R and Shiny Training: If you find this blog to be interesting, please note that I offer personalized and group-based training sessions that may be reserved through Buy me a Coffee. Additionally, I provide training services in the Spanish language and am available to discuss means by which I may contribute to your Shiny project.

< section id="motivation" class="level1">

Motivation

Recently, the R for Data Science in Spanish repository (R Para Ciencia de Datos) has shown some errors when users send pull requests (i.e., see the most recent at the time of writing).

A few weeks ago I stopped using Netlify to render the R4DS-ES website, and I went back to GitHub Pages. I explained all the steps in a previous post.

I need to keep the project and contributions as simple as possible, so I moved the book to Quarto.

< section id="renaming-files" class="level1">

Renaming files

To rename all the files, I wrote a small bash script 99-rename-rmd-to-qmd.sh in the root of the project:

#!/bin/bash

# rename all Rmd files to qmd
for f in *.Rmd; do
    mv -- "$f" "${f%.Rmd}.qmd"
done

Then I run it with bash 99-rename-rmd-to-qmd.sh and all the files were renamed from Rmd to qmd.

< section id="installing-quarto" class="level1">

Installing Quarto

As I explained in a previous post, I created my own APT to be able to run apt install quarto on Linux Mint (i.e., Ubuntu).

< section id="rendering-the-book-locally" class="level1">

Rendering the book locally

From Visual Studio Code, I opened index.qmd and I was able to render the book and see it in the preview pane after doing the following changes:

1. Rename bookdown.yml to _quarto.yml.
2. Delete _output.tml.
3. Delete the header from index.qmd
4. Downgrade to knitr version 1.42 (i.e., see this bug report)

On the left side, I could see some entries that read “(PART) Explorar”. To fix this I deleted the headers of the form # (PART) Explorar from QMD files, and modified the _quarto.yml file to list the parts of the book:

project:
  type: book
  output-dir: docs
  
book:
  title: "R Para Ciencia de Datos"
  date: "2023-06-06"
  chapters:
    - index.qmd
    - 01-intro.qmd
    - part: "Explorar"
      chapters:
      - 02-explore.qmd
      - 03-visualize.qmd
      - 04-workflow-basics.qmd
      - 05-transform.qmd
      - 06-workflow-scripts.qmd
      - 07-eda.qmd
      - 08-workflow-projects.qmd
    - part: "Manejar o domar datos"
      chapters:
      - 09-wrangle.qmd
      - 10-tibble.qmd
      - 11-import.qmd
      - 12-tidy.qmd
      - 13-relational-data.qmd
      - 14-strings.qmd
      - 15-factors.qmd
      - 16-datetimes.qmd
    - part: "Programar"
      chapters:
      - 17-program.qmd
      - 18-pipes.qmd
      - 19-functions.qmd
      - 20-vectors.qmd
      - 21-iteration.qmd
    - part: "Modelar"
      chapters:
      - 22-model.qmd
      - 23-model-basics.qmd
      - 24-model-building.qmd
      - 25-model-many.qmd
    - part: "Comunicar"
      chapters:
      - 26-communicate.qmd
      - 27-rmarkdown.qmd
      - 28-communicate-plots.qmd
      - 29-rmarkdown-formats.qmd
      - 30-rmarkdown-workflow.qmd
format:
  html:
    theme: cosmo

In order to process changes on the fly, I added a file _metadata.yml with the following content:

# freeze computational output
# (see https://quarto.org/docs/projects/code-execution.html#freeze)
freeze: auto

< section id="rendering-the-book-on-github-actions" class="level1">

Rendering the book on GitHub Actions

Because R4DS-ES used a peer review process, I had to create the infrastructure to render the book on GitHub Actions. This is because it kept things simple, and the website could be updated when I merged a pull request.

I modified the file .github/workflows/render-book.yml to the following:

on:
  push:
    branches: traduccion

name: Render and Publish

jobs:
  build-deploy:
    runs-on: ubuntu-latest
    steps:
      - name: Check out repository
        uses: actions/checkout@v3
      
      - name: Setup R
        uses: r-lib/actions/setup-r@v2

      - name: Setup pandoc
        uses: r-lib/actions/setup-pandoc@v2

      - name: Get R version
        id: get-r-version
        run: echo "version=$(Rscript -e 'cat(as.character(getRversion()))')" >> $GITHUB_OUTPUT
        shell: bash  

      - name: Cache R packages
        uses: actions/cache@v3
        with:
          path: ${{ env.R_LIBS_USER }}
          key: ${{ runner.os }}-${{ steps.get-r-version.outputs.version }}-3-

      - name: Install pak
        run: |
          install.packages("pak", repos = "https://r-lib.github.io/p/pak/dev/")
        shell: Rscript {0}

      - name: Install dependencies
        run: |
          pak::local_install_dev_deps()
        shell: Rscript {0}

      - name: Install knitr 1.42 to avoid bug in 1.43
        run: |
          devtools::install_version("knitr", version = "1.42")
        shell: Rscript {0}

      - name: Set up Quarto
        uses: quarto-dev/quarto-actions/setup@v2
        with:
          tinytex: false
      
      - name: Publish to GitHub Pages (and render)
        uses: quarto-dev/quarto-actions/publish@v2
        with:
          target: gh-pages
        env:
          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }} # this secret is always available for github actions

Note the line that downgrades knitr to version 1.42 for the reasons explained above. The file is adapted from Quarto Actions.

< section id="optional-steps" class="level1">

Optional steps

Because I moved a bit from the configurations of the original book, I renamed the metapackage r4ds we use to install the book’s dependencies to r4dses and I did some spring cleaning in the DESCRIPTION, .Rbuildignore, and I removed some files that were not needed anymore. The result was to obtain zero errors, warnings, and notes when running devtools::check().

It is important to mention that the book repository is, in part an R package, and the command devtools::install_github("cienciadedatos/r4ds") shall install the book’s dependencies by reading the DESCRIPTION file on it.

< section id="books-repository" class="level1">

Book’s repository

The main changes consisted in moving everything, adding rmarkdown as a dependency, and downgrading knitr.

You can see the final result here.