Fix slicer re #363

[prisae]

This is really an edge case, not high priority at all (more of cosmetic nature than anything else). Closes #363.

[codecov]

Codecov Report

Attention: Patch coverage is 85.96491% with 8 lines in your changes missing coverage. Please review.

Project coverage is 86.25%. Comparing base (ffd4e04) to head (d44f849).
Report is 9 commits behind head on main.

Files with missing lines	Patch %	Lines
examples/plot_slicer_demo.py	0.00%	7 Missing ⚠️
discretize/mixins/mpl_mod.py	90.90%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #364      +/-   ##
==========================================
+ Coverage   85.76%   86.25%   +0.49%     
==========================================
  Files          88       89       +1     
  Lines       18654    18695      +41     
  Branches     2959     2959              
==========================================
+ Hits        15998    16125     +127     
+ Misses       1996     1884     -112     
- Partials      660      686      +26     

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

[jcapriot]

Hey @prisae, the logic for this choosing of vmin and vmax was getting a bit convoluted to me. So, I just changed it myself here :). This mimics what matplotlib does internally (Creating a Normalize() object if there is no norm, and then calling autoscale_None on the norm. This should make it so all three plots have a consistent norm and catch all these odd edge cases.

It's worked in my quick tests, but feel free to check for yourself first.

[prisae]

That looks much better @jcapriot, great! I tested it and it works on my side. Not sure why the documentation test is failing... 🤔

[jcapriot]

Ok, I’ll get the doc test fixed and then merge it.

[jcapriot]

Okay, a few things.

• I moved the plt.show out of the plot_3d_slicer call (Made you unable to interact with the figure before showing it when running from a terminal, or running it as a script).
• Did need to filter out the nans when using autoscale.
• Needed to make sure it was a 1D array going in to it (for some norms... odd that this behavior is not consistent on matplotlib's end).

[jcapriot]

Ok!

[prisae]

Thanks Joe! I will check tomorrow at work with all my examples, but it looks logic, might just need some adjustments in the examples regarding the plt.show(). Might be worth to make a note in that regard in the release notes.