Use astropy pixel_to_pixel in calc_pixmap

[mcara]

Closes #51

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 97.28%. Comparing base (0104065) to head (8affa48).
Report is 19 commits behind head on main.

Additional details and impacted files

@@             Coverage Diff             @@
##             main     #161       +/-   ##
===========================================
+ Coverage   75.29%   97.28%   +21.99%     
===========================================
  Files           7        3        -4     
  Lines         344      221      -123     
===========================================
- Hits          259      215       -44     
+ Misses         85        6       -79     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[jdavies-st]

Looks good. I guess I had the thought when I opened the issue that it would completely replace calc_pixmap, but I see that one still needs to bundle it into a pixmap that the drizzle C code can use. Bummer.

Anyway, looks good. Hopefully jwst.resample can now use the calc_pixmap function here. I recall it had been copied over to jwst.resample because for along time the one here did not support gwcs and APE 14 did not exist.

👍

[nden]

This will need careful consideration, and as a minimum a regression test run. Looking at the astropy pixel_to_pixel function, it uses the high level API pixel_to_world which generates SkyCoord objects. It's supposed to be efficient with views of arrays. All I'm saying is it needs some more testing specifically for the use in drizzle.

[mcara]

There are unit tests testing calc_pixmap in drizzle. calc_pixmap is not used in the pipelines and so running regression tests would not test this helper function.