One nice feature in RStudio is that the notebook code runs in the console. That gives you access to variables for testing and playing around. It's useful.
There are some problems with this:
I was able to create a linked console by opening a terminal and using jupyter console --existing. It worked as intended.
Sounds a bit crazy, but what if after each cell runs the notebook environment can be copied to a console. You can play around with variables but not pollute the notebook environment.
I was able to create a linked console by opening a terminal and using
jupyter console --existing. It worked as intended.
You can create a console linked to the same kernel as a notebook by right-clicking in the notebook and choosing "New console for notebook". (This is a console hooked up to the same kernel as the notebook, so you get a new console for each notebook you do this for).
Sounds a bit crazy, but what if after each cell runs the notebook environment can be copied to a console. You can play around with variables but not pollute the notebook environment.
It sounds like you are asking to fork the kernel after every cell execution, and open it up as a separate kernel not connected to the notebook, so you can play around with the state at the time of that execution? That sounds pretty tricky. Can you give some concrete scenarios walking through what this workflow might be like?
Most helpful comment
You can create a console linked to the same kernel as a notebook by right-clicking in the notebook and choosing "New console for notebook". (This is a console hooked up to the same kernel as the notebook, so you get a new console for each notebook you do this for).
It sounds like you are asking to fork the kernel after every cell execution, and open it up as a separate kernel not connected to the notebook, so you can play around with the state at the time of that execution? That sounds pretty tricky. Can you give some concrete scenarios walking through what this workflow might be like?