2.08. Extending Wikibase with terminology capabilities to meet the needs of NFDI4Culture stakeholders

Project proposal context

The Antelope (Annotation, Terminology Lookup and Personalization) service, developed by the NFDI4Culture task areas "Data capture and enrichment of digital cultural assets" (TA1) and "Overarching technical, ethical and legal activities" (TA5) team based at TIB, addresses the specific needs of the culture research communities when it comes to the use of vocabularies and ontologies. Antelope extends the functionalities of existing tools (Falcon 2.0; iArt; TIB Terminology Service) into a common framework with the goal of closing the gaps between:

  1. using semantic concepts in describing digitized cultural objects,
  2. annotating different types of media, and
  3. introducing automation in assisted data curation and annotation workflows.

The service framework is accessible both via a frontend web portal and as a separate data service, with a special focus on integration into third-party research data management (RDM) systems.

This project implements the first such pilot integration, namely via a plug-in for Wikibase. Wikibase is an open source software suite by Wikimedia Germany, allowing the storage and management of Linked Open Data (LOD), besides featuring other common characteristics of wikis, such as collaboration and version control features. The Wikibase4Research and Semantic Kompakkt services already utilize customised Wikibase instances for the NFDI4Culture community. The new pilot integration between Antelope and Wikibase will empower the culture research community to utilize standard terminologies directly in the RDM environment of Wikibase without having to perform manual terms search across separate platforms.

Deliverables

1.) Core development of the Antelope plug-in for Wikibase

The development of the plug-in involved outlining several possible use cases and workflows for data annotation within Wikibase. The main integration workflow focused on enabling Antelope search directly within the content-creation UI (user interface) of Wikibase, so that users can annotate their data with standard terms from ontologies and vocabularies of their choosing.

Leveraging Antelope’s API within the Special Pages feature of the Wikibase UI, the plug-in offers users to

  1. select a Wikibase item they wish to annotate;
  2. run Antelope search based on a keyword associated with this item, or using the item label itself;
  3. display results matching the search input amongst the terminologies available in Antelope (list is configurable and includes Wikidata, GND, Iconclass, Getty AAT, among many others from the TIB Terminology service);
  4. allow users to select a result that will be used to automatically create an annotation statement on the Wikibase item page with appropriate provenance and terminology source. The annotations are saved in the Wikibase triplestore as LOD statements and can be queried via the SPARQL endpoint.

2.) Adaption of the Plug-in approach for Wikibase vs Semantic MediaWiki

The basic plug-in functionalities described above are available for any standard Wikibase instance. For users of the Wikibase4Research service from NFDI4Culture, there is the option to also enable additional features in the plug-in UI which are only available for the Semantic Wikibase extension available with the standard Wikibase4Research Docker Image. These features rely on Semantic MediaWiki UI components not available in plain Wikibase instances, supporting more complex ui components such as form elements with fulltext search and autofill for Wikibase items and properties. Another feature is the possibility to see existing annotations for the selected items before making new annotations.

3.) Testing and release

The plug-in is tested with two projects affiliated with NFDI4Culture:

  1. Herrenhäuser des Ostseeraums from the University of Greifswald; and
  2. GESAH+ (Collection Albrecht Haupt) from the TIB.

The testing is aimed to improve the plug-in functionalities based on real world user requirements. Following successful completion of the testing framework, the plug-in will be released openly to the wider Wikibase community following best practices already established with previous releases of Wikibase4Research and related tooling.

4.) Further integration with TS4NFDI

The development of this plug-in feeds into ongoing contributions of the Antelope tooling to the TS4NFDI Base service Integration Phase. Significant extension of the functionalities and workflow scenarios possible with annotation from the Antelope framework into the Wikibase system, including bulk operations, fine-tuning and customisation of dictionaries used for entity linking, and more, are planned to be completed until the end of 2026.

Lozana Rossenova and Kolja Bailly