Fid selection should consider guide star selection

[jeanconn]

In obsid 22030 in JUN0319, which was guide star constrained, it looks to me like we lost a bright guide star due to the fid trap effect. I think this is happening more often than we realize (separate issue #299). I also think it would not have hurt this observation to use a different set of fid lights, there is just the issue that fid light selection is not informed by guide selection (and it seems like we'd need another loop to do this).

{'fid_dist_to_trap': 303.231991941503,
'fid_id': 6,
'id': 548932512,
'stage': 0,
'star_dist_to_register': 312.05753388327059,
'text': 'Cand 548932512 in trap zone for fid 6',
'type': 'fid trap effect'},

This related to #208 in that at least if I had been able to manually specify fid lights going into proseco we could have avoided this by hand without too much fuss.