fix: pandas DataFrame with non-default Index was resulting in scrambled data in scatter_matrix

[MarcoGorelli]

These issue get closed "for free" with the latest Narwhals release, as a result of narwhals-dev/narwhals#1474 having been addressed:

• closes plotly.express.scattermatrix displays no color if supplied a pandas.DataFrame with a DataTime index #4917
• closes Using a DataFrame with a scrambled index and color in px.scatter_matrix() yields mismatched/null colors #4788
• closes scatter_matrix is incorrect when using color and non-sorted index #3577
• closes parallel_coordinates is incorrect when using color and non-sorted index #3576

Furthermore, it allows for the Plotly codebase to be simplified a bit, as maybe_reset_index isn't necessary any more. So, I've:

• made that simplification
• bumped the minimum Narwhals
• added a test to check this stays fixed

Demo:

Latest stable plotly release:

This branch:

[emilykl]

Thanks for the fix, @MarcoGorelli. Sorry for the delayed review, I wanted to take the time to make sure I understood the differences between the Pandas and Narwhals index behavior. Looks good! 🚀