Graphics Array needs suboptions to show

[kcrisman]

p1=plot(x^2, (x, -pi, pi), color='blue',legend_label="f") 
G=graphics_array(((p1,p1),(p1,p1))) 
G.show(figsize=[8,8]) 
Traceback (click to the left of this block for traceback) 
... 
KeyError: 'pop(): dictionary is empty' 

This is reported at this sage-support thread.

I think the fix should be the same as at #10244, but for the GraphicsArray class.

Related tickets:

• GraphicsArray() should be a Graphics() object, or have functionality to return one #10656 - Graphics Arrays should be able to become Graphics objects
• options for items in a GraphicsArray() should be set independently #10657 - options should be passed to all graphics, and they should be set separately if desired

CC: @jasongrout @johanrosenkilde @sagetrac-jhonrubia6 @fchapoton

Component: graphics

Issue created by migration from https://trac.sagemath.org/ticket/11160

[kcrisman]

comment:1

Hmm, but that doesn't work... but it must be a similar issue, anyway.

[kcrisman]

comment:2

Adding the suboptions we usually send to save to the normal !Graphics class matplotlib removes the error, but does the following - see attached graphic. I don't have time to do any more with this now, though.

[kcrisman]

Note the second graph has two legends for the same function!

[kcrisman]

comment:3

Attachment: tmp_0.png

Replying to @kcrisman:

Adding the suboptions we usually send to save to the normal !Graphics class matplotlib removes the error, but does the following - see attached graphic. I don't have time to do any more with this now, though.

I should add that my code was slightly simpler:

sage: p = plot(x,legend_label='f')
sage: G = graphics_array((p,p))
sage: G

[johanrosenkilde]

comment:5

This is a bug which might be related:

sage: g = Graphics()
sage: g += plot(x,legend_label='f')
sage: g.set_legend_options(title="Nice")
sage: G = graphics_array((g,g))
sage: G.show()
...
KeyError: 'font_family'

It seems that this part of the problem might be completely solved by adding the suboptions usually added when show()'ing a graphics object.

[egourgoulhon]

comment:8

The issue described here is fixed in #27865.

[egourgoulhon]

comment:9

Since #27865 has been merged in Sage 8.9.beta0, we could close this ticket as duplicate.

[johanrosenkilde]

comment:11

All the examples listed here work fine in the newest develop, at least.

[egourgoulhon]

comment:12

Replying to @johanrosenkilde:

All the examples listed here work fine in the newest develop, at least.

Then maybe you could set this ticket to positive review, given its status is "sage-duplicate/invalid/wontfix" (the positive review is necessary to close it).

[kcrisman]

comment:13

Any doctests necessary or are they all included in #27865 as I suspect?

[egourgoulhon]

comment:14

Replying to @kcrisman:

Any doctests necessary or are they all included in #27865 as I suspect?

Yes the doctests relevant to this ticket are included in #27865; in Sage 8.9.beta6, you can find them in lines 3481 - 3499 of src/sage/plot/plot.py:

        sage: L = [plot(sin(k*x), (x,-pi,pi)) for k in [1..3]]
        sage: ga = graphics_array(L)
        sage: ga.show(figsize=[5,3])  # smallish and compact

        sage: ga.show(figsize=[5,7])  # tall and thin; long time

        sage: ga.show(figsize=4)  # width=4 inches, height fixed from default aspect ratio

as well as in lines 555-613 of src/sage/plot/multigraphics.py:

        sage: g1 = plot(sin(x^2), (x, 0, 6), axes_labels=['$x$', '$y$'],
        ....:           axes=False, frame=True, gridlines='minor')
        sage: y = var('y')
        sage: g2 = streamline_plot((sin(x), cos(y)), (x,-3,3), (y,-3,3),
        ....:                      aspect_ratio=1)
        sage: g3 = graphs.DodecahedralGraph().plot()
        sage: g4 = polar_plot(sin(5*x)^2, (x, 0, 2*pi), color='green',
        ....:                 fontsize=8) \
        ....:      + circle((0,0), 0.5, rgbcolor='red', fill=True, alpha=0.1,
        ....:               legend_label='pink')
        sage: g4.set_legend_options(loc='upper right')

        sage: G = graphics_array([g1, g2, g3, g4])

        sage: G.show(figsize=[8, 3])

[kcrisman]

comment:15

I only asked because maybe that is why someone hasn't hit positive review. So I guess jsrn should :-) or actually you could as well, I think.

[egourgoulhon]

comment:16

Similarly #10466 and #10657 are set to "duplicate" and are waiting for a positive review to be closed...