@BeforeClass
public static void setUpServices() throws IOException {
openNLP = new OpenNLP(new ClasspathDataFileProvider("DUMMY_SYMBOLIC_NAME"));
searcher = new TestSearcherImpl(NAME, SimpleTokenizer.INSTANCE);
// add some terms to the searcher
Representation rep = factory.createRepresentation("urn:test:PatrickMarshall");
rep.addNaturalText(NAME, "Patrick Marshall");
rep.addReference(TYPE, OntologicalClasses.DBPEDIA_PERSON.getUnicodeString());
searcher.addEntity(rep);
rep = factory.createRepresentation("urn:test:Geologist");
rep.addNaturalText(NAME, "Geologist");
rep.addReference(TYPE, NamespaceEnum.skos + "Concept");
rep.addReference(REDIRECT, "urn:test:redirect:Geologist");
searcher.addEntity(rep);
// a redirect
rep = factory.createRepresentation("urn:test:redirect:Geologist");
rep.addNaturalText(NAME, "Geologe (redirect)");
rep.addReference(TYPE, NamespaceEnum.skos + "Concept");
searcher.addEntity(rep);
rep = factory.createRepresentation("urn:test:NewZealand");
rep.addNaturalText(NAME, "New Zealand");
rep.addReference(TYPE, OntologicalClasses.DBPEDIA_PLACE.getUnicodeString());
searcher.addEntity(rep);
rep = factory.createRepresentation("urn:test:UniversityOfOtago");
rep.addNaturalText(NAME, "University of Otago");
rep.addReference(TYPE, OntologicalClasses.DBPEDIA_ORGANISATION.getUnicodeString());
searcher.addEntity(rep);
rep = factory.createRepresentation("urn:test:University");
rep.addNaturalText(NAME, "University");
rep.addReference(TYPE, NamespaceEnum.skos + "Concept");
searcher.addEntity(rep);
rep = factory.createRepresentation("urn:test:Otago");
rep.addNaturalText(NAME, "Otago");
rep.addReference(TYPE, OntologicalClasses.DBPEDIA_PLACE.getUnicodeString());
searcher.addEntity(rep);
// add a 2nd Otago (Place and University
rep = factory.createRepresentation("urn:test:Otago_Texas");
rep.addNaturalText(NAME, "Otago (Texas)");
rep.addNaturalText(NAME, "Otago");
rep.addReference(TYPE, OntologicalClasses.DBPEDIA_PLACE.getUnicodeString());
searcher.addEntity(rep);
rep = factory.createRepresentation("urn:test:UniversityOfOtago_Texas");
rep.addNaturalText(NAME, "University of Otago (Texas)");
rep.addReference(TYPE, OntologicalClasses.DBPEDIA_ORGANISATION.getUnicodeString());
searcher.addEntity(rep);
}
/**
* @param literalFactory the LiteralFactory to use
* @param graph the MGraph to use
* @param contentItemId the contentItemId the enhancement is extracted from
* @param relatedEnhancements enhancements this textAnnotation is related to
* @param suggestion the entity suggestion
* @param nameField the field used to extract the name
* @param lang the preferred language to include or <code>null</code> if none
*/
public static UriRef writeEntityAnnotation(
EnhancementEngine engine,
LiteralFactory literalFactory,
MGraph graph,
UriRef contentItemId,
Collection<NonLiteral> relatedEnhancements,
Suggestion suggestion,
String nameField,
String lang) {
Representation rep = suggestion.getEntity().getRepresentation();
// 1. extract the "best label"
// Start with the matched one
Text label = suggestion.getMatchedLabel();
// if the matched label is not in the requested language
boolean langMatch =
(lang == null && label.getLanguage() == null)
|| (label.getLanguage() != null && label.getLanguage().startsWith(lang));
// search if a better label is available for this Entity
if (!langMatch) {
Iterator<Text> labels = rep.getText(nameField);
while (labels.hasNext() && !langMatch) {
Text actLabel = labels.next();
langMatch =
(lang == null && actLabel.getLanguage() == null)
|| (actLabel.getLanguage() != null && actLabel.getLanguage().startsWith(lang));
if (langMatch) { // if the language matches ->
// override the matched label
label = actLabel;
}
}
} // else the matched label will be the best to use
Literal literal;
if (label.getLanguage() == null) {
literal = new PlainLiteralImpl(label.getText());
} else {
literal = new PlainLiteralImpl(label.getText(), new Language(label.getLanguage()));
}
// Now create the entityAnnotation
UriRef entityAnnotation =
EnhancementEngineHelper.createEntityEnhancement(graph, engine, contentItemId);
// first relate this entity annotation to the text annotation(s)
for (NonLiteral enhancement : relatedEnhancements) {
graph.add(new TripleImpl(entityAnnotation, DC_RELATION, enhancement));
}
UriRef entityUri = new UriRef(rep.getId());
// add the link to the referred entity
graph.add(new TripleImpl(entityAnnotation, ENHANCER_ENTITY_REFERENCE, entityUri));
// add the label parsed above
graph.add(new TripleImpl(entityAnnotation, ENHANCER_ENTITY_LABEL, literal));
if (suggestion.getScore() != null) {
graph.add(
new TripleImpl(
entityAnnotation,
ENHANCER_CONFIDENCE,
literalFactory.createTypedLiteral(suggestion.getScore())));
}
Iterator<Reference> types = rep.getReferences(RDF_TYPE.getUnicodeString());
while (types.hasNext()) {
graph.add(
new TripleImpl(
entityAnnotation, ENHANCER_ENTITY_TYPE, new UriRef(types.next().getReference())));
}
// add the name of the ReferencedSite that manages the Entity
if (suggestion.getEntity().getSite() != null) {
graph.add(
new TripleImpl(
entityAnnotation,
new UriRef(RdfResourceEnum.site.getUri()),
new PlainLiteralImpl(suggestion.getEntity().getSite())));
}
return entityAnnotation;
}