check-html-metadata.yaml

name: Validate HTML Metadata on PRs

on:
  pull_request:
    types: [opened, synchronize, reopened]

jobs:
  validate-metadata:
    runs-on: ubuntu-latest
    steps:
      # Checkout the PR branch
      - name: Checkout PR branch
        uses: actions/checkout@v2
        with:
          ref: ${{ github.event.pull_request.head.ref }}
          repository: ${{ github.event.pull_request.head.repo.full_name }}

      - name: Validate all HTML files in the PR branch
        id: validate
        run: |
          ERROR_MESSAGE=""
          VALIDATION_PASSED=false
          
          # Find all HTML files in the public/submissions/ directory
          HTML_FILES=$(find public/submissions -name "*.html" || true)

          # If no HTML files are found, skip validation
          if [ -z "$HTML_FILES" ]; then
            echo "No HTML files found in public/submissions/. Skipping validation."
            exit 0
          else
            echo "Validating the following HTML files:"
            echo "$HTML_FILES"
          fi

          # Loop through all HTML files and validate them
          for file in $HTML_FILES; do
            echo "Processing file: $file"

            # Check for <meta name="author"> tag
            AUTHOR=$(grep -Poz '(?s)<meta[^>]*name=["'\'']author["'\''][^>]*>' "$file" || true)
            if [ -z "$AUTHOR" ]; then
              ERROR_MESSAGE+="File $file is missing <meta name='author'> tag.\n"
            fi

            # Check for <title> tag
            TITLE=$(grep -Poz '(?s)<title[^>]*>.*?</title>' "$file" || true)
            if [ -z "$TITLE" ]; then
              ERROR_MESSAGE+="File $file is missing <title> tag.\n"
            fi
          done

          # Pass the error message to the next step
          echo "ERROR_MESSAGE=$ERROR_MESSAGE" >> $GITHUB_ENV

          # If there are no errors, mark validation as passed
          if [ -z "$ERROR_MESSAGE" ]; then
            echo "VALIDATION_PASSED=true" >> $GITHUB_ENV
          else
            echo "VALIDATION_PASSED=false" >> $GITHUB_ENV
            exit 1  # Fail the step if validation fails
          fi

      # Always delete the existing validation comment (whether it passed or failed)
      - name: Delete existing validation comment
        uses: actions/github-script@v5
        with:
          script: |
            const { data: comments } = await github.rest.issues.listComments({
              owner: context.repo.owner,
              repo: context.repo.repo,
              issue_number: context.issue.number
            });

            const validationComment = comments.find(comment => comment.user.login === 'github-actions[bot]' && (comment.body.includes('HTML Metadata Validation Failed') || comment.body.includes('HTML Validation Passed')));

            if (validationComment) {
              console.log("Deleting previous comment...");
              await github.rest.issues.deleteComment({
                owner: context.repo.owner,
                repo: context.repo.repo,
                comment_id: validationComment.id
              });
            } else {
              console.log("No existing comment found.");
            }

      # Post a new validation comment (either pass or fail)
      - name: Post validation comment
        uses: actions/github-script@v5
        env:
          PR_AUTHOR: ${{ github.event.pull_request.user.login }}
          ERROR_MESSAGE: ${{ env.ERROR_MESSAGE }}
          VALIDATION_PASSED: ${{ env.VALIDATION_PASSED }}
        with:
          script: |
            const issueNumber = context.payload.pull_request.number;
            const prAuthor = `@${process.env.PR_AUTHOR}`;
            
            let message;
            if (process.env.VALIDATION_PASSED === 'true') {
              message = `**HTML Validation Passed**\n\nGreat job, ${prAuthor}! All required metadata is present.`;
            } else {
              message = `**HTML Metadata Validation Failed:**\n\n${prAuthor}, the following issues were found:\n\n${process.env.ERROR_MESSAGE}`;
            }

            await github.rest.issues.createComment({
              issue_number: issueNumber,
              owner: context.repo.owner,
              repo: context.repo.repo,
              body: message
            })