WhyIs: a nanoscale knowledge graph framework

This framework is interesting: used for a dozen projects of various sizes in RPI’s Tetherless World consortium.

They’re using Blazegraph, SPARQL for interaction, Celery queues for inference triggers.
Their notions of assertion + [para]graphs, provenance depth and granularity, and federation are familiar. They distinguish source-provenance (before it reaches the graph) from publishing-provenance (the algorithmic context, code versions, servers, &c of the publishing process)