Rethink sky-based calibration meta-data in UVCal

[jpober]

Nsources for sky-based calibration is a useful concept but doesn’t really capture everything we want. We should make sure that this and several other sky-optional parameters really capture the full-provenance of a calibration.

[bhazelton]

Also baseline_range. Should this have units? Or should it be a string describing the baseline selection criteria (to accommodate weighting, etc)?