mpld3 support

[berceanu]

trafficstars

Not sure weather this belongs here or in IJulia, but is there any forthcoming support for http://mpld3.github.io?

[stevengj]

I haven't tried MPLD3, but I don't see any particular reason why calling these functions now, via PyCall, wouldn't work already. Try it and see what happens.

i.e. try something like

using PyPlot
using PyCall
@pyimport mpld3
plot(rand(10))
mpld3.save_html()

Hopefully mpld3.show() will work, and perhaps even mpld3.display() will work to display it inline in the notebook; the latter depends precisely on how they did it, though (e.g. it sounds like mpld3.display() just returns an object with a repr_html method, in which case it should work in IJulia too).

[berceanu]

save_html() needs a figure instance as input. i tried save_html(gcf()) but it didn't work because gcf() fails with

PyError (PyObject_Call) <class 'AssertionError'>
AssertionError()
  File "/usr/lib/python3.4/site-packages/matplotlib/backend_bases.py", line 2169, in print_figure
    bbox_inches = self.figure.get_tightbbox(renderer)
  File "/usr/lib/python3.4/site-packages/matplotlib/figure.py", line 1565, in get_tightbbox
    _bbox = Bbox.union([b for b in bb if b.width != 0 or b.height != 0])
  File "/usr/lib/python3.4/site-packages/matplotlib/transforms.py", line 714, in union
    assert(len(bboxes))

 in pyerr_check at /home/berceanu/.julia/v0.3/PyCall/src/exception.jl:58
 in pycall at /home/berceanu/.julia/v0.3/PyCall/src/PyCall.jl:85
 in fn at /home/berceanu/.julia/v0.3/PyCall/src/conversions.jl:181
 in writemime at /home/berceanu/.julia/v0.3/PyPlot/src/PyPlot.jl:145
 in base64 at base64.jl:125
 in display_dict at /home/berceanu/.julia/v0.3/IJulia/src/execute_request.jl:34

[stevengj]

Not sure what is going on there. Did you try mpld3.show() or mpld3.display()?

[berceanu]

yup mpld3.show() -> kernel becomes unresponsive `mpld3.display()' gives

`PyError (PyObject_Call) <class 'AssertionError'>
AssertionError()
  File "/usr/lib/python3.4/site-packages/matplotlib/backend_bases.py", line 2169, in print_figure
    bbox_inches = self.figure.get_tightbbox(renderer)
  File "/usr/lib/python3.4/site-packages/matplotlib/figure.py", line 1565, in get_tightbbox
    _bbox = Bbox.union([b for b in bb if b.width != 0 or b.height != 0])
  File "/usr/lib/python3.4/site-packages/matplotlib/transforms.py", line 714, in union
    assert(len(bboxes))

 in pyerr_check at /home/berceanu/.julia/v0.3/PyCall/src/exception.jl:58
 in pycall at /home/berceanu/.julia/v0.3/PyCall/src/PyCall.jl:85
 in fn at /home/berceanu/.julia/v0.3/PyCall/src/conversions.jl:181
 in writemime at /home/berceanu/.julia/v0.3/PyPlot/src/PyPlot.jl:145
 in base64 at base64.jl:125
 in display_dict at /home/berceanu/.julia/v0.3/IJulia/src/execute_request.jl:34
 in display at /home/berceanu/.julia/v0.3/IJulia/src/inline.jl:35
 in display at multimedia.jl:149
 in display at /home/berceanu/.julia/v0.3/IJulia/src/inline.jl:56

[jverzani]

Was curious if this works now. The following did (a straight translation of http://mpld3.github.io/examples/scatter_tooltip.html). I couldn't get this brushing in this one to work though: http://mpld3.github.io/examples/linked_brush.html.

using PyPlot
using PyCall
@pyimport mpld3
pygui(true) # suppresses drawing in IJulia, but must be better way


fig, ax = subplots(subplot_kw={:axisbg=>"#EEEEEE"})
N = 100

scatter = ax[:scatter](randn(N),
                       randn(N),
                       c=rand(N),
                       s=1000 * rand(N),
                       alpha=0.3,
                       cmap=PyPlot.cm[:jet])
ax[:grid](color="white", linestyle="solid")

ax[:set_title]("Scatter Plot (with tooltips!)", size=20)

labels = ["point $i" for i in 1:N]
tooltip = mpld3.plugins[:PointLabelTooltip](scatter, labels=labels)
mpld3.plugins[:connect](gcf(), tooltip)

mpld3.display(gcf())