Create new config files and keywords for relative alignment

[stscijgbot-hstdp]

Issue HLA-1267 was created on JIRA by Jennifer Mack:

Proposed update to the alignment software logic for both HST ipppssoot*.fits and HAP hst_ipppss*.fits data products:

--> Use {}separate config file settings for relative versus absolute (catalog) alignment{}.  This is important in order to preserve any small, sub-pixel dithers which optimally sample the PSF when there are few point sources in the image.

Summary:  FIT-REL-catalog solutions are currently derived from relative alignment between exposures in an ASN, followed by a catalog fit. In most cases, the relative pointing of HST is already very accurate and the relative alignment is not necessary. In cases where there are few matched sources, this step can introduce errors and degrade the quality of the combined drizzled products significantly, as shown in the attached PNG file which is based on the notebook tutorial by Roberto Avila for WFC3/IR dataset {}ib1f19010{}.

https://spacetelescope.github.io/hst_notebooks/notebooks/DrizzlePac/optimize_image_sampling/optimize_image_sampling.html

 

PROPOSAL:

 1.)  Add new header keywords nmatches_rel and fit_rms_rel  to capture the accuracy of the relative alignment for FIT-REL-* and FIT-SVM- solutions{*}.  The current keyword values for nmatches{} and fit_rms currently are overwritten during the second alignment step, e.g. to the reference catalog.

2.) For the relative alignment (image-to-image), we would like to use different criteria, e.g. Nmatches>=20 and fitgeometry='rscale', recommended by astrometry expert Varun Bajaj.   When there are not enough matches, the relative alignment step can be skipped, the new keywords nmatches_rel and fit_rms_rel* are set to N/A and the catalog alignment is performed for the drizzled combined product instead (see step 3).   If successful, the WCSNAME can still be FIT-REL-catalog, and the additional keywords will let us know what was done by the software.

3.) Keep the current config file settings for {}catalog alignment, e.g.  Nmatches>=10 and fitgeometry='rscale{}' (see HLA-1233) and confirm whether any logic on the maximum fit RMS is folded in somewhere as an additional criteria.

___________

The relevant code appears to be at line 822 in https://github.com/spacetelescope/drizzlepac/blob/main/drizzlepac/haputils/align_utils.py