From: Jean-Michel Philippe <jean-michel.philippe@ir...> - 2004-09-14 06:41:47
> figure(num = 1, figsize=(8, 6), dpi=80, facecolor='w', edgecolor='k')
With matplotlib 0.62.4 I wasn't able to detect any effect of the
'edgecolor' property. I tried both string color values ('w', 'k' and
'r') and floating point color values (1.0, 0.95 and 0.5). What is it
supposed to change exactly, the window border, the subplot border? Are
my values wrong?
From: John Hunter <jdhunter@ac...> - 2004-09-14 14:41:41
>>>>> "Jean-Michel" == Jean-Michel Philippe <jean-michel.philippe@...> writes:
Jean-Michel> With matplotlib 0.62.4 I wasn't able to detect any
Jean-Michel> effect of the 'edgecolor' property. I tried both
Jean-Michel> string color values ('w', 'k' and 'r') and floating
Jean-Michel> point color values (1.0, 0.95 and 0.5). What is it
Jean-Michel> supposed to change exactly, the window border, the
Jean-Michel> subplot border? Are my values wrong?
The window border.
The figure background frame is a rectangle. Like all rectangles in
matplotlib, it has a facecolor and an edgecolor. Eg, the default axes
facecolor is white and edgecolor is black. The figure facecolor in a
GUI backend is gray, which is modeled after matlab and in my opinion
is more pleasing to the eye than a field of white. The default
edgecolor is white, but it's easier to see if you use 'r'. I just
tested and it is visible in GTKAgg and WXAgg but appears to be clipped
in tkagg. Not sure why.
The other gotcha is that the default is different for savefig. That
is because when you want to save hardcopy, you typically do not want
that gray background so the face and edgecolor for savefig default to
white. Thus you would need to do
to see it in hardcopy. Both the figure and savefig defaults are
exposed in the .matplotlibrc file.
I personally have never used the edgecolor property of figures. I
just added the attribute for completeness to expose the rectangle