fix(models): retry after pydantic schema rebuild error

[MukundaKatta]

Summary

• catch PydanticUserError when reading deferred pydantic core schema metadata
• retry after model_rebuild(force=True, raise_errors=False) before inspecting extras schema
• add a regression test that exercises the rebuild fallback path with a synthetic model class

Testing

• python3 -m py_compile src/openai/_models.py tests/test_models.py
• python3 -m pytest tests/test_models.py -k rebuilds_after_pydantic_user_error (fails locally: ModuleNotFoundError: No module named 'pytest_asyncio')

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: b0cd49aa3f

ℹ️ About Codex in GitHub

Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

[chatgpt-codex-connector]

Guard v2-only PydanticUserError import

Importing PydanticUserError at module load time breaks the declared pydantic>=1.9.0,<3 compatibility path: in Pydantic v1, pydantic.errors does not export PydanticUserError, so importing openai._models raises ImportError before the PYDANTIC_V1 branch can short-circuit. This would fail all v1 environments (including the repo’s test-pydantic-v1 nox session) even though the new retry logic is only intended for v2.

Useful? React with 👍 / 👎.