IPython Comm backend

I've been working on a backend based on the webagg backend, but that uses the IPython Comm architecture at https://github.com/ipython/ipython/pull/4195 to send messages instead of starting a server and opening websocket connections. I have an initial version in my github ipython-comm branch (see https://github.com/jasongrout/matplotlib/compare/ipython-comm). I'm getting confused about how the backend infrastructure works, though, like what the purpose for the FigureManager class is, etc. I'm running out of time to work on this now, and I'm hoping that someone will take what work I've done here and get it working properly with the matplotlib architecture. If not, I'll probably tinker with this more later.

Thanks,

Jason

···

--
Jason Grout

Sorry to take so long to get to this. This is a nice piece of work.

The most obvious thing is that this is a copy-and-paste of the existing WebAgg backend -- and maintaining the two is going to be much harder than building both out of the same pieces. As of 6389d14f, the WebAgg backend was refactored so that the transport that it uses to communicate to the browser is no longer hard coded. This was done in large part to support working with IPyhton in this way. (That is, it used to only communicate with the browser through Tornado, but now it can be anything that can send bits back and forth). There's an example of this in `examples/user_interfaces/embedding_webagg.py` that shows how to do this (using Tornado, but again, it doesn't have to be). There's no guarantees that this interface is sufficient, so it may require some back and forth on this to make it all work.

I think the first thing I would do would be to refactor this to use that. It's a little hard to tell what you've changed from the original WebAgg backend to get it to support IPython. If it were built on top of, rather than in addition to, WebAgg, that would be more obvious.

Mike

···

On 10/10/2013 06:08 PM, Jason Grout wrote:

I've been working on a backend based on the webagg backend, but that
uses the IPython Comm architecture at
https://github.com/ipython/ipython/pull/4195 to send messages instead of
starting a server and opening websocket connections. I have an initial
version in my github ipython-comm branch (see
https://github.com/jasongrout/matplotlib/compare/ipython-comm). I'm
getting confused about how the backend infrastructure works, though,
like what the purpose for the FigureManager class is, etc. I'm running
out of time to work on this now, and I'm hoping that someone will take
what work I've done here and get it working properly with the matplotlib
architecture. If not, I'll probably tinker with this more later.

Thanks,

Jason

--
Jason Grout

------------------------------------------------------------------------------
October Webinars: Code for Performance
Free Intel webinars can help you accelerate application performance.
Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most from
the latest Intel processors and coprocessors. See abstracts and register >
http://pubads.g.doubleclick.net/gampad/clk?id=60134071&iu=/4140/ostg.clktrk
_______________________________________________
Matplotlib-devel mailing list
Matplotlib-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/matplotlib-devel

--
                    _

\/|o _|_ _. _ | | \.__ __|__|_|_ _ _ ._ _
>>(_| |(_|(/_| |_/|(_)(/_|_ |_|_)(_)(_)| | |

http://www.droettboom.com

Thanks for the feedback. I was thinking that a refactor to pull out the communication layer would be really nice.

I didn't change the WebAgg backend because I figured you wanted it around still. I figured a plain old diff with the file would reveal changes.

Anyways, thanks for the pointer to the refactor commit. I hope to look at this again sometime soon.

Jason

···

On 10/16/13 1:58 PM, Michael Droettboom wrote:

Sorry to take so long to get to this. This is a nice piece of work.

The most obvious thing is that this is a copy-and-paste of the existing
WebAgg backend -- and maintaining the two is going to be much harder
than building both out of the same pieces. As of 6389d14f, the WebAgg
backend was refactored so that the transport that it uses to communicate
to the browser is no longer hard coded. This was done in large part to
support working with IPyhton in this way. (That is, it used to only
communicate with the browser through Tornado, but now it can be anything
that can send bits back and forth). There's an example of this in
`examples/user_interfaces/embedding_webagg.py` that shows how to do this
(using Tornado, but again, it doesn't have to be). There's no guarantees
that this interface is sufficient, so it may require some back and forth
on this to make it all work.

I think the first thing I would do would be to refactor this to use
that. It's a little hard to tell what you've changed from the original
WebAgg backend to get it to support IPython. If it were built on top
of, rather than in addition to, WebAgg, that would be more obvious.

Following a very helpful conversation with Michael this morning in the dev hangout, I got this working with the current master (of matplotlib and ipython). The refactoring made the code much better; thanks!

I updated the pull request at https://github.com/matplotlib/matplotlib/pull/2524

To test this, run IPython (master branch, to get the comm commits), and put this in a cell: https://github.com/matplotlib/matplotlib/pull/2524#issuecomment-28539813

Then you can execute something like:

from matplotlib.figure import Figure
import numpy as np
fig = Figure()
a = fig.add_subplot(111)
t = np.arange(0.0, 3.0, 0.01)
s = np.sin(2 * np.pi * t)
a.plot(t, s)
CommFigure(fig)

and get a live figure in the IPython notebook that uses the comm messaging infrastructure.

Michael---do you have time to take it from here?

Thanks,

Jason

···

On 10/16/13 3:46 PM, Jason Grout wrote:

On 10/16/13 1:58 PM, Michael Droettboom wrote:

Sorry to take so long to get to this. This is a nice piece of work.

The most obvious thing is that this is a copy-and-paste of the existing
WebAgg backend -- and maintaining the two is going to be much harder
than building both out of the same pieces. As of 6389d14f, the WebAgg
backend was refactored so that the transport that it uses to communicate
to the browser is no longer hard coded. This was done in large part to
support working with IPyhton in this way. (That is, it used to only
communicate with the browser through Tornado, but now it can be anything
that can send bits back and forth). There's an example of this in
`examples/user_interfaces/embedding_webagg.py` that shows how to do this
(using Tornado, but again, it doesn't have to be). There's no guarantees
that this interface is sufficient, so it may require some back and forth
on this to make it all work.

I think the first thing I would do would be to refactor this to use
that. It's a little hard to tell what you've changed from the original
WebAgg backend to get it to support IPython. If it were built on top
of, rather than in addition to, WebAgg, that would be more obvious.

Thanks for the feedback. I was thinking that a refactor to pull out the
communication layer would be really nice.

I didn't change the WebAgg backend because I figured you wanted it
around still. I figured a plain old diff with the file would reveal
changes.

Anyways, thanks for the pointer to the refactor commit. I hope to look
at this again sometime soon.

I also got this initial experimental demo working on the Sage cell server:

http://sagecell.sagemath.org/?q=ilpajg (for a sage plotting example)

http://sagecell.sagemath.org/?q=spadja (for one of the examples from the
matplotlib docs)

Thanks,

Jason

···

On 11/14/13 7:24 PM, Jason Grout wrote:

Following a very helpful conversation with Michael this morning in the
dev hangout, I got this working with the current master (of matplotlib
and ipython).

Very nice.
Thanks for sharing Jason.

···

On 15 November 2013 13:18, Jason Grout <jason-sage@…691…> wrote:

On 11/14/13 7:24 PM, Jason Grout wrote:

Following a very helpful conversation with Michael this morning in the

dev hangout, I got this working with the current master (of matplotlib

and ipython).

I also got this initial experimental demo working on the Sage cell server:

http://sagecell.sagemath.org/?q=ilpajg (for a sage plotting example)

http://sagecell.sagemath.org/?q=spadja (for one of the examples from the

matplotlib docs)

Thanks,

Jason


DreamFactory - Open Source REST & JSON Services for HTML5 & Native Apps

OAuth, Users, Roles, SQL, NoSQL, BLOB Storage and External API Access

Free app hosting. Or install the open source package on any LAMP server.

Sign up and see examples for AngularJS, jQuery, Sencha Touch and Native!

http://pubads.g.doubleclick.net/gampad/clk?id=63469471&iu=/4140/ostg.clktrk


Matplotlib-devel mailing list

Matplotlib-devel@lists.sourceforge.net

https://lists.sourceforge.net/lists/listinfo/matplotlib-devel

This is great. I can see what next steps are needed, but probably not for a few days...

Mike

···

On 11/14/2013 08:24 PM, Jason Grout wrote:

On 10/16/13 3:46 PM, Jason Grout wrote:

On 10/16/13 1:58 PM, Michael Droettboom wrote:

Sorry to take so long to get to this. This is a nice piece of work.

The most obvious thing is that this is a copy-and-paste of the existing
WebAgg backend -- and maintaining the two is going to be much harder
than building both out of the same pieces. As of 6389d14f, the WebAgg
backend was refactored so that the transport that it uses to communicate
to the browser is no longer hard coded. This was done in large part to
support working with IPyhton in this way. (That is, it used to only
communicate with the browser through Tornado, but now it can be anything
that can send bits back and forth). There's an example of this in
`examples/user_interfaces/embedding_webagg.py` that shows how to do this
(using Tornado, but again, it doesn't have to be). There's no guarantees
that this interface is sufficient, so it may require some back and forth
on this to make it all work.

I think the first thing I would do would be to refactor this to use
that. It's a little hard to tell what you've changed from the original
WebAgg backend to get it to support IPython. If it were built on top
of, rather than in addition to, WebAgg, that would be more obvious.

Thanks for the feedback. I was thinking that a refactor to pull out the
communication layer would be really nice.

I didn't change the WebAgg backend because I figured you wanted it
around still. I figured a plain old diff with the file would reveal
changes.

Anyways, thanks for the pointer to the refactor commit. I hope to look
at this again sometime soon.

Following a very helpful conversation with Michael this morning in the
dev hangout, I got this working with the current master (of matplotlib
and ipython). The refactoring made the code much better; thanks!

I updated the pull request at
https://github.com/matplotlib/matplotlib/pull/2524

To test this, run IPython (master branch, to get the comm commits), and
put this in a cell:
https://github.com/matplotlib/matplotlib/pull/2524#issuecomment-28539813

Then you can execute something like:

from matplotlib.figure import Figure
import numpy as np
fig = Figure()
a = fig.add_subplot(111)
t = np.arange(0.0, 3.0, 0.01)
s = np.sin(2 * np.pi * t)
a.plot(t, s)
CommFigure(fig)

and get a live figure in the IPython notebook that uses the comm
messaging infrastructure.

Michael---do you have time to take it from here?

--
                    _

\/|o _|_ _. _ | | \.__ __|__|_|_ _ _ ._ _
>>(_| |(_|(/_| |_/|(_)(/_|_ |_|_)(_)(_)| | |

http://www.droettboom.com