Skip to content

UnicodeDecodeError when making a plot using the 'classic' style and text.usetex=True #8423

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Closed
ngoldbaum opened this issue Apr 3, 2017 · 7 comments · Fixed by #14567
Closed

UnicodeDecodeError when making a plot using the 'classic' style and text.usetex=True #8423

ngoldbaum opened this issue Apr 3, 2017 · 7 comments · Fixed by #14567
Labels
topic: text/usetex
Milestone
v3.2.0

Comments

@ngoldbaum
Copy link
Contributor

ngoldbaum commented Apr 3, 2017 •
edited
Loading

Bug report

When making a plot using the 'classic' style, matplotlib overrides the default value for text.latex.unicode, setting it to False. This causes a problem for any plot that has negative numbers in any label, because matplotlib now uses a unicode symbol to represent the negative sign.

Code for reproduction

import numpy as np
import matplotlib.pyplot as plt
from matplotlib import rcParams
rcParams['text.usetex'] = True

# Example data
t = np.arange(-1.0, 1.0 + 0.01, 0.01)
s = np.cos(4 * np.pi * t) + 2

plt.plot(t, s)

with plt.style.context('classic'):
    plt.savefig('tex_demo')

Actual outcome

The following traceback:

  File "test_mpl.py", line 14, in <module>
    plt.savefig('tex_demo')
  File "/Users/goldbaum/Documents/matplotlib-git/lib/matplotlib/pyplot.py", line 707, in savefig
    res = fig.savefig(*args, **kwargs)
  File "/Users/goldbaum/Documents/matplotlib-git/lib/matplotlib/figure.py", line 1763, in savefig
    self.canvas.print_figure(*args, **kwargs)
  File "/Users/goldbaum/Documents/matplotlib-git/lib/matplotlib/backend_bases.py", line 2234, in print_figure
    **kwargs)
  File "/Users/goldbaum/Documents/matplotlib-git/lib/matplotlib/backends/backend_agg.py", line 548, in print_png
    FigureCanvasAgg.draw(self)
  File "/Users/goldbaum/Documents/matplotlib-git/lib/matplotlib/backends/backend_agg.py", line 467, in draw
    self.figure.draw(self.renderer)
  File "/Users/goldbaum/Documents/matplotlib-git/lib/matplotlib/artist.py", line 68, in draw_wrapper
    return draw(artist, renderer, *args, **kwargs)
  File "/Users/goldbaum/Documents/matplotlib-git/lib/matplotlib/figure.py", line 1240, in draw
    renderer, self, artists, self.suppressComposite)
  File "/Users/goldbaum/Documents/matplotlib-git/lib/matplotlib/image.py", line 139, in _draw_list_compositing_images
    a.draw(renderer)
  File "/Users/goldbaum/Documents/matplotlib-git/lib/matplotlib/artist.py", line 68, in draw_wrapper
    return draw(artist, renderer, *args, **kwargs)
  File "/Users/goldbaum/Documents/matplotlib-git/lib/matplotlib/axes/_base.py", line 2386, in draw
    mimage._draw_list_compositing_images(renderer, self, artists)
  File "/Users/goldbaum/Documents/matplotlib-git/lib/matplotlib/image.py", line 139, in _draw_list_compositing_images
    a.draw(renderer)
  File "/Users/goldbaum/Documents/matplotlib-git/lib/matplotlib/artist.py", line 68, in draw_wrapper
    return draw(artist, renderer, *args, **kwargs)
  File "/Users/goldbaum/Documents/matplotlib-git/lib/matplotlib/axis.py", line 1118, in draw
    tick.draw(renderer)
  File "/Users/goldbaum/Documents/matplotlib-git/lib/matplotlib/artist.py", line 68, in draw_wrapper
    return draw(artist, renderer, *args, **kwargs)
  File "/Users/goldbaum/Documents/matplotlib-git/lib/matplotlib/axis.py", line 268, in draw
    self.label1.draw(renderer)
  File "/Users/goldbaum/Documents/matplotlib-git/lib/matplotlib/artist.py", line 68, in draw_wrapper
    return draw(artist, renderer, *args, **kwargs)
  File "/Users/goldbaum/Documents/matplotlib-git/lib/matplotlib/text.py", line 797, in draw
    mtext=mtext)
  File "/Users/goldbaum/Documents/matplotlib-git/lib/matplotlib/backends/backend_agg.py", line 255, in draw_tex
    Z = texmanager.get_grey(s, size, self.dpi)
  File "/Users/goldbaum/Documents/matplotlib-git/lib/matplotlib/texmanager.py", line 585, in get_grey
    pngfile = self.make_png(tex, fontsize, dpi)
  File "/Users/goldbaum/Documents/matplotlib-git/lib/matplotlib/texmanager.py", line 508, in make_png
    dvifile = self.make_dvi(tex, fontsize)
  File "/Users/goldbaum/Documents/matplotlib-git/lib/matplotlib/texmanager.py", line 400, in make_dvi
    texfile = self.make_tex(tex, fontsize)
  File "/Users/goldbaum/Documents/matplotlib-git/lib/matplotlib/texmanager.py", line 311, in make_tex
    if rcParams['text.latex.unicode']:
UnicodeEncodeError: 'ascii' codec can't encode character '\u2212' in position 298: ordinal not in range(128)

Expected outcome

The plot should be saved to disk without erroring. The script runs without error on matplotlib 1.5.3.

Matplotlib version

Matplotlib 2.0.0 on OSX, installed via pip.
@ngoldbaum
Copy link
Contributor Author

ngoldbaum commented Apr 3, 2017 •
edited
Loading

Ah, it works if the bits that interact with matplotlib look like this:

with plt.style.context('classic'):
    plt.plot(t, s)
    plt.savefig('tex_demo')

instead of this:

plt.plot(t, s)
with plt.style.context('classic'):
    plt.savefig('tex_demo')

So I guess that explains why no one noticed until now.

Unfortunately the way the script runs in the description is more or less how yt is using matplotlib. It's kind of a hack, but it's the only way I could figure out to work around #6518 and make sure yt plots have the same fonts they had under matplotlib 1.5.3.

@tacaswell
Copy link
Member

I can reproduce this, but do not understand why....

I think the important thing is if the figure (and hence axes and such) are created in the context manager or not.

The three obviously relavent rcparams ('text.latex.unicode', 'text.usetex', 'axes.unicode_minus') are the same in the new default and classic mode (or I can not read which is also possible).

A workaround might be to set 'axes.unicode_minus' to False

@tacaswell
Copy link
Member 

import numpy as np
import matplotlib.pyplot as plt
from matplotlib import rcParams
rcParams['text.usetex'] = True

# Example data
t = np.arange(-1.0, 1.0 + 0.01, 0.01)
s = np.cos(4 * np.pi * t) + 2


plt.close('all')
plt.plot(t, s)

with plt.style.context({'text.usetex': True}, 'classic'):

    plt.savefig('tex_demo')

This behaves as expected, but again, not quite sure why.

@tacaswell tacaswell added this to the 2.0.1 (next bug fix release) milestone Apr 4, 2017
@ngoldbaum
Copy link
Contributor Author

According to the API docs, plt.style.context({'text.usetex': True}, 'classic') is equivalent to plt.style.context({'text.usetex': True}) (the second argument to context is a boolean). But I see what you're getting at, this also triggers a unicode error:

import matplotlib.pyplot as plt
# Example data
t = np.arange(-1.0, 1.0 + 0.01, 0.01)
s = np.cos(4 * np.pi * t) + 2

from matplotlib import rcParams
rcParams['text.usetex'] = True

plt.close('all')
plt.plot(t, s)

with plt.style.context({'text.usetex': False}):
    plt.savefig('tex_demo')

@tacaswell
Copy link
Member

hmm, I mis-remembered that API then...It is a list of styles, not *args of styles:

import numpy as np
import matplotlib.pyplot as plt
from matplotlib import rcParams
rcParams['text.usetex'] = True

# Example data
t = np.arange(-1.0, 1.0 + 0.01, 0.01)
s = np.cos(4 * np.pi * t) + 2


plt.close('all')
plt.plot(t, s)

with plt.style.context(['classic', {'text.usetex': True}]):

    plt.savefig('tex_demo')

and that order makes more sense!

@ngoldbaum
Copy link
Contributor Author

ngoldbaum commented Apr 4, 2017 •
edited
Loading

FWIW I've mitigated this on the yt side with yt PR 2573 (https://bitbucket.org/yt_analysis/yt/pull-requests/2573). I guess using the classic style just to get the fonts I want was a bit of a blunt instrument.

@QuLogic QuLogic modified the milestones: 2.0.1 (next bug fix release), 2.0.2 (next bug fix release) May 3, 2017
@tacaswell tacaswell modified the milestones: 2.1.1 (next bug fix release), 2.2 (next feature release) Oct 9, 2017
@anntzer
Copy link
Contributor

anntzer commented Jun 18, 2019

text.latex.unicode is now on by default, per #11381 (not on the 2.2.x branch, FWIW).
However this now fails with

RuntimeError: latex was not able to process the following string:
b'$\\u22121.5$'

Here is the full report generated by latex:
This is pdfTeX, Version 3.14159265-2.6-1.40.20 (TeX Live 2019) (preloaded format=latex)
 restricted \write18 enabled.
entering extended mode
(/tmp/mpld/tex.cache/ce3437ef20f4f97917585a4bd387f953.tex
LaTeX2e <2018-12-01>

<elided>

! Package inputenc Error: Unicode character − (U+2212)
(inputenc)                not set up for use with LaTeX.

See the inputenc package documentation for explanation.
Type  H <return>  for immediate help.
 ...                                              
                                                  
l.14 ...tsize{10.000000}{12.500000}{\sffamily $−
                                                  1.5$}
No pages of output.
Transcript written on ce3437ef20f4f97917585a4bd387f953.log.

but I have a fix coming up for that...

@anntzer anntzer mentioned this issue Jun 18, 2019
Merged
6 tasks
@QuLogic QuLogic modified the milestones: needs sorting, v3.2.0 Jun 19, 2019
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
topic: text/usetex
Projects
None yet
Milestone
v3.2.0
Development

Successfully merging a pull request may close this issue.

Fix unicode_minus + usetex. anntzer/matplotlib
5 participants
@jkseppan @tacaswell @QuLogic @anntzer @ngoldbaum