Repository navigation
Conversation
…xclusion Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Codex Review SummaryThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 78da16ccab
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".
| "_models/timm/*": "timm", | ||
| "_models/ultralytics/*": "ultralytics", | ||
| } | ||
| collect_ignore_glob = [ |
There was a problem hiding this comment.
Preserve skips for explicitly selected optional-package tests
When an optional package is absent, these globs only suppress files found while recursively collecting a directory; pytest still collects a file or node ID passed explicitly. Consequently, commands such as pytest tests/_models/timm/test_timm.py now fail during import instead of reporting skips, while invoking one of these directories alone exits with “no tests collected” (status 5). This is reproducible with pytest 9.0.3 and breaks the documented workflow for running a specific test file, so the optional-package handling needs to cover explicit selections as well.
Useful? React with 👍 / 👎.
…ctories Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
What has changed and why?
Phase 1 of the CI speedup: replace module-level
pytest.skip(..., allow_module_level=True)withpytestmark = pytest.mark.requires(...).Stacked on #995. Merge that first, then change this PR's base to
main.requires("torchmetrics>=1.5")in the 10_metricsmodules andrequires(f"torchmetrics>={TORCHMETRICS_MIN_VERSION}")in the template tests.test_classwise_metric_collection.pyimported torchmetrics classes at module level; those imports moved into the test.test_train_task.py:requires("albumentations>=1.4.0", "torchmetrics>=1.5", "transformers").tests/_models/{rfdetr,super_gradients,timm,ultralytics}: package imports moved from module level into the tests, plus a module-widerequires(...).test_get_modelintest_super_gradients_package.pyparametrizes class names instead of classes.Markers only apply after a module is imported, so optional packages must not be imported at module level.
Only remaining module-level skip: ffmpeg in
test_extract_video_frames.py(system binary, not a Python package).How has it been tested?
uv run --isolated --with torchmetrics==0.8.2):tests/_metricsandtest_train_task.pycollect without errors, 102 skipped withRequires torchmetrics>=1.5.Requires <package>.main, except that the super_gradients tests are now collected and reported as skipped when super-gradients is missing (they were hidden by the module-level skip before).mypy testspasses.Did you update the documentation?
🤖 Generated with Claude Code