feat: make plotly-express dataframe agnostic via narwhals

[FBruzzesi]

Description

This PR migrates plotly-express module logic from pandas to narwhals. In this way, pandas is not a required dependency for plotly-express (or at least for its entirety - e.g. trendlines will still require pandas for now) and users coming with polars,
pyarrow or other eager dataframes supported in narwhals do not need to depend on pandas in the first place.

Closes #4749

Code PR

• I have read through the contributing notes and understand the structure of the package. In particular, if my PR modifies code of plotly.graph_objects, my modifications concern the codegen files and not generated files.
• I have added tests (if submitting a new feature or correcting a bug) or
  modified existing tests.
• For a new feature, I have added documentation examples in an existing or
  new tutorial notebook (please see the doc checklist as well).
• I have added a CHANGELOG entry if fixing/changing/adding anything substantial.
• For a new feature or a change in behaviour, I have updated the relevant docstrings in the code to describe the feature or behaviour (please see the doc checklist as well).

Out of scope for the PR

• No documentation change has been done so far
• Adapt plotly data accordingly included in this PR
• timeseries/trendlines

cc: @MarcoGorelli @LiamConnors

[FBruzzesi]

For context, although trace_data.get_column(attr_value) is a Narwhals Series, which is an iterable, explicitly calling a to_list() makes sure that the elements we loop over are python objects in all cases.
Calling to_py_scalar for each element is significantly more time consuming for the cases in which elements are not python scalars (such as pyarrow).

I will make sure to add an in-line comment

[LiamConnors]

This is an example of a graph that renders differently each time in polars. Looks like the order for the values of x, color and facet_col change.

import plotly.express as px
df = px.data.tips(return_type="polars")
fig = px.bar(df, x="day", y="total_bill", color="smoker", barmode="group", facet_col="sex",
            )
fig.show()

But for pandas the order is consistent and seems to depend on the order of the data. Would it be possible to make it consistent for polars too. cc @emilykl

[MarcoGorelli]

But for pandas the order is consistent and seems to depend on the order of the data. Would it be possible to make it consistent for polars too. cc @emilykl

yup, addressed in the latest commit 👌

latest timings: https://www.kaggle.com/code/marcogorelli/visualise-timings?scriptVersionId=205986199

[emilykl]

@FBruzzesi @MarcoGorelli All looks good from our side. You can go ahead and merge when ready.

This was a massive effort, thank you very much for the contribution! 🚀

[FBruzzesi]

Hey @emilykl thanks a ton! We really appreciated the collaborative effort! As mention on a few meetings, it led us to investigate deeper and improve narwhals codebase significantly as well, it has been such a win-win!

Regarding merging, the branch keeps going out of sync with master, and I will need another approval to merge it myself (and hopefully I need to time it correctly 😇) or you can approve and merge whenever you want. There are no other changes on this PR from our side 🚀

[emilykl]

@FBruzzesi @MarcoGorelli Merged! 🚀 🎉 💥