Bright stars may not have large enough mag_err for guide star selection

[jeanconn]

Related to #67

For stars brighter than 6.1 mags, we probably want to set the lower bound on the applied mag_err to 0.1 mags for this check

proseco/proseco/guide.py

Line 176 in 52145e0

bad_mag = (((cand_guides['mag'] - n_sigma * cand_guides['mag_err']) < bright_lim) |

Reviewing the candidate stars, it looks like there are 647 possible candidate stars with MAG_ACA_ERR < 0.1 and MAG_ACA > 5.9 and MAG_ACA < 6.1 (where stars is the whole proseco AGASC).

In [63]: worry = stars[(stars['MAG_ACA'] > 5.9) & (stars['MAG_ACA'] < 6.1) & (stars['MAG_ACA_ERR'] < 10) & (stars['CLASS'] == 0) & (stars['ASPQ1'] < 20) 
    ...: & (stars['POS_ERR'] < 3000)]

In [64]: len(worry)
Out[64]: 637

[taldcroft]

In StarsTable we probably need a mag_aca_err value which is explicitly the best estimate of the true systematic catalog error in addition to mag_err which is the per-readout error. This is for post-release, nothing is changing right now.

[jeanconn]

I think #254 basically closed this issue for me, but @taldcroft comments about needing a new mag_aca_err have not been addressed. Not sure if that needs a new issue.