options for items in a GraphicsArray() should be set independently

[nilesjohnson]

In particular, one ought to be able to specify axes=True/False on a per-graphic basis, if desired.

Currently, the following results in 8 plots with axes, and a text frame which also has axes, even though axes=False is explicitly given.

A = [plot(sin(t+k*pi/4),(-pi,pi)) for k in range(8)] + [text('hello world',(0,0), axes=False)]
graphics_array(A,3,3)

See the post at asksage about this issue for more details.

Related tickets:

• Graphics Array needs suboptions to show #11160 - Graphics Arrays need suboptions somehow
• GraphicsArray() should be a Graphics() object, or have functionality to return one #10656 - Graphics Arrays should be able to become Graphics objects

CC: @kcrisman @paulmasson @fchapoton

Component: graphics

Keywords: GraphicsArray, matplotlib

Issue created by migration from https://trac.sagemath.org/ticket/10657

[eviatarbach]

comment:3

This is also a problem with other options, such as tick_formatter, ticks, and axes_labels. They are only applied to the last graphic; see attached image produced by the following code:

graphics_array([plot(sin(B*x), xmin=-2*pi, xmax=2*pi, tick_formatter=pi, ticks=pi/2, axes_labels=('', '%s'%B)) for B in [-2,-1,-1/2,1/2,1,2]], 3, 2)

[eviatarbach]

Attachment: sage0.png

Picture of bug

[kcrisman]

Description changed:

--- 
+++ 
@@ -8,3 +8,7 @@
 ```
 
 See the post at asksage about [this issue](http://ask.sagemath.org/question/308/can-i-convert-a-graphicsarray-object-to-a-graphics) for more details.
+
+Related tickets: 
+* #11160 - Graphics Arrays need suboptions somehow
+* #10656 - Graphics Arrays should be able to become Graphics objects

[kcrisman]

comment:5

Replying to @eviatarbach:

This is also a problem with other options, such as tick_formatter, ticks, and axes_labels. They are only applied to the last graphic; see attached image produced by the following code:

> graphics_array([plot(sin(B*x), xmin=-2*pi, xmax=2*pi, tick_formatter=pi, ticks=pi/2, axes_labels=('', '%s'%B)) for B in [-2,-1,-1/2,1/2,1,2]], 3, 2)

#12591 is this. I'm going to suggest we keep this ticket for the original issue, though clearly they are related.

[ppurka]

comment:6

Workaround from sage-support:

p = plot(x, axes_labels=['x', 'x'])
q = plot(x^2, axes_labels=['x', 'x^2'])
r = plot(exp, axes_labels=['x', 'exp'])
# We want the order p,q,r; so we start with cyclic shifts 
graphics_array([q, r, p]).show() # cyclic permutation
graphics_array([r, p, q]).show() # cyclic permutation of above
graphics_array([p, q, r]).show() # another shift to get our order.

[kcrisman]

comment:7

Somewhere in this set of related tickets we should also unify the usage of figsize between graphics objects and graphics arrays. Maybe graphics arrays need to be a separate file eventually..

[kcrisman]

comment:8

Here is another report, by a seasoned Sage user.

[kcrisman]

comment:9

This should include setting axes for each plot; currently you can just set them all at once...

[kcrisman]

comment:10

Okay, the problem seems to be that we use the matplotlib save command for each (sub)figure, which doesn't seem to set them, but then the Sage save command for the very last one - which gives it the options as well as puts the image in the right file to appear. At #12591 there is some work at fixing a related issue that is basically the same bug, but fixing them together will be a little tricky, though a good thing in the long run.

[kcrisman]

comment:11

This post is probably also the same thing.