Make fitting_parallel.py private to make sure function parallel_fit_dask is imported from astropy.modeling.fitting - #16974

Thank you for your contribution to Astropy! 🌌 This checklist is meant to remind the package maintainers who will review this pull request of some common things to look for.

• Do the proposed changes actually accomplish desired goals?
• Do the proposed changes follow the Astropy coding guidelines?
• Are tests added/updated as required? If so, do they follow the Astropy testing guidelines?
• Are docs added/updated as required? If so, do they follow the Astropy documentation guidelines?
• Is rebase and/or squash necessary? If so, please provide the author with appropriate instructions. Also see instructions for rebase and squash.
• Did the CI pass? If no, are the failures related? If you need to run daily and weekly cron jobs as part of the PR, please apply the "Extra CI" label. Codestyle issues can be fixed by the bot.
• Is a change log needed? If yes, did the change log check pass? If no, add the "no-changelog-entry-needed" label. If this is a manual backport, use the "skip-changelog-checks" label unless special changelog handling is necessary.
• Is this a big PR that makes a "What's new?" entry worthwhile and if so, is (1) a "what's new" entry included in this PR and (2) the "whatsnew-needed" label applied?
• At the time of adding the milestone, if the milestone set requires a backport to release branch(es), apply the appropriate "backport-X.Y.x" label(s) before merge.

Broke RTD

WARNING: py:func reference target not found: astropy.modeling.fitting.parallel_fit_dask

pre-commit.ci autofix

Huh, any idea why this would have broken RTD?

The ref does not exist anymore because module is private?

But the ref is pointing to the fitting.py file not parallel_fitting.py, so it should be fine?

I am not the one you have to convince...

Could you try explicitly re-exporting it ?

diff --git a/astropy/modeling/fitting.py b/astropy/modeling/fitting.py
index 40f0bac01d..dee76ed6ce 100644
--- a/astropy/modeling/fitting.py
+++ b/astropy/modeling/fitting.py
@@ -33,7 +33,7 @@ import numpy as np
 from astropy.units import Quantity
 from astropy.utils.exceptions import AstropyUserWarning
 
-from ._fitting_parallel import parallel_fit_dask
+from ._fitting_parallel import parallel_fit_dask as parallel_fit_dask  # noqa: PLC0414
 from .optimizers import DEFAULT_ACC, DEFAULT_EPS, DEFAULT_MAXITER, SLSQP, Simplex
 from .spline import (
     SplineExactKnotsFitter,

@neutrinoceros - no that didn't seem to work either.

I've pushed a hack that does fix it. The issue is that sphinx-automodapi uses function.module to determine the fully qualified name of a function and excludes it by default if it is not local to the submodule being documented (which is sensible).

@nstarman - this is going to be a major issue if we do use this pattern to make submodules private. Do you know of a way around it?

This may be a configuration issue? I think Sphinx needs something like:

.. automodule:: fitting
   :members:
   :undoc-members:
   :show-inheritance:

# Automatically document members imported in modules
autodoc_default_options = {
    'members': True,
    'imported-members': True,  # Include imported members
    'undoc-members': True,  # Include members w/out docstrings
    'show-inheritance': True,
}

I think the solution has to be different here because we are using sphinx-automodapi which lists the individual classes and functions. We usually don't have this issue when we import up one level, I think this is because it is a 'sideways' import. I wonder if it is a bug in sphinx-automodapi's smart_resolver, will investigate.

Working on it in astropy/sphinx-automodapi#190

Ok so astropy/sphinx-automodapi#190 should fix this, would appreciate some reviews of that PR, then I can release a new version of sphinx-automodapi and try it here.

Triggering a new build to pick up sphinx-automodapi v0.18.0

Seems to work! 🎉