I want to create some plots using sage to then save them in my computer. I don't want the background to be white (or transparent). Is there a way of specifying the background of a plot inside sage?
(I'm looking for an option of the form:
sage: t=var('t')
sage: p=parametric_plot((t,2*t), (t, -5, 5))
sage: p.show(background=(.1,.2,.3))
or something like that.)
Using PIL offers a further way. PIL is included in Sage.
def show_bg(gobj,bgcolor=(100,200,200),**kwargs):
"""
add background color to a transparent plot
gobj any 2D graphics object created with option transparent=True
bgcolor tuple (r,g,b) r,g,b in range 0..255
or string with proper color name
kwargs all other keyword args are passed to save
EXAMPLES:
sage: P = plot(sin,-pi,pi,color='red',thickness=2,transparent=True)
sage: P += parametric_plot((2*x,x),(x,-2,2),color='gold',thickness=3)
sage: show_bg(P,aspect_ratio=3/2,figsize=5,bgcolor='mistyrose')
"""
# create temp file in DATA folder of this worksheet
gobj.save(DATA+'tmp_plot.png',**kwargs)
from PIL import Image
im = Image.open('./data/tmp_plot.png')
# create background file of same size and given bgcolor
bg = Image.new('RGB',im.size,bgcolor)
# paste im into bg using im as mask
bg.paste(im,(0,0),im)
bg.save('output.png')
I tried to use your example in green, but I got this answer: File "/tmp/tmpGDjbwd/___code___.py", line 5, in <module> show_bg(P,aspect_ratio=_sage_const_3 /_sage_const_2 ,figsize=_sage_const_5 ,bgcolor='mistyrose') NameError: name 'show_bg' is not defined I guess I should include some package before using this?
Create a plot with option transparent=True and save to DATA folder.
G = circle((2,2),3,transparent=True)
G.save(DATA+'circle.png')
Now you may create a backround file using imagemagick command-line tool and compose both files.
%sh
convert -size 600x600 xc:#ffffaa ./data/bg.png
convert ./data/bg.png ./data/circle.png -composite output.png
This SO question also seems relevant. However, even with using the relevant Sage functionality
sage: P = plot(sin(x),(x,-pi,pi),color='green',linestyle='--',thickness=10)
sage: P = plot(sin(x),(x,-pi,pi),color='green',linestyle='--',thickness=10)
sage: M = P.matplotlib()
sage: M.add_subplot(111,axisbg='r')
<matplotlib.axes.AxesSubplot at 0x115234610>
sage: from matplotlib.backends.backend_agg import FigureCanvasAgg
sage: M.set_canvas(FigureCanvasAgg(M))
sage: M.savefig('/Users/.../test.png',dpi=100,transparent=False)
it covers up/replaces the previous Axes object. It doesn't seem (easily) possible to change this afterward. (That is apparently a limitation in matplotlib and/or a feature of how they use the subplot instances.) However, it would be relatively straightforward to add this in Sage, though one would have to make sure it didn't mess anything up.
def show_bg(gobj,bgcolor=(100,200,200),**kwargs):
"""
add new background color to a plot
gobj any 2D graphics object
bgcolor tuple (r,g,b) r,g,b in range 0..255
or string with proper color name
kwargs all other keyword args are passed to save
EXAMPLES:
sage: P = plot(sin,-pi,pi,color='red',thickness=2)
sage: P += parametric_plot((2*x,x),(x,-2,2),color='gold',thickness=3)
sage: show_bg(P,aspect_ratio=3/2,figsize=5,bgcolor='mistyrose')
"""
# make gobj transparent
# the first element of graphics-array has to be transparent
# so let's draw a dummy point with option transparent=True
D = gobj.get_minmax_data()
gobj_t = point((D['xmin'],D['ymin']),color=bgcolor,transparent=True)
gobj_t += gobj
# create temp file in DATA folder of this worksheet
gobj_t.save(DATA+'tmp_plot.png',**kwargs)
from PIL import Image
im = Image.open('./data/tmp_plot.png')
# create background file of same size and given bgcolor
bg = Image.new('RGB',im.size,bgcolor)
# paste im into bg using im as mask
bg.paste(im,(0,0),im)
bg.save('output.png')
Interesting that mpl doesn't seem to support this... I assume this still works with "Pillow"? See http://trac.sagemath.org/ticket/15539 - I can't test this with newest devel version of Sage, I destroyed my install last night.
Asked: 2014-04-25 14:21:00 +0100
Seen: 3,468 times
Last updated: Apr 29 '14
Copyright Sage, 2010. Some rights reserved under creative commons license. Content on this site is licensed under a Creative Commons Attribution Share Alike 3.0 license.