Setting up VSCode for Python, R and Quarto

Author

Klaus G.

Published

September 21, 2024

This article guides you through installing Python, R, and Quarto in VSCode on a Windows machine.

Downloads

Download the installers for:

First we install Python. You can have multiple Python versions installed in parallel. On Windows machines Python can be installed in different locations, depending if it was installed via the app store or as an installable. Open e cmd.exe or powershell.exe and type python --version. You will get the Python interpreter version of the python.exe that is in the %PATH% variable. It can be handy to have the folders with python.exe in the PATH variable. Normally pip is installed with Python. pip is the python package installer. You can type pip --version on the command line. If you dont get a response then you might have to install pip separately.

Also install R, Quarto, Git and VSCode. The order does not matter

The paths of the executables of all should be in the PATH environment variable of Windows. echo %PATH% on cmd.exe. Or/and you should know the respective paths so we can add them to the VSCode extenstions manually.

VSCode extensions

In VSCode we need some extensions to interact properly with the above installed. The extensions names are listed below.

  • For Python:
    • Python
    • Python debugger
    • Pylance
  • For R:
    • R
    • R debugger
  • Quarto
Tip

You can set VSCode to autosave.

Python virtual environment

Create a folder for a Python/R project with a Windows explorer and open it with File –> Open Folder. Now we create a virtual Python environment in this folder. With Ctrl + shift + p open the VSCode command palette. Type “create” and you should see Python: Create Environment. VSCode will find a the python interpreter that is in %PATH% or you can enter a path to a different python.exe (different version). From now onward, if you use pip to install a python library, it will be installed locally in the ./.venv folder of your project. You should have a .venv folder in your project folder.

With Terminal –> New Terminal open a new terminal window (basically a OS command line). Lets install the python library radian. radian makes the R console colorful and better readable. Open a terminal and pip install -U radian installs the required files. We need some more Python libraries. pip install pandas installs the pandas library. pip list lists all locally installed libraries and pip freeze > requirements.txt creates a list with all installed libraries plus version numbers. A great interactive python console is ipython, pip install ipython. Other useful packages are plotnine, statsmodels, scipy.

Customize VSCode extensions

We still have to adapt a few settings in the VSCode extensions. Click on the Extensions symbol on the very left side, then on the sproket, select Extension settings. Look for R > Rterm: Windows and enter the path for radian ({workspaceFolder}\.venv\Scripts\radian.exe … copy this path as it is. You dont have to replace {workspaceFolder} with an explicit path). Set R: Bracketed Paste. And in the R debugger extension we need to set the the R interpreter path, R > Rpath Windows to, e.g. C:\Program Files\R\R-4.4.1\bin\x64\R.exe. Note that the extension settings are global to all VSCode projects.

Getting R ready

For R to work properly we need first three R packages: languageserver and httpg. Simplest way is to click on the R symbol on the very left of VSCode (somewhere below the extensions symbol). Under ‘help pages’ click on ‘Install CRAN Package’ or on the plus sign next to it and install * httpg and * languageserver * renv

At this point we have a working quarto setup with a virtual Python environment and globally installed R packages. We can also setup a virtual R environment, (renv)[https://rstudio.github.io/renv/articles/renv.html], to keep R packages local in the project folder. Virtual environments allow different projects using different interpreters and package versions.

Using renv

Installing renv: More information can be found here renv. With renv::init() on the R console we create a R environment in the current project folder. To open the R console just type radian on a VSCode terminal command line (Terminal -> New Terminal). You can also type R.exe for a plain R console. After restarting the R session (exit the console with q() )we should also install the following packages from the R console.

  • install.packages("languageserver")
  • install.packages("httpgd")
  • install.packages("reticulate")
  • install.packages("tidyverse")
  • install.packages("ggplot2")

Hint: renv.lock only reflects packages used in the project, see. To be sure that renv.lock is up to date, type renv::snapshot() after writing some R code.

Quarto

Quarto commands are available from the terminal command line. To render a .qmd you would type quarto render myScrip.qmd for example. Tip: quarto install tinytex from the terminal command line, installs a library that allows .pdf rendering from quarto.

Important

Both Python and R? I would say if you only need Python then do not install a mixed environment. You only need the Python part. Same for R only.

Let’s try it out!

my_R_and_python_example.qmd shows how to use R and Python in the same Quarto document, how to exchange dataframes and how multi-indexing is translated. Try it with your VSCode setup. You have to uncomment the use_virtualenv() at the beginning … or we execute and render this .qmd file also with a a GitHub Action workflow … here the result

Resources

Python download https://www.python.org/downloads/

VSCode download https://code.visualstudio.com/download

Quarto download https://quarto.org/docs/download/

Git download https://git-scm.com/download/win

Quarto HTML basics https://quarto.org/docs/output-formats/html-basics.html

Quarto Markdown basics https://quarto.org/docs/authoring/markdown-basics.html

Quarto GitHub Pages https://quarto.org/docs/publishing/github-pages.html

renv https://rstudio.github.io/renv/articles/renv.html

some tweak for renv

Helpful videos

Git in VSCode https://www.youtube.com/watch?v=i_23KUAEtUM

How this page was published on GitHub Pages?

In .github.yml a GitHub Actions workflow was defined. A how-to can be found here. If you use Actions then you dont need to render to a ./doc directory. In the GitHub Pages settings you do NOT set pages from Actions but keep from gh-pages branch. Also there is a misleading mistake in the how-to:

Before configuring the publishing action, it’s important that you run quarto publish gh-pages locally, once. This will create the _publish.yml configuration required by the subsequent invocations of the GitHub Action. To do this, run the following from within your project:

… No _publish.yml will ever be created.

And have a look at the source files for this page!

Note

This is still a work in progress! Any feedback highly welcome!

Caution

You can also contribute on Github directly! Become a contributor, make changes and create a pull request.

The panda eats, shoots and leaves.

The panda eats shoots and leaves.