EuroScienceGateway MS6: Integrated EuroScienceGateway knowledge graph

The Workflowhub Knowledge Graph has been improved and its generation made more robust.

When this work was last reported, a complete knowledge graph had been generated but several criticisms were made. The previous graph was:

• Verbose and hard for a human to read or navigate

• Had unresolvable URIs as root data entities

• Contained many duplicate entries

• Contained sparse metadata from only a single source

Work has successfully been undertaken to address all of these points. The graph now uses partially resolvable, more human readable, URIs for root data entities. Steps have been added to the generation software to add metadata from additional sources (enrichment) and to remove duplicate entries (consolidation).

Several areas of the codebase have been refactored and improved, to help ensure repeatability and longevity. 

The new knowledge graph still has areas that could be improved. Partially resolvable URIs should be migrated to fully resolvable alternatives. Further enrichment processes should be added which affords greater de-duplication.