The AIMS community is pleased to announce that the Ontology Plug-in has been integrated into the Eprints Software. The plug-in has been adapted for maximum interoperability, and also ease for new thesauri. The fully functioning version is constructed by installing an Eprints instance adding the newly created package (Eprints Bazaar)- which adapts Eprints with the ontology plug-in. The Ontology Plug-in can be installed separately on the same or different server allowing higher scalability and interoperability. Additionallty, the Ontology Plug-in can work as a broker for different repositories at the same time.

The Ontology Plug-in tool integrated in Eprints Software

Once the Ontology Package is installed, the Eprints database scheme is changed, so every item like an article, a book or a paper can be stored with agricultural terms (label + URI). Each chosen term is displayed in the language specified by the user when viewing the item, and not the language that was used to query the term.

The package also provides an interface that allows Eprints to communicate with any instance of the Ontology Plug-in given the URI - where the plug-in is hosted, not necessarily the Eprints server.

The Ontology Plug-in works as a broker and harvests different thesauri using adapters created for each thesaurus. The functionality in the Eprints Ontology Package to communicate with the thesauri was not added, so all communication with the thesauri goes through the Ontology Plug-in whether it is a query or a resolve request. This makes it easier to add new thesauri by only adding a new “Thesaurus Adapter” to the Ontology Plug-in.

The Ontology Plug-in does not have to be hosted on the same server as the Eprints repository and after installing the Eprints Repository and then the package, the Ontology Plug-in can be linked-to by specifying a URI which points to the server where the Plug-in is hosted.

The RDF/XML export feature has also been adapted to generate the metadata associated with the documents in different formats, such as RDF/XML and RDF/N3. Not it also contains agricultural terms chosen when the item was uploaded. The terms are represented as RDF triples and can be visualized using any linked data visualization tool.  Eprints uses the Dublin Core Ontology to export triples representing the subjects of each item.  The exported linked-data representation also includes properties taken from other ontologies like FOAF.

The Agro Ontology Plug-in Project

The Agro Ontology Plug-in was developed by Nawar Halabi at the University of Southampton, with the assistance of Adam Field, Sebastien Francois, Tim Miles-Board, and other members of EPrints Services, in association with Imma Subirats Coll of the Food and Agriculture Organization (FAO) of the United Nations, and Marc Goovaerts and Dirk Leinders from Hasselt University (Belgium). A paper describing this work was presented at Open Repositories 2014.

When installed, the plugin will (1) replace the subject stage with the ontology, (2) record the URIs selected with the tool in the subject tree, (3) resolve the labels for the URIs in the language of the repository for displaying on the abstract page, and (4) expose the URIs via the repository linked data.

By default, this package uses the ontology plugin (broker) hosted by Hasselt University, Belgium. In case you are hosting your own ontology plugin or want to use a different ontology plugin, please edit the zz_agro_ontology_cfg.pl file in the cfg/cfg.d/ sub-directory inside your repository's directory. Both the "agro_ontology_broker_uri" and "broker_interface_uri" configuration variables need to change to point to the broker's search URI (to which search requests are sent) and the broker's interface URI (which provides the web interface for the search) accordingly.

For more information contact us at [email protected].