Weird 'image size is too large' issue in matplotlib 3.4+

Description

Some geographic plots cannot be plotted due to the image size being too large. This happens when a colorbar is plotted with pcolormesh. It works when the image is plotted with contourf.

Steps to reproduce

import numpy as np
import xarray as xr
import proplot as plot

rng = np.random.default_rng()
x = rng.standard_normal((100, 100))

da = xr.DataArray(x,
     dims=['lat', 'lon'],
     coords={'lat': np.linspace(-90, 90, 100),
             'lon': np.linspace(0, 360, 100)})

fig, ax = plot.subplots(proj='eqearth')
ax.format(coast=True)
m = ax.pcolormesh(da)
ax.colorbar(m)

Actual behavior: [What actually happened]

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
~/miniconda3/envs/basicf4/lib/python3.8/site-packages/IPython/core/formatters.py in __call__(self, obj)
    339                 pass
    340             else:
--> 341                 return printer(obj)
    342             # Finally look for special method names
    343             method = get_real_method(obj, self.print_method)

~/miniconda3/envs/basicf4/lib/python3.8/site-packages/IPython/core/pylabtools.py in <lambda>(fig)
    250         png_formatter.for_type(Figure, lambda fig: print_figure(fig, 'png', **kwargs))
    251     if 'retina' in formats or 'png2x' in formats:
--> 252         png_formatter.for_type(Figure, lambda fig: retina_figure(fig, **kwargs))
    253     if 'jpg' in formats or 'jpeg' in formats:
    254         jpg_formatter.for_type(Figure, lambda fig: print_figure(fig, 'jpg', **kwargs))

~/miniconda3/envs/basicf4/lib/python3.8/site-packages/IPython/core/pylabtools.py in retina_figure(fig, **kwargs)
    140 def retina_figure(fig, **kwargs):
    141     """format a figure as a pixel-doubled (retina) PNG"""
--> 142     pngdata = print_figure(fig, fmt='retina', **kwargs)
    143     # Make sure that retina_figure acts just like print_figure and returns
    144     # None when the figure is empty.

~/miniconda3/envs/basicf4/lib/python3.8/site-packages/IPython/core/pylabtools.py in print_figure(fig, fmt, bbox_inches, **kwargs)
    132         FigureCanvasBase(fig)
    133 
--> 134     fig.canvas.print_figure(bytes_io, **kw)
    135     data = bytes_io.getvalue()
    136     if fmt == 'svg':

~/miniconda3/envs/basicf4/lib/python3.8/site-packages/proplot/figure.py in _preprocess(self, *args, **kwargs)
    108         with rc_context, fig_context:
    109             fig.auto_layout()
--> 110             result = func(self, *args, **kwargs)
    111         return result
    112 

~/miniconda3/envs/basicf4/lib/python3.8/site-packages/matplotlib/backend_bases.py in print_figure(self, filename, dpi, facecolor, edgecolor, orientation, format, bbox_inches, pad_inches, bbox_extra_artists, backend, **kwargs)
   2253                 # force the figure dpi to 72), so we need to set it again here.
   2254                 with cbook._setattr_cm(self.figure, dpi=dpi):
-> 2255                     result = print_method(
   2256                         filename,
   2257                         facecolor=facecolor,

~/miniconda3/envs/basicf4/lib/python3.8/site-packages/matplotlib/backend_bases.py in wrapper(*args, **kwargs)
   1667             kwargs.pop(arg)
   1668 
-> 1669         return func(*args, **kwargs)
   1670 
   1671     return wrapper

~/miniconda3/envs/basicf4/lib/python3.8/site-packages/matplotlib/backends/backend_agg.py in print_png(self, filename_or_obj, metadata, pil_kwargs, *args)
    506             *metadata*, including the default 'Software' key.
    507         """
--> 508         FigureCanvasAgg.draw(self)
    509         mpl.image.imsave(
    510             filename_or_obj, self.buffer_rgba(), format="png", origin="upper",

~/miniconda3/envs/basicf4/lib/python3.8/site-packages/matplotlib/backends/backend_agg.py in draw(self)
    399     def draw(self):
    400         # docstring inherited
--> 401         self.renderer = self.get_renderer(cleared=True)
    402         # Acquire a lock on the shared font cache.
    403         with RendererAgg.lock, \

~/miniconda3/envs/basicf4/lib/python3.8/site-packages/matplotlib/backends/backend_agg.py in get_renderer(self, cleared)
    415                           and getattr(self, "_lastKey", None) == key)
    416         if not reuse_renderer:
--> 417             self.renderer = RendererAgg(w, h, self.figure.dpi)
    418             self._lastKey = key
    419         elif cleared:

~/miniconda3/envs/basicf4/lib/python3.8/site-packages/matplotlib/backends/backend_agg.py in __init__(self, width, height, dpi)
     89         self.width = width
     90         self.height = height
---> 91         self._renderer = _RendererAgg(int(width), int(height), dpi)
     92         self._filter_renderers = []
     93 

ValueError: Image size of 68279018x33565395 pixels is too large. It must be less than 2^16 in each direction.

When plot.subplots(proj='cyl') is used, there is no error but there is a large space around the figure.

Proplot version

matplotlib 3.4.2 proplot 0.7.0

[kinyatoride]

Thanks! The error disappeared when I downgraded the matplotlib. Don’t overwork yourself 😊

[lukelbd]

Looks like another weird auto_layout issue. Maybe related to #244?

This appears to be a problem in matplotlib 3.4+ only. For the time being downgradiing to 3.3.4 or earlier should work. Evidently #251 did not fix all the 3.4-specific bugs.