deploy-data.yml

name: Validate/Deploy Data

on:
  push: {branches: [master]}

jobs:
  validate-then-deploy:
    name: Validate, then deploy data
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - uses: jdx/mise-action@v2

      - run: npm ci

      - run: mise run validate

      # Having validated the data, we now prepare a bundle.  This script creates
      # files in a docs/ directory.
      - run: mise run bundle-data

      # Notes: actions/checkout@v2 no longer fetches entire history nor enters
      # detached HEAD state.  We really just need to check out an (orphaned)
      # branch and then add, commit, and push the appropriate directory.
      - name: Commit the data
        run: |
          git config --global user.name "github-actions[bot]"
          git config --global user.email "41898282+github-actions[bot]@users.noreply.github.com"
          git checkout --orphan "gh-pages-$GITHUB_SHA"
          git --work-tree=docs add .
          git commit -m "Automated data deployment at $(date -Is)"
          git show --stat HEAD

      # If the previous commit successfully happened, download the latest state
      # of the remote branch gh-pages.
      - name: Deploy the data
        run: |
          git fetch --prune --unshallow origin gh-pages
          git push --force-with-lease origin "gh-pages-$GITHUB_SHA:gh-pages"