使用Pellet作为推理器的OutOfMemoryError

Question

我尝试使用Pellet推断数据，但我的应用程序在推理过程中总是因为"OutOfMemoryError: Java heap space“而崩溃。

正如其他地方所建议的那样，我已经尝试增加堆大小，目前我正在使用这些VM参数："-XX:MaxPermSize=256m -Xmx6144m“。此外，我必须设置'PelletOptions.USE_CONTINUOUS = true'，但到目前为止，我只能将崩溃延迟几分钟。(到目前为止最长的跑步时间: 30分钟)。

我将数据存储在Jena TDB triplestore中。它的总大小是170MB，由962117条语句组成。本体已经在Protege中建模，当与Hermit或Fact++进行推理时，没有不一致之处(当然，在这一点上没有数据)。我想要推断的语句已经使用等价的类(例如:A和(hasX some X))、逆属性(hasX <--> isXOf)和一些传递属性进行了建模。有55个类: 14个根类，11个等价类(一个根类的所有子类型)。目前，已将14个类的实例添加到tripestore中。

我也使用了pellint来分析我的本体，但我不确定从这个结果中得出哪些结论：

[Untyped classes]
- http://www.w3.org/2002/07/owl#AllDisjointClasses

[Untyped object roles]
- http://www.w3.org/2002/07/owl#topObjectPriority
- http://www.w3.org/2002/07/owl#members

[Untyped datatype roles]
- http://www.w3.org/2002/07/owl#qualifiedCardinality

[Untyped individuals]
- 4 BNodes

下面是我创建InfModel的代码：

 Dataset dataset = TDBFactory.createDataset(..);
 Model model = dataset.getDefaultModel();

 OntModel ont = ModelFactory.createOntologyModel(PelletReasonerFactory.THE_SPEC,model);
 ont.read("file:.." , "RDF/XML");

 PelletOptions.USE_CONTINUOUS_RULES = true;
 Reasoner reasoner = PelletReasonerFactory.theInstance().create();
 reasoner = reasoner.bindSchema(ont);

 InfModel infModel = infModel = ModelFactory.createInfModel(reasoner, model);

这就是我查询数据的基本方式：

 QueryExecution qe = QueryExecutionFactory.create("SELECT ...", infModel);
 Resultset rs = qe.execSelect();

 while(rs.hasNext())
 {
     QuerySolution sol = rs.nextSolution();
     ...
 }
 qe.close();

正如您所看到的，在查询数据、准备或验证InfModel时，我收到了各种异常。你有什么想法，我可以尝试解决这个错误吗？

 java.lang.OutOfMemoryError: Java heap space
 at java.util.Arrays.copyOf(Unknown Source)
 at java.lang.AbstractStringBuilder.expandCapacity(Unknown Source)
 at java.lang.AbstractStringBuilder.append(Unknown Source)
 at java.lang.StringBuilder.append(Unknown Source)
 at java.net.URLStreamHandler.parseURL(Unknown Source)
 at sun.net.www.protocol.file.Handler.parseURL(Unknown Source)
 at java.net.URL.<init>(Unknown Source)
 at java.net.URL.<init>(Unknown Source)
 at sun.misc.URLClassPath$FileLoader.getResource(Unknown Source)
 at sun.misc.URLClassPath.getResource(Unknown Source)
 at java.net.URLClassLoader$1.run(Unknown Source)
 at java.security.AccessController.doPrivileged(Native Method)
 at java.net.URLClassLoader.findClass(Unknown Source)
 at java.lang.ClassLoader.loadClass(Unknown Source)
 at sun.misc.Launcher$AppClassLoader.loadClass(Unknown Source)
 at java.lang.ClassLoader.loadClass(Unknown Source)
 at com.hp.hpl.jena.sparql.engine.iterator.QueryIteratorCheck.warn(QueryIteratorCheck.java:114)
 at com.hp.hpl.jena.sparql.engine.iterator.QueryIteratorCheck.dump(QueryIteratorCheck.java:95)
 at com.hp.hpl.jena.sparql.engine.iterator.QueryIteratorCheck.checkForOpenIterators(QueryIteratorCheck.java:68)
 at com.hp.hpl.jena.sparql.engine.iterator.QueryIteratorCheck.close(QueryIteratorCheck.java:50)
 at com.hp.hpl.jena.sparql.engine.iterator.QueryIteratorWrapper.closeIterator(QueryIteratorWrapper.java:50)
 at com.hp.hpl.jena.sparql.engine.iterator.QueryIteratorBase.close(QueryIteratorBase.java:184)
 at com.hp.hpl.jena.sparql.engine.iterator.QueryIteratorCloseable.close(QueryIteratorCloseable.java:39)
 at com.hp.hpl.jena.sparql.engine.QueryExecutionBase.close(QueryExecutionBase.java:180)


 java.lang.OutOfMemoryError: Java heap space
 at shared.SharedObjectFactory.build(SharedObjectFactory.java:303)
 at aterm.pure.PureFactory.makeAppl(PureFactory.java:191)
 at aterm.pure.PureFactory.makeAppl(PureFactory.java:185)
 at aterm.pure.PureFactory.makeAppl(PureFactory.java:228)
 at org.mindswap.pellet.utils.ATermUtils.makeLiteral(ATermUtils.java:245)
 at org.mindswap.pellet.ABox.addLiteral(ABox.java:1696)
 at org.mindswap.pellet.tableau.completion.rule.SomeValuesRule.applySomeValuesRule(SomeValuesRule.java:174)
 at org.mindswap.pellet.tableau.completion.rule.SomeValuesRule.apply(SomeValuesRule.java:64)
 at org.mindswap.pellet.tableau.completion.rule.AbstractTableauRule.apply(AbstractTableauRule.java:64)
 at org.mindswap.pellet.tableau.completion.SROIQStrategy.complete(SROIQStrategy.java:157)
 at org.mindswap.pellet.ABox.isConsistent(ABox.java:1423)
 at org.mindswap.pellet.ABox.isConsistent(ABox.java:1260)
 at org.mindswap.pellet.KnowledgeBase.consistency(KnowledgeBase.java:1987)
 at org.mindswap.pellet.KnowledgeBase.isConsistent(KnowledgeBase.java:2061)
 at org.mindswap.pellet.jena.PelletInfGraph.prepare(PelletInfGraph.java:247)
 at org.mindswap.pellet.jena.PelletInfGraph.prepare(PelletInfGraph.java:230)
 at com.hp.hpl.jena.rdf.model.impl.InfModelImpl.prepare(InfModelImpl.java:89)


 java.lang.OutOfMemoryError: Java heap space
 at org.mindswap.pellet.tableau.branch.DisjunctionBranch.tryBranch(DisjunctionBranch.java:197)
 at org.mindswap.pellet.tableau.branch.Branch.tryNext(Branch.java:117)
 at org.mindswap.pellet.tableau.completion.rule.ChooseRule.apply(ChooseRule.java:82)
 at org.mindswap.pellet.tableau.completion.rule.ChooseRule.apply(ChooseRule.java:56)
 at org.mindswap.pellet.tableau.completion.rule.AbstractTableauRule.apply(AbstractTableauRule.java:64)
 at org.mindswap.pellet.tableau.completion.SROIQStrategy.complete(SROIQStrategy.java:157)
 at org.mindswap.pellet.ABox.isConsistent(ABox.java:1423)
 at org.mindswap.pellet.ABox.isConsistent(ABox.java:1260)
 at org.mindswap.pellet.KnowledgeBase.consistency(KnowledgeBase.java:1987)
 at org.mindswap.pellet.KnowledgeBase.isConsistent(KnowledgeBase.java:2061)
 at org.mindswap.pellet.jena.PelletInfGraph.prepare(PelletInfGraph.java:247)
 at org.mindswap.pellet.jena.PelletInfGraph.prepare(PelletInfGraph.java:230)
 at org.mindswap.pellet.jena.PelletInfGraph.validate(PelletInfGraph.java:564)
 at com.hp.hpl.jena.rdf.model.impl.InfModelImpl.validate(InfModelImpl.java:111)

Answer 1

对infModel的查询显然导致了Pellet生成成百上千条owl:differentFrom语句。解决方案是使用ModelExtractor

ModelExtractor ext = new ModelExtractor(inf); 
Model m = ext.extractModel();
QueryExecution qe = QueryExecutionFactory.create("SELECT ...", m);