JP-3343: Add bounding_box to imaging WCS objects without one

[tapastro]

Resolves JP-3343

Closes #7820

This PR addresses a help desk inquiry regarding missing FITS SIP approximations for a NIRSpec TAcq image. These values were not filled during assign_wcs because the bounding_box was never defined. This PR is an experiment to see if saving a bounding box to imaging WCS objects of size {data.shape}, when one does not already exist, will break things.

Checklist for maintainers

• added entry in CHANGES.rst within the relevant release section
• updated or added relevant tests
• updated relevant documentation
• added relevant milestone
• added relevant label(s)
• ran regression tests, post a link to the Jenkins job below.
  How to run regression tests on a PR
• Make sure the JIRA ticket is resolved properly

[tapastro]

Checking regtests here: https://plwishmaster.stsci.edu:8081/job/RT/job/JWST-Developers-Pull-Requests/855/

[codecov]

Codecov Report

Patch coverage has no change and project coverage change: -0.01% ⚠️

Comparison is base (589917d) 76.52% compared to head (4f9a351) 76.51%.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #7809      +/-   ##
==========================================
- Coverage   76.52%   76.51%   -0.01%     
==========================================
  Files         456      456              
  Lines       36964    36965       +1     
==========================================
  Hits        28285    28285              
- Misses       8679     8680       +1     

Flag	Coverage Δ		*Carryforward flag
nightly	77.39% <ø> (ø)		Carriedforward from 589917d

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

Files Changed	Coverage Δ
jwst/assign_wcs/util.py	83.70% <0.00%> (-0.16%)	⬇️

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

[hbushouse]

Since this is in response to a Help Desk query, it would be good to retroactively create a JP ticket and link it here. Also helps with the bean counting of how many tickets/bugs we fix! ;-)

[hbushouse]

Seems simple and straightforward enough to me. But really need the blessing of the WCS czar (@nden ).

[hbushouse]

Regtest results show mainly the differences due to a recent update to the refpix step for NRS IRS2 data. One test, for an NRS2 image, does show that the FITS SIP keywords are now included in the cal product header and fully populated. So that's a good sign.

[nden]

I suggest we add the code in assign_wcs_step, just before the SIP computation. It doesn't feel right to mix it with the computation of S_REGION.

[tapastro]

New regtest run here: https://plwishmaster.stsci.edu:8081/job/RT/job/JWST-Developers-Pull-Requests/867/

Strange failure has disappeared from regtests.

I moved the bbox generation to right before the S_REGION calculations, because most of them check for bbox presence and mock one if it didn't exist. This isn't quite where @nden wanted it - I can move it to assign_wcs_step if that's still preferred.

[nden]

Initially I was suggesting to assign a bounding_box to all imaging modes in assign_wcs_step. Looking at the current changes I realized this will not cover resampled products. Now I think your first solution is best. Sorry for the flip flop.

[hbushouse]

Is everyone agreed that this is all OK now? @nden