import os
import hashlib
import json
from datetime import datetime
from pathlib import Path

ROM_ROOT = 'R:\\'
OUTPUT_DIR = Path('C:/Users/admin/.openclaw/workspace/rom-inventory')
HASH_LIMIT = 100 * 1024 * 1024

def get_file_hash(filepath, limit_bytes=None):
    hasher = hashlib.md5()
    try:
        with open(filepath, 'rb') as f:
            if limit_bytes:
                chunk = f.read(limit_bytes)
                hasher.update(chunk)
            else:
                for chunk in iter(lambda: f.read(8192), b''):
                    hasher.update(chunk)
        return hasher.hexdigest()
    except Exception as e:
        return None

def scan_roms():
    print(f'Starting full ROM scan at {datetime.now()}')
    print(f'Root: {ROM_ROOT}')
    
    all_files = []
    hash_map = {}
    stats = {
        'total_files': 0,
        'total_size': 0,
        'hashed_files': 0,
        'skipped_large': 0,
        'errors': 0,
        'by_system': {},
        'by_manufacturer': {}
    }
    
    organized_path = Path(ROM_ROOT) / 'Rom Sets (Organized)'
    if organized_path.exists():
        for manufacturer_dir in organized_path.iterdir():
            if manufacturer_dir.is_dir():
                manufacturer = manufacturer_dir.name
                stats['by_manufacturer'][manufacturer] = {'files': 0, 'size': 0}
                
                for system_dir in manufacturer_dir.iterdir():
                    if system_dir.is_dir():
                        system = system_dir.name
                        if system not in stats['by_system']:
                            stats['by_system'][system] = {'files': 0, 'size': 0}
                        
                        for root, dirs, files in os.walk(system_dir):
                            for f in files:
                                filepath = Path(root) / f
                                try:
                                    size = filepath.stat().st_size
                                    all_files.append({
                                        'path': str(filepath),
                                        'name': f,
                                        'size': size,
                                        'system': system,
                                        'manufacturer': manufacturer
                                    })
                                    stats['total_files'] += 1
                                    stats['total_size'] += size
                                    stats['by_manufacturer'][manufacturer]['files'] += 1
                                    stats['by_manufacturer'][manufacturer]['size'] += size
                                    stats['by_system'][system]['files'] += 1
                                    stats['by_system'][system]['size'] += size
                                    
                                    if size <= HASH_LIMIT:
                                        file_hash = get_file_hash(filepath)
                                        if file_hash:
                                            all_files[-1]['hash'] = file_hash
                                            stats['hashed_files'] += 1
                                            if file_hash not in hash_map:
                                                hash_map[file_hash] = []
                                            hash_map[file_hash].append(str(filepath))
                                    else:
                                        stats['skipped_large'] += 1
                                except Exception as e:
                                    stats['errors'] += 1
    
    duplicates = {h: files for h, files in hash_map.items() if len(files) > 1}
    
    print(f'\nScan complete at {datetime.now()}')
    print(f'Total files: {stats["total_files"]:,}')
    print(f'Total size: {stats["total_size"] / (1024**3):.2f} GB')
    print(f'Hashed: {stats["hashed_files"]:,}')
    print(f'Skipped (large): {stats["skipped_large"]:,}')
    print(f'Duplicates found: {len(duplicates)}')
    
    OUTPUT_DIR.mkdir(parents=True, exist_ok=True)
    
    output = {
        'scan_date': datetime.now().isoformat(),
        'stats': stats,
        'duplicates': duplicates,
        'files': all_files[:10000]
    }
    
    with open(OUTPUT_DIR / 'rom-full-scan.json', 'w') as f:
        json.dump(output, f, indent=2)
    
    print(f'\nResults saved to {OUTPUT_DIR / "rom-full-scan.json"}')
    return stats

if __name__ == '__main__':
    scan_roms()