Use q_att_raw for ASPSOL quaternions if available

[taldcroft]

Description

Make get_atts and friends work with modern ASPSOL data where the desired column is q_att_raw not q_att.

Testing

• Passes unit tests on MacOS (only ran asp_l1 tests) [TESTS RUNNING at this point]
• Functional testing

Via a temporary print statement, confirmed that q_att_raw is being used for obsid 23315 (recent) and q_att is being used for obsid 8008 (old).

[jeanconn]

Yeah, that looks like the right thing. I can't think of any code that is going to be dependent on the "wrong" behavior.

I suppose there are a couple of other problems:

1. The mica vv code is doing its own version of the CentroidResiduals code (so the mica.vv code was tested with the aimpoint correction pipeline change but the residual code was not)
2. The residuals code doesn't have modern testing against a recent file.

Neither needs to be addressed in this PR.

[taldcroft]

I'm slightly confused about your point 1, just in the sense that the mica VV report for e.g. 23315 looks A-OK. Is that not what you are referring to? When I was searching around in the mica code I ran into that mapping that accounts for the various "raw" versions of columns, so I was under the impression it is all just working.

[jeanconn]

My point was that IIRC the code to do residuals is duplicated. That's not ideal and why I missed this for the centroid dashboard use case.