Chad Dombrova

This is ready for a new round of CI and (hopefully) final review

Good news! The CI properly alerted me to the fact that the stubs needed to be rebuilt due to changes recently pushed to main. Rebased and rebuilt.

This is ready for another run of the CI.

The linux jobs with DOC=ON are failing because the dotted path to the documented objects is different than expected: ```diff [INFO] frozendoc: -.. py:function:: FixedFunctionStyleFromString(str: str) -> PyOpenColorIO.PyOpenColorIO.FixedFunctionStyle [INFO] frozendoc:...

> Hey @chadrik -- so, we discussed this PR at the most recent TSC meeting, and I think we'd like to automate the stub generation entirely in CI, as opposed...

> Is the idea to push the changes as a commit to the branch, so that they can be reviewed as part of the MR? If so, this may offer...

I made some progress on an automated-commit solution: https://github.com/AcademySoftwareFoundation/OpenImageIO/pull/4802#issuecomment-2973480004 That said, it would be great if we could separate these two concerns: is merging this contingent upon solving the auto-commit...

> Maybe we can separate out the CI / GHA stuff, and reframe this PR as a process that should take place before release (like the frozen-docs stuff), as opposed...

> Adding this here for reference. Bernard Kwok's approach to [generating stubs for MaterialX](https://github.com/kwokcb/types-MaterialX). Looking for an approach that avoids putting an ongoing burden on all contributors. @doug-walker if what...

Happy to help. I saw that my new changes to `stubgen` were released in mypy 1.7, and those changes included some improvements to parsing recursive types. @rwgk If possible, it...