structure-validation.yml

name: 📋 Repository Structure Validation

on:
  pull_request:
    branches: [ main ]
  push:
    branches: [ main ]

jobs:
  validate-structure:
    runs-on: ubuntu-latest

    steps:
    - name: 📥 Checkout code
      uses: actions/checkout@v4
      with:
        fetch-depth: 0

    - name: 🔍 Get changed files
      id: changed-files
      uses: tj-actions/changed-files@v40

    - name: 📁 Validate directory structure
      run: |
        echo "📁 Validating repository structure..."

        # Define expected structure for each language
        declare -A language_structure=(
          ["C"]="algorithms data_structures dynamic_programming"
          ["CPP"]="algorithms data_structures dynamic_programming object_oriented_programming"
          ["Java"]="algorithms data_structures dynamic_programming projects"
          ["Python"]="algorithms data_structures dynamic_programming projects"
        )

        # Define expected algorithm subdirectories
        algorithm_dirs="searching sorting graph_algorithms arrays strings mathematical"
        data_structure_dirs="trees stacks queues graphs heaps linked_lists"

        structure_issues=()

        # Check existing languages
        for lang in C CPP Java Python; do
          if [ -d "$lang" ]; then
            echo "🔍 Checking $lang structure..."

            # Check main directories
            for dir in ${language_structure[$lang]}; do
              if [ ! -d "$lang/$dir" ]; then
                structure_issues+=("Missing directory: $lang/$dir/")
              fi
            done

            # Check algorithms subdirectories
            if [ -d "$lang/algorithms" ]; then
              for alg_dir in $algorithm_dirs; do
                if [ -d "$lang/algorithms/$alg_dir" ]; then
                  echo "✅ Found $lang/algorithms/$alg_dir/"
                fi
              done
            fi

            # Check data structures subdirectories
            if [ -d "$lang/data_structures" ]; then
              for ds_dir in $data_structure_dirs; do
                if [ -d "$lang/data_structures/$ds_dir" ]; then
                  echo "✅ Found $lang/data_structures/$ds_dir/"
                fi
              done
            fi

            # Check for README
            if [ ! -f "$lang/README.md" ]; then
              structure_issues+=("Missing $lang/README.md")
            else
              echo "✅ Found $lang/README.md"
            fi
          fi
        done

        # Check for new languages
        new_languages=($(find . -maxdepth 1 -type d -name "[A-Z]*" ! -name "C" ! -name "CPP" ! -name "Java" ! -name "Python" | sed 's|./||'))

        for lang in "${new_languages[@]}"; do
          echo "🌍 Checking new language: $lang"

          # Check if README exists for new language
          if [ ! -f "$lang/README.md" ]; then
            structure_issues+=("New language $lang/ is missing README.md with setup instructions")
          else
            echo "✅ Found $lang/README.md"
          fi

          # Check basic structure
          required_dirs=("algorithms" "data_structures" "dynamic_programming")
          for req_dir in "${required_dirs[@]}"; do
            if [ ! -d "$lang/$req_dir" ]; then
              echo "💡 Consider creating $lang/$req_dir/ directory"
            else
              echo "✅ Found $lang/$req_dir/"
            fi
          done

          # Validate file types in new language directory
          lang_files=($(find "$lang" -type f -name "*.*" 2>/dev/null))
          for file in "${lang_files[@]}"; do
            ext="${file##*.}"
            case "$ext" in
              c|cpp|java|py|js|ts|go|rs|kt|swift|php|rb|cs|dart|scala|md)
                echo "✅ Valid file: $file"
                ;;
              *)
                echo "⚠️ Unknown file type: $file"
                ;;
            esac
          done
        done

        # Report issues
        if [ ${#structure_issues[@]} -gt 0 ]; then
          echo "⚠️ Structure issues found:"
          printf '%s\n' "${structure_issues[@]}"
        else
          echo "✅ Repository structure looks good!"
        fi

    - name: 🏗️ Validate algorithm organization
      run: |
        echo "🏗️ Validating algorithm organization..."

        # Check for algorithms in correct directories
        misplaced_algorithms=()

        # Search for common algorithm patterns
        search_patterns=(
          "search:searching"
          "sort:sorting"
          "graph:graph_algorithms"
          "tree:trees"
          "stack:stacks"
          "queue:queues"
          "heap:heaps"
          "link:linked_lists"
          "dynamic:dynamic_programming"
        )

        for pattern in "${search_patterns[@]}"; do
          keyword="${pattern%%:*}"
          expected_dir="${pattern##*:}"

          # Find files containing the keyword
          files_with_keyword=($(find . -name "*.c" -o -name "*.cpp" -o -name "*.java" -o -name "*.py" -o -name "*.js" -o -name "*.go" -o -name "*.rs" | xargs grep -l -i "$keyword" 2>/dev/null || true))

          for file in "${files_with_keyword[@]}"; do
            # Check if file is in expected directory
            if [[ "$file" != *"/$expected_dir/"* ]] && [[ "$file" != *"$expected_dir"* ]]; then
              # Skip if it's just a comment or documentation
              if grep -q -i "$keyword" "$file" && ! grep -q -E "//.*$keyword|#.*$keyword|/\*.*$keyword" "$file"; then
                misplaced_algorithms+=("$file: Contains '$keyword' but not in $expected_dir directory")
              fi
            fi
          done
        done

        if [ ${#misplaced_algorithms[@]} -gt 0 ]; then
          echo "💡 Algorithm organization suggestions:"
          printf '%s\n' "${misplaced_algorithms[@]}"
        else
          echo "✅ Algorithm organization looks good!"
        fi

    - name: 📝 Check documentation completeness
      run: |
        echo "📝 Checking documentation completeness..."

        missing_docs=()

        # Check main repository documentation
        required_docs=("README.md" "CONTRIBUTING.md" "HALL_OF_FAME.md")
        for doc in "${required_docs[@]}"; do
          if [ ! -f "$doc" ]; then
            missing_docs+=("Missing main documentation: $doc")
          else
            echo "✅ Found $doc"
          fi
        done

        # Check language-specific documentation
        for lang_dir in */; do
          lang_name="${lang_dir%/}"

          # Skip hidden directories and non-language directories
          if [[ "$lang_name" =~ ^\. ]] || [[ "$lang_name" =~ ^(node_modules|\.git|\.github)$ ]]; then
            continue
          fi

          # Check for language README
          if [ -d "$lang_dir" ] && [ ! -f "$lang_dir/README.md" ]; then
            # Only report if directory contains code files
            code_files=($(find "$lang_dir" -name "*.c" -o -name "*.cpp" -o -name "*.java" -o -name "*.py" -o -name "*.js" -o -name "*.go" -o -name "*.rs" 2>/dev/null))
            if [ ${#code_files[@]} -gt 0 ]; then
              missing_docs+=("Language directory $lang_dir is missing README.md")
            fi
          elif [ -f "$lang_dir/README.md" ]; then
            echo "✅ Found $lang_dir/README.md"
          fi
        done

        if [ ${#missing_docs[@]} -gt 0 ]; then
          echo "📝 Documentation suggestions:"
          printf '%s\n' "${missing_docs[@]}"
        else
          echo "✅ Documentation is complete!"
        fi

    - name: 🎯 Generate structure report
      run: |
        echo "🎯 Repository Structure Report"
        echo "=============================="
        echo ""

        echo "📊 Language Distribution:"
        for dir in */; do
          if [[ -d "$dir" ]] && [[ ! "$dir" =~ ^\. ]] && [[ ! "$dir" =~ ^(node_modules|\.git|\.github)$ ]]; then
            lang_name="${dir%/}"
            file_count=$(find "$dir" -name "*.c" -o -name "*.cpp" -o -name "*.java" -o -name "*.py" -o -name "*.js" -o -name "*.go" -o -name "*.rs" -o -name "*.kt" -o -name "*.swift" -o -name "*.php" -o -name "*.rb" -o -name "*.cs" -o -name "*.dart" -o -name "*.scala" 2>/dev/null | wc -l)
            echo "  📁 $lang_name: $file_count files"
          fi
        done

        echo ""
        echo "🏗️ Algorithm Categories:"
        categories=("searching" "sorting" "graph_algorithms" "arrays" "strings" "mathematical")
        for category in "${categories[@]}"; do
          count=$(find . -path "*/$category/*" -type f \( -name "*.c" -o -name "*.cpp" -o -name "*.java" -o -name "*.py" -o -name "*.js" -o -name "*.go" -o -name "*.rs" \) 2>/dev/null | wc -l)
          echo "  🔍 $category: $count implementations"
        done

        echo ""
        echo "🏗️ Data Structure Categories:"
        ds_categories=("trees" "stacks" "queues" "graphs" "heaps" "linked_lists")
        for category in "${ds_categories[@]}"; do
          count=$(find . -path "*/$category/*" -type f \( -name "*.c" -o -name "*.cpp" -o -name "*.java" -o -name "*.py" -o -name "*.js" -o -name "*.go" -o -name "*.rs" \) 2>/dev/null | wc -l)
          echo "  🏗️ $category: $count implementations"
        done

        echo ""
        echo "✅ Structure validation completed!"