Apply new validation to metrics.pairwise_distances

[csadorf]

Convert the dense paths in cuml.metrics.pairwise_distances and nan_euclidean_distances to the new input-validation infrastructure.

Part of #7998.

[copy-pr-bot]

Auto-sync is disabled for draft pull requests in this repository. Workflows must be run manually.

Contributors can view more details about this message here.

[coderabbitai]

📝 Walkthrough

Walkthrough

This PR migrates pairwise_distances.pyx from the legacy input_to_cuml_array preprocessing to the newer check_array validation workflow. It updates X and Y normalization, adjusts pointer extraction, derives memory layout flags, and updates kernel invocation signatures. A test exception type is also corrected.

Changes

Input Validation & Kernel Integration Migration

Layer / File(s)	Summary
Import & Validation Setup python/cuml/cuml/metrics/pairwise_distances.pyx (lines 13–18)	Removed input_to_cuml_array import and added check_array import to establish the new validation pathway.
X and Y Preprocessing python/cuml/cuml/metrics/pairwise_distances.pyx (lines 180–200, 355–365)	X and Y are now normalized via check_array into X_m and Y_m; shapes (n_samples_x, n_features_x) and dtypes are extracted from the checked arrays.
Memory Layout Derivation python/cuml/cuml/metrics/pairwise_distances.pyx (lines 366–370, 421–422)	is_row_major flag is derived from X_m layout; destination array dest_m is created with explicit row-major or Fortran-order specification.
Pointer Extraction & Kernel Calls python/cuml/cuml/metrics/pairwise_distances.pyx (lines 423–424, 433–436, 441–447)	Data pointers d_X_ptr and d_Y_ptr are extracted from X_m.data.ptr and Y_m.data.ptr; pairwise_distance kernel is invoked with explicit scalar dimension parameters (n_samples_x, n_samples_y, n_features_x).
Test Exception Correction python/cuml/tests/test_metrics.py (lines 1319–1320)	Test expectation for dtype mismatch with convert_dtype=False updated from TypeError to ValueError.

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~12 minutes

Possibly related PRs

• rapidsai/cuml#8024: Applies the same input_to_cuml_array to check_array migration pattern across the Lars module with equivalent pointer and layout handling updates.
• rapidsai/cuml#8006: Performs parallel input-validation utility migration from legacy input handling to the new check_array/check_inputs framework.
• rapidsai/cuml#8050: Applies identical input-validation migration pattern in cuml.metrics modules, replacing legacy preprocessing with check_array-based normalization.

Suggested labels

Cython / Python, improvement, non-breaking, sklearn-api-compat

Suggested reviewers

• betatim
• jcrist
• viclafargue

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title 'Apply new validation to metrics.pairwise_distances' directly describes the main change: applying new validation infrastructure to pairwise_distances.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
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.
Description check	✅ Passed	The pull request description clearly describes the changes made: replacing input_to_cuml_array with check_array, modifying layout derivation logic, changing dtype-mismatch error type from TypeError to ValueError, and updating corresponding tests. All changes are directly related to the changeset.

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

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

Warning

Review ran into problems

🔥 Problems

Git: Failed to clone repository. Please run the @coderabbitai full review command to re-trigger a full review. If the issue persists, set path_filters to include or exclude specific files.

---

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

[coderabbitai]

🧹 Nitpick comments (1)

python/cuml/tests/test_metrics.py (1)

1320-1323: ⚡ Quick win

Add a regression for the n_features == 1 layout path.

This PR added explicit degenerate-layout handling, but the current order regression at Lines 1230-1295 still repeats the single-sample (1, 4) case twice. Please add a (n, 1) variant here as well so the new y_order="A" / contiguity propagation branch is actually locked down against scikit-learn.

As per coding guidelines, "Test files must validate numerical correctness by comparing with scikit-learn, include edge case coverage (empty datasets, single sample, high-dimensional data)".

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@python/cuml/tests/test_metrics.py` around lines 1320 - 1323, Add a regression
that covers the degenerate n_features==1 layout by adding a test variant using X
arrays shaped (n, 1) (e.g., 4x1) alongside the existing (1, 4) case so the new
y_order="A" / contiguity propagation branch is exercised; construct X_double
(float64) and X_float (float32) with shape (n,1) and compare
pairwise_distances(...) results against sklearn.metrics.pairwise_distances
(using the same metric and convert_dtype settings) to assert numerical
equality/consistency as the other order tests do.

🤖 Prompt for all review comments with AI agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Nitpick comments:
In `@python/cuml/tests/test_metrics.py`:
- Around line 1320-1323: Add a regression that covers the degenerate
n_features==1 layout by adding a test variant using X arrays shaped (n, 1)
(e.g., 4x1) alongside the existing (1, 4) case so the new y_order="A" /
contiguity propagation branch is exercised; construct X_double (float64) and
X_float (float32) with shape (n,1) and compare pairwise_distances(...) results
against sklearn.metrics.pairwise_distances (using the same metric and
convert_dtype settings) to assert numerical equality/consistency as the other
order tests do.

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Enterprise

Run ID: c1bf9ce4-247f-4f6e-8a54-4c6bcbcbacbc

📥 Commits

Reviewing files that changed from the base of the PR and between 7ba22e6 and 1303f61.

📒 Files selected for processing (2)

• python/cuml/cuml/metrics/pairwise_distances.pyx
• python/cuml/tests/test_metrics.py