New QL calibration QAs

[rstaten]

This PR includes two new QuickLook QA algorithms developed by @eguzman that examine the final outputs from processing calibration exposures. Specifically, they:

• Add a QA for arc processing that checks 'w' coefficients after resolution fit. These metrics are meant as a 'quick' assessment of whether calibration exposures provide final calibration results - in this case the wavelength resolutions.
• Add a dummy QA for flat processing to validate continuum lamp exposures (to be updated with metrics in a future branch).
• Take monthly psf files as input for resolution fit instead of deprecated bootcalib files.
• Note: Further optimization of the included metrics will benefit from simulations with real world effects and commissioning data.

@Srheft, @rkehoe Please review these edits.

[rkehoe]

This is valuable to complete the last QAs for calibration processing. The update looks good, but I don't favor the use of an extra 'Data' dictionary out of the run method into the plotting. The coefficients should just be taken by the plotting from the psf object input.

However, that is under-the-hood and not important at the moment. I'd rather merge this now and then can clean that up later, maybe when fleshing out the fiber flat metric.

@Srheft Thanks ahead for following up with your test to a merge.

[Srheft]

@rstaten, @eguzman thanks for your work on this. As discussed internally:

• The currently executed idea for this QA has been suggested in ticket Adding a new QA for check resolution  #657 to which no comments were received thus we moved on with place holders.

• This PR adds two QAs called Check_FiberFlat with scalar metric called FFLMEAN and Check_Resolution with three metrics: Medians, RMS, NBADCOEFFS (see qa_quicklook.py )

• The feasibility of the suggested quantities to be reported in these QAs is not obvious to me/some. The metrics in these QAs will be completed and/or altered upon access to faulty exposures where the traces of the bad fibers exhibit measurable characteristic behavior; one of which could be an out-of-norm set of best-fit Legendre polynomial or a flat-out failure to find a solution for the fit, etc. The metric to check the quality of fiber flat also needs a full revision.

• Ticket New PSF format fails QL #681 addresses the recently inflicted failure in arc processing.