Is it possible to have an interactive console when working with Jupyter notebooks?

[chklee30]

I'm a user who is very used to using RStudio.

My typical workflow in RStudio involves writing code in an RMarkdown script, and then using the R console to interactively perform trivial tasks that I don't necessarily want to write in my main script (for ex. printing out the contents of a list variable for a quick eye check).

I thought that Positron would allow me to seamlessly copy this exact workflow for Python as well, where my RMarkdown script is replaced with a .ipynb jupyter notebook file, and the Python console would act analogously to the R console, but it seems like this may not be the case? Please confirm this for me.

In the attached screenshot, you can see that I've defined a 'data' variable in a jupyter notebook, this notebook is using a base conda env which is the same environment that is loaded in the Python console, but I cannot access the variable from the console. Moreoever, the 'variables' pane has a separate tab for an R environment, the base conda environment, and the actual jupyter notebook itself.

Is there no way to access this jupyter layer (the environment that holds my 'data' variable) interactively through a console?

[juliasilge]

I believe we have this feature request outlined in #3801, if you'd like to take a look, give a 👍, and add any additional details on your use case.

In the meantime, you might try out using a Quarto file (.qmd) for your Python analyses together with the console. It sounds like you (like me) might be accustomed to that style of working, and that setup transitions very well to Python.

[psomhorst]

Until this is supported, I can recommend the pypercent alternative approach (see the VSCode docs). In normal Python script files, you can create 'code cells' using #%% that can be run one at a time and out of order, just like notebooks. They run in the console, and you can use the console between cell runs.