Re: Using Emacs to dsiplay interactive python plots with matplotlib

[Eduardo Ochs]

On Fri, 19 May 2023 at 07:43, CNRS <valentin.poncet@cnrs.fr> wrote:
>
> Hello all,
>
>
> I posted the following on stackoverflow
> [<https://stackoverflow.com/questions/76180861/is-it-possible-to-have-interactive-plots-in-emacs>],
> I thought maybe it’s more appropriate to ask it here, as you might
> have more knowledge on that matter and might not be using stack.
>
>
> I was trying to use Emacs as my editor for my research (I followed all
> steps in Jon Fincher’s guide on Realpython :
> [<https://realpython.com/emacs-the-best-python-editor/>] and learnt the
> basics with the tutorial).
>
> I use jupyter notebooks a lot for data visualization, so I installed
> ein [<https://github.com/millejoh/emacs-ipython-notebook>]. I can use
> my virtual environment in Emacs (pyvenv-workon), open my notebook run
> all the cells fine : until I launch a cell with a maptlotlib plot.
>
> In jupyter to have an interactive plot I put « %matplotlib nbagg » in
> my code cell and it works just fine. In Emacs it prints «
> <IPython.core.display.Javascript object » but it does not display it.
> If I do « %matplotlib inline », it opens a .png file in my image
> viewer but I do not have the features of an interactive plot.
>
> In an ideal world I would also like to be able to see a matplotlib
> anmation.
>
> I browsed a bit and I think I understood that Emacs could not work with
> Javascript objects, is that correct ? A possible workaround for me
> would be to have the /interactive/ plot in a seperate window but I did
> not find how to do that, is there some wiki I missed ?
>
> Cheers
>
>
> What I tried :
>
> · %matplotlib nbagg returns an interactive figure in jupyter notebook
> browser interface, seems not to work in Emacs.
>
> · plt.ion() should give interactive access does nothing.
>
> · %matplotlb inline opens a non-interactive .png file with the plot


A related question:

are there any packages that 1) let Emacs talk to external programs
that do interactive plots and animations, and that 2) are
"hacker-friendly" in the sense that their innards are easy to
understand and easy to play with? In 2020 or 2021 I spent a few days
trying to learn how to use Jupyter Notebooks from Emacs, but I found
that there were too many layers (for me - my mental buffers are small,
sorry!) between the user interface and the low-level parts that I
mostly interested in...

  Thanks in advance!
    Eduardo Ochs
    http://anggtwu.net/#eev
    http://anggtwu.net/eepitch.html
    http://anggtwu.net/2021-org-for-non-users.html