Verbalising OWL Ontologies in IsiZulu with Python

Abstract

Ontologies as a component of Semantic Web technologies are used in Sub-Saharan Africa mainly as part of ontology-driven information systems that may include an interface in a local language. IsiZulu is one such local language, which is spoken by about 23 million people in South Africa, and for which verbalisation patterns to verbalise an ontology exist. We have implemented the algorithms corresponding to these patterns in Python so as to link it most easily to the various technologies that use ontologies and for other NLP tasks. This was linked to Owlready, a new Python-based OWL API, so as to verbalise an ontology in isiZulu. The verbaliser can run in ‘ontology inside’ mode, outputting the sentences in the terminal for further processing in an ontology-driven information system, and in GUI mode that displays colour-coded natural language sentences for users such as domain experts and linguists. The demo will showcase its features.