FIX: create_violin() compatible with NumPy 2.4 percentile API

[AJSTYLE-lab]

Summary

This PR fixes the create_violin() function in plotly.figure_factory._violin that was failing on NumPy 2.4+ due to the removal of the interpolation argument in np.percentile.

Previously, the code used:
python np.percentile(x, q, interpolation="linear")

which raises TypeError in NumPy 2.4.

This PR introduces a safe wrapper _percentile() that uses method="linear" for NumPy >= 2.4 and falls back to interpolation="linear" for older versions.

Fix

• Added _percentile() helper function in _violin.py.

• Updated calc_stats() to use _percentile() for all percentile calculations (q1, q2, q3).

Checklist

• Added or updated code to handle NumPy 2.4+ API

• Verified that TestViolin.test_violin_fig now passes

• Ran all unit tests locally (pytest) to confirm no regressions

• Committed changes with clear message

• PR targets main branch of plotly/plotly.py

• Followed contribution guidelines

Tests

• Ran locally:
  python -m pytest tests/test_optional/test_figure_factory/test_figure_factory.py::TestViolin::test_violin_fig
• Test result: ✅ Passed
• Compatible with all supported NumPy versions (>=1.17, including 2.4).

Additional Notes

• This PR only modifies _violin.py; no other files affected.

• Code formatting checked with ruff.

• Closes [BUG]: create_violin() failure with numpy==2.4.0: TypeError: percentile() got an unexpected keyword argument 'interpolation' #5461

Screenshot of Testing:

[camdecoster]

Thanks for the PR. For linting, it can be hard to discern which changes need to be looked at. Could you please rebase this branch to move all linting commits into the first commit, and the bug fix changes into the second? That way we'll be able to review this more quickly.