JP-3588: Implement new reference file datamodel for PASTASOSS transformations

[tapastro]

Partially resolves JP-3588

This PR adds a new reference file datamodel for the PASTASOSS algorithm, replacing the existing ATOCA transformation algorithm. This effectively deprecates the spectrace and wavemap datamodels - TBD on how best to remove them from here and/or CRDS.

Checklist

• added entry in CHANGES.rst (either in Bug Fixes or Changes to API)
• updated relevant tests
• updated relevant documentation
• updated relevant milestone(s)
• added relevant label(s)

[braingram]

🍝

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 66.44%. Comparing base (8a63a5f) to head (f78fb11).
Report is 12 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #320      +/-   ##
==========================================
+ Coverage   66.41%   66.44%   +0.03%     
==========================================
  Files         101      102       +1     
  Lines        5434     5439       +5     
==========================================
+ Hits         3609     3614       +5     
  Misses       1825     1825              

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

[braingram]

The suggestions I added don't yet account for the ndim validators. I assume these properties are arrays? If so, the schema will have to be updated to accommodate that.

Would you add a unit test that makes an instance of the new model with expected junk data? That would help me understand what should go in those spots.

[braingram]

There one item in the schema that causes the following to fail (which I think is worth adding as a generic test-all-model-schemas test):

import asdf
import stdatamodels.jwst.datamodels as dm
s = dm.PastasossModel()
asdf.schema.check_schema(s.schema)

Otherwise looks good to me.

Note that by default all ndim validation is disabled due to validate_arrays being default False. I have a hunch this is off as an attempt to avoid data loading or something but it is suprising to me that this disables ndim validation.

[braingram]

LGTM!