On 11/30/2012 04:41 PM, Tiago Peixoto [via Main discussion list for the
graph-tool project] wrote:
> But this is correct. Does it really take five hours if you supply pos =
> random_layout() ?!
>
The one that took 6 hours was without supplying "pos". When I supplied a
random one, it took more than 2 hours and I stopped it.
Strangely, unlike the former it uses only one CPU. I believe graph_draw
is taking too long, because random_layout finishes almost immediatly.

When using your code to filter the largest component it takes 5 minutes,
most of the time processing the layout (as shown with the verbose option
you suggested).
But graph_draw is not wasting time calculating the graph layout when
supplied random positions, I don't know what it could be doing.

Does it also happen to you:
    pos = gt.random_layout(polysemy)
    print "here"
    gt.graph_draw(polysemy, pos=pos, output='ejercicio_5_gv.png')

> > Also I would like to see nodes expanded through the canvas, because
> they are
> > densely packed into a corner. What is the option for controlling this?
>
> The bounding box is calculated automatically if fit_view=True. If you
> set it False, the positions will be considered to correspond to the
> output format units (pixels in case of png, points in case of pdf, etc).
>
> If you are getting strange-looking results with fit_view=True, then
> please send me an example.
I've attached a screenshot, which shows unused space in the image.

I would like to separate nodes as much as possible so that edges are
visible, or strech them to fit the picture at least.

Cheers,
Alejandro

Screenshot from 2012-11-30 20:42:02.png (518K) Download Attachment


View this message in context: Re: Drawing large graphs
Sent from the Main discussion list for the graph-tool project mailing list archive at Nabble.com.