ScritpInputFormat for wikipedia RDF format

script-input-format-wikipedia-rdf.groovy

/*
Parses the long_abstracts format:
<http://dbpedia.org/resource/Anarchism> <http://dbpedia.org/ontology/abstract> "Anarchism is a collection of movements and ideologies that ..."@en .
 */
import org.openrdf.rio.*
import org.openrdf.rio.helpers.*

def parse(line, factory) {

    def reader = new StringReader(line);
    def model = new org.openrdf.model.impl.LinkedHashModel();

    def rdfParser = Rio.createParser(RDFFormat.TURTLE);
    rdfParser.setRDFHandler(new StatementCollector(model));
    rdfParser.parse(reader, "http://dbpedia.org/");

    def v1;

    model.forEach{
        def name = it.getSubject().stringValue().split("/").last();
        def text = it.getObject().stringValue();

        v1 = factory.vertex(name, "article");
        v1.property("article.name", name);
        v1.property("article.text", text);
    }

    return v1;

}