make_fitswcs_transform to handle header with no PC about gwcs

[jdavies-st]

Yes, i.e. handle the case where there's a CD matrix instead of a PC matrix.

[pllim]

For comparison, that same header that crashed GWCS has the following response from astropy.wcs:

>>> from astropy.wcs import WCS
>>> w = WCS(header)
WARNING: FITSFixedWarning: RADECSYS= 'FK5 ' 
the RADECSYS keyword is deprecated, use RADESYSa. [astropy.wcs.wcs]
WARNING: FITSFixedWarning: 'unitfix' made the change 'Changed units: 'degree' -> 'deg', 'degree' -> 'deg''. [astropy.wcs.wcs]
>>> w
WCS Keywords

Number of WCS axes: 2
CTYPE : 'RA---TAN'  'DEC--TAN'  
CRVAL : 299.91736667  22.68769444  
CRPIX : 5276.0  25.0  
CD1_1 CD1_2  : -5.611e-05  0.0  
CD2_1 CD2_2  : 0.0  5.611e-05  
NAXIS : 2272  4273

[sosey]

something like datamodel.get_fits_wcs() ?

[pllim]

Why do I need datamodel here? This is not JWST data.

[sosey]

get_fits_wcs() pulls the fits wcs info from datamodels, I've used it with jwst data, but a datamodel doesn't have to be jwst science or reference files. Was just posting as an example of what you might want. I used datamodel as a placeholder for any data model with header info and data info, since people are usu pulling the header that's associated with some data.

[pllim]

Re: get_fits_wcs() -- jwst dependency is a deal-breaker for this use case. FYI.

[pllim]

Update: I am investigating a fix. PR forthcoming...

[stscijgbot]

This ticket is now being tracked at AL-42