pr-link-check.yml

name: Link Check PR Changes

on:
  pull_request:
    paths:
      - 'content/**/*.md'
      - 'data/**/*.yml'
      - 'layouts/**/*.html'
    types: [opened, synchronize, reopened]

jobs:
  link-check:
    name: Check links in affected files
    runs-on: ubuntu-latest
    
    steps:
      - name: Checkout repository
        uses: actions/checkout@v4
        with:
          fetch-depth: 0
          
      - name: Detect content changes
        id: detect
        run: |
          echo "🔍 Detecting changes between ${{ github.base_ref }} and ${{ github.sha }}"
          
          # For PRs, use the GitHub Files API to get changed files
          if [[ "${{ github.event_name }}" == "pull_request" ]]; then
            echo "Using GitHub API to detect PR changes..."
            curl -s -H "Authorization: token ${{ secrets.GITHUB_TOKEN }}" \
              "https://api.github.com/repos/${{ github.repository }}/pulls/${{ github.event.number }}/files" \
              | jq -r '.[].filename' > all_changed_files.txt
          else
            echo "Using git diff to detect changes..."
            git diff --name-only ${{ github.event.before }}..${{ github.sha }} > all_changed_files.txt
          fi
          
          # Filter for content markdown files
          CHANGED_FILES=$(grep '^content/.*\.md$' all_changed_files.txt || true)
          
          echo "📁 All changed files:"
          cat all_changed_files.txt
          echo ""
          echo "📝 Content markdown files:"
          echo "$CHANGED_FILES"
          
          if [[ -n "$CHANGED_FILES" ]]; then
            echo "✅ Found $(echo "$CHANGED_FILES" | wc -l) changed content file(s)"
            echo "has-changes=true" >> $GITHUB_OUTPUT
            echo "changed-content<<EOF" >> $GITHUB_OUTPUT
            echo "$CHANGED_FILES" >> $GITHUB_OUTPUT
            echo "EOF" >> $GITHUB_OUTPUT
            
            # Check if any shared content files were modified
            SHARED_CHANGES=$(echo "$CHANGED_FILES" | grep '^content/shared/' || true)
            if [[ -n "$SHARED_CHANGES" ]]; then
              echo "has-shared-content=true" >> $GITHUB_OUTPUT
              echo "🔄 Detected shared content changes: $SHARED_CHANGES"
            else
              echo "has-shared-content=false" >> $GITHUB_OUTPUT
            fi
          else
            echo "❌ No content changes detected"
            echo "has-changes=false" >> $GITHUB_OUTPUT
            echo "has-shared-content=false" >> $GITHUB_OUTPUT
          fi
          
      - name: Skip if no content changes
        if: steps.detect.outputs.has-changes == 'false'
        run: |
          echo "No content changes detected in this PR - skipping link check"
          echo "✅ **No content changes detected** - link check skipped" >> $GITHUB_STEP_SUMMARY
          
      - name: Setup Node.js
        if: steps.detect.outputs.has-changes == 'true'
        uses: actions/setup-node@v4
        with:
          node-version: '20'
          cache: 'yarn'
          
      - name: Install dependencies
        if: steps.detect.outputs.has-changes == 'true'
        run: yarn install --frozen-lockfile
        
      - name: Build Hugo site
        if: steps.detect.outputs.has-changes == 'true'
        run: npx hugo --minify
        
      - name: Download link-checker binary
        if: steps.detect.outputs.has-changes == 'true'
        run: |
          echo "Downloading link-checker binary from docs-v2 releases..."
          
          # Download from docs-v2's own releases (always accessible)
          curl -L -H "Accept: application/vnd.github+json" \
            -H "Authorization: Bearer ${{ secrets.GITHUB_TOKEN }}" \
            -o link-checker-info.json \
            "https://api.github.com/repos/influxdata/docs-v2/releases/tags/link-checker-v1.2.3"
          
          # Extract download URL for linux binary
          DOWNLOAD_URL=$(jq -r '.assets[] | select(.name | test("link-checker.*linux")) | .url' link-checker-info.json)
          
          if [[ "$DOWNLOAD_URL" == "null" || -z "$DOWNLOAD_URL" ]]; then
            echo "❌ No linux binary found in release"
            echo "Available assets:"
            jq -r '.assets[].name' link-checker-info.json
            exit 1
          fi
          
          echo "📥 Downloading: $DOWNLOAD_URL"
          curl -L -H "Accept: application/octet-stream" \
            -H "Authorization: Bearer ${{ secrets.GITHUB_TOKEN }}" \
            -o link-checker "$DOWNLOAD_URL"
          
          chmod +x link-checker
          ./link-checker --version
        
      - name: Verify link checker config exists
        if: steps.detect.outputs.has-changes == 'true'
        run: |
          if [[ ! -f .ci/link-checker/production.lycherc.toml ]]; then
            echo "❌ Configuration file .ci/link-checker/production.lycherc.toml not found"
            echo "Please copy production.lycherc.toml from docs-tooling/link-checker/"
            exit 1
          fi
          echo "✅ Using configuration: .ci/link-checker/production.lycherc.toml"
        
      - name: Map changed content to public files
        if: steps.detect.outputs.has-changes == 'true'
        id: mapping
        run: |
          echo "Mapping changed content files to public HTML files..."
          
          # Create temporary file with changed content files
          echo "${{ steps.detect.outputs.changed-content }}" > changed-files.txt
          
          # Map content files to public files
          PUBLIC_FILES=$(cat changed-files.txt | xargs -r ./link-checker map --existing-only)
          
          if [[ -n "$PUBLIC_FILES" ]]; then
            echo "Found affected public files:"
            echo "$PUBLIC_FILES"
            echo "public-files<<EOF" >> $GITHUB_OUTPUT
            echo "$PUBLIC_FILES" >> $GITHUB_OUTPUT
            echo "EOF" >> $GITHUB_OUTPUT
            
            # Count files for summary
            FILE_COUNT=$(echo "$PUBLIC_FILES" | wc -l)
            echo "file-count=$FILE_COUNT" >> $GITHUB_OUTPUT
          else
            echo "No public files found to check"
            echo "public-files=" >> $GITHUB_OUTPUT
            echo "file-count=0" >> $GITHUB_OUTPUT
          fi
        
      - name: Run link checker
        if: steps.detect.outputs.has-changes == 'true' && steps.mapping.outputs.public-files != ''
        id: link-check
        run: |
          echo "Checking links in ${{ steps.mapping.outputs.file-count }} affected files..."
          
          # Create temporary file with public files list
          echo "${{ steps.mapping.outputs.public-files }}" > public-files.txt
          
          # Run link checker with detailed JSON output
          set +e  # Don't fail immediately on error
          
          cat public-files.txt | xargs -r ./link-checker check \
            --config .ci/link-checker/production.lycherc.toml \
            --format json \
            --output link-check-results.json
          
          EXIT_CODE=$?
          
          if [[ -f link-check-results.json ]]; then
            # Parse results
            BROKEN_COUNT=$(jq -r '.summary.broken_count // 0' link-check-results.json)
            TOTAL_COUNT=$(jq -r '.summary.total_checked // 0' link-check-results.json)
            SUCCESS_RATE=$(jq -r '.summary.success_rate // 0' link-check-results.json)
            
            echo "broken-count=$BROKEN_COUNT" >> $GITHUB_OUTPUT
            echo "total-count=$TOTAL_COUNT" >> $GITHUB_OUTPUT
            echo "success-rate=$SUCCESS_RATE" >> $GITHUB_OUTPUT
            
            if [[ $BROKEN_COUNT -gt 0 ]]; then
              echo "❌ Found $BROKEN_COUNT broken links out of $TOTAL_COUNT total links"
              echo "check-result=failed" >> $GITHUB_OUTPUT
            else
              echo "✅ All $TOTAL_COUNT links are valid"
              echo "check-result=passed" >> $GITHUB_OUTPUT
            fi
          else
            echo "❌ Link check failed to generate results"
            echo "check-result=error" >> $GITHUB_OUTPUT
          fi
          
          exit $EXIT_CODE
          
      - name: Process and report results
        if: always() && steps.detect.outputs.has-changes == 'true' && steps.mapping.outputs.public-files != ''
        run: |
          if [[ -f link-check-results.json ]]; then
            # Create detailed error annotations for broken links
            if [[ "${{ steps.link-check.outputs.check-result }}" == "failed" ]]; then
              echo "Creating error annotations for broken links..."
              
              jq -r '.broken_links[]? | 
                "::error file=\(.file // "unknown"),line=\(.line // 1)::Broken link: \(.url) - \(.error // "Unknown error")"' \
                link-check-results.json || true
            fi
            
            # Generate summary comment
            cat >> $GITHUB_STEP_SUMMARY << 'EOF'
          ## Link Check Results
          
          **Files Checked:** ${{ steps.mapping.outputs.file-count }}  
          **Total Links:** ${{ steps.link-check.outputs.total-count }}  
          **Broken Links:** ${{ steps.link-check.outputs.broken-count }}  
          **Success Rate:** ${{ steps.link-check.outputs.success-rate }}%
          
          EOF
          
            if [[ "${{ steps.link-check.outputs.check-result }}" == "failed" ]]; then
              echo "❌ **Link check failed** - see annotations above for details" >> $GITHUB_STEP_SUMMARY
            else
              echo "✅ **All links are valid**" >> $GITHUB_STEP_SUMMARY
            fi
          else
            echo "⚠️ **Link check could not complete** - no results file generated" >> $GITHUB_STEP_SUMMARY
          fi
          
      - name: Upload detailed results
        if: always() && steps.detect.outputs.has-changes == 'true' && steps.mapping.outputs.public-files != ''
        uses: actions/upload-artifact@v4
        with:
          name: link-check-results
          path: |
            link-check-results.json
            changed-files.txt
            public-files.txt
          retention-days: 30