Tutorial pyplot_scales.py crashes when used with plt.tight_layout() #6789

afvincent · 2016-07-18T08:46:52Z

Trying to replace plt.subplots_adjust with plt.tight_layout in the pyplot tutorial pyplot_scales.py (see #6779), seems to results in errors.

Matplotlib version, Python version and Platform (Windows, OSX, Linux ...):
Linux (at least Archlinux and CentOs 7)
Tested on mpl 1.5.1 from Anaconda and mpl 2.0.0b3.post1799+ga4fdd60 from github
MWE (kind of…):
Based on pyplot_scales.py:

import numpy as np
import matplotlib.pyplot as plt
plt.style.use('default')

# make up some data in the interval ]0, 1[
y = np.random.normal(loc=0.5, scale=0.4, size=1000)
y = y[(y > 0) & (y < 1)]
y.sort()
x = np.arange(len(y))

plt.figure(1)
plt.subplot(221)
plt.subplot(222)
plt.subplot(223)

# logit
plt.subplot(224)
plt.plot(x, y)
plt.yscale('logit')
#plt.subplots_adjust(top=0.92, bottom=0.08, left=0.10, right=0.95, hspace=0.25,
#                    wspace=0.35)
plt.tight_layout()

plt.show()

When I run this example with plt.tight_layout instead of plt.subplots_adjust (or simply no subplot layout adjusting at all), I get these outputs (mpl 2.0.0b3.post1799+ga4fdd60):

with Qt4Agg backend

In [1]: run pyplot_scales_mwe.py  # that's the MWE posted above
/home/adrien/matplotlib/lib/matplotlib/transforms.py:380: RuntimeWarning: invalid value encountered in double_scalars
  return points[1, 0] - points[0, 0]
/home/adrien/matplotlib/lib/matplotlib/tight_layout.py:199: RuntimeWarning: invalid value encountered in double_scalars
  kwargs["wspace"] = hspace / h_axes
/home/adrien/matplotlib/lib/matplotlib/axis.py:1008: UserWarning: Unable to find pixel distance along axis for interval padding of ticks; assuming no interval padding needed.
  warnings.warn("Unable to find pixel distance along axis "
/home/adrien/matplotlib/lib/matplotlib/axis.py:1018: UserWarning: Unable to find pixel distance along axis for interval padding of ticks; assuming no interval padding needed.
  warnings.warn("Unable to find pixel distance along axis "

Traceback (most recent call last):
  File "/home/adrien/matplotlib/lib/matplotlib/backends/backend_qt5agg.py", line 182, in __draw_idle_agg
    FigureCanvasAgg.draw(self)
  File "/home/adrien/matplotlib/lib/matplotlib/backends/backend_agg.py", line 464, in draw
    self.figure.draw(self.renderer)
  File "/home/adrien/matplotlib/lib/matplotlib/artist.py", line 68, in draw_wrapper
    return draw(artist, renderer, *args, **kwargs)
  File "/home/adrien/matplotlib/lib/matplotlib/figure.py", line 1262, in draw
    renderer, self, dsu, self.suppressComposite)
  File "/home/adrien/matplotlib/lib/matplotlib/image.py", line 139, in _draw_list_compositing_images
    a.draw(renderer)
  File "/home/adrien/matplotlib/lib/matplotlib/artist.py", line 68, in draw_wrapper
    return draw(artist, renderer, *args, **kwargs)
  File "/home/adrien/matplotlib/lib/matplotlib/axes/_base.py", line 2383, in draw
    mimage._draw_list_compositing_images(renderer, self, dsu)
  File "/home/adrien/matplotlib/lib/matplotlib/image.py", line 139, in _draw_list_compositing_images
    a.draw(renderer)
  File "/home/adrien/matplotlib/lib/matplotlib/artist.py", line 68, in draw_wrapper
    return draw(artist, renderer, *args, **kwargs)
  File "/home/adrien/matplotlib/lib/matplotlib/axis.py", line 1109, in draw
    ticks_to_draw = self._update_ticks(renderer)
  File "/home/adrien/matplotlib/lib/matplotlib/axis.py", line 942, in _update_ticks
    tick_tups = [t for t in self.iter_ticks()]
  File "/home/adrien/matplotlib/lib/matplotlib/axis.py", line 885, in iter_ticks
    majorLocs = self.major.locator()
  File "/home/adrien/matplotlib/lib/matplotlib/ticker.py", line 1721, in __call__
    return self.tick_values(vmin, vmax)
  File "/home/adrien/matplotlib/lib/matplotlib/ticker.py", line 1726, in tick_values
    locs = self._raw_ticks(vmin, vmax)
  File "/home/adrien/matplotlib/lib/matplotlib/ticker.py", line 1674, in _raw_ticks
    nbins = max(min(self.axis.get_tick_space(), 9),
  File "/home/adrien/matplotlib/lib/matplotlib/axis.py", line 2008, in get_tick_space
    return int(np.floor(length / size))
ValueError: cannot convert float NaN to integer

which results in

(NB: it's a screenshot because savefig complained about the “ValueError: cannot convert float NaN to integer” stuff)

with TkAgg backend

In [1]: run pyplot_scales_mwe.py  # that's again the MWE posted above
/home/adrien/matplotlib/lib/matplotlib/transforms.py:380: RuntimeWarning: invalid value encountered in double_scalars
  return points[1, 0] - points[0, 0]
/home/adrien/matplotlib/lib/matplotlib/tight_layout.py:199: RuntimeWarning: invalid value encountere
8000
d in double_scalars
  kwargs["wspace"] = hspace / h_axes
/home/adrien/matplotlib/lib/matplotlib/axis.py:1008: UserWarning: Unable to find pixel distance along axis for interval padding of ticks; assuming no interval padding needed.
  warnings.warn("Unable to find pixel distance along axis "
/home/adrien/matplotlib/lib/matplotlib/axis.py:1018: UserWarning: Unable to find pixel distance along axis for interval padding of ticks; assuming no interval padding needed.
  warnings.warn("Unable to find pixel distance along axis "

Exception in Tkinter callback
Traceback (most recent call last):
  File "/home/adrien/anaconda/envs/Python2.7_matplotlib-git/lib/python2.7/lib-tk/Tkinter.py", line 1536, in __call__
    return self.func(*args)
  File "/home/adrien/matplotlib/lib/matplotlib/backends/backend_tkagg.py", line 283, in resize
    self.show()
  File "/home/adrien/matplotlib/lib/matplotlib/backends/backend_tkagg.py", line 354, in draw
    FigureCanvasAgg.draw(self)
  File "/home/adrien/matplotlib/lib/matplotlib/backends/backend_agg.py", line 464, in draw
    self.figure.draw(self.renderer)
  File "/home/adrien/matplotlib/lib/matplotlib/artist.py", line 68, in draw_wrapper
    return draw(artist, renderer, *args, **kwargs)
  File "/home/adrien/matplotlib/lib/matplotlib/figure.py", line 1262, in draw
    renderer, self, dsu, self.suppressComposite)
  File "/home/adrien/matplotlib/lib/matplotlib/image.py", line 139, in _draw_list_compositing_images
    a.draw(renderer)
  File "/home/adrien/matplotlib/lib/matplotlib/artist.py", line 68, in draw_wrapper
    return draw(artist, renderer, *args, **kwargs)
  File "/home/adrien/matplotlib/lib/matplotlib/axes/_base.py", line 2383, in draw
    mimage._draw_list_compositing_images(renderer, self, dsu)
  File "/home/adrien/matplotlib/lib/matplotlib/image.py", line 139, in _draw_list_compositing_images
    a.draw(renderer)
  File "/home/adrien/matplotlib/lib/matplotlib/artist.py", line 68, in draw_wrapper
    return draw(artist, renderer, *args, **kwargs)
  File "/home/adrien/matplotlib/lib/matplotlib/axis.py", line 1109, in draw
    ticks_to_draw = self._update_ticks(renderer)
  File "/home/adrien/matplotlib/lib/matplotlib/axis.py", line 942, in _update_ticks
    tick_tups = [t for t in self.iter_ticks()]
  File "/home/adrien/matplotlib/lib/matplotlib/axis.py", line 885, in iter_ticks
    majorLocs = self.major.locator()
  File "/home/adrien/matplotlib/lib/matplotlib/ticker.py", line 1721, in __call__
    return self.tick_values(vmin, vmax)
  File "/home/adrien/matplotlib/lib/matplotlib/ticker.py", line 1726, in tick_values
    locs = self._raw_ticks(vmin, vmax)
  File "/home/adrien/matplotlib/lib/matplotlib/ticker.py", line 1674, in _raw_ticks
    nbins = max(min(self.axis.get_tick_space(), 9),
  File "/home/adrien/matplotlib/lib/matplotlib/axis.py", line 2008, in get_tick_space
    return int(np.floor(length / size))
ValueError: cannot convert float NaN to integer

which produces an empty figure.

If this is an image generation bug attach a screenshot demonstrating the issue:
See above.
If this is a regression (Used to work in an earlier version of Matplotlib), please
note where it used to work:
The behavior seems to be the same with mpl 1.5.1 and (at least) both TkAgg and Qt4Agg backends.

The text was updated successfully, but these errors were encountered:

dstansby · 2017-01-19T21:53:01Z

Here's a more minimalist example that throws the same error as above:

import matplotlib.pyplot as plt

fig, axs = plt.subplots(2, 2)

axs[1, 1].set_yscale('logit')
plt.tight_layout()
plt.show()

dstansby · 2017-01-28T12:47:15Z

Todo (see #7880):

Make tight_layout ignore artists at non-finite positions

janmedlock · 2017-03-01T23:39:52Z

I've hit this bug, too. A slightly different minimal example, that doesn't depend on tight_layout:

import matplotlib.pyplot as plt

plt.plot([0.1, 0.9], [1, 2])
plt.xscale('logit')
plt.xlabel('x')
plt.show()

Rendering with QT5Agg causes a similar error as in the original report:

Traceback (most recent call last):
  File "/usr/lib/python3/dist-packages/matplotlib/backends/backend_qt5agg.py", g
    FigureCanvasAgg.draw(self)
  File "/usr/lib/python3/dist-packages/matplotlib/backends/backend_agg.py", linw
    self.figure.draw(self.renderer)
  File "/usr/lib/python3/dist-packages/matplotlib/artist.py", line 63, in draw_r
    draw(artist, renderer, *args, **kwargs)
  File "/usr/lib/python3/dist-packages/matplotlib/figure.py", line 1143, in draw
    renderer, self, dsu, self.suppressComposite)
  File "/usr/lib/python3/dist-packages/matplotlib/image.py", line 139, in _draws
    a.draw(renderer)
  File "/usr/lib/python3/dist-packages/matplotlib/artist.py", line 63, in draw_r
    draw(artist, renderer, *args, **kwargs)
  File "/usr/lib/python3/dist-packages/matplotlib/axes/_base.py", line 2409, inw
    mimage._draw_list_compositing_images(renderer, self, dsu)
  File "/usr/lib/python3/dist-packages/matplotlib/image.py", line 139, in _draws
    a.draw(renderer)
  File "/usr/lib/python3/dist-packages/matplotlib/artist.py", line 63, in draw_r
    draw(artist, renderer, *args, **kwargs)
  File "/usr/lib/python3/dist-packages/matplotlib/axis.py", line 1150, in draw
    self.label.draw(renderer)
  File "/usr/lib/python3/dist-packages/matplotlib/artist.py", line 63, in draw_r
    draw(artist, renderer, *args, **kwargs)
  File "/usr/lib/python3/dist-packages/matplotlib/text.py", line 762, in draw
    raise ValueError("posx and posy should be finite values")
ValueError: posx and posy should be finite values

I tried all the backends on my machine, with the same result (but perhaps slightly different error messages).

The script works fine if plt.xlabel() is removed, so I investigated the labeling code. Before rendering the label's position (using plt.gca().xaxis.get_label().get_position()) is (0.5, 0) and after rendering it is (0.5, -inf).

Setting the y position to -inf happens in matplotlib.axis.XAxis._update_label_position(), which gets called during rendering. Both before and after (partial?) rendering, plt.gca().spines['bottom'].get_path() is Path(array([[13., 0.], [13., 0.]]), None). But the transformed path,

spine = plt.gca().spines['bottom']
spinepath = spine.get_transform().transform_path(spine.get_path())

is Path(array([[7373.45454545, 52.8], [7373.45454545, 52.8]]), None) before and Path(array([[nan, nan], [nan, nan]]), None) after (partial?) rendering.

And I'm over my head at this point.

janmedlock · 2017-03-02T00:11:22Z

... and calling gca().spines['bottom']._adjust_location() ('bottom' for logit x scale, 'left' for logit y scale) before rendering prevents the error:

import matplotlib.pyplot as plt

plt.plot([0.1, 0.9], [1, 2])
plt.xscale('logit')
plt.xlabel('x')
gca().spines['bottom']._adjust_location()
plt.show()

It also works with dstansby's tight_layout example:

import matplotlib.pyplot as plt

fig, axs = plt.subplots(2, 2)

axs[1, 1].set_yscale('logit')
axs[1, 1].spines['left']._adjust_location()
plt.tight_layout()
plt.show()

dstansby · 2018-09-24T20:33:43Z

Seems to be fixed on master, feel free to re-open if it turns out to still be a problem.

afvincent mentioned this issue Jul 18, 2016

Fix tutorial pyplot scales (issue #6775) #6779

Merged

tacaswell added this to the 2.0.1 (next bug fix release) milestone Jul 22, 2016

dstansby mentioned this issue Jan 19, 2017

If text location isn't finite, set it to not visible #7880

Closed

dstansby self-assigned this Feb 11, 2017

dstansby removed their assignment Apr 9, 2017

QuLogic modified the milestones: 2.0.1 (next bug fix release), 2.0.2 (next bug fix release) May 3, 2017

tacaswell modified the milestones: 2.1.1 (next bug fix release), 2.2 (next feature release) Oct 9, 2017

colasri mentioned this issue Dec 11, 2017

error in plotter.plot Samreay/ChainConsumer#42

Closed

willvousden mentioned this issue Jan 12, 2018

Logit scale is broken mwaskom/seaborn#1357

Closed

Samreay mentioned this issue Apr 11, 2018

Errors with matplotlib 2.2.2 Samreay/ChainConsumer#58

Closed

dstansby closed this as completed Sep 24, 2018

Zaharid mentioned this issue Oct 11, 2018

Using nested reports in validphys appears to shift critical error from compile time to run time NNPDF/nnpdf#302

Closed

Zaharid mentioned this issue Oct 27, 2018

Regression when calling annotate with nan values for the position #12648

Closed

story645 removed this from the future releases milestone Oct 6, 2022

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Uh oh!

Tutorial pyplot_scales.py crashes when used with plt.tight_layout() #6789

Tutorial pyplot_scales.py crashes when used with plt.tight_layout() #6789

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Tutorial pyplot_scales.py crashes when used with plt.tight_layout() #6789

Tutorial pyplot_scales.py crashes when used with plt.tight_layout() #6789

Comments

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!