pkgdown.yml

---
name: Pkgdown

on:
  workflow_call:

concurrency:
  group: pkgdown-${{ github.event.pull_request.number || github.ref }}
  cancel-in-progress: true

env:
  TEMPLATE_REF: main
  TEMPLATE_REPO: boehringer-ingelheim/dv.templates

jobs:
  pkgdown:
    name: ${{ vars.CI_IMAGE }}
    runs-on: ubuntu-22.04
    container:
      image: ${{ vars.CI_IMAGE }}    
    env:
      GITHUB_PAT: ${{ secrets.GITHUB_TOKEN }}

    steps:
      - name: Checkout project ⬇️
        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 #@v4

      - name: Checkout templates repo ⬇️  
        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 #@v4
        with:
          ref: ${{env.TEMPLATE_REF}}
          repository: ${{env.TEMPLATE_REPO}}
          sparse-checkout: |
            scripts/
            inst/validation
          path: templates

      - name: Overwrite quality control scripts with those on dv.templates
        shell: bash
        run: |
          TEMPLATE_DIR="templates/inst/validation"
          REPO_DIR="inst/validation"

          if test ! -d "${REPO_DIR}"; then
            echo "### Quality Control documentation is not present"
            echo "### Include quality control documentation or skip it by creating following file 'inst/validation/skip_qc'"
            exit 1
          fi

          if test -f "${REPO_DIR}/skip_qc"; then
            echo "skip_qc file found skipping qc documentation"
            exit 0
          fi
          
          # Conserve specs.R file
          cp "${REPO_DIR}/specs.R" "${TEMPLATE_DIR}/specs.R"

          set +e  # Disable immediate exit in conflicting parts on error for diff command
          DIFF_OUTPUT=$(diff -qr "${TEMPLATE_DIR}" "${REPO_DIR}")
          set -e  # Enable immediate exit on error after diff command

          echo "# Quality control files" | tee -a "${GITHUB_STEP_SUMMARY}"
          
          if [ -n "${DIFF_OUTPUT}" ]; then
            echo "Warning: The contents of the directories are different!" | tee -a "${GITHUB_STEP_SUMMARY}"
            echo "Modified files:" | tee -a "${GITHUB_STEP_SUMMARY}"
            echo "$DIFF_OUTPUT" | tee -a "${GITHUB_STEP_SUMMARY}"
          else
            echo "The directories are identical." | tee -a "${GITHUB_STEP_SUMMARY}"
          fi

          rm -rf "${REPO_DIR}"
          mv "${TEMPLATE_DIR}" "${REPO_DIR}"

      - name: Install package dependencies
        uses: boehringer-ingelheim/dv.templates/.github/actions/dependencies@main

      - name: Get validation results
        uses: actions/download-artifact@v4
        with:
          name: val_results
          path: inst/validation/results

      - name: Build site
        run: |
          # Pay the hadley tax: https://github.com/r-lib/roxygen2/issues/905
          desc <- read.dcf("DESCRIPTION")
          desc <- desc[, setdiff(colnames(desc), "RoxygenNote"), drop=FALSE]
          write.dcf(x = desc, file = "DESCRIPTION")
          # Generate docs
          rox_messages <- capture.output(roxygen2::roxygenize(clean = TRUE), type = "message")
          message(paste(rox_messages, collapse = '\n'))
          error_mask <- startsWith(rox_messages, '✖')
          if(any(error_mask)) {
            stop(paste("Errors while rendering roxygen documentation:", paste(rox_messages[error_mask], collapse = "\n")))
          }
          # Render site
          pkgdown::build_site(new_process = FALSE) # otherwise errors are not logged
        shell: Rscript {0}

      - name: Remove Quality Control vignette from Articles dropdown
        run: |
          # This removal is highly dependent on the pkgdown template, but we've tried to make the regexp patterns
          # fairly restrictive to prevent deletions from happening if the template changes. If that were to happen, 
          # the QC vignette would reappear under Articles, which is annoying, but not the end of the world.

          # For reference, see https://regex101.com/ and https://stackoverflow.com/a/56927135
          PATTERN_QC_ITEM='^\s*<a class="dropdown-item" href=".*/qc.html">.*</a>\s*$'
          PATTERN_DROPDOWN_WITH_ITEMS='(?s)<li class=\"(active )?nav-item dropdown\">.*?(?:<\/li>(*SKIP)(*FAIL)|.)dropdown-articles.*?dropdown-item.*?<\/li>'
          PATTERN_DROPDOWN_WITHOUT_ITEMS='(?s)<li class=\"(active )?nav-item dropdown\">.*?(?:<\/li>(*SKIP)(*FAIL)|.)dropdown-articles.*?<\/li>'

          readarray -t FILES < <(find docs -name "*.html" -print0 | xargs -0 grep -l "$PATTERN_QC_ITEM")
          for f in "${FILES[@]}"; do
            sed -i "\#$PATTERN_QC_ITEM#d" "$f"
            grep -Pzoq "$PATTERN_DROPDOWN_WITH_ITEMS" "$f"
            if [ $? -ne 0 ]; then
              perl -0 -i -p -e "s/$PATTERN_DROPDOWN_WITHOUT_ITEMS//gs" "$f"
            fi
          done
        shell: bash

      - name: Check URLs 🌐
        run: |
          lychee . --format markdown --verbose --no-progress --exclude "https://boehringer-ingelheim.github.io*" >> $GITHUB_STEP_SUMMARY
        shell: bash

      - name: Checkout gh-pages branch ⬇️
        if: github.ref_name == 'main'
        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 #@v4
        with:
          path: gh-pages
          ref: gh-pages

      - name: Deploy to GitHub pages
        if: github.ref_name == 'main'
        uses: JamesIves/github-pages-deploy-action@65b5dfd4f5bcd3a7403bbc2959c144256167464e # v4.5.0 but immutable
        with:
          clean: true
          branch: gh-pages
          folder: docs