CI matrix, linting, CLI parsing, and modernize type hints + flowfield/numba fixes

[sergioald]

Motivation

• Modernize the codebase for newer Python versions, improve developer experience with linting/formatting, and make CLI case selection more robust and user-friendly.
• Improve performance-safety and clarity in the flow-field / numba codepaths and ensure deterministic parallel behaviour and clearer warnings.
• Add a console entry point and tighten a few numeric/validation failure modes to avoid surprising runtime errors.

Description

• CI and tooling: expand the GitHub Actions smoke-tests job to a Python-version matrix (3.10, 3.11, 3.12) and parameterize actions/setup-python; add ruff as a pre-commit hook with --fix enabled and register a ldsfl-run script entry in pyproject.toml.
• CLI and runner: replace run_ldsfl.parse_cases with a stricter parser that supports ranges (e.g. 1,3-5), de-duplicates and validates input, expose it as an argparse type, and update run_ldsfl.main to use it; add more robust parsing tests and update smoke/integration tests accordingly.
• Type hints and API modernisation: migrate many typing annotations to PEP 604/explicit tuple[...] and X | None styles, remove unused imports, and modernise return annotations (e.g. -> tuple[...]) across flowfield, periodic flowfield, semiana, geometry, and other modules.
• Flow-field / numba and numerical fixes: rearrange imports, add stricter checks and clearer RuntimeWarning usage (with stacklevel), avoid large temporary allocations, use zip(..., strict=True) where appropriate, and tidy ThreadPoolExecutor worker typing and behaviour; add safer checks for non-finite migration speed in dxdy2.
• Miscellaneous: reorder and tidy imports in GUI modules, fix small bugs/unused variables in resistance.py, move some numpy/pandas imports to top-level modules, and add a more informative final-warning stacklevel when saving snapshots fails.

Testing

• Ran pre-commit run --all-files locally to apply formatting/lint fixes (hook added to repo); this completed successfully.
• Executed the test-suite with pytest -q and coverage (python -m pytest --cov=ldsfl --cov-report=term-missing --cov-report=xml), and the updated unit tests (including tests/test_cli_config.py) passed.
• CI will run the smoke-tests matrix across Python 3.10/3.11/3.12 (configured in .github/workflows/tests.yml) which runs pre-commit and pytest as described above.

[chatgpt-codex-connector]

💡 Codex Review

LDSFL_Meander/.github/workflows/tests.yml

Line 42 in b615dbd

name: coverage-xml

Use a matrix-specific artifact name

In this workflow, all three python-version matrix jobs run the actions/upload-artifact@v4 step and try to create the same coverage-xml artifact. The upload-artifact v4 documentation says artifact names must be unique and matrix jobs with the same name will hit conflict errors (https://github.com/actions/upload-artifact#not-uploading-to-the-same-artifact), so two of the three CI legs will fail after tests pass unless the name includes ${{ matrix.python-version }} or similar.

ℹ️ About Codex in GitHub

Your team has set up Codex to 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 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

Ship the module targeted by the console script

The new ldsfl-run entry point imports run_ldsfl:main, but this project’s setuptools config only discovers packages matching ldsfl* under [tool.setuptools.packages.find], so the top-level run_ldsfl.py module is not included in a normal wheel. Setuptools only performs automatic py_modules discovery when both packages and py_modules are left unspecified (https://setuptools.pypa.io/en/latest/userguide/package_discovery.html#automatic-discovery), so installed users will get ModuleNotFoundError: run_ldsfl when invoking ldsfl-run unless the module is moved into ldsfl or explicitly listed as a py-module.

Useful? React with 👍 / 👎.