Add MinMaxScaler, MaxAbsScaler, and PolynomialFeatures to cuml.accel#8032
Merged
rapids-bot[bot] merged 6 commits intorapidsai:mainfrom Apr 30, 2026
Merged
Add MinMaxScaler, MaxAbsScaler, and PolynomialFeatures to cuml.accel#8032rapids-bot[bot] merged 6 commits intorapidsai:mainfrom
MinMaxScaler, MaxAbsScaler, and PolynomialFeatures to cuml.accel#8032rapids-bot[bot] merged 6 commits intorapidsai:mainfrom
Conversation
- Convert all non-parameter attributes set on an array-api estimator, not just public ones - Add support for `_params_from_cpu` overrides on `ArrayAPIProxyBase`. - Use a `classproperty` to define `_parameter_constraints`
jcrist
added
improvement
|
No actionable comments were generated in the recent review. 🎉 ℹ️ Recent review info⚙️ Run configurationConfiguration used: Path: .coderabbit.yaml Review profile: CHILL Plan: Enterprise Run ID: 📒 Files selected for processing (1)
🚧 Files skipped from review as they are similar to previous changes (1)
📝 WalkthroughSummary by CodeRabbit
WalkthroughAdds cuml.accel ArrayAPI proxy support for three sklearn preprocessing estimators (MinMaxScaler, MaxAbsScaler, PolynomialFeatures), updates estimator-proxy plumbing for class-level properties and CPU param extraction, extends docs with CPU-fallback details, and expands/adjusts related tests and xfail list. Changes
Estimated code review effort🎯 3 (Moderate) | ⏱️ ~25 minutes Possibly related PRs
Suggested labels
Suggested reviewers
🚥 Pre-merge checks | ✅ 4 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (4 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches🧪 Generate unit tests (beta)
Review rate limit: 9/10 reviews remaining, refill in 6 minutes. Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Caution
Some comments are outside the diff and can’t be posted inline due to platform limitations.
⚠️ Outside diff range comments (1)
python/cuml/cuml_accel_tests/integration/test_preprocessing.py (1)
82-96:⚠️ Potential issue | 🟠 MajorBug: Parameterized test doesn't use the
clsparameter.The test is parameterized over
StandardScaler,MinMaxScaler, andMaxAbsScaler, but the test body hardcodesStandardScaler()on lines 86 and 88. Theclsparameter is never used.🐛 Proposed fix to use the parameterized class
`@pytest.mark.parametrize`("cls", [StandardScaler, MinMaxScaler, MaxAbsScaler]) def test_scaler_partial_fit(cls): X, _ = make_blobs(n_samples=100, centers=3, random_state=42) - model = StandardScaler().fit(X) + model = cls().fit(X) - model2 = StandardScaler() + model2 = cls() model2.partial_fit(X[:25]) assert model2.n_samples_seen_ == 25 model2.partial_fit(X[25:]) assert model2.n_samples_seen_ == X.shape[0] sol = model.transform(X) res = model2.transform(X) np.testing.assert_allclose(sol, res)🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@python/cuml/cuml_accel_tests/integration/test_preprocessing.py` around lines 82 - 96, The test_scaler_partial_fit is parameterized with cls but incorrectly instantiates StandardScaler directly; replace both hardcoded StandardScaler() calls (the one assigned to model and the one assigned to model2) with cls() so the test uses the parameterized class, ensuring model = cls().fit(X) and model2 = cls() before calling model2.partial_fit and model2.transform to validate behavior across StandardScaler, MinMaxScaler, and MaxAbsScaler.
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Outside diff comments:
In `@python/cuml/cuml_accel_tests/integration/test_preprocessing.py`:
- Around line 82-96: The test_scaler_partial_fit is parameterized with cls but
incorrectly instantiates StandardScaler directly; replace both hardcoded
StandardScaler() calls (the one assigned to model and the one assigned to
model2) with cls() so the test uses the parameterized class, ensuring model =
cls().fit(X) and model2 = cls() before calling model2.partial_fit and
model2.transform to validate behavior across StandardScaler, MinMaxScaler, and
MaxAbsScaler.
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Enterprise
Run ID: 09c9503c-10cc-4c29-a2f4-dfb39e93fc54
📒 Files selected for processing (7)
docs/source/cuml-accel/faq.rstdocs/source/cuml-accel/limitations.rstpython/cuml/cuml/accel/_overrides/sklearn/preprocessing.pypython/cuml/cuml/accel/estimator_proxy.pypython/cuml/cuml_accel_tests/integration/test_preprocessing.pypython/cuml/cuml_accel_tests/test_estimator_proxy.pypython/cuml/cuml_accel_tests/upstream/scikit-learn/xfail-list.yaml
💤 Files with no reviewable changes (1)
- python/cuml/cuml_accel_tests/upstream/scikit-learn/xfail-list.yaml
Member
Author
|
/merge |
csadorf
reviewed
Apr 30, 2026
|
|
||
| ``MaxAbsScaler`` will fall back to CPU in the following cases: | ||
|
|
||
| - If ``X`` is sparse. |
Contributor
There was a problem hiding this comment.
(potential follow-up) Delegate to cuML's implementation for sparse inputs.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
This adds support for
MinMaxScaler,MaxAbsScalerandPolynomialFeaturestocuml.accel, through sklearn's array-api (using the recently addedArrayAPIProxyBase). I had to make a few additional modifications to support these without regressions:_parameter_constraintsas a class attribute on the proxy classes. This fixed a few existing xfails._ArrayAPIWrapper._internal_modelinstance, not just the public ones._params_from_cpuonArrayAPIProxyBase, for cases where certain parameter combinations don't support the array-api.Fixes #8014.
Fixes #8030.
Fixes #8031.