Fix sample_weight handling in KernelRidge

[jcrist]

This fixes a bug introduced in the validation port to KernelRidge. This solver needs to special case scalar sample weights; the naive application of check_inputs here resulted in them being treated differently than before.

This PR fixes the validation flow to treat scalar weights as they were before. It also adjusts the hypothesis test to hit this case in an example.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Enterprise

Run ID: b3d99446-99b3-4744-bca0-3931d0e8bcc5

📥 Commits

Reviewing files that changed from the base of the PR and between f1a1092 and d72d569.

📒 Files selected for processing (1)

• python/cuml/cuml/kernel_ridge/kernel_ridge.py

🚧 Files skipped from review as they are similar to previous changes (1)

• python/cuml/cuml/kernel_ridge/kernel_ridge.py

---

📝 Walkthrough

Summary by CodeRabbit

• Bug Fixes

  • Improved sample-weight handling to correctly accept scalar finite weights and normalize/validate non-scalar or non-finite weights.
  • Increased robustness of kernel computations to improve reliability across kernel types.

• Tests

  • Expanded tests for column-vector targets, scalar sample weights, and target reshaping; updated precision coverage for certain cases.

Walkthrough

KernelRidge.fit now conditionally validates sample_weight (allowing finite scalars unchanged, normalizing others via check_sample_weight and check_consistent_length), and _get_kernel is wrapped with run_in_internal_context. Tests update gradient handling for 1D targets and add/adjust parametrizations.

Changes

Cohort / File(s)	Summary
KernelRidge Implementation python/cuml/cuml/kernel_ridge/kernel_ridge.py	fit: sample_weight is no longer passed through check_inputs; finite scalar weights are allowed unchanged, non-scalar/non-finite weights are normalized with check_sample_weight (using estimator dtype and convert_dtype) and validated with check_consistent_length. _get_kernel decorated with run_in_internal_context. Minor comment removed in _solve_cholesky_kernel.
KernelRidge Tests python/cuml/tests/test_kernel_ridge.py	gradient_norm now treats 1D y as 2D by reshaping before CuPy conversion and per-target gradient computation. Test parametrization: one dtype changed from np.float32 to np.float64, and an explicit example added for linear kernel with column-vector y and scalar sample_weight. File header year updated.

Estimated code review effort

🎯 3 (Moderate) | ⏱️ ~20 minutes

Possibly related PRs

• Apply new input validation to cuml.kernel_ridge #8007: Both PRs touch KernelRidge input validation and kernel computation context.

Suggested labels

Cython / Python, sklearn-api-compat

Suggested reviewers

• dantegd
• csadorf

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 0.00% which is insufficient. The required threshold is 80.00%.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title 'Fix sample_weight handling in KernelRidge' directly and specifically describes the main change—fixing a bug in sample_weight validation introduced during a validation port.
Description check	✅ Passed	The description explains the bug context (validation port issue), the specific problem (scalar sample weights handled incorrectly), the fix applied (special-casing scalar weights), and testing improvements, all directly related to the changeset.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

_{Review rate limit: 9/10 reviews remaining, refill in 6 minutes.}

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[coderabbitai]

Actionable comments posted: 1

🤖 Prompt for all review comments with AI agents

Verify each finding against the current code and only fix it if needed.

Inline comments:
In `@python/cuml/cuml/kernel_ridge/kernel_ridge.py`:
- Around line 307-314: The scalar fast-path for sample_weight bypasses
validation and allows invalid scalars (0, NaN, Inf) into _solve_cholesky_kernel;
change the cp.isscalar(sample_weight) branch in kernel_ridge.fit (around the
sample_weight handling) to validate scalars before skipping array checks: either
call check_sample_weight on a 1-element array-wrapped scalar or perform the same
checks manually (finite, non-negative, non-zero if required) and raise the same
errors as check_sample_weight, so that invalid scalar values are rejected
consistently with the non-scalar path and with scikit-learn behavior.

🪄 Autofix (Beta)

Fix all unresolved CodeRabbit comments on this PR:

• Push a commit to this branch (recommended)
• Create a new PR with the fixes

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Enterprise

Run ID: 93b3bf5d-e721-4ba5-aca5-5d0c0195630b

📥 Commits

Reviewing files that changed from the base of the PR and between 016840c and f1a1092.

📒 Files selected for processing (2)

• python/cuml/cuml/kernel_ridge/kernel_ridge.py
• python/cuml/tests/test_kernel_ridge.py