savefig() as PNG get a different result than the image shown in ipython notebook(SVG probably) #5859

gepcel · 2016-01-15T15:10:32Z

I'm using matplotlib to plot some images in ipython notebook. But after saving fig as PNG, I get a different image than the one shown in the notebook. I've come across this issue multiple times. Mostly savefig() lose the font info, but there maybe others. Can anyone take a look?

The code is as following:

%matplotlib inline
%config InlineBackend.figure_format = 'svg'

mpl.rcParams['font.sans-serif'] = 'SimHei'  #to render some Chinese character properly 

d = [19, 14, 8, 21, 17, 10, 16, 11, 8]
cmap=plt.get_cmap('Set1', 6)

fig = plt.figure(dpi=400)
for i, x in enumerate(d):
    plt.bar(i, x, color=cmap(i%3))

ax = gca()
xlim(-0.2, 9)
width = ax.patches[0].get_width()

ax.xaxis.set_minor_locator(FixedLocator(np.arange(9)+width/2))
ax.xaxis.set_minor_formatter(FixedFormatter(['合计', '鱼卵', '仔稚鱼']*3))
ax.xaxis.set_major_locator(FixedLocator(np.array([1, 4, 7])+width/2))
ax.xaxis.set_major_formatter(FixedFormatter(['2012年6月', '2013年6月', '2014年6月']))
ax.xaxis.set_tick_params(which='major', pad=17)
ax.xaxis.set_tick_params(which='both', tick1On=False, tick2On=False)
ax.yaxis.set_tick_params(which='both', tick1On=False, tick2On=False, label1On=False)

for r in ax.patches:
    h = r.get_height()
    ax.text(r.get_x()+w/2, h+0.4, int(h), ha='center', va='center')
legend(ax.patches[:3], ['合计', '鱼卵', '仔稚鱼'], loc='upper right')
plt.tight_layout()
fig.savefig('a.png', dpi=fig.dpi)

I got the fig as following in notebook:

And the 'a.png' from 'savefig()' is as following:

I'm doing some quick plot and formatting, using both 'pyplot' and 'pylab' sometimes, sorry for the dirty code.
Some backgroud:

Windows 10;
matplotlib  '1.5.1rc1'
python 3.4
ipython 3.2.1
matplotlibrc file is set as default

Some test:

1. Result of `fig.savefig('a.svg')` is the same as shown in notebook.
2. If manually edit the `matplotlibrc` as following, `savefig` get the same result as shown in notebook.

    font.family         : sans-serif
    font.sans-serif: SimHei

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

tacaswell · 2016-01-15T15:20:28Z

To clairfy, if you edit matplotlibrc the png comes out correct?

gepcel · 2016-01-15T15:33:29Z

@tacaswell Double checked, Yes.

Change the two lines as said in matplotlibrc -> restarte the kernel -> remove the line mpl.rcParams['font.sans-serif'] = 'SimHei' #to render some Chinese character properly in the code.

Then I can get the same image both by savefig as PNG and the image shown in notebook.

gepcel · 2016-01-15T15:37:07Z

Sorry, I misclosed the issue by clicking "Close and comment", I mean “Close the Preview and comment"

gepcel · 2016-01-18T06:28:16Z

@tacaswell @pelson
As I said before, I'm not sure the font info missing is the only defference between the png file from fig.savefig('a.png') and the image shown in notebook. But I don't know they are relative or not. And I don't know the issue is in my system environment or matplotlib or cartopy. This is very frustrating. Can anybody help? Let me show by an example:

%matplotlib inline
%config InlineBackend.figure_format = 'svg'

import cartopy.crs as ccrs
from cartopy.mpl.ticker import LongitudeFormatter, LatitudeFormatter
import numpy as np
import matplotlib.pyplot as plt
import pandas as pd

#prepare the data
data = pd.DataFrame(np.array([[123.5, 39.2, 0, 1, 2, 3, 4],
                              [123.5, 39.3, 5, 6, 0, 1, 2],
                              [121.5, 40.7, 3, 4, 5, 6, 0],
                              [121.5, 40.3, 1, 2, 3, 4, 5]]),
                    index=['D01', 'D02', 'D03', 'D04'],
                    columns=['x', 'y', 'g_3', 'g_4', 'g_5', 'g_6', 'g_7'])
labels = ['0', '0-0.5', '0.5-1', '1-2', '2-3', '3-4', '4-5']
groups = [data.groupby('g_'+str(i)) for i in [3, 5, 6, 4, 7]]

x0, x1, y0, y1 = extent = 121, 124, 38.8, 41.1
width = 5
height = width * (y1-y0)/(x1-x0)

# creating 3*2 maps. Actually I only need 5, but I don't how to do that while x&y 
# being both shared. So let's put a legend on the 6th one.
fig, axs = plt.subplots(3, 2, sharex=True, sharey=True, 
                        figsize=(width*2, height*3), 
                        dpi=300, subplot_kw={'projection': ccrs.PlateCarree()})
# for everymap
for i, ax in enumerate(axs.flatten()):
    #some settings of everymap
    ax.set_extent(extent)
    ax.set_xticks([122, 123], crs=ccrs.PlateCarree())
    ax.set_yticks([39, 40, 41], crs=ccrs.PlateCarree())
    ax.xaxis.set_major_formatter(LongitudeFormatter())
    ax.yaxis.set_major_formatter(LatitudeFormatter())
    [tick.label.set_fontsize(11) for tick in ax.yaxis.get_major_ticks()]
    [tick.label.set_fontsize(11) for tick in ax.xaxis.get_major_ticks()]

    #plot the point for the first 5th map
    if i < 5:
        for n, g in groups[i]:
            ax.scatter(g.x, g.y, s=10*n**2, c='k', transform=ccrs.PlateCarree())

# I need a legend for all 5 maps, let's put it on the 6th map
# I don't need to plot on the 6th map, so, I put it outside the extent
for n in range(len(labels)):
    ax.scatter(120, 38, s=10*n**2, c='k', transform=ccrs.PlateCarree(),
              label=labels[n])
ax.legend(loc='upper right', scatterpoints=1)
fig.tight_layout()
fig.savefig('a.png', dpi=300)

By running the code in notebook, I got an image shown as inline in notebook. I did a screen clipping. The image is as:

And the a.png file saved by fig.savefig('a.png', dpi=300) is as:

Difference:

Of the a.png file, do notice that the first plot is not the same size as the others, and it isn't arange well.
The scatter point of zero size (s=10*n**2 when n=0)
- The image shown in the notebook, the scatter point of size s=0 also got a small black point. But since I've zoomed out the page to get a full clipping, it might not be seen. It is really there in my notebook.
- The a.png file got no point of size zero, even zoomed in.

Besides those, what's the right way to do if I only need 3 or 5 (any odd number other than 1) axes? Because by my way, I have to erase the 6th axis in Photoshop and move the 6th's xaxis up to the 4th axis.

Please tell me if any more information is needede.
Thanks.

tacaswell · 2016-05-16T20:30:42Z

@gepcel Can you open a new issue for misplaced subplots?

mdboom · 2016-05-16T20:31:38Z

Have you tried deleting your font cache in ~/.matplotlib or ~/.config/matplotlib? matplotlib may not be picking up the font.

gepcel · 2016-05-17T13:11:07Z

@tacaswell yes, I can open a new issue for misplaced subplots tomorrow. I'll check if I can still produce this in my current environment.

Thanks for replying @mdboom . Tomorrow, I'll check that. It has been a long time.

gepcel · 2016-05-22T01:45:13Z

@tacaswell I have opened a new issue about the misplaced subplots as #6457.

gepcel · 2016-05-22T09:26:43Z

@mdboom.
Thanks for helping, it works.
I can still reproduce this after a very long time, even after a total reinstall of windows os.
And I have searched issues like #6326, and #4492.
Deleting the ~/.matplotlib works.

gepcel · 2016-05-23T07:19:19Z

For me, "deleting the ~/.matplotlib" can solve my problem, I'm fine if anyone close this issue, or anyone still want to do something.

Elgins · 2016-10-14T09:50:52Z

@gepcel Hi, gepcel, I have the same puzzle with you, I am user Mac, the plt.show() display the image fine, but when I savefig() it went wrong whatever png or svg. I delete the cache file mentioned above but it still didn't work, could you show me more detial about how you solve this, many thks if you help me.

tacaswell · 2016-10-14T12:06:53Z

This section of the docs http://matplotlib.org/devdocs/users/text_props.html#text-with-non-latin-glyphs maybe helpful here.

Elgins · 2016-10-16T09:24:33Z

Thank you very much, I have figure this problem out. For anyone who maybe count this, here is my suggestion: no matter Windows or Mac OS, just find the path that matplotlib use for font store, and replace the default font "Vera" with your right font.

tacaswell · 2016-10-19T16:11:43Z

@Elgins Replacing the 'vera' files with your preferred font is an extremely bad idea.

Elgins · 2016-10-21T06:05:38Z

@tacaswell Yeah, I am not professional, this should be the last method to try:)

QuLogic · 2017-02-06T05:37:36Z

Reporter says this is working.

tacaswell added the topic: text/unicode label Jan 15, 2016

tacaswell added this to the Critical bug fix release (1.5.2) milestone Jan 15, 2016

gepcel closed this as completed Jan 15, 2016

gepcel reopened this Jan 15, 2016

mdboom self-assigned this May 16, 2016

mdboom modified the milestones: 2.0.1 (next bug fix release), 1.5.2 (Critical bug fix release) May 16, 2016

QuLogic closed this as completed Feb 6, 2017

QuLogic modified the milestones: unassigned, 2.0.1 (next bug fix release) Feb 6, 2017

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Uh oh!

savefig() as PNG get a different result than the image shown in ipython notebook(SVG probably) #5859

savefig() as PNG get a different result than the image shown in ipython notebook(SVG probably) #5859

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

savefig() as PNG get a different result than the image shown in ipython notebook(SVG probably) #5859

savefig() as PNG get a different result than the image shown in ipython notebook(SVG probably) #5859

Comments

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!