Kelly BodwinAug 1, 2023

Does anyone in my #rstats and #python crew have good advice for a Colab -> Quarto render workflow?

Context: we'd like our students to work in Colab, but turn in a final Quarto-rendered file.

The workflow of "download ipynb -> run quarto render" feels like bad reproducibility. As students tweak the notebook for render, this will leave the Colab copy un-updated.

Ideas?

Show threadKelly Bodwin

Gonna be annoying and tag you, @minecr and @Posit.

And just to clarify - we're not looking for an intermediate stop in qmd or RStudio. With infinite class time we'd do that, but for this one we're looking for a Colab-centric process that still gives us that beautiful Quarto rendering. :)

Aug 1, 2023 at 10:06pmWeb
Show threadScott LaForestAug 2, 2023
@kellybodwin @minecr @Posit I don’t know if this would work for you, but you can run multiline shell commands in Colab code blocks using the %%shell command. See the attached screenshot of a quarto rendered pdf using the shell on Colab to install quarto and render it.
Show threadJune ChoeAug 2, 2023

@scottyla @kellybodwin @minecr @Posit Just gave this a try because I'm interested in doing this too for the upcoming semester.

Sadly I don't think this works as straightforwardly in notebooks with *R runtime*. The %%shell macro doesn't work, though I could get it working with `system()`. But the bigger problem is with mounting the drive to point to the current notebook. There's no equivalent of google.colab.drive in R, and clicking the "Mount Drive" icon says it's only available for Python 😢