Open rknop opened 3 months ago
Right now, it looks like associate_object will only associate bad objects with bad measurements, and good objects with good measurements, so the failure case I described above won't happen.
Still, when pulling out all measurements of an object, you often want to pull the good and the bad, so it still makes sense to get rid of the is_bad field of Object.
Right now, in Measurements.associate_object, if a new object is created from a Measurements that has is_bad True, that object's is_bad is set True.
We don't want this. Something that could well happen: a real SN is detected, but is low enough S/N that for whatever reason (shape, bad subtraction of nearby host, etc.) that the measurement is bad. Later measurements might be just fine. But, the whole object will have been marked bad just because the first measurement happened to be bad.
Probably the Object class shouldn't have an is_bad at all.