Fix issue in constructing skycell WCS.

[schlafly]

This PR fixes some issues in the of the skycell WCS.

Formerly, the code used the center pixel of the WCS to mark the edge of the bounding box. That meant that the bounding box only covered a quarter of the image, leading to issues.

The "shift" used to move from the projection center was also using the edge of the bounding box. We now use the projection center in x/y coordinates.

This is intended to address an issue that @snbianco raised. I've placed a new example L3 image at
/grp/roman/eschlafly/misc/r0099101001001001001_F158_visit_r274dp63x31y81_i2d.asdf
that should have this issue fixed, but please check!

[codecov]

Codecov Report

Attention: Patch coverage is 20.00000% with 4 lines in your changes missing coverage. Please review.

Project coverage is 79.37%. Comparing base (8a7f497) to head (dedc0cc).
Report is 232 commits behind head on main.

Files with missing lines	Patch %	Lines
romancal/pipeline/mosaic_pipeline.py	20.00%	4 Missing ⚠️

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #1297   +/-   ##
=======================================
  Coverage   79.37%   79.37%           
=======================================
  Files         117      117           
  Lines        8086     8086           
=======================================
  Hits         6418     6418           
  Misses       1668     1668           

Flag	Coverage Δ		*Carryforward flag
nightly	62.84% <ø> (ø)		Carriedforward from 55eed26

*This pull request uses carry forward flags. Click here to find out more.

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

[ddavis-stsci]

LGTM

[snbianco]

This resolves the problem I was running into, thank you!