Fixes for astroNN

[hopkins9942]

download.astroNNDistances and download.astroNNAges previously only downloaded DR14 distance and age files, regardless of dr argument, causing incorrect file to be saved as the combined astroNN file for DR16 and DR17. Now fixed so download.astroNN is called for DR16 and DR17, downloading correct combined astroNN file.

Judging by previous code in read._add_astroNN_orbits, the DR16 astroNN FITS file used to have a field Energy_Err, however in most up-to-date version (v1) this field has been changed to Energy_err to match DR17. Now fixed so fields_to_append is the same for both DR16 and DR17. N.B. this bug only affects users with fitsio installed, as with fitsio FITS fields are case-sensitive whereas with astropy.io.fits they are not.

[jobovy]

Thanks for the fixes! Just so I understand correctly:

• You are saying that the DR16 file was updated to also have Energy_err (the v1 file)? This whole v1 issue passed me by somewhat, so we're still catching up with bugs introduced by it...
• I think your changes to the download functions are correct, but I don't think think we'd run into this issue, because as long as the main astroNN file was downloaded first, all the astroNN paths in DR>16 resolved to that one, so no additional downloads were triggered. And we generally use astroNN through the allStar interface, which first accesses the main astroNN file. But good to fix!

[hopkins9942]

• Yes it was, I've just found the v0 file and confirmed it has Energy_Err where v1 has Energy_err.
• That's all true, unless allStar is first called with either use_astroNN_distances=True or use_astroNN_ages=True. In those cases the DR14 distance or age file would have been downloaded first and saved incorrectly as the main astroNN file for whatever DR is being used, preventing the correct file from being downloaded later on.