ci-pr.yml

# This action runs when a pull request to main is created or updated
# (but not when it is merged into main).

name: CI on pull request create or update
on:
  # Using "pull_request" is not sufficient to allow commits to PR branches in
  # forks. Since the workflow only accesses the xlsx and turtle files, but does
  # not run any code, it is safe to use pull_request_target. More info:
  # https://securitylab.github.com/resources/github-actions-preventing-pwn-requests/
  pull_request_target:
    branches:
      - main
    types: [opened, reopened, synchronize]
  workflow_dispatch:

env:
  FORCE_COLOR: "1"  # Make tool output pretty.
  PIP_DISABLE_PIP_VERSION_CHECK: "1"
  PIP_PROGRESS_BAR: "off"
  LOGLEVEL: "DEBUG"
  RUN_URL: $GITHUB_SERVER_URL/$GITHUB_REPOSITORY/actions/runs/$GITHUB_RUN_ID

jobs:
  build_vocab:
    name: Merge & build vocabulary
    # Run on all pull request types except merge which has its own workflow.
    if: ${{ !github.event.pull_request.merged }}
    runs-on: ubuntu-latest
    timeout-minutes: 10

    permissions:
      # Give the default GITHUB_TOKEN write permission to commit to the PR branch.
      contents: write

    steps:
      - name: Checkout PR branch into cwd
        uses: actions/checkout@1af3b93b6815bc44a9784bd300feb67ff0d1eeb3
        with:
          # Checkout the fork/head-repository to push changes to the fork.
          # Without this the base repository will be checked out and committed to.
          repository: ${{ github.event.pull_request.head.repo.full_name }}

          # Checkout the branch made in the fork.
          ref: ${{ github.head_ref }}

      - name: Checkout main branch into directory _main_branch/
        uses: actions/checkout@1af3b93b6815bc44a9784bd300feb67ff0d1eeb3
        with:
          ref: main
          path: _main_branch
          # This sparse-checkout includes all toplevel files and everything below vocabularies/
          sparse-checkout: |
            vocabularies/
          fetch-depth: 1

      - uses: actions/setup-python@83679a892e2d95755f2dac6acb0bfd1e9ac5d548
        with:
          python-version: '3.13'

      - name: Install dependencies
        run: |
          python -VV
          python -m pip install --upgrade pip setuptools wheel

          # install tagged version
          python -m pip install git+https://github.com/nfdi4cat/voc4cat-tool.git@v0.10.0
          # python -m pip install git+https://github.com/nfdi4cat/voc4cat-tool.git@main

          # install custom pylode 2.x  (adds sorted collections,
          #    uses URIs as fragment IDs and is git-installable)
          python -m pip install git+https://github.com/dalito/pyLODE.git@nfdi4cat-2.x

          # Example for installing from private repo:
          # python -m pip install git+https://anon:${{ SECRETS.READ_PAT_VOC4CAT }}@github.com/dalito/voc4cat-tool.git@main

      - name: Set dynamic environment variables.
        run: |
          echo "RUN_DATE=$(date +'%Y-%m-%dT%H%M')" >> $GITHUB_ENV

      - name: Run voc4cat (pre-convert checks)
        run: |
          voc4cat --version
          # check inbox file names
          voc4cat check --config _main_branch/idranges.toml --logfile outbox/voc4cat.log --ci-pre inbox-excel-vocabs/ _main_branch/vocabularies
          # check xlsx file(s). If the check fails, write annotated file to outbox.
          voc4cat check --config _main_branch/idranges.toml --logfile outbox/voc4cat.log --outdir outbox inbox-excel-vocabs/

      - name: Run voc4cat (convert)
        run: |
          # convert file(s) from xlsx in inbox to turtle in outbox
          voc4cat convert --config _main_branch/idranges.toml --logfile outbox/voc4cat.log --outdir outbox inbox-excel-vocabs/
          if [ ! -f outbox/*.ttl ]; then
            echo "No ttl file in outbox. Building joined vocabulary ttl-file from individual ttl-files in vocabulary."
            voc4cat transform --join --logfile outbox/voc4cat.log --outdir outbox/ vocabularies/
          fi

      - name: Run voc4cat (post-convert checks)
        run: |
          # check all ttl file(s) in outbox
          voc4cat check --config _main_branch/idranges.toml --logfile outbox/voc4cat.log outbox/
          # check if vocabulary changes are allowed
          voc4cat check --config _main_branch/idranges.toml --logfile outbox/voc4cat.log --ci-post _main_branch/vocabularies outbox/

      - name: Run voc4cat (build HTML documentation)
        run: |
          voc4cat docs --logfile outbox/voc4cat.log --force outbox/

      - name: Split vocabulary and merge into vocabulary dir (in PR branch)
        run: |
          voc4cat transform --split --inplace --logfile outbox/voc4cat.log --outdir outbox_new_voc outbox/
          voc4cat-merge --logfile outbox/voc4cat.log  outbox_new_voc/ vocabularies/
          # copy xlsx to outbox so that the xlsx file is part of the artifact
          find inbox-excel-vocabs -name '*.xlsx' -exec cp {} -t outbox/ \;
          git status

      - name: Run voc4cat (re-build updated Excel file and joined turtle vocabulary file)
        # Passing the config is important to make use of the prefixes therein.
        run: |
          mkdir -p outbox/updated-xlsx-ttl
          voc4cat convert --config idranges.toml --logfile outbox/voc4cat.log --template templates/voc4cat_template_043.xlsx outbox/updated-xlsx-ttl
          voc4cat transform --config idranges.toml --logfile outbox/voc4cat.log -O outbox/updated-xlsx-ttl --join vocabularies/

      - name: Store artifacts
        if: ${{ always() }}
        uses: actions/upload-artifact@330a01c490aca151604b8cf639adc76d48f6c5d4
        with:
          name: voc4cat-outbox_${{ env.RUN_DATE }}_run${{ GITHUB.RUN_ID }}
          path: outbox/

      - name: Add/stage vocabulary changes
        # Pinning of action managed by dependabot
        uses: stefanzweifel/git-auto-commit-action@28e16e81777b558cc906c8750092100bbb34c5e3
        with:
          # glob pattern of files which should be added to the commit
          file_pattern: 'vocabularies/\*.ttl'

      - name: Remove Excel vocabulary files from inbox
        run: |
          git rm inbox-excel-vocabs/\*.xlsx || echo "No xlsx files to remove."
          git status

      - name: Commit vocabulary changes & xlsx cleanup on behalf of PR-author
        # Pinning of action managed by dependabot
        uses: stefanzweifel/git-auto-commit-action@28e16e81777b558cc906c8750092100bbb34c5e3
        with:
          commit_user_name: voc4cat-CI-bot
          commit_message: 'CI: vocabulary update in ${{ GITHUB.SERVER_URL }}/${{ GITHUB.REPOSITORY }}/actions/runs/${{ GITHUB.RUN_ID }}'