createSample.yml

# This is a basic workflow to help you get started with Actions

name: Create sample files

# Controls when the action will run. 
on:

  pull_request_target:
    branches: [ data ]
    types: [closed] # synchronize is temporary for debugging

  # Allows you to run this workflow manually from the Actions tab
  # workflow_dispatch:

# A workflow run is made up of one or more jobs that can run sequentially or in parallel
jobs:
  Changes:
    runs-on: ubuntu-latest
    if: ${{ github.repository_owner == 'clarin-eric' && github.event_name == 'pull_request_target' && github.event.pull_request.merged == true }}
    outputs:
      parla_changed:  '${{ steps.detect-changes.outputs.parla_changed }}'
    steps:
      - name: Checkout ParlaMint
        uses: actions/checkout@v3
        with:
          path: ParlaMint
          fetch-depth: 2 # when called a pull_request - test is run on test merge -> full history is not needed to detect changes
      - name: Setup # initialize dependencies cache - used in next steps
        uses: ./ParlaMint/.github/actions/ParlaMintEnvSetup
      - name: ParlaMint status (Detect changed files and get list of parliaments that should be processed)
        id: detect-changes
        uses: ./ParlaMint/.github/actions/ParlaMintStatus

  # Creates sample files and commits them
  CreateSample:
    needs: [Changes]
    runs-on: ubuntu-latest
    if: ${{ needs.Changes.outputs.parla_changed != '[]' && github.event_name == 'pull_request_target' && github.event.pull_request.merged == true }}
    steps:
      - name: Checkout ParlaMint
        uses: actions/checkout@v3
        with:
          path: ParlaMint
      - name: Setup
        uses: ./ParlaMint/.github/actions/ParlaMintEnvSetup
        with:
          repo_token: ${{secrets.GITHUB_TOKEN}}

      - name: Create Parliaments Samples ${{needs.Changes.outputs.parla_changed}}
        run: |
          cd $GITHUB_WORKSPACE/ParlaMint
          for parla in $(jq -r '.[]' <<< '${{needs.Changes.outputs.parla_changed}}' ); do
            echo "::group::Processing ParlaMint-$parla"
            DIR="${{env.SAMPLE_DIR}}/$parla"
            mkdir $DIR
            echo "::notice::New sample files [$parla] TEXT"
            java -jar $GITHUB_WORKSPACE/Saxon.jar outDir=$DIR revRespPers='GitHub Action' -xsl:${{env.SAMPLE_SCRIPT}} Samples/ParlaMint-$parla/ParlaMint-$parla.xml

            echo "::notice::New sample files [$parla] ANNOTATED"
            if [ -f "Samples/ParlaMint-$parla/ParlaMint-$parla.ana.xml" ] ; then
              java -jar $GITHUB_WORKSPACE/Saxon.jar outDir=$DIR revRespPers='GitHub Action' -xsl:${{env.SAMPLE_SCRIPT}} Samples/ParlaMint-$parla/ParlaMint-$parla.ana.xml
            else
              echo "::warning::skipping annotated conversion - missing corpus root file"
            fi
            echo "::notice::Move new sample files to Samples/ParlaMint-$parla"
            ls $DIR|grep "ParlaMint-$parla[\.\_]"|xargs -I {} mv $DIR/{} Samples/ParlaMint-$parla/
            echo "::endgroup::"
          done
      - name: Remove unused data from repository
        run: |
          cd $GITHUB_WORKSPACE/ParlaMint
          shopt -s globstar
          for parla in $(jq -r '.[]' <<< '${{needs.Changes.outputs.parla_changed}}' ); do
            for file in $(echo Samples/ParlaMint-$parla/**/ParlaMint-$parla\_*) ; do
              echo "testing $file"
              xmlfile=$(echo $file|sed -E 's/(-meta\.tsv|\.conllu|\.txt|\.vert)$/.xml/;s/^Samples\/ParlaMint-[^\/]*\///')
              cat Samples/ParlaMint-$parla/ParlaMint-$parla{,.ana}.xml | grep -Fq "$xmlfile" || git rm "$file"
            done
          done
      - name: Create Derived files and Validate ${{needs.Changes.outputs.parla_changed}}
        uses: ./ParlaMint/.github/actions/ParlaMintValidate
        with:
          parlas: '${{needs.Changes.outputs.parla_changed}}'

      - name: Commit new sample files ${{needs.Changes.outputs.parla_changed}}
        if: ${{ success() }}
        run: |
          cd $GITHUB_WORKSPACE/ParlaMint
          for parla in $(jq -r '.[]' <<< '${{needs.Changes.outputs.parla_changed}}' ); do
            git add Samples/ParlaMint-$parla/ParlaMint-*.{txt,tsv,conllu,vert,xml} || echo "::warning:: $parla suppress fatal: pathspec '<FILE>' did not match any files"
            git diff --name-only Samples/ParlaMint-$parla
          done
          git status
          git commit -a -m "action: generating ParlaMint-${{needs.Changes.outputs.parla_changed}} sample files with #${{github.event.number}}"
          git push
      - name: Commit new sample root
        if: ${{ success() }}
        uses: ./ParlaMint/.github/actions/ParlaMintSample2root