Automatic directional offsets for NIRISS SOSS

[jbrande]

Sometimes, given challenging position angle requirements for targets in multiple systems, observers might apply extra offsets to their NIRISS pointings. This PR includes code to grab the relevant offset information from the FITS headers and applies those to the PASTASOSS modeled traces to ensure correct straightening of the observed spectral trace and correct spectral extraction. This PR also includes code to extrapolate the trace position and wavelength solution out to longer (or shorter, I guess) wavelengths depending on the direction of the offset, so that observers don't lose any flux.

Currently, this works for X-offsets, but Y-offsets are also potentially allowed and we may want to add code to handle these as well. As these are observatory-level operations, I've currently chosen to make this an automatic, header-driven step in S3, but we might want to also allow for passing in custom offset values in the S3 ECFs.

[jbrande]

While messing around with this, it also occurred to me that it might be useful to plot the nominal spectral trace over the median flux in Fig. 3308. Colors and styles naturally preliminary.

Thoughts?

[codecov]

Codecov Report

❌ Patch coverage is 42.10526% with 22 lines in your changes missing coverage. Please review.
✅ Project coverage is 59.82%. Comparing base (fd9d4ff) to head (6b50a72).

Files with missing lines	Patch %	Lines
src/eureka/S3_data_reduction/niriss.py	33.33%	18 Missing ⚠️
src/eureka/S3_data_reduction/plots_s3.py	63.63%	4 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #826      +/-   ##
==========================================
- Coverage   59.88%   59.82%   -0.06%     
==========================================
  Files         107      107              
  Lines       14029    14064      +35     
==========================================
+ Hits         8401     8414      +13     
- Misses       5628     5650      +22     

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

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[kevin218]

Yes, I support adding a trace to the plot!

[jbrande]

That last commit fixes a minor oversight, where the default pastasoss trace retrieval assumes the trace is at the nominal position first, and then rotates it based on the pupil wheel position. Since rotations and translations don't commute, it's not proper to then shift that pre-rotated trace by the planned X-offset. This caused a difference of a few pixels out at the long wavelength end compared to the observed spectral trace, so I've edited my initial code to translate first, then rotate. I'm not convinced this makes a major difference in the final lightcurve stability, but we'll see.

Otherwise, unless anyone has any other suggestions for trace positioning or plotting edits (or whether we should support manual X-direction offsets in the ECFs), I don't think I've got anything else to change here.