add TimesSeries.from_group_dataframe parallel mode

[BohdanBilonoh]

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.

Summary

TimeSeries.from_group_dataframe creates time series in for loop. It is suboptimal if the number of groups is large or very large.

[dennisbader]

Thanks @BohdanBilonoh for this neat PR 🚀
This will be a great improvement when dealing with a large number of time series groups :)

I had some minor suggestions on how we could potentially simplify some things here and there.

Also, we're soon to merge PR #2284 which separates the TimeSeries specific utils (will be moved into a new ts_utils.py) from the non-specific utils.

After that, you can import _parallel_apply and _build_tqdm_iterator in timeseries.py, and also import them in other places as before. We can then remove these changes from this PR.

So I'd say let's ignore my comment about this for now and let's address the other comments first. After we've merged the other PR, we can apply the final changes here.

[dennisbader]

Thanks for already applying the changes @BohdanBilonoh, it looks great now! 🚀
I'll reach out to you once we've merge the other PR. I can then also apply the changes myself.

[BohdanBilonoh]

Thanks for already applying the changes @BohdanBilonoh, it looks great now! 🚀 I'll reach out to you once we've merge the other PR. I can then also apply the changes myself.

Thanks for review

[dennisbader]

#2284 was merged, so we can apply the changes here now 🚀

[dennisbader]

Thanks @BohdanBilonoh, for applying the last changes. Great job, this is exactly how I imagined it 🚀 🚀 🚀