Keep track of source versions

[cthoyt]

This is a bit of a pain for HGNC since the data is updated on a continuous basis, but for ontologies, there are version numbers available directly in the data itself. Same is sort of true for UniProt (I wrote a version getter in bioversions).

This would allow us to benchmark Gilda's performance over time (which I expect for the most part should improve)

Unfortunately, this means that we'd have to make most of the updates in INDRA since most of the resources are imported from there