include matplotlib html5 canvas rendering to sage

[williamstein]

See http://code.google.com/p/mplh5canvas/

Let's include this in Sage soon. I just tried it out for fun on my laptop:

(1) Grab and extract the tarball from here:
http://code.google.com/p/mplh5canvas/downloads/list
then install it with "sage -python setup.py install"

(2) Install netifaces:
sage -sh
easy_install netifaces
exit

(3) In the Sage notebook, or command line, paste something like this:

import matplotlib
matplotlib.use('module://mplh5canvas.backend_h5canvas')
from pylab import *
import time

"""Simple static plot, mostly for testing zooming...
"""
clf()

theta = arange(0,8*pi,0.1)
a=1
b=.2

for dt in arange(0,2*pi,pi/2.0):

     x = a*cos( theta+dt )*exp(b*theta)
     y = a*sin( theta+dt )*exp(b*theta)

     dt = dt+pi/4.0

     x2 = a*cos( theta+dt )*exp(b*theta)
     y2 = a*sin( theta+dt )*exp(b*theta)

     xf = concatenate( (x,x2[::-1]) )
     yf = concatenate( (y,y2[::-1]) )

     p1=fill(xf,yf)

show()

Basically, you can put nearly any example from the matplotlib website before show (and after clf()) and it will "just work".

What happens above is that a new server is started and the plot displayed in it. Figuring out how to use our existing twisted server for interaction could be a nontrivial but very import challenge.

CC: @jasongrout

Component: graphics

Issue created by migration from https://trac.sagemath.org/ticket/9471