/** Sets up the default matching configuration for the ontologies */
public void defaultConfig() {
bkSources = new Vector<String>();
File ontRoot = new File(BK_PATH);
if (ontRoot.exists()) {
FileFilter ont =
new ExtensionFilter(
"Ontology Files (*.owl, *.rdf, *.rdfs, *.xml)",
new String[] {".owl", ".rdf", ".rdfs", ".xml"},
false);
File[] ontFiles = ontRoot.listFiles(ont);
for (File f : ontFiles) bkSources.add(f.getName());
FileFilter lex =
new ExtensionFilter("Lexicon Files (*.lexicon)", new String[] {".lexicon"}, false);
File[] lexFiles = ontRoot.listFiles(lex);
for (File f : lexFiles) bkSources.add(f.getName());
bkSources.add("WordNet");
} else {
System.out.println("WARNING: 'store/knowledge' directory not found!");
}
selectedSources = new Vector<String>(bkSources);
size = SizeCategory.getSizeCategory();
lang = LanguageSetting.getLanguageSetting();
languages = new HashSet<String>();
for (String s : source.getLexicon().getLanguages())
if (target.getLexicon().getLanguages().contains(s)) languages.add(s);
if (languages.contains("en") || languages.size() == 0) language = "en";
else language = languages.iterator().next();
matchSteps = new Vector<MatchStep>();
if (lang.equals(LanguageSetting.TRANSLATE)) matchSteps.add(MatchStep.TRANSLATE);
matchSteps.add(MatchStep.LEXICAL);
if (lang.equals(LanguageSetting.SINGLE)) matchSteps.add(MatchStep.BK);
if (!size.equals(SizeCategory.HUGE)) matchSteps.add(MatchStep.WORD);
matchSteps.add(MatchStep.STRING);
if (size.equals(SizeCategory.SMALL) || size.equals(SizeCategory.MEDIUM))
matchSteps.add(MatchStep.STRUCT);
double sourceRatio =
(source.dataPropertyCount() + source.objectPropertyCount()) * 1.0 / source.classCount();
double targetRatio =
(target.dataPropertyCount() + target.objectPropertyCount()) * 1.0 / target.classCount();
if (sourceRatio >= 0.05 && targetRatio >= 0.05) matchSteps.add(MatchStep.PROPERTY);
if (size.equals(SizeCategory.HUGE)) matchSteps.add(MatchStep.OBSOLETE);
matchSteps.add(MatchStep.SELECT);
matchSteps.add(MatchStep.REPAIR);
hierarchic = true;
wms = WordMatchStrategy.AVERAGE;
ssm = StringSimMeasure.ISUB;
primaryStringMatcher = size.equals(SizeCategory.SMALL);
nss = NeighborSimilarityStrategy.DESCENDANTS;
directNeighbors = false;
sType = SelectionType.getSelectionType();
structuralSelection = size.equals(SizeCategory.HUGE);
flagSteps = new Vector<Problem>();
for (Problem f : Problem.values()) flagSteps.add(f);
}
public void translateOntologies() {
Vector<String> sLangs = new Vector<String>(source.getLexicon().getLanguages());
Vector<String> tLangs = new Vector<String>(target.getLexicon().getLanguages());
for (String l1 : sLangs) {
for (String l2 : tLangs) {
if (!l1.equals(l2)) {
Dictionary d = new Dictionary(l1, l2);
d.translateLexicon(source.getLexicon());
d = new Dictionary(l2, l1);
d.translateLexicon(target.getLexicon());
}
}
}
languages = new HashSet<String>();
for (String s : source.getLexicon().getLanguages())
if (target.getLexicon().getLanguages().contains(s)) languages.add(s);
}
public void openOntologies(URI src, URI tgt) throws OWLOntologyCreationException {
closeOntologies();
// Initialize the URIMap and RelationshipMap
uris = new URIMap();
rels = new RelationshipMap();
if (useReasoner) PropertyConfigurator.configure(dir + LOG);
long time = System.currentTimeMillis() / 1000;
System.out.println("Loading source ontology");
source = new Ontology2Match(src);
time = System.currentTimeMillis() / 1000 - time;
System.out.println(source.getURI() + " loaded in " + time + " seconds");
System.out.println("Classes: " + source.classCount());
System.out.println("Names: " + source.getLexicon().size());
System.out.println("Individuals: " + source.individualCount());
System.out.println(
"Properties: " + (source.dataPropertyCount() + source.objectPropertyCount()));
time = System.currentTimeMillis() / 1000;
System.out.println("Loading target ontology");
target = new Ontology2Match(tgt);
time = System.currentTimeMillis() / 1000 - time;
System.out.println(target.getURI() + " loaded in " + time + " seconds");
System.out.println("Classes: " + target.classCount());
System.out.println("Names: " + target.getLexicon().size());
System.out.println("Individuals: " + target.individualCount());
System.out.println(
"Properties: " + (target.dataPropertyCount() + target.objectPropertyCount()));
System.out.println("Direct Relationships: " + rels.relationshipCount());
time = System.currentTimeMillis() / 1000;
System.out.println("Running transitive closure on RelationshipMap");
rels.transitiveClosure();
time = System.currentTimeMillis() / 1000 - time;
System.out.println("Transitive closure finished in " + time + " seconds");
System.out.println("Extended Relationships: " + rels.relationshipCount());
System.out.println("Disjoints: " + rels.disjointCount());
// Reset the alignment, mapping, and evaluation
a = null;
activeMapping = -1;
evaluation = null;
// Refresh the user interface
if (userInterface != null) userInterface.refresh();
defaultConfig();
System.out.println("Finished!");
}
/** @return whether there are properties of corresponding types in both active ontologies */
public boolean hasProperties() {
return hasOntologies()
&& ((source.dataPropertyCount() > 0 && target.dataPropertyCount() > 0)
|| (source.objectPropertyCount() > 0 && target.objectPropertyCount() > 0));
}
/** @return whether there are classes in both active ontologies */
public boolean hasClasses() {
return hasOntologies() && source.classCount() > 0 && target.classCount() > 0;
}
public static Alignment tfIdfMainParse(String sourcePath, String targetPath)
throws OWLOntologyCreationException {
double sim = 0.0d;
csMaps = new Alignment();
source = new Ontology2Match(sourcePath);
target = new Ontology2Match(targetPath);
String sName, tName;
List<double[]> docsrcVector, doctgtVector = new ArrayList<double[]>();
// Get class Names of the source and target
Set<Integer> sourceKeys = source.getClasses();
Set<Integer> targetKeys = target.getClasses();
// Parse Class Names of Source and Target
for (Integer i : sourceKeys) {
for (Integer j : targetKeys) {
sName = source.getName(i);
tName = target.getName(j);
sim = calcCosSim(sName, tName);
csMaps.add(new Mapping(i, j, sim));
}
}
// Get Data Properties of the source and target
Set<Integer> sDProp = source.getDataProperties();
Set<Integer> tDProp = target.getDataProperties();
// Parse Data Properties of Source and Target
for (Integer i : sDProp) {
for (Integer j : tDProp) {
sName = source.getName(i);
tName = target.getName(j);
sim = calcCosSim(sName, tName);
csMaps.add(new Mapping(i, j, sim));
}
}
// Get Object Properties of the source and target
Set<Integer> sOProp = source.getObjectProperties();
Set<Integer> tOProp = target.getObjectProperties();
// Parse Object Properties of Source and Target
for (Integer i : sOProp) {
for (Integer j : tOProp) {
sName = source.getName(i);
tName = target.getName(j);
sim = calcCosSim(sName, tName);
csMaps.add(new Mapping(i, j, sim));
}
}
return csMaps;
}