Replace plotnine implementation of barplot with seaborn/matplotlib

[kcArtemis]

Removed the dependency (plotnine)

PR Checklist

• Referenced issue is linked
• If you've fixed a bug or added code that should be tested, add tests!
• Documentation in docs is updated

Description of changes

In this pull request, I addressed the issue #425 by removing the dependency on plotnine. Instead, I have implemented the use of matplotlib and seaborn for plotting. Specifically, I updated the barplot function and the perturbscore function to utilize these libraries, providing a more lightweight solution for the project.

Technical details

While making these changes, I ensured that the functionality remains consistent, and the transition to matplotlib and seaborn is seamless. The choice of these libraries was made to minimize dependencies and provide a more straightforward solution for users. No additional dependencies are introduced, and the overall performance is maintained.

Additional context

[codecov]

Codecov Report

Merging #441 (7a1edfe) into main (e17ade7) will decrease coverage by 0.41%.
The diff coverage is 1.81%.

❗ Current head 7a1edfe differs from pull request most recent head 79f4d20. Consider uploading reports for the commit 79f4d20 to get more accurate results

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #441      +/-   ##
==========================================
- Coverage   58.23%   57.83%   -0.41%     
==========================================
  Files          40       40              
  Lines        4899     4933      +34     
==========================================
  Hits         2853     2853              
- Misses       2046     2080      +34     

Files	Coverage Δ
pertpy/plot/_mixscape.py	9.61% <1.81%> (-1.88%)	⬇️

[Zethson]

Great!
Thank you very much!

1. Could you please remove plotnine from the dependencies from pyproject.toml?
2. Could you please paste a before and after plot here? Like one with the old plotnine implementation and one with your new implementation
3. We have a lot of hardcoded plot parameters now (like s, alpha,, and more). Which ones do you think we should expose to the user? Should we populate them at least with kwargs?

Thanks!

[kcArtemis]

Great! Thank you very much!

1. Could you please remove plotnine from the dependencies from pyproject.toml?
2. Could you please paste a before and after plot here? Like one with the old plotnine implementation and one with your new implementation
3. We have a lot of hardcoded plot parameters now (like s, alpha,, and more). Which ones do you think we should expose to the user? Should we populate them at least with kwargs?

Thanks!

we can expose s and alpha to the user since s:marker size and alpha: transparency parameter, for that we can make the use of kwargs.

[Zethson]

images_pertpy_mixscale.pdf

Before and after by @kcArtemis