ParasiteAxesAuxTrans pcolor/pcolormesh and contour/contourf broken

[kmuehlbauer]

Bug report

Bug summary

Calling pcolormesh/pcolor or contour/contourf from a parasite axes created with ParasiteAxesAuxTrans breaks. In the call to create the mesh using super_pcolor (or super_contour) an additional self is added to the *args:

matplotlib/lib/mpl_toolkits/axes_grid1/parasite_axes.py

Line 114 in 9a46926

mesh = super_pcolor(self, gx, gy, C, **kwargs)

This breaks subsequent pcolor/pcolormesh/pcontour/pcontourf processing.

Code for reproduction

import numpy as np
import matplotlib.pyplot as plt
from mpl_toolkits.axisartist import SubplotHost, ParasiteAxesAuxTrans
from matplotlib.projections import PolarAxes
from matplotlib.transforms import Affine2D

fig = plt.figure(1, figsize=(7, 4))
ax1 = SubplotHost(fig, 1, 1, 1)
fig.add_subplot(ax1)
tr = Affine2D().scale(np.pi/180., 1.) + PolarAxes.PolarTransform()
ax2 = ParasiteAxesAuxTrans(ax1, tr, "equal")
ax1.parasites.append(ax2)
data = np.random.random((360,10))
x = np.arange(data.shape[0])
y = np.arange(data.shape[1])
xx, yy = np.meshgrid(x, y)
ax2.pcolormesh(xx, yy, data.T)
ax1.set_xlim(-10, 10)
ax1.set_ylim(-10, 10)
plt.show()

Actual outcome

---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-1-3448b36e4131> in <module>()
     15 y = np.arange(data.shape[1])
     16 xx, yy = np.meshgrid(x, y)
---> 17 ax2.pcolormesh(xx, yy, data.T)
     18 ax1.set_xlim(-10, 10)
     19 ax1.set_ylim(-10, 10)

/home/kai/miniconda3/envs/wradlib/lib/python3.6/site-packages/mpl_toolkits/axes_grid1/parasite_axes.py in pcolormesh(self, *XYC, **kwargs)
    120 
    121     def pcolormesh(self, *XYC, **kwargs):
--> 122         return self._pcolor(super().pcolormesh, *XYC, **kwargs)
    123 
    124     def pcolor(self, *XYC, **kwargs):

/home/kai/miniconda3/envs/wradlib/lib/python3.6/site-packages/mpl_toolkits/axes_grid1/parasite_axes.py in _pcolor(self, super_pcolor, *XYC, **kwargs)
    114             gx = wxy[:, 0].reshape(orig_shape)
    115             gy = wxy[:, 1].reshape(orig_shape)
--> 116             mesh = super_pcolor(self, gx, gy, C, **kwargs)
    117             mesh.set_transform(self._parent_axes.transData)
    118 

/home/kai/miniconda3/envs/wradlib/lib/python3.6/site-packages/matplotlib/__init__.py in inner(ax, data, *args, **kwargs)
   1783                         "the Matplotlib list!)" % (label_namer, func.__name__),
   1784                         RuntimeWarning, stacklevel=2)
-> 1785             return func(ax, *args, **kwargs)
   1786 
   1787         inner.__doc__ = _add_data_doc(inner.__doc__,

/home/kai/miniconda3/envs/wradlib/lib/python3.6/site-packages/matplotlib/axes/_axes.py in pcolormesh(self, alpha, norm, cmap, vmin, vmax, shading, antialiased, *args, **kwargs)
   5957         allmatch = (shading == 'gouraud')
   5958 
-> 5959         X, Y, C = self._pcolorargs('pcolormesh', *args, allmatch=allmatch)
   5960         Ny, Nx = X.shape
   5961         X = X.ravel()

/home/kai/miniconda3/envs/wradlib/lib/python3.6/site-packages/matplotlib/axes/_axes.py in _pcolorargs(funcname, allmatch, *args)
   5527         else:
   5528             raise TypeError(
-> 5529                 'Illegal arguments to %s; see help(%s)' % (funcname, funcname))
   5530 
   5531         Nx = X.shape[-1]

TypeError: Illegal arguments to pcolormesh; see help(pcolormesh)

Expected outcome

Works before PR #11678.

Matplotlib version

• Operating system: Linux
• Matplotlib version: 3.0.0
• Matplotlib backend (print(matplotlib.get_backend())): Qt5Agg
• Python version: 3.6.6
• Jupyter version (if applicable):
• Other libraries:

Running conda environment using conda-forge channel.

[kmuehlbauer]

Since I do not have that much insight into the matplotlib machinery, I can't tell, if this commit is sufficient to fix this. At least it works as expected on my machine. Happy to create a PR.

[ImportanceOfBeingErnest]

PRs are welcome. Especially if it could also add a test for this.

Maybe @anntzer can comment on why he put self in there in the first place in #11678

[anntzer]

Because I was hoping that errors like this would be caught by CI? :p
Anyways it's clearly a mistake on my part, the fix looks good, thanks.