Plotting functions

[mumichae]

Provide python function for plotting metrics

[codecov]

Codecov Report

Merging #325 (d9e1d70) into main (ed3e284) will increase coverage by 3.84%.
The diff coverage is 98.24%.

❗ Current head d9e1d70 differs from pull request most recent head eeb8693. Consider uploading reports for the commit eeb8693 to get more accurate results

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #325      +/-   ##
==========================================
+ Coverage   53.29%   57.13%   +3.84%     
==========================================
  Files          22       40      +18     
  Lines        1852     2158     +306     
==========================================
+ Hits          987     1233     +246     
- Misses        865      925      +60     

Flag	Coverage Δ
unittest	57.13% <98.24%> (?)
unittests	?

Flags with carried forward coverage won't be shown. Click here to find out more.

Files Changed	Coverage Δ
scib/plotting.py	97.82% <97.82%> (ø)
scib/__init__.py	100.00% <100.00%> (ø)
tests/plots/test_plots.py	100.00% <100.00%> (ø)

... and 36 files with indirect coverage changes

[LuckyMD]

Oooh, you are building a python version of Marta's plotting functionality? Or is this a different visualization? Do you have an example?

[mumichae]

It's a simplified version of Marta's plot. I thought that some basic plotting functions should be part of the package, but was starting out with something simple

Overall score is optional:

[LuckyMD]

Very nice! Would be good to check if the overall score calculation is the same as for marta's code :)

[mumichae]

The overall score is just a simple mean of all metrics and equally weighted. I haven't yet addressed more sophisticated weighting, but I think that would be good to address as a separate problem.

[LuckyMD]

Okay... I tested different weighting approaches and converged on min-max scaling in the paper... I wouldn't test too many for now, but should be consistent with what we have for the R function. Can keep mean as well as that's useful for when you have only few methods tested.