KReS functionality demonstrations
This video shows KReS functionalities through two demonstrations.
The first one shows an application for performing integrity check on remote data from the LOD before storing such data on a local knowledge base e.g. to the aim of enriching it. The input is text, FISE enhancers are used for identifying terms in the text that can be associated with DBPedia and Geonamens entities. KReS (ontonet) creates a dedicated scope for this task and retrieves from the LOD the RDF graph describing such entities, such graphs are stored in the scope. The user defines a rule (through KReS rules module) that expresses validity constraints for the integrity check. KReS (reasoners) performs the integrity check and gives as result all valid entities (identified by their URIs): first a DL reasoner is launched in order to perform integrity on the inferred knowledge and improve recall e.g. all sameAs entities are retrieved. Then the rule is executed and the results are returned.
The second demonstration shows how KReS can be used for automatically producing RDF data compliant with the Google Rich Snippet vocabularies. The demo features an interactive interface that has been used as early inspiring prototype for VIE development. The input is text, FISE NLP-based enhancers are used for identifying terms in the text that can be associated with DBPedia and Geonamens entities. A FISE engine called “dulcifier” based-on KReS is in charge of retrieving the RDF graphs describing such entities and transform them (through KReS refactor component) in order to produce a RDF graph compliant with Google Rich Snippet vocabularies. The resulting RDF is then used for enriching the HTML source of the page with RDFa annotations. Finally the resulting enriched HTML is tested against the Google Rich Snippet testing tool.