solves #726: bug in default inverse mapping

[0Navin0]

# =============================================
# Following the same example described in the issue #726  
# after bug fix
# =============================================

In [23]: wa
Out[23]:
<SkyCoord (ICRS): (ra, dec) in deg
    (5.46165002, -72.12183946)>


In [29]: wcsobj.backward_transform
---------------------------------------------------------------------------
NotImplementedError                       Traceback (most recent call last)
File ~/Softwares/gwcs/gwcs/wcs/_pipeline.py:639, in Pipeline.backward_transform(self)
    638 try:
--> 639     backward = self.forward_transform.inverse
    640 except NotImplementedError as err:

File ~/anaconda3/envs/rdm_env_latest/lib/python3.14/site-packages/astropy/modeling/core.py:1383, in Model.inverse(self)
   1382 elif self._inverse is not None:
-> 1383     result = self._inverse()
   1384     if result is not NotImplemented:
.
.
.
.
NotImplementedError: No analytical or user-supplied inverse transform has been implemented for this model.

# Even when backward transform is not implemented in the original definition of the `wcsobj`, it should default successfully to `numerical_inverse` as shown below:
In [24]: wcsobj.world_to_pixel(wa)
Out[24]: (np.float64(3.999999999638021), np.float64(3.999999999540023))

In [25]: wcsobj.numerical_inverse(wa.ra.deg, wa.dec.deg)
Out[25]: (3.999999999638021, 3.999999999540023)

This PR addresses a bug that triggered when a backward_transform is not defined and default numerical_inverse needs to be used.

Tasks

• no-changelog-entry-needed

News fragment change types:

• changes/727.bugfix.rst: fixes an issue

[mcara]

I suppose tests added in #600 are missing the important block of code that was added in that PR, as well as romancal regression tests do not test for this usage.

Could you, please, update or add unit tests to cover this code?

[mcara]

Also, you are using main branch.

[0Navin0]

Also, you are using main branch.

Oh sorry. I am still learning my way around Github, so this happened by mistake. Should I create a new feature branch, and then again create a pull request? Any suggestion is appreciated.

[0Navin0]

This PR has been replaced by #729

[codecov]

Codecov Report

❌ Patch coverage is 0% with 1 line in your changes missing coverage. Please review.
✅ Project coverage is 85.78%. Comparing base (51d19be) to head (065b461).
⚠️ Report is 34 commits behind head on main.

Files with missing lines	Patch %	Lines
gwcs/wcs/_wcs.py	0.00%	1 Missing ⚠️

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #727   +/-   ##
=======================================
  Coverage   85.78%   85.78%           
=======================================
  Files          41       41           
  Lines        4340     4340           
=======================================
  Hits         3723     3723           
  Misses        617      617           

☔ 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.