Narwhalify_from_group_dataframe

[cnhwl]

Checklist before merging this PR:

• Mentioned all issues that this PR fixes or addresses.
• Summarized the updates of this PR under Summary.
• Added an entry under Unreleased in the Changelog.

Fixes #2730.

Summary

A first draft of refactoring the from_group_dataframe() according to #2661. Since I'm not very familiar with the Narwhals library yet, more help would be appreciated!

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 95.19%. Comparing base (2ab62bd) to head (a8340cf).
⚠️ Report is 1 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #2766      +/-   ##
==========================================
- Coverage   95.23%   95.19%   -0.05%     
==========================================
  Files         146      146              
  Lines       15576    15588      +12     
==========================================
+ Hits        14834    14839       +5     
- Misses        742      749       +7     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

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

[dennisbader]

Hi @cnhwl, and thanks a lot for this PR! We were a bit busy over the last week with the release of Darts 0.35.0.
Next week we should be able to review it :)

[dennisbader]

Hi @cnhwl and thanks for this PR, it's a very good start 🚀

I oushed some updates for this PR:

• adapt from_group_dataframe() logic to work with polars
• add polars to from_group_dataframe() tests

The Narwhals team is current fixing an issue related to parallelization using polars DFs wrapped by narwhals (see here). Our tests are expected to fail until then.

I will also have to test how this PR affects the performance of the current from_group_dataframe(). But I'm not too worried.

[cnhwl]

Hi @cnhwl and thanks for this PR, it's a very good start 🚀

I made some updates locally to finalize the PR:

• adapt from_group_dataframe() logic to work with polars
• add polars to from_group_dataframe() tests

I would like to push these changes into your branch. It seems like I don't have permission to do so. Could you grant me access to your forked Darts repository?

Of course! @dennisbader I have been added you as a collaborator on my fork repository 🤝.

[dennisbader]

Hi @cnhwl and sorry for the long wait. We haven't forgotten about this PR :) We're currently working on a big refactor of our TimeSeries in #2807. After that is merged, we're coming back to this one!

[cnhwl]

@dennisbader Got it, thanks for the heads-up. No problem at all, happy to wait for the refactor to be completed 🚀

[dennisbader]

Hi @cnhwl and thanks for this great PR 🚀 It looks really good!

I finally found the time now to review and make a couple of adjustment for performance reasons :)
I also compared the performance against the old version, and now we're even a bit more efficient after some updates 💯

Thanks again and sorry for the long wait!

[dennisbader]

I applied a last change to get the same efficiency as we had it before for pandas DataFrames. Will merge after the tests have passed :)