Jena - 从任何语言的本体类中获取标签

Question

For user interface purposes I would like to find a label for a given ontology class within Jena. The label should be in the language preferred by the user, if possible, but if no label is available in the user's language, I would like to show just any label, and hope the user manages to understand. To do that I would like to loop over all available labels, and then pick one (maybe just the first one for the first attempt).

So far the code looks like:

import com.hp.hpl.jena.ontology.OntClass;
public class LabelUtils {
    public static String getLabel(OntClass ontClass, String preferredLanguage) {
        String label = ontClass.getLabel(preferredLanguage);
        if (label == null) {
            // here find another label, in any language, whatever
        }
        return label;
    }
}

How do I list all labels, no matter what the language? I only found a method ontClass.listLabels(String language) which - of course - only returns labels for that language.

I can just assume that an English label always exist, or loop over a list of hard wired locales that are checked for labels, but that seems really lame.

Answer 1

Rob Hall's answer works, but if you want to use the OntResource interface methods, you can use something closer to your original approach. Note what the Javadoc for OntResource#listLabels says (emphasis added) the follwing.

listLabels

com.hp.hpl.jena.util.iterator.ExtendedIterator listLabels(String lang)

Answer an iterator over all of the label literals for this resource.

Parameters:

• lang - The language tag to restrict the listed comments to, or null to select all [labels]

Answer 2

Labels are related to elements in OWL through the rdfs:label property.

To verify this, we can start with a minimalistic model created through the following means:

final OntModel model = ModelFactory.createOntologyModel(OntModelSpec.OWL_DL_MEM);
final OntClass clazz = model.createClass("urn:ex:class");
clazz.addLabel("className1", "en");
clazz.addLabel("className2", "fr");

Writing out the model as N3, we get the following content:

<urn:ex:class>
      a       owl:Class ;
      rdfs:label "className1"@en , "className2"@fr .

Now, because OntClass inherits listPropertyValues from OntResource , we can work with each declared label programmatically like this:

final NodeIterator labels = clazz.listPropertyValues(RDFS.label);
while( labels.hasNext() ) {
    final RDFNode labelNode = labels.next();
    final Literal label = labelNode.asLiteral();
    System.out.println( label.getLanguage() );
}

For out example application, the following output should hopefully demonstrate that you can reflect on the language tag explicitly.

fr
en

Answer 3

I found this question today, and I thought the Rob Hall`s solution interesting because, using the method listLabels, you can list only the rdfs: labels. Using that approach, you can list, for example, the labels defined with the vocabulary SKOS (skos: altLabel and skos: prefLabel).

Posting this as an answer because I dont have reputation to comment...