Support coercing host arrays in mlfunc/ReflectedAttr

[jcrist]

In #8339 we added mlfunc and ReflectedAttr as new output coercion mechanisms. As part of that, we only implemented coercion paths from device matrices (cupy.ndarray/cupyx.scipy.sparse.spmatrix) since cuml generally does all computations on device and any conversion to host will happen later.

However, while porting cuml.fil I realized there are a few cases (cuml.fil included) where we natively output host memory but may still want to coerce it to device later as part of reflection.

This PR:

• Adds some code to handle numpy.ndarray/scipy.sparse.spmatrix as source arrays for coercion in mlfunc/ReflectedAttr
• Adds tests for the same
• Ports cuml.fil to use mlfunc/ReflectedAttr instead of the legacy equivalents
• Also addresses a few suggestions for added comments leftover from Unify reflection system & decorators #8339.

Followup to #8339.
Part of #8177.

[csadorf]

Am I correct to interpret this as a must-fix follow-up?

[jcrist]

No, this is required to continue porting to the new system, but is not a bug in anything that's already been ported.

[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: 23006302-09cc-4a2b-84e8-b987e46d4dd7

📥 Commits

Reviewing files that changed from the base of the PR and between e751c15 and d24ab14.

📒 Files selected for processing (4)

• python/cuml/cuml/__init__.py
• python/cuml/cuml/fil/compat.py
• python/cuml/cuml/internals/outputs.py
• python/cuml/tests/test_reflection.py

🚧 Files skipped from review as they are similar to previous changes (2)

• python/cuml/cuml/init.py
• python/cuml/cuml/fil/compat.py

---

📝 Walkthrough

Summary by CodeRabbit

• Bug Fixes
  • Improved FIL inference compatibility across CPU/GPU by adjusting how outputs are produced for different memory types.
  • Enhanced convert_arrays handling for dense and sparse data, including NumPy, pandas, SciPy, and CuPy sparse—preserving requested sparse formats.
  • Improved validation and error messaging for unsupported return types.
• Tests
  • Expanded conversion tests across backends (NumPy/CuPy/SciPy/cupyx), memory orders, sparse formats, and indexed pandas conversions.
• Documentation
  • Clarified array conversion behavior and updated an internal workaround note.
• Chores
  • Refreshed copyright header text.

Walkthrough

Changes

FIL output compatibility

Layer / File(s)	Summary
Array conversion and reflection handling python/cuml/cuml/internals/outputs.py, python/cuml/tests/test_reflection.py	convert_arrays now handles NumPy and sparse inputs explicitly, reflection checks cover additional array types, and tests cover input backends, sparse formats, memory orders, and indexes.
nvforest inference output routing python/cuml/cuml/fil/compat.py	FIL inference uses mlfunc, detects CPU or GPU-backed models, selects host or device preprocessing, and returns NumPy or CuPy arrays based on FIL memory settings.
CuPy setup maintenance python/cuml/cuml/__init__.py	The copyright declaration and CuPy workaround comment were updated.

Estimated code review effort: 3 (Moderate) | ~20 minutes

Suggested reviewers: divyegala

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title accurately summarizes the main change: adding host-array coercion support for mlfunc and ReflectedAttr.
Description check	✅ Passed	The description matches the change set by describing host-array coercion, tests, FIL porting, and follow-up comments.
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.

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

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

[coderabbitai]

Actionable comments posted: 1

🤖 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.

Inline comments:
In `@python/cuml/tests/test_reflection.py`:
- Around line 349-375: Rename the test_convert_arrays_sparse_array parameter
format to sparse_format, and update all references in the parametrization,
sparse.random calls, and output format assertion while preserving the existing
test 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: b42916a0-f6d9-4113-8a50-9e539a2e3da1

📥 Commits

Reviewing files that changed from the base of the PR and between 89b5804 and e751c15.

📒 Files selected for processing (4)

• python/cuml/cuml/__init__.py
• python/cuml/cuml/fil/compat.py
• python/cuml/cuml/internals/outputs.py
• python/cuml/tests/test_reflection.py

[csadorf]

LGTM, just one non-blocking question.

[jcrist]

/merge