[MRG]Add UnifOrtho sampling of slicing directions for Sliced Wasserstein

[Samuel-Vangu]

Types of changes

• New feature (non-breaking change which adds functionality)
• Documentation (added or improved documentation for existing code)

Motivation and context / Related issue

The Sliced Wasserstein distance only supports a low-discrepancy sampling
alternative to plain Monte Carlo (QSW/RQSW, PR #838) for dim == 3. This
leaves no principled alternative in higher dimension, where variance
reduction matters most.

Adds UnifOrtho sampling: independent blocks of mutually orthogonal
directions, drawn from the Haar measure on O(dim), defined for any
dim and recommended specifically for large dimension by a recent
numerical and theoretical study. Reuses the existing get_random_rotations
directly, requiring no new randomization primitive.

Closes #850 .

How has this been tested (if it applies)

• New unit tests covering: unit-norm output, exact mutual orthogonality
  within a block, correctness for arbitrary dimensions (including
  non-multiples of dim), seed reproducibility, cross-backend consistency
  (NumPy, PyTorch, JAX, TensorFlow), and reduced approximation error
  compared to uniform sampling in high dimension against a closed-form
  reference distance.
• Full existing test suite re-run to confirm no regression.
• New gallery example (plot_uniortho.py) demonstrating both the
  orthogonal block structure and the convergence gain in high dimension.

PR checklist

• I have read the CONTRIBUTING document.
• The documentation is up-to-date with the changes I made (check build artifacts).
• All tests passed, and additional code has been covered with new tests.
• I have added the PR and Issue fix to the RELEASES.md file.

[Samuel-Vangu]

Hi @rflamary @clbonet,

this one's ready for review whenever you have time.
Not a big PR -- I've essentially reused existing code
(get_random_rotations) to add UnifOrtho sampling.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 96.85%. Comparing base (6372a66) to head (20a01a6).

Additional details and impacted files

@@           Coverage Diff            @@
##           master     #853    +/-   ##
========================================
  Coverage   96.85%   96.85%            
========================================
  Files         128      128            
  Lines       26160    26268   +108     
========================================
+ Hits        25337    25443   +106     
- Misses        823      825     +2     

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.