耶拿;属性的域和范围因所选推理器而异

时间:2015-03-04 11:56:31

标签: java jena owl reasoning pellet

使用Jena,我正在尝试获取属性的域和范围。

让我们考虑以下本体

 @prefix : <http://www.semanticweb.org/eng.medianhilal/ontologies/2015/2/untitled-ontology-82#> .
 @prefix owl: <http://www.w3.org/2002/07/owl#> .
 @prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> .
 @prefix xml: <http://www.w3.org/XML/1998/namespace> .
 @prefix xsd: <http://www.w3.org/2001/XMLSchema#> .
 @prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> .
 @base <http://www.semanticweb.org/eng.medianhilal/ontologies/2015/2/untitled-ontology-82> .

<http://www.semanticweb.org/eng.medianhilal/ontologies/2015/2/untitled-ontology-82> rdf:type owl:Ontology .

:P rdf:type owl:ObjectProperty ;   
   rdfs:domain :A1 ;   
   rdfs:range :B1 .

:A rdf:type owl:Class .

:A1 rdf:type owl:Class ;    
    rdfs:subClassOf :A .

:A2 rdf:type owl:Class ;    
    rdfs:subClassOf :A1 .

:B rdf:type owl:Class .

:B1 rdf:type owl:Class ;
    rdfs:subClassOf :B .

:B2 rdf:type owl:Class ;    
    rdfs:subClassOf :B1 .

我们可以看到,A1是P的域,B1是它的范围。根据OWL语义,我们可以推断A也是P的域,而B也是它的范围see here

然而,使用Jena和推理员并不总是给出预期的行为。让我们区分两种情况,第一种是使用Pellet推理器,第二种是使用OWL_DL_MEM_RULE_INF

代码

import org.mindswap.pellet.jena.PelletReasonerFactory;
import com.hp.hpl.jena.ontology.OntClass;
import com.hp.hpl.jena.ontology.OntModel;
import com.hp.hpl.jena.ontology.OntProperty;
import com.hp.hpl.jena.rdf.model.ModelFactory;
import com.hp.hpl.jena.rdf.model.Resource;
import com.hp.hpl.jena.util.iterator.ExtendedIterator;

public class Test{
    public static void main (String [] args)
{                       

    OntModel ontModel = ModelFactory.createOntologyModel(PelletReasonerFactory.THE_SPEC);
    /*OntModel ontModel = ModelFactory.createOntologyModel(OntModelSpec.OWL_DL_MEM_RULE_INF);*/
    ontModel.read("path_to_ontology",  "RDF/XML");                              
    ontModel.setStrictMode(false); 


    String myNS = ontModel.getNsPrefixURI("");              

Resource r = ontModel.getResource(myNS + "P" );     
    OntProperty prop = (OntProperty) r.as( OntProperty.class);
    ExtendedIterator <OntClass> opDomains = (ExtendedIterator <OntClass>) prop.listDomain();                        
    while(opDomains.hasNext()){
        OntClass domain = opDomains.next();
        System.out.println("DOMAIN: " + domain.getURI());
    }       

    ExtendedIterator <OntClass> opRanges = (ExtendedIterator <OntClass>) prop.listRange();
    while(opRanges.hasNext()){
        OntClass ran = opRanges.next();
        System.out.println("RANGE: " + ran.getURI());
    }
}
}

使用Pellet:这会提供以下输出:

DOMAIN: http://www.semanticweb.org/eng.medianhilal/ontologies/2015/2/untitled-ontology-82#A1
RANGE: http://www.semanticweb.org/eng.medianhilal/ontologies/2015/2/untitled-ontology-82#B1

使用OWL_DL_MEM_RULE_INF:这会提供以下输出:

DOMAIN: http://www.w3.org/2002/07/owl#Thing
DOMAIN: http://www.semanticweb.org/eng.medianhilal/ontologies/2015/2/untitled-ontology-82#A1
DOMAIN: http://www.w3.org/2000/01/rdf-schema#Resource
DOMAIN: http://www.semanticweb.org/eng.medianhilal/ontologies/2015/2/untitled-ontology-82#A
RANGE: http://www.w3.org/2002/07/owl#Thing
RANGE: http://www.semanticweb.org/eng.medianhilal/ontologies/2015/2/untitled-ontology-82#B1
RANGE: http://www.w3.org/2000/01/rdf-schema#Resource
RANGE: http://www.semanticweb.org/eng.medianhilal/ontologies/2015/2/untitled-ontology-82#B   

问题:

1-为什么会出现这种差异?

2-正确的结论是什么?

3-是否有某种方法可以强制Pellet提供类似于OWL_DL_MEM_RULE_INF的结果?

1 个答案:

答案 0 :(得分:0)

不是枚举域和范围,而是尝试询问属性是否具有给定的域或范围:

import static com.hp.hpl.jena.rdf.model.ResourceFactory.createResource;
...
Collection<Resource> classes = Arrays.asList(
    createResource("http://www.semanticweb.org/eng.medianhilal/ontologies/2015/2/untitled-ontology-82#A"),
    createResource("http://www.semanticweb.org/eng.medianhilal/ontologies/2015/2/untitled-ontology-82#B")
);
...
for (Resource theClass: classes) {
    if (prop.hasRange(theClass) System.out.printf("RANGE: %s\n", theClass);
    if (prop.hasDomain(theClass) System.out.printf("DOMAIN: %s\n", theClass);
}

我认为您会发现,pellet会按预期报告范围和域。这是reasoners工作方式的不同之处:jena的内置推理器是一个在rdf上工作的混合规则引擎,而pellet是一个OWL推理器。在实践中,这意味着推断的三元组显然不存在。

Pellet有an FAQ,它解释了一些差异,以及如何提取所有推论。