[enhancement] accelerate array_api inputs for sklearnex's validate_data and _check_sample_weight

[icfaust]

Description

Continues dlpack work from #2275. validate_data and _check_sample_weight do not follow standard sklearnex offloading practice, namely they compute always wherever the data is (as the data movement could ruin any speedup provided by oneDAL, the algorithm is extraordinarily simple), and they do not patch out sklearn functions. Therefore, they must be enabled separately for array_api support. Since they are to be included in every zero-copy array_api supported algorithm, it is a prerequisite for enabling every other estimator.

Previously this aspect was controlled by the looking for the flags attribute, which is not in the array_api standard. The array api standard does not include python-facing attributes or methods which can show if C-contiguous or F-contiguous. However, the array_api standard requires dlpack support. The attributes of from a DLPack tensor can be checked for the memory layout instead. This PR introduces a special onedal backend function which extracts and checks the necessary memory layout (without taking ownership of the tensor). A python function is created which first checks and queries the flags or __dlpack__ attributes. If neither are available, it will return False triggering the sklearn _assert_all_finite. This is done as to_table will attempt to convert to a contiguous memory layout, which again will ruin the performance gain.

---

PR should start as a draft, then move to ready for review state after CI is passed and all applicable checkboxes are closed.
This approach ensures that reviewers don't spend extra time asking for regular requirements.

You can remove a checkbox as not applicable only if it doesn't relate to this PR in any way.
For example, PR with docs update doesn't require checkboxes for performance while PR with any change in actual code should have checkboxes and justify how this code change is expected to affect performance (or justification should be self-evident).

Checklist to comply with before moving PR from draft:

PR completeness and readability

• I have reviewed my changes thoroughly before submitting this pull request.
• I have commented my code, particularly in hard-to-understand areas.
• I have updated the documentation to reflect the changes or created a separate PR with update and provided its number in the description, if necessary.
• Git commit message contains an appropriate signed-off-by string (see CONTRIBUTING.md for details).
• I have added a respective label(s) to PR if I have a permission for that.
• I have resolved any merge conflicts that might occur with the base branch.

Testing

• I have run it locally and tested the changes extensively.
• All CI jobs are green or I have provided justification why they aren't.
• I have extended testing suite if new functionality was introduced in this PR.

Performance

• I have measured performance for affected algorithms using scikit-learn_bench and provided at least summary table with measured data, if performance change is expected.
• I have provided justification why performance has changed or why changes are not expected.
• I have provided justification why quality metrics have changed or why changes are not expected.
• I have extended benchmarking suite and provided corresponding scikit-learn_bench PR if new measurable functionality was introduced in this PR.

[codecov]

Codecov Report

Attention: Patch coverage is 57.14286% with 6 lines in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
onedal/datatypes/table.cpp	40.00%	0 Missing and 3 partials ⚠️
onedal/utils/validation.py	66.66%	1 Missing and 1 partial ⚠️
onedal/datatypes/numpy/data_conversion.cpp	0.00%	0 Missing and 1 partial ⚠️

Flag	Coverage Δ
azure	77.96% <75.00%> (-0.01%)	⬇️
github	71.08% <57.14%> (+0.07%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Files with missing lines	Coverage Δ
sklearnex/utils/validation.py	58.75% <100.00%> (+0.52%)	⬆️
onedal/datatypes/numpy/data_conversion.cpp	50.24% <0.00%> (ø)
onedal/utils/validation.py	62.17% <66.66%> (+0.12%)	⬆️
onedal/datatypes/table.cpp	58.13% <40.00%> (-2.39%)	⬇️

... and 1 file with indirect coverage changes