JP-3659: Fix for output WCS when first data set is empty

[melanieclarke]

Resolves JP-3659

Closes #8561

Fix a small bug resulting in incorrect output spectral WCS when the first input data set contains only NaN.

Checklist for PR authors (skip items if you don't have permissions or they are not applicable)

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

[melanieclarke]

Spot-checking regression tests locally, I'm expecting these failures:

FAILED jwst/regtest/test_nirspec_mos_spec3.py::test_nirspec_mos_spec3[b000000031-crf] FAILED jwst/regtest/test_nirspec_mos_spec3.py::test_nirspec_mos_spec3[b000000031-s2d] FAILED jwst/regtest/test_nirspec_mos_spec3.py::test_nirspec_mos_spec3[b000000031-x1d] FAILED jwst/regtest/test_nirspec_mos_spec3.py::test_nirspec_mos_spec3[v000000049-s2d] FAILED jwst/regtest/test_nirspec_mos_spec3.py::test_nirspec_mos_spec3[v000000049-x1d]

These MOS sources all have data that trigger the error condition: there is a small spectral scrap containing only NaN values used as the reference model for the output WCS. The b31 s2d is particularly egregious: because of the error, the good spectra do not align when resampled together, so the slit looks doubled in the output. After this change, the resampled data looks more sensible.

[codecov[bot]]

Codecov Report

Attention: Patch coverage is 92.30769% with 1 line in your changes missing coverage. Please review.

Project coverage is 59.51%. Comparing base (dba65ae) to head (28d8796). Report is 380 commits behind head on master.

Files with missing lines	Patch %	Lines
jwst/resample/resample_spec.py	92.30%	1 Missing :warning:

Additional details and impacted files

```diff @@ Coverage Diff @@ ## master #8562 +/- ## ========================================== + Coverage 58.63% 59.51% +0.88% ========================================== Files 391 391 Lines 39081 39254 +173 ========================================== + Hits 22914 23362 +448 + Misses 16167 15892 -275 ```

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.

[melanieclarke]

Regression tests started here but later aborted: https://plwishmaster.stsci.edu:8081/job/RT/job/JWST-Developers-Pull-Requests/1532/

My first fix didn't work for the reason I thought it did; trying something else.

[melanieclarke]

Okay, I think I have my fix the right way around now. These are the regtest failures I'm now expecting:

FAILED jwst/regtest/test_nirspec_mos_spec3.py::test_nirspec_mos_spec3[b000000030-crf] FAILED jwst/regtest/test_nirspec_mos_spec3.py::test_nirspec_mos_spec3[b000000030-s2d] FAILED jwst/regtest/test_nirspec_mos_spec3.py::test_nirspec_mos_spec3[b000000030-x1d] FAILED jwst/regtest/test_nirspec_mos_spec3.py::test_nirspec_mos_spec3[b000000031-s2d] FAILED jwst/regtest/test_nirspec_mos_spec3.py::test_nirspec_mos_spec3[b000000031-x1d] FAILED jwst/regtest/test_nirspec_mos_spec3.py::test_nirspec_mos_spec3[v000000048-crf] FAILED jwst/regtest/test_nirspec_mos_spec3.py::test_nirspec_mos_spec3[v000000048-s2d] FAILED jwst/regtest/test_nirspec_mos_spec3.py::test_nirspec_mos_spec3[v000000048-x1d]

The b31 s2d that looked weirdly doubled to me was actually closer to correct than the stacked background in b30: it is doubled because it is a background for 2 dithers that don't overlap. The third dither for the same slit is marked as a virtual source instead (v49). The b30 s2d is now also appropriately doubled for the same reason; it is associated with virtual source v48.

Regression tests restarted here: https://plwishmaster.stsci.edu:8081/job/RT/job/JWST-Developers-Pull-Requests/1533/

[melanieclarke]

Regression test results look as expected, so I will take this out of draft.