The AGROVOC multilingual thesaurus is currently deployed into two main distributions, both in RDF:
1. AGROVOC Core includes AGROVOC in all languages. It is an RDF dataset, available in RDF/XML and NT formats.
2. AGROVOC LOD includes AGROVOC in all languages. It is currently published in NT and NQ formats. This is the distribution also accessible from the SPARQL endpoint and web services. The materialization of SKOS core labels from the skosxl and trivially inferred triples (due to symmetric and inverseOf properties, no transitive closure computed) are contained as well, along with Agrontology vocabulary.
The AGROVOC VoID provides metadata about the AGROVOC Linked Dataset.
Date neutral links to the latest AGROVOC release are also available on request.
Technical information about the two deployments
1. AGROVOC Core. Content is composed of:
- data dump from VocBench
- timestamp on the AGROVOC SKOS Concept Scheme, indicating the date/time of release
- AGROVOC mapping to other triple stores (now stored in the default graph, previously (until 2014) stored in the named graph;
See AGROVOC VoID descriptor to get the list of all mapped files.
2. AGROVOC LOD is saved both in N-Quads (including provenance information through named graphs) and N-Triples format. Its content includes:
- AGROVOC Core (above)
- pointer to void:inDataset triple, linking to the AGROVOC VoID descriptor, for each skos:Concept, skosxl:Label and skos:ConceptScheme
- materialization of SKOS core labels from the skosxl reified Labels (skosxl labels are maintained too)
- materialization of trivially inferred triples (due to symmetric and inverseOf properties, no transitive closure computed)
- Agrontology vocabulary (stored in the named graph)