Using a dual_canvas in a separate window

[felixludos]

I would like to develop a relatively simple interactive interface that can be accessed without any code or editor from jupyter. Ideally, just by running a python script from the terminal, a browser window would open with the interface (built in a dual_canvas) running.
I understand that this is meant to be used inside jupyter, but how much more difficult would it be to open a page with all widgets working.
Because if there is a way to get that working - this library could basically replace most GUI packages for python (which tend to be very old and annoying).

Anyway, this library is fantastic - keep up the great work!

[AaronWatters]

Dual canvas does not require the Jupyter backend.
It would be pretty easy to generate and launch a static HTML page which did not communicate back to the Python interpreter.
The hard part is adding interactions that call back to Python.

There is an interesting approach called StreamLit.
As I understand it this approach "reruns the script from the top" whenever there is a change in the UI which communicates with the Python back-end.

https://www.streamlit.io/

You could do a lot with this (and I'm thinking of writing an interface for jp_doodle to streamlet) but you could not (for example)
update the dots in a dot-plot without redrawing the whole canvas. So any interaction like that would be at least "jerky".

On my list of things to look into: I would like to develop a simple platform which uses electron and web sockets which could provide a simple mechanism for implementing HTML based GUIs for Python processes. So far it's just a concept.

Thanks for the suggestion.