wiki:RdfsImporter

RDFS Importer

User Documentation

  • go to wiki page RdfsImporter
  • attach the ontology RDFS file to that page
  • enter a name for the ontology and select the appropriate file from the dropdown box
  • upload it
  • a new wikipage will be generated with name RDFS_ontologyname
  • if you want, you can edit the ontology afterwards.
  • now, you can use ontology-based autocompletion as described on Autocompletion

You can include the upload form for ontologies into every page you want, you just have to insert

[{UploadOntology}]

Developer Documentation

The RDFS importer is realized as a plugin: de.opendfki.kaukoluwiki.plugin.UploadOntology. The plugin also relies on functionality provided by de.opendfki.kaukoluwiki.importer.RDFSImporter.


Design Thoughts

The wiki should be able to take advantage of ontologies, so we implement the possibility to import ontologies in RDFS format. The working principles are as follows:

  • The user defines a name for the ontology and selects the RDFS file (GUI)
  • The RDFS file is loaded into a temporary MemoryStore and all relevant triples are extracted. These are the triples with one of the following predicates:
    • rdf:type
    • rdfs:label
    • rdfs:comment
    • rdfs:subClassOf
    • rdfs:domain
    • rdfs:range
  • The extracted triples are passed to a text generator that turns them into a text, that will result in the given triples, when parsed by the triple parser.
  • Additionally, a defining triple will be created for every rdfs:Class and rfd:Property (Note: There may be some problems with unclean ontologies that use owl properties that are not recognized by the importer. If something like that happens, please write a detailed report and make a ticket. Someone will fix it.)
  • The returned text is stored on a page called RDFS_Ontologyname through the SesameProvider

Note: No inferencing is done at this time, this is implemented in the TripleStore itself.

Open questions

  • When a page is beeing imported, all the triples are written with fully qualified URIs. That makes the triples quite hard to read, but since the default SemanticKeywords for subjects and predicates may be changed by the user, this is the only way to ensure that the triples will be parsed correctly. A possible solution might be to enhance the mechanism of defining or renaming SemanticKeywords
Last modified 10 years ago Last modified on 08/18/07 11:16:50