String referenceSparqlQuery, RDFResourceTree tree, do
 		// apply some filters
 		QueryTreeUtils.removeVarLeafs(tree);
 		QueryTreeUtils.prune(tree, null, Entailment.RDF);
-
+
+		// remove redundant rdf:type triples
+		QueryTreeUtils.keepMostSpecificTypes(tree, dataset.getReasoner());
+
+		//
+		PredicateExistenceFilter filter = new PredicateExistenceFilterDBpedia(null);
+		tree = filter.filter(tree);
+
 		String learnedSPARQLQuery = QueryTreeUtils.toSPARQLQueryString(tree, dataset.getBaseIRI(), dataset.getPrefixMapping());
 		logger.info("learned SPARQL query:{}", learnedSPARQLQuery);
 

diff --git a/...s-core/src/test/resources/org/dllearner/algorithms/qtl/dbpedia_meaningless_properties.txt b/...s-core/src/test/resources/org/dllearner/algorithms/qtl/dbpedia_meaningless_properties.txt
@@ -477,9 +477,9 @@ http://dbpedia.org/ontology/dbnlCodeDutch
 http://dbpedia.org/ontology/dcc
 http://dbpedia.org/ontology/deadInFightDate
 http://dbpedia.org/ontology/deadInFightPlace
-http://dbpedia.org/ontology/deathAge
-http://dbpedia.org/ontology/deathDate
-http://dbpedia.org/ontology/deathYear
+#http://dbpedia.org/ontology/deathAge
+#http://dbpedia.org/ontology/deathDate
+#http://dbpedia.org/ontology/deathYear
 http://dbpedia.org/ontology/debut
 http://dbpedia.org/ontology/dec
 http://dbpedia.org/ontology/decay
@@ -2052,8 +2052,8 @@ http://dbpedia.org/ontology/daylightSavingTimeZone
 http://dbpedia.org/ontology/deFactoLanguage
 http://dbpedia.org/ontology/dean
 http://dbpedia.org/ontology/deanery
-http://dbpedia.org/ontology/deathCause
-http://dbpedia.org/ontology/deathPlace
+#http://dbpedia.org/ontology/deathCause
+#http://dbpedia.org/ontology/deathPlace
 http://dbpedia.org/ontology/debutTeam
 http://dbpedia.org/ontology/debutWork
 http://dbpedia.org/ontology/decoration
@@ -2742,7 +2742,7 @@ http://dbpedia.org/ontology/sportCountry
 http://dbpedia.org/ontology/sportDiscipline
 http://dbpedia.org/ontology/sportGoverningBody
 http://dbpedia.org/ontology/sportSpecialty
-http://dbpedia.org/ontology/spouse
+#http://dbpedia.org/ontology/spouse
 http://dbpedia.org/ontology/spurOf
 http://dbpedia.org/ontology/spurType
 http://dbpedia.org/ontology/stadium