Fix Slicer matplotlib-warning.

[prisae]

Closes #215

[prisae]

What I am not sure about is... I think I included clim as a parameter because it is also there in plotSlice and plotImage. However, this way it can be provided as clim but as well in pcolor_opts. Not sure if it would be better to deprecate clim, and push the use of pcolor_opts, to make the discretize-code slimmer and leave these things to matplotlib?

What is your @jcapriot, @lheagy ?

[codecov]

Codecov Report

Merging #217 into master will increase coverage by 0.01%.
The diff coverage is 90.90%.

@@            Coverage Diff             @@
##           master     #217      +/-   ##
==========================================
+ Coverage   81.39%   81.41%   +0.01%     
==========================================
  Files          24       24              
  Lines        5197     5202       +5     
==========================================
+ Hits         4230     4235       +5     
  Misses        967      967              

Impacted Files	Coverage Δ
discretize/View.py	70.83% <90.90%> (+0.16%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update a4ead91...a0edf08. Read the comment docs.

[prisae]

That needs more testing, because vmin/vmax are log if the norm is log... so the current version will fail. I think we need to talk about this, and maybe get rid of clim to simplify things.

[prisae]

Good to go and ready for review. Works also with norms.

I made another small change. We should not patch matplotlib, and propagate deprecation warnings. We should simply not cause matplotlib errors.

So now, these two possibilities run without warnings:

mesh.plot_3d_slicer(res, pcolor_opts={'norm': LogNorm(vmin=0.3, vmax=200)})

and

mesh.plot_3d_slicer(np.log10(res), clim=[np.log10(0.3), np.log10(200)])

and these two cases will raise a matplotlib deprecation warning, because they should be adjusted (to the above versions):

mesh.plot_3d_slicer(res, clim=[0.3, 200], pcolor_opts={'norm': LogNorm()})

and

mesh.plot_3d_slicer(res, clim=[0.3, 10], pcolor_opts={'norm': LogNorm(vmin=0.3, vmax=200)})

Here a minimum working example to test:

import warnings
import discretize
import numpy as np
import matplotlib.pyplot as plt
from matplotlib.colors import LogNorm
from matplotlib.cbook import MatplotlibDeprecationWarning

# ALWAYS SHOW THE MATPLOTLIB WARNINGS
warnings.simplefilter('always', MatplotlibDeprecationWarning)

mesh = discretize.TensorMesh([5, 1, 1])
res = 10**np.arange(5)

# 1: Good: Provide vmin/vmax in the norm
mesh.plot_3d_slicer(res, pcolor_opts={'norm': LogNorm(vmin=1, vmax=1000)})

# 2: Good: Provide clim, but NO norm
mesh.plot_3d_slicer(np.log10(res), clim=[0, 3])

# 3: Deprecated: Provide clim with a norm (new deprecated since matplotlib 3.3) 
mesh.plot_3d_slicer(res, clim=[1, 1000], pcolor_opts={'norm': LogNorm()})

# 4: Deprecated: Provide clim with a norm AND norm-v{min;max} (new deprecated since matplotlib 3.3) 
mesh.plot_3d_slicer(res, clim=[1, 1000], pcolor_opts={'norm': LogNorm(vmin=1, vmax=1000)})

[jcapriot]

This looks good for the slicer