Fixes: #581
Fixes: #580

This PR changes how nodes wrap/unwrap values:

• removing wrapping/unwrapping of tagged scalars (which we no longer use in new files) (remove scalar node wrapping #580)
• wrap/unwrap DNode/LNode values so setattr with a DNode doesn't result in an invalid model (Assigning m.meta to m.meta leads to validation errors #581)

With the PR old models can be made:

>> m = rdm.ImageModel.create_fake_data(tag="asdf://stsci.edu/datamodels/roman/tags/wfi_image-1.0.0")
>> m.validate()  # fails on main
>> m.save("foo.asdf")  # works with this PR

However if a user modifies an old model they will need to wrap values.

>> type(m.meta.filename)
roman_datamodels.stnode.Filename
>> m.meta.filename = "foo"
>> m.validate()
Failed validating 'tag' in schema['properties']['meta']['allOf'][0]['allOf'][0]['properties']['filename']:
>> m.meta.filename = stnode.Filename("foo")
>> m.validate()  # succeeds

Regtests: https://github.com/spacetelescope/RegressionTests/actions/runs/18410870474
show 11 differences all from tag differences for meta.filename where:

• for tvac input pipeline runs meta.filename is now the expected str
• for non-tvac runs meta.filename is the type that was provided as an input. As we have some uncal files with Filename instead of str. I suggest that we regenerate the regtest files. Perhaps we want to wait until we ferret out a few more tagged node uses in romancal. However if that work takes longer than review of this PR I suggest we regenerate the files with this PR.

Tasks

• Update or add relevant roman_datamodels tests.
• Update relevant docstrings and / or docs/ page.
• Does this PR change any API used downstream? (If not, label with no-changelog-entry-needed.)
  • Write news fragment(s) in changes/: echo "changed something" > changes/<PR#>.<changetype>.rst (see below for change types).
  • Start a romancal regression test (https://github.com/spacetelescope/RegressionTests/actions/workflows/romancal.yml) with this branch installed ("git+https://github.com/<fork>/rad@<branch>").