Fix setting vmin vmax

[melonora]

closes #303

Fixes the problem in #303. Main problem is that clipping for the Normalize object is set to False by default in matplotlib. I have set this value to True and added tests. I don't think we currently have a usecase for setting it to False, but feel free to comment if there is a usecase for us currently.

[timtreis]

Code lgtm, but can you add the fix to the changelog? Also, what do you think of plotting a 2x1 grid with the non-clipped and clipped image so it's visually easy see the difference?

[melonora]

Code lgtm, but can you add the fix to the changelog? Also, what do you think of plotting a 2x1 grid with the non-clipped and clipped image so it's visually easy see the difference?

added to changelog. Regarding 2 x 1 grid, I would if ncols was working xD. I don't want to create a fig object here now.

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 80.35%. Comparing base (1b9b577) to head (88c018b).
Report is 1 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #307      +/-   ##
==========================================
+ Coverage   80.21%   80.35%   +0.13%     
==========================================
  Files          11       11              
  Lines        1739     1746       +7     
==========================================
+ Hits         1395     1403       +8     
+ Misses        344      343       -1     

Files	Coverage Δ
src/spatialdata_plot/pl/utils.py	70.21% <100.00%> (+0.34%)	⬆️

[timtreis]

Why the aversion against the fig object? Used it f.e. in the 0-transparency cmap PR: https://github.com/scverse/spatialdata-plot/pull/302/files#diff-59872f651a8fef967621d774a8b7441219dc474f1d9604424c54fe19a162bbfd

[melonora]

I would just solve the issue with ncols as well:) like I would not add more code in a test because of ncols, though behavior in this case is a bit tricky

[melonora]

added the baseline to the test for now and will open an issue on ncols as argument

[timtreis]

LGTM, thanks for adjusting the test!