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-v1.5.0"

          # 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 using actual v1.3.1 JSON structure
            ERROR_COUNT=$(jq -r '.summary.error_count // 0' link-check-results.json)
            WARNING_COUNT=$(jq -r '.summary.warning_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)

            # Reclassify file-not-found warnings as errors
            # link-checker classifies missing local files as warnings (no HTTP status code),
            # but these represent genuinely broken internal links and should fail CI.
            FILE_NOT_FOUND_COUNT=$(jq '[.warnings[] | select(.error | test("Cannot find file"))] | length' link-check-results.json 2>/dev/null || echo 0)
            if [[ $FILE_NOT_FOUND_COUNT -gt 0 ]]; then
              echo "⚠️ Found $FILE_NOT_FOUND_COUNT missing local file(s) — reclassifying as errors"
              # Move file-not-found entries from warnings to errors
              jq '
                .errors += [.warnings[] | select(.error | test("Cannot find file")) | .severity = "error"]
                | .warnings = [.warnings[] | select(.error | test("Cannot find file") | not)]
                | .summary.error_count = (.errors | length)
                | .summary.warning_count = (.warnings | length)
              ' link-check-results.json > link-check-results-fixed.json
              mv link-check-results-fixed.json link-check-results.json

              ERROR_COUNT=$(jq -r '.summary.error_count // 0' link-check-results.json)
              WARNING_COUNT=$(jq -r '.summary.warning_count // 0' link-check-results.json)
            fi

            echo "error-count=$ERROR_COUNT" >> $GITHUB_OUTPUT
            echo "warning-count=$WARNING_COUNT" >> $GITHUB_OUTPUT
            echo "total-count=$TOTAL_COUNT" >> $GITHUB_OUTPUT
            echo "success-rate=$SUCCESS_RATE" >> $GITHUB_OUTPUT

            if [[ $ERROR_COUNT -gt 0 ]]; then
              echo "❌ Found $ERROR_COUNT broken links out of $TOTAL_COUNT total links"
              echo "check-result=failed" >> $GITHUB_OUTPUT
            else
              echo "✅ All $TOTAL_COUNT links are valid ($WARNING_COUNT warnings)"
              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 != ''
        env:
          FILE_COUNT: ${{ steps.mapping.outputs.file-count }}
          TOTAL_COUNT: ${{ steps.link-check.outputs.total-count }}
          ERROR_COUNT: ${{ steps.link-check.outputs.error-count }}
          WARNING_COUNT: ${{ steps.link-check.outputs.warning-count }}
          SUCCESS_RATE: ${{ steps.link-check.outputs.success-rate }}
          CHECK_RESULT: ${{ steps.link-check.outputs.check-result }}
        run: |
          if [[ -f link-check-results.json ]]; then
            # Generate summary header
            echo "## Link Check Results" >> $GITHUB_STEP_SUMMARY
            echo "" >> $GITHUB_STEP_SUMMARY
            echo "| Metric | Value |" >> $GITHUB_STEP_SUMMARY
            echo "|--------|-------|" >> $GITHUB_STEP_SUMMARY
            echo "| Files Checked | ${FILE_COUNT} |" >> $GITHUB_STEP_SUMMARY
            echo "| Total Links | ${TOTAL_COUNT} |" >> $GITHUB_STEP_SUMMARY
            echo "| Errors | ${ERROR_COUNT} |" >> $GITHUB_STEP_SUMMARY
            echo "| Warnings | ${WARNING_COUNT} |" >> $GITHUB_STEP_SUMMARY
            echo "| Success Rate | ${SUCCESS_RATE}% |" >> $GITHUB_STEP_SUMMARY
            echo "" >> $GITHUB_STEP_SUMMARY

            # Report broken links (errors) with annotations
            if [[ "${CHECK_RESULT}" == "failed" ]]; then
              echo "### Broken Links" >> $GITHUB_STEP_SUMMARY
              echo "" >> $GITHUB_STEP_SUMMARY
              echo "| Severity | Source File | Broken URL | Error |" >> $GITHUB_STEP_SUMMARY
              echo "|----------|------------|------------|-------|" >> $GITHUB_STEP_SUMMARY

              # Process errors (these fail CI)
              jq -c '.errors[]?' link-check-results.json 2>/dev/null | while read -r entry; do
                URL=$(echo "$entry" | jq -r '.url // "unknown"')
                ERROR=$(echo "$entry" | jq -r '.error // "Unknown error"')
                FILE=$(echo "$entry" | jq -r '.file // "unknown"')
                LINE=$(echo "$entry" | jq -r '.line // empty')

                # Map public path to content path for annotations
                CONTENT_FILE=$(echo "$FILE" | sed 's|.*/public/|content/|' | sed 's|/index\.html$|/_index.md|')

                # Create GitHub annotation
                if [[ -n "$LINE" && "$LINE" != "null" ]]; then
                  echo "::error file=${CONTENT_FILE},line=${LINE}::Broken link: ${URL} (${ERROR})"
                else
                  echo "::error file=${CONTENT_FILE}::Broken link: ${URL} (${ERROR})"
                fi

                # Add row to summary table
                SAFE_URL=$(echo "$URL" | sed 's/|/\\|/g')
                SAFE_ERROR=$(echo "$ERROR" | sed 's/|/\\|/g' | cut -c1-80)
                echo "| 🔴 error | \`${CONTENT_FILE}\` | ${SAFE_URL} | ${SAFE_ERROR} |" >> $GITHUB_STEP_SUMMARY
              done

              echo "" >> $GITHUB_STEP_SUMMARY
              echo "---" >> $GITHUB_STEP_SUMMARY
              echo "❌ **Link check failed** — fix the broken links listed above before merging." >> $GITHUB_STEP_SUMMARY
            else
              echo "✅ **All links are valid**" >> $GITHUB_STEP_SUMMARY
            fi

            # Report warnings (don't fail CI, but useful context)
            WARNING_ARRAY_LEN=$(jq '.warnings | length' link-check-results.json 2>/dev/null || echo 0)
            if [[ "$WARNING_ARRAY_LEN" -gt 0 ]]; then
              echo "" >> $GITHUB_STEP_SUMMARY
              echo "<details>" >> $GITHUB_STEP_SUMMARY
              echo "<summary>⚠️ ${WARNING_ARRAY_LEN} warning(s) (do not fail CI)</summary>" >> $GITHUB_STEP_SUMMARY
              echo "" >> $GITHUB_STEP_SUMMARY
              echo "| Source File | URL | Issue |" >> $GITHUB_STEP_SUMMARY
              echo "|------------|-----|-------|" >> $GITHUB_STEP_SUMMARY

              jq -c '.warnings[]?' link-check-results.json 2>/dev/null | head -20 | while read -r entry; do
                URL=$(echo "$entry" | jq -r '.url // "unknown"')
                ERROR=$(echo "$entry" | jq -r '.error // "Unknown"')
                FILE=$(echo "$entry" | jq -r '.file // "unknown"')
                CONTENT_FILE=$(echo "$FILE" | sed 's|.*/public/|content/|' | sed 's|/index\.html$|/_index.md|')
                SAFE_URL=$(echo "$URL" | sed 's/|/\\|/g')
                SAFE_ERROR=$(echo "$ERROR" | sed 's/|/\\|/g' | cut -c1-80)
                echo "| \`${CONTENT_FILE}\` | ${SAFE_URL} | ${SAFE_ERROR} |" >> $GITHUB_STEP_SUMMARY
              done

              if [[ "$WARNING_ARRAY_LEN" -gt 20 ]]; then
                echo "" >> $GITHUB_STEP_SUMMARY
                echo "_Showing first 20 of ${WARNING_ARRAY_LEN} warnings. Download the artifact for full results._" >> $GITHUB_STEP_SUMMARY
              fi
              echo "" >> $GITHUB_STEP_SUMMARY
              echo "</details>" >> $GITHUB_STEP_SUMMARY
            fi

            # Add helpful tips
            echo "" >> $GITHUB_STEP_SUMMARY
            echo "<details>" >> $GITHUB_STEP_SUMMARY
            echo "<summary>💡 Troubleshooting Tips</summary>" >> $GITHUB_STEP_SUMMARY
            echo "" >> $GITHUB_STEP_SUMMARY
            echo "- **404 errors**: The linked page doesn't exist. Check for typos or update the link." >> $GITHUB_STEP_SUMMARY
            echo "- **Relative links**: Use relative paths starting with \`/\` for internal links." >> $GITHUB_STEP_SUMMARY
            echo "- **Anchors**: Ensure heading anchors match the linked fragment exactly." >> $GITHUB_STEP_SUMMARY
            echo "- **Warnings**: External sites may be temporarily unavailable — these don't fail CI." >> $GITHUB_STEP_SUMMARY
            echo "" >> $GITHUB_STEP_SUMMARY
            echo "</details>" >> $GITHUB_STEP_SUMMARY
          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