Rectify now accepts str and HDUList + stop mangling METISLMSSPectralTrace params dict

[astronomyk]

Human-readbale summary:

Bug 1: rectify_cube() and traces() ignore their own checks for filename or HDUList at the top of the functions.
Bug 2: map_spectra_to_focal_plane continues even when chucking an error on a XiLamImage call. Fails again two lines later (line 235) on xilam_wcs = xilam.wcs
Bug 3: Copy of params dict, to avoid altering the original contents of the dict being passed to the function

For more information, see the toaster's explanation below the line. But I think the code changes are more informative anyway

---

Three independent, few-line fixes, kept as separate commits for review:

1. rectify_traces / rectify_cube crash when given a filename

Both methods document accepting "str or fits.HDUList" and open the file into inhdul — but then passed the original argument on to make_image_interpolations() and SpectralTrace.rectify(). With a filename this raised AttributeError: 'str' object has no attribute 'header'; with an in-memory HDUList it happened to work, masking the bug. Now inhdul is used consistently after opening.

2. map_spectra_to_focal_plane turned a handled error into a NameError

A ValueError raised while building the XiLamImage was caught and logged, but execution then continued with the unbound local variable:

NameError: name 'xilam' is not defined

The method now returns None — the same "no contribution from this trace" contract as the footprint-outside-FoV and empty-footprint paths, which all callers already handle.

3. MetisLMSSpectralTrace wrote into the trace list's meta dict

MetisLMSSpectralTraceList.make_spectral_traces passes params=self.meta by reference to each of the 28 slice traces; each trace then updated that shared dict with its slice-specific values (slice, aperture_id, plus construction kwargs). After building all traces the list's meta carried whichever slice came last. The constructor now copies the dict at the boundary.

Testing

One regression test per fix:

• TestRectifyTracesInput::test_rectify_traces_opens_filename_before_use
• TestMapSpectraErrorPath::test_returns_none_when_xilamimage_fails
• TestMetisLMSSpectralTraceInit::test_does_not_mutate_caller_params

Notebooks demonstrating each fix on the METIS IRDB (including the full LSS observe→readout→rectify-from-file chain for fix 1) are attached below.

[codecov]

Codecov Report

❌ Patch coverage is 80.00000% with 1 line in your changes missing coverage. Please review.
✅ Project coverage is 77.09%. Comparing base (beb191d) to head (ae14246).

Files with missing lines	Patch %	Lines
scopesim/effects/metis_lms_trace_list.py	50.00%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #985      +/-   ##
==========================================
+ Coverage   76.46%   77.09%   +0.62%     
==========================================
  Files          69       69              
  Lines        9025     9027       +2     
==========================================
+ Hits         6901     6959      +58     
+ Misses       2124     2068      -56     

☔ View full report in Codecov by Harness.
📢 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.

[teutoburg]

What practical issues did this cause?

[oczoske]

I think this was a debugging thing. The idea was to see all cases where a trace gave a ValueError. If we want to finish on the first one, we can simply remove the try...except.

[astronomyk]

I'd be happy with simply removing the try/except, as that forces it to fail loudly... right?