Complex data formatting

[foranonymoussubmissions2022]

Hi,

I am currently working on a knowledge graph embedding project where each triple in my dataset is associated with a topic, effectively adding a fourth column to the data.

Is there a recommended approach or a feasible workaround within PyKEEN's architecture that would allow me to access the topic indices at the level of the Interaction function in the pipeline?