Stirling-PDF/.github/workflows/file_hash_generation.yml

name: Generate Template Hashes

on:
  push:
    branches:
      - main
      - hashs
    paths:
      - 'src/main/resources/templates/**'
      - 'src/main/resources/static/**'
  workflow_dispatch:  # Allow manual triggering

jobs:
  generate-hash:
    runs-on: ubuntu-latest
    
    steps:
      - name: Checkout code
        uses: actions/checkout@v3
      
      - name: Calculate template hashes
        id: hash
        run: |
          # Create a script to calculate individual file hashes
          cat > calculate-hashes.sh << 'EOF'
          #!/bin/bash
          set -e
          
          # Directories to hash
          TEMPLATE_DIR="src/main/resources/templates"
          STATIC_DIR="src/main/resources/static"
          OUTPUT_FILE="src/main/resources/reference-hash.json"
          
          # Create output directory if it doesn't exist
          mkdir -p $(dirname "$OUTPUT_FILE")
          
          # Text file extensions that need normalization
          TEXT_EXTENSIONS=("html" "htm" "css" "js" "txt" "md" "xml" "json" "csv" "properties")
          
          # Function to check if a file is a text file
          is_text_file() {
            ext="${1##*.}"
            for text_ext in "${TEXT_EXTENSIONS[@]}"; do
              if [ "$ext" = "$text_ext" ]; then
                return 0
              fi
            done
            return 1
          }
          
          # Function to calculate normalized hash for text files
          calculate_text_hash() {
            # Normalize line endings to LF and calculate CRC32
            tr -d '\r' < "$1" | cksum | awk '{print $1}'
          }
          
          # Function to calculate hash for binary files
          calculate_binary_hash() {
            cksum "$1" | awk '{print $1}'
          }
          
          # Start JSON
          echo "{" > "$OUTPUT_FILE"
          
          # Find all files and calculate CRC32 hash
          FIRST=true
          find "$TEMPLATE_DIR" "$STATIC_DIR" -type f | sort | while read file; do
            # Get relative path from src/main/resources
            REL_PATH=$(echo "$file" | sed 's|^src/main/resources/||')
            
            # Calculate hash based on file type
            if is_text_file "$file"; then
              HASH=$(calculate_text_hash "$file")
            else
              HASH=$(calculate_binary_hash "$file")
            fi
            
            # Add to JSON
            if [ "$FIRST" = true ]; then
              FIRST=false
            else
              echo "," >> "$OUTPUT_FILE"
            fi
            
            echo "  \"$REL_PATH\": \"$HASH\"" >> "$OUTPUT_FILE"
          done
          
          # End JSON
          echo "}" >> "$OUTPUT_FILE"
          
          echo "Generated hashes for $(grep -c ":" "$OUTPUT_FILE") files"
          EOF
          
          chmod +x calculate-hashes.sh
          ./calculate-hashes.sh
      
      - name: Commit and push if changed
        run: |
          git config --local user.email "github-actions[bot]@users.noreply.github.com"
          git config --local user.name "GitHub Actions"
          
          git add src/main/resources/reference-hash.json
          
          # Only commit if there are changes
          if git diff --staged --quiet; then
            echo "No changes to commit"
          else
            git commit -m "Update template reference hashes [skip ci]"
            git push
          fi