Currently the user can only guess from the notebooks that he/she needs to use ipympl as matplotlib backend to get things working. We should consider using a context manager to temporarily set ipympl as backend and return to the users default. This would have the advantage of minimal interference and will not use a lot of memory for plots which are intended not to be interactive.
Currently the user can only guess from the notebooks that he/she needs to use ipympl as matplotlib backend to get things working. We should consider using a context manager to temporarily set ipympl as backend and return to the users default. This would have the advantage of minimal interference and will not use a lot of memory for plots which are intended not to be interactive.