We should load Earth Rotation Angle into VisitInfo, though it may require some calculations based on TAI/UT1 and/or the LST value read from the headers. Once we have ERA, we should convert our LST values to be computed from ERA.
Russell's comment about it on slack:
this seems to have the info: https://books.google.com/books?id=MM1LDzQIJ9kC&pg=PA317&lpg=PA317&dq=difference+between+earth+rotation+angle+and+greenwich+mean+sidereal+time&source=bl&ots=Hhg2i-DTVZ&sig=FzjubnSJpPYnALZ2aDtGKVwaJS8&hl=en&sa=X&ved=0ahUKEwiatubqperPAhWCgFQKHbbHBUAQ6AEIUTAH#v=onepage&q=difference%20between%20earth%20rotation%20angle%20and%20greenwich%20mean%20sidereal%20time&f=false
And, furthermore, it looks like ERA is equal to GMST within smallish terms.
None of the camera headers I know of for our obs_ packages specify UT1. Most seem to specify LST (I have no idea if that is mean or apparent). So it might make sense to just assume this is LMST, adjust for longitude and write it as an approximate ERA.
A different option, which relies on an internet connection, is to use astropy.Time to compute UT1 from TAI. This will download a file from IERS that has UT1-TAI from 1970 through a few months in the future. That file will quickly go out of date and thus need to be re-downloaded as we process newer data. (Dates outside the range of the file are handled by assuming things are static past the ends of the file, which is reasonable). It makes me uneasy to use such a file, but it would be very nice to have UT1 and ERA is trivial to compute from it.