transformers/.github/workflows/pr-repo-consistency-bot.yml

name: PR Repo. Consistency Bot

on:
  issue_comment:
    types:
      - created
    branches-ignore:
      - main
concurrency:
  group: ${{ github.workflow }}-${{ github.event.issue.number }}-${{ startsWith(github.event.comment.body, '@bot /repo') || startsWith(github.event.comment.body, '@bot /style') }}
  cancel-in-progress: true
permissions:
  contents: read


jobs:
  get-pr-number:
    name: Get PR number
    if: ${{ github.event.issue.state == 'open' && contains(fromJSON('["ydshieh", "ArthurZucker", "zucchini-nlp", "molbap", "gante", "LysandreJik", "Cyrilvallez", "Rocketknight1", "SunMarc", "eustlb", "MekkCyber", "vasqu", "ivarflakstad", "stevhliu", "ebezzam", "remi-or", "itazap", "3outeille", "IlyasMoutawwakil", "tarekziade"]'), github.actor) && (startsWith(github.event.comment.body, '@bot /repo') || startsWith(github.event.comment.body, '@bot /style')) }}
    uses: ./.github/workflows/get-pr-number.yml

  get-pr-info:
    name: Get PR commit SHA
    needs: get-pr-number
    if: ${{ needs.get-pr-number.outputs.PR_NUMBER != ''}}
    uses: ./.github/workflows/get-pr-info.yml
    with:
      pr_number: ${{ needs.get-pr-number.outputs.PR_NUMBER }}

  check-timestamps:
    name: Check timestamps (security check)
    runs-on: ubuntu-22.04
    needs: get-pr-info
    outputs:
      VERIFIED_PR_HEAD_SHA: ${{ needs.get-pr-info.outputs.PR_HEAD_SHA }}
    steps:
      - name: Verify `merge_commit` timestamp is older than the issue comment timestamp
        env:
          COMMENT_DATE: ${{ github.event.comment.created_at }}
          PR_MERGE_COMMIT_TIMESTAMP: ${{ needs.get-pr-info.outputs.PR_MERGE_COMMIT_TIMESTAMP }}
        run: |
            COMMENT_TIMESTAMP=$(date -d "${COMMENT_DATE}" +"%s")
            echo "COMMENT_DATE: $COMMENT_DATE"
            echo "COMMENT_TIMESTAMP: $COMMENT_TIMESTAMP"
            if [ $COMMENT_TIMESTAMP -le $PR_MERGE_COMMIT_TIMESTAMP ]; then
              echo "Last commit on the pull request is newer than the issue comment triggering this run! Abort!";
              exit -1;
            fi

  init_comment_with_url:
    name: Init Comment on PR
    runs-on: ubuntu-22.04
    needs: [get-pr-number, check-timestamps]
    outputs:
       comment_id: ${{ steps.init_comment.outputs.comment_id }}
    permissions:
      pull-requests: write
    steps:
      - name: Delete existing bot comment if it exists
        env:
          PR_NUMBER: ${{ needs.get-pr-number.outputs.PR_NUMBER }}
        uses: actions/github-script@d7906e4ad0b1822421a7e6a35d5ca353c962f410 # v6.4.1
        with:
          script: |
            const PR_NUMBER = parseInt(process.env.PR_NUMBER, 10);
            
            // Get all comments on the PR
            const { data: comments } = await github.rest.issues.listComments({
              owner: context.repo.owner,
              repo: context.repo.repo,
              issue_number: PR_NUMBER
            });
            
            // Find existing bot comments that start with "Repo. Consistency" or "Style fix"
            const existingComments = comments.filter(comment => 
              comment.user.login === 'github-actions[bot]' && 
              (comment.body.startsWith('Repo. Consistency') || comment.body.startsWith('Style fix'))
            );

            if (existingComments.length > 0) {
              // Get the most recent comment
              const mostRecentComment = existingComments
                .sort((a, b) => new Date(b.created_at) - new Date(a.created_at))[0];
              
              console.log(`Deleting most recent comment #${mostRecentComment.id}`);
              await github.rest.issues.deleteComment({
                owner: context.repo.owner,
                repo: context.repo.repo,
                comment_id: mostRecentComment.id
              });
            }

      - name: Comment on PR with workflow run link
        id: init_comment
        env:
          PR_NUMBER: ${{ needs.get-pr-number.outputs.PR_NUMBER }}
          COMMENT_BODY: ${{ github.event.comment.body }}
        uses: actions/github-script@d7906e4ad0b1822421a7e6a35d5ca353c962f410 # v6.4.1
        with:
          script: |
            const PR_NUMBER = parseInt(process.env.PR_NUMBER, 10);
            const COMMENT_BODY = process.env.COMMENT_BODY;
            const runUrl = `${process.env.GITHUB_SERVER_URL}/${process.env.GITHUB_REPOSITORY}/actions/runs/${process.env.GITHUB_RUN_ID}`
            
            // Determine which command was used
            const isStyleFix = COMMENT_BODY.startsWith('@bot /style');
            const messagePrefix = isStyleFix ? 'Style fix' : 'Repo. Consistency fix';

            const { data: botComment } = await github.rest.issues.createComment({
              owner: context.repo.owner,
              repo: context.repo.repo,
              issue_number: PR_NUMBER,
              body: `${messagePrefix} is beginning .... [View the workflow run here](${runUrl}).`
            });
            core.setOutput('comment_id', botComment.id);

  run-repo-consistency-checks:
    runs-on: ubuntu-22.04
    needs: [get-pr-info, check-timestamps, init_comment_with_url]
    permissions:
      contents: read
      pull-requests: read
    outputs:
      changes_detected: ${{ steps.run_repo_checks.outputs.changes_detected || steps.run_style_checks.outputs.changes_detected }}
      util_scripts_modified: ${{ steps.check_util_scripts.outputs.util_scripts_modified }}
    steps:
      # Checkout the trusted base repository (main branch) - this is safe
      - name: Checkout base repository
        uses: actions/checkout@34e114876b0b11c390a56381ad16ebd13914f8d5 # v4.3.1
        with:
          ref: main
          persist-credentials: false

      - name: Set up Python
        uses: actions/setup-python@7f4fc3e22c37d6ff65e88745f38bd3157c663f7c # v4.9.1
        with:
          python-version: "3.10"

      - name: Install dependencies from trusted main branch
        run: |
          python -m pip install --upgrade pip
          pip install -e ".[quality]"
          pip install --no-cache-dir --upgrade 'torch' 'torchaudio' 'torchvision' --index-url https://download.pytorch.org/whl/cpu

      - name: Fetch and checkout PR code manually
        env:
          PR_HEAD_REPO_FULL_NAME: ${{ needs.get-pr-info.outputs.PR_HEAD_REPO_FULL_NAME }}
          PR_HEAD_REF: ${{ needs.get-pr-info.outputs.PR_HEAD_REF }}
          PR_HEAD_SHA: ${{ needs.check-timestamps.outputs.VERIFIED_PR_HEAD_SHA }}
        run: |
          # Create separate directory for PR code
          mkdir -p pr-repo
          cd pr-repo
          
          # Initialize git and fetch with full history
          git init
          git remote add pr-origin "https://github.com/${PR_HEAD_REPO_FULL_NAME}.git"
          git fetch pr-origin "${PR_HEAD_REF}"
          git checkout "${PR_HEAD_SHA}"

          # Also fetch main branch from upstream for comparison (required by `check_modular_conversion`)
          git remote add upstream https://github.com/${{ github.repository }}.git
          git fetch upstream main:main

      - name: Check if util scripts are modified in PR
        id: check_util_scripts
        uses: actions/github-script@d7906e4ad0b1822421a7e6a35d5ca353c962f410 # v6.4.1
        with:
          script: |
            // Re-fetch the PR file list from the API rather than using get-pr-info's PR_FILES
            // output, to avoid template injection and E2BIG issues (see pr_slow_ci_suggestion.yml).
            const UTIL_SCRIPTS = new Set([
              'setup.py',
              'utils/custom_init_isort.py',
              'utils/sort_auto_mappings.py',
              'utils/check_doc_toc.py',
              'utils/check_copies.py',
              'utils/check_modular_conversion.py',
              'utils/check_dummies.py',
              'utils/check_pipeline_typing.py',
              'utils/check_doctest_list.py',
              'utils/check_docstrings.py',
              'utils/add_dates.py',
            ]);

            const files = await github.paginate(github.rest.pulls.listFiles, {
              owner: context.repo.owner,
              repo: context.repo.repo,
              pull_number: context.payload.issue.number,
            });

            const modified = files.some(f => UTIL_SCRIPTS.has(f.filename));
            core.setOutput('util_scripts_modified', modified ? 'true' : 'false');

      - name: Install editable transformers from PR branch
        if: steps.check_util_scripts.outputs.util_scripts_modified != 'true'
        run: |
          cd pr-repo
          pip install -e .

      - name: Run repo consistency checks with trusted script
        id: run_repo_checks
        if: steps.check_util_scripts.outputs.util_scripts_modified != 'true' && startsWith(github.event.comment.body, '@bot /repo')
        run: |
          # Continue on errors (like Makefile's - prefix)
          set +e
          
          # Run commands in PR directory (with the copied trusted scripts)
          cd pr-repo
          
          # Run style commands
          ruff check examples tests src utils scripts benchmark benchmark_v2 setup.py conftest.py --fix
          ruff format examples tests src utils scripts benchmark benchmark_v2 setup.py conftest.py
          python utils/custom_init_isort.py
          python utils/sort_auto_mappings.py
          
          # Run fix-repo commands
          python setup.py deps_table_update
          python utils/check_doc_toc.py --fix_and_overwrite
          python utils/check_copies.py --fix_and_overwrite
          python utils/check_modular_conversion.py --fix_and_overwrite
          python utils/check_dummies.py --fix_and_overwrite
          python utils/check_pipeline_typing.py --fix_and_overwrite
          python utils/check_doctest_list.py --fix_and_overwrite
          python utils/check_docstrings.py --fix_and_overwrite
          python utils/add_dates.py
          
          # Check if there are changes
          if [ -n "$(git status --porcelain)" ]; then
            echo "changes_detected=true" >> $GITHUB_OUTPUT
          else
            echo "changes_detected=false" >> $GITHUB_OUTPUT
          fi

      - name: Run style checks with trusted script
        id: run_style_checks
        if: steps.check_util_scripts.outputs.util_scripts_modified != 'true' && startsWith(github.event.comment.body, '@bot /style')
        run: |
          # Continue on errors (like Makefile's - prefix)
          set +e
          
          # Run commands in PR directory (with the copied trusted scripts)
          cd pr-repo
          
          # Run style commands
          ruff check examples tests src utils scripts benchmark benchmark_v2 setup.py conftest.py --fix
          ruff format examples tests src utils scripts benchmark benchmark_v2 setup.py conftest.py
          python utils/sort_auto_mappings.py
          
          # Run fix-repo commands
          python setup.py deps_table_update
          python utils/check_doc_toc.py --fix_and_overwrite
          python utils/check_docstrings.py --fix_and_overwrite
          
          # Check if there are changes
          if [ -n "$(git status --porcelain)" ]; then
            echo "changes_detected=true" >> $GITHUB_OUTPUT
          else
            echo "changes_detected=false" >> $GITHUB_OUTPUT
          fi

      - name: Save modified files
        if: steps.check_util_scripts.outputs.util_scripts_modified != 'true' && (steps.run_repo_checks.outputs.changes_detected == 'true' || steps.run_style_checks.outputs.changes_detected == 'true')
        run: |
          cd pr-repo
          mkdir -p ../artifact-staging
          git diff --name-only > ../artifact-staging/modified-files.txt
          # Copy each modified file
          while IFS= read -r file; do
            mkdir -p "../artifact-staging/pr-repo/$(dirname "$file")"
            cp "$file" "../artifact-staging/pr-repo/$file"
          done < ../artifact-staging/modified-files.txt

      - name: Upload modified files
        if: steps.check_util_scripts.outputs.util_scripts_modified != 'true' && (steps.run_repo_checks.outputs.changes_detected == 'true' || steps.run_style_checks.outputs.changes_detected == 'true')
        uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # v4.6.2
        with:
          name: modified-files
          path: artifact-staging/

  commit-and-comment:
    runs-on: ubuntu-22.04
    needs: [get-pr-number, get-pr-info, check-timestamps, init_comment_with_url, run-repo-consistency-checks]
    if: always()
    permissions:
      pull-requests: write
      contents: write
    steps:
      - name: Download modified files
        if: needs.run-repo-consistency-checks.outputs.changes_detected == 'true'
        uses: actions/download-artifact@d3f86a106a0bac45b974a628896c90dbdf5c8093 # v4.3.0
        with:
          name: modified-files

      - name: Push changes to fork using git
        if: needs.run-repo-consistency-checks.outputs.changes_detected == 'true'
        env:
          PR_HEAD_REF: ${{ needs.get-pr-info.outputs.PR_HEAD_REF }}
          PR_HEAD_SHA: ${{ needs.check-timestamps.outputs.VERIFIED_PR_HEAD_SHA }}
          PR_HEAD_REPO_FULL_NAME: ${{ needs.get-pr-info.outputs.PR_HEAD_REPO_FULL_NAME }}
          GITHUB_TOKEN: ${{ secrets.HF_STYLE_BOT_ACTION }}
        run: |
          # Initialize a fresh git repository for pushing
          mkdir push-repo
          cd push-repo
          git init
          
          # Configure git
          git config user.name "github-actions[bot]"
          git config user.email "github-actions[bot]@users.noreply.github.com"
          
          # Add fork as remote with token
          git remote add origin "https://x-access-token:${GITHUB_TOKEN}@github.com/${PR_HEAD_REPO_FULL_NAME}.git"

          # Fetch only the specific branch
          git fetch origin "${PR_HEAD_REF}"

          # Checkout the branch
          git checkout -b "${PR_HEAD_REF}" "origin/${PR_HEAD_REF}"
          
          # Verify we're on the correct SHA
          current_sha=$(git rev-parse HEAD)
          if [ "$current_sha" != "$PR_HEAD_SHA" ]; then
            echo "❌ Error: Branch has been updated since workflow started"
            echo "Expected SHA: $PR_HEAD_SHA"
            echo "Current SHA: $current_sha"
            exit 1
          fi
          
          # Copy modified files from artifact
          echo "Copying modified files..."
          while IFS= read -r file; do
            if [ -n "$file" ]; then
              echo "  - $file"
              mkdir -p "$(dirname "$file")"
              cp "../pr-repo/$file" "$file"
            fi
          done < ../modified-files.txt
          
          # Check if there are changes
          if [ -n "$(git status --porcelain)" ]; then
            git add .
            git commit -m "Apply repo consistency fixes"
            git push origin "HEAD:${PR_HEAD_REF}"
            echo "✅ Changes pushed successfully"
          else
            echo "No changes to commit"
          fi

      - name: Prepare final comment message
        id: prepare_final_comment
        if: needs.init_comment_with_url.result == 'success'
        env:
          CHANGES_DETECTED: ${{ needs.run-repo-consistency-checks.outputs.changes_detected }}
          UTIL_SCRIPTS_MODIFIED: ${{ needs.run-repo-consistency-checks.outputs.util_scripts_modified }}
          COMMENT_BODY: ${{ github.event.comment.body }}
        run: |
          # Determine which command was used
          if [[ "$COMMENT_BODY" == "@bot /style"* ]]; then
            MESSAGE_PREFIX="Style fix"
          else
            MESSAGE_PREFIX="Repo. Consistency"
          fi

          if [ "$UTIL_SCRIPTS_MODIFIED" = 'true' ]; then
            echo "final_comment=${MESSAGE_PREFIX}: \`setup.py\` or some script files under the \`utils/\` directory are modified in this PR. Please run style/repo. checks/fixes locally." >> $GITHUB_OUTPUT
          elif [ "$CHANGES_DETECTED" = 'true' ]; then
            echo "final_comment=${MESSAGE_PREFIX} bot fixed some files and pushed the changes." >> $GITHUB_OUTPUT
          else
            echo "final_comment=${MESSAGE_PREFIX} fix runs successfully without any file modified." >> $GITHUB_OUTPUT
          fi

      - name: Comment on PR
        if: needs.init_comment_with_url.result == 'success'
        uses: actions/github-script@d7906e4ad0b1822421a7e6a35d5ca353c962f410 # v6.4.1
        env:
          PR_NUMBER: ${{ needs.get-pr-number.outputs.PR_NUMBER }}
          COMMENT_ID: ${{ needs.init_comment_with_url.outputs.comment_id }}
          FINAL_COMMENT: ${{ steps.prepare_final_comment.outputs.final_comment }}
        with:
          script: |
            const pr_number = parseInt(process.env.PR_NUMBER, 10);
            const comment_id = parseInt(process.env.COMMENT_ID, 10);
            const body = process.env.FINAL_COMMENT;
            await github.rest.issues.updateComment({
              owner: context.repo.owner,
              repo: context.repo.repo,
              comment_id,
              body,
            });