name: Check Links in README

on:
  workflow_dispatch:

jobs:
  check-links:
    runs-on: ubuntu-latest
    
    steps:
      - name: Checkout repository
        uses: actions/checkout@v4

      - name: Check all links in markdown files
        run: |
          #!/bin/bash
          set -o pipefail
          
          echo ""
          echo "━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━"
          echo "🔍 Checking ALL Links in All Markdown Files"
          echo "━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━"
          echo ""
          
          # Find all markdown files
          find . -name "*.md" -type f | sort > all_markdown_files.txt
          total_files=$(wc -l < all_markdown_files.txt)
          echo "📊 Found $total_files markdown files to check"
          echo ""
          
          has_errors=0
          
          # Process each markdown file
          while IFS= read -r md_file; do
            echo "━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━"
            echo "📄 Checking: $md_file"
            echo "━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━"
            
            file_has_errors=0
            
            # Extract ALL links from the markdown file
            # Match both []() and direct URLs - using Perl regex for better handling
            {
              grep -oP '\]\(([^\)]+)\)' "$md_file" 2>/dev/null | sed 's/](\(.*\))/\1/' || true
              grep -oP '!\[[^\]]*\]\(([^\)]+)\)' "$md_file" 2>/dev/null | sed 's/!\[.*\](\(.*\))/\1/' || true
            } > /tmp/links_$$.txt
            
            link_count=$(wc -l < /tmp/links_$$.txt 2>/dev/null || echo "0")
            
            if [ $link_count -eq 0 ]; then
              echo "  ℹ️  No links found in this file"
              echo ""
              continue
            fi
            
            echo "  📊 Found $link_count links to check"
            echo ""
            
            # Check each link
            while IFS= read -r link; do
              [ -z "$link" ] && continue
              
              # Decode URL-encoded characters
              decoded_link=$(echo "$link" | sed 's/%20/ /g' | sed 's/%23/#/g' | sed 's/%28/(/g' | sed 's/%29/)/g' | sed 's/%E2%80%A2/•/g' | sed 's/%E1%9A%96/ᚖ/g' | sed 's/%E3%80%8A/《/g' | sed 's/%E3%80%8B/》/g' | sed 's/%E3%80%8E/『/g' | sed 's/%E3%80%8F/』/g' | sed 's/%E2%9B%94/⛔/g' | sed 's/%E2%9C%A8/✨/g' | sed 's/%7B/{/g' | sed 's/%7D/}/g' | sed 's/%2B/+/g' | sed 's/%E3%83%86/テ/g' | sed 's/%E3%83%B3/ン/g' | sed 's/%E3%83%8D/ネ/g' | sed 's/%E3%82%B9/ス/g' | sed 's/%E3%82%A4/イ/g' | sed 's/%E3%83%BB/・/g' | sed 's/%E3%83%95/フ/g' | sed 's/%E3%83%AA/リ/g' | sed 's/%E3%83%BC/ー/g' | sed 's/%E3%83%8A/ナ/g' | sed 's/%5B/[/g' | sed 's/%5D/]/g')
              
              # Check if it's an external URL
              if [[ "$decoded_link" =~ ^https?:// ]]; then
                # URL-encode the link for curl (convert spaces and special chars)
                # Use the original encoded link, not the decoded one
                encoded_url="$link"
                
                # Check external URL with curl (with User-Agent header for better compatibility)
                http_code=$(curl -s -o /dev/null -w "%{http_code}" --max-time 10 \
                  -H "User-Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36" \
                  "$encoded_url" 2>/dev/null || echo "000")
                
                # Accept 2xx and 3xx status codes as valid
                if [[ "$http_code" =~ ^[23][0-9][0-9]$ ]]; then
                  echo "  ✅ $decoded_link"
                else
                  echo "  ❌ $decoded_link (HTTP $http_code)"
                  file_has_errors=1
                  has_errors=1
                fi
              else
                # Local file - remove leading slash if present
                if [[ "$decoded_link" =~ ^/ ]]; then
                  file_path="${decoded_link#/}"
                else
                  # Relative path - resolve from markdown file location
                  md_dir=$(dirname "$md_file")
                  file_path="$md_dir/$decoded_link"
                fi
                
                # Check if file exists (using test -f which handles spaces better)
                if [ -f "$file_path" ]; then
                  echo "  ✅ $decoded_link"
                else
                  echo "  ❌ $decoded_link (file not found at: $file_path)"
                  file_has_errors=1
                  has_errors=1
                fi
              fi
            done < /tmp/links_$$.txt
            
            rm -f /tmp/links_$$.txt
            
            if [ $file_has_errors -eq 0 ]; then
              echo ""
              echo "  ✅ All links valid in this file"
            else
              echo ""
              echo "  ❌ Some links are broken in this file"
            fi
            echo ""
            
          done < all_markdown_files.txt
          
          echo "━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━"
          
          if [ $has_errors -eq 0 ]; then
            echo "✅ FINAL RESULT: All links are valid across all markdown files!"
            echo "━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━"
            exit 0
          else
            echo "❌ FINAL RESULT: Some links are broken. Please review the output above."
            echo "━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━"
            exit 1
          fi