Fresh start - excluded large ROM JSON files

tools/rom-inventory.py

@@ -0,0 +1,156 @@
+#!/usr/bin/env python3
+"""ROM Library Inventory Scanner"""
+
+import os
+import json
+import hashlib
+from pathlib import Path
+from datetime import datetime
+from collections import defaultdict
+
+ROOT = r"R:\"
+OUTPUT_DIR = r"C:\Users\admin\.openclaw\workspace\rom-inventory"
+
+def ensure_output_dir():
+    os.makedirs(OUTPUT_DIR, exist_ok=True)
+
+def get_file_hash(filepath, block_size=65536):
+    try:
+        hasher = hashlib.md5()
+        with open(filepath, "rb") as f:
+            for block in iter(lambda: f.read(block_size), b""):
+                hasher.update(block)
+        return hasher.hexdigest()
+    except:
+        return None
+
+def get_file_size(filepath):
+    try:
+        return os.path.getsize(filepath)
+    except:
+        return 0
+
+ROM_EXTENSIONS = {
+    ".sfc", ".nes", ".z64", ".v64", ".n64", ".gba", ".gbc", ".gb",
+    ".nds", ".3ds", ".iso", ".bin", ".cue", ".cso", ".pbp",
+    ".rom", ".md", ".sms", ".gg", ".sg", ".col", ".vec",
+    ".a26", ".a52", ".a78", ".lnx", ".j64", ".jag",
+    ".pce", ".ngp", ".ngc", ".ws", ".wsc", ".vb",
+    ".zip", ".7z", ".rar"
+}
+
+def is_rom_file(filename):
+    return Path(filename).suffix.lower() in ROM_EXTENSIONS
+
+def scan_console(console_path, console_name):
+    roms = []
+    total_size = 0
+    file_count = 0
+    
+    console_path = Path(console_path)
+    if not console_path.exists():
+        return {"console": console_name, "error": "Path not found"}
+    
+    for root, dirs, files in os.walk(console_path):
+        for file in files:
+            if is_rom_file(file):
+                filepath = Path(root) / file
+                size = get_file_size(filepath)
+                total_size += size
+                file_count += 1
+                
+                file_hash = None
+                if size < 100 * 1024 * 1024:
+                    file_hash = get_file_hash(filepath)
+                
+                roms.append({
+                    "name": file,
+                    "path": str(filepath),
+                    "size": size,
+                    "hash": file_hash
+                })
+    
+    return {
+        "console": console_name,
+        "file_count": file_count,
+        "total_size_bytes": total_size,
+        "total_size_gb": round(total_size / (1024**3), 2),
+        "roms": roms[:100]
+    }
+
+def scan_all():
+    ensure_output_dir()
+    
+    organized_path = Path(ROOT) / "Rom Sets (Organized)"
+    
+    results = {
+        "scan_date": datetime.now().isoformat(),
+        "root": ROOT,
+        "organized": {},
+        "duplicates": [],
+        "summary": {}
+    }
+    
+    print("Scanning Rom Sets (Organized)...")
+    for manufacturer in organized_path.iterdir():
+        if manufacturer.is_dir():
+            print(f"  {manufacturer.name}...")
+            results["organized"][manufacturer.name] = {}
+            for console in manufacturer.iterdir():
+                if console.is_dir():
+                    console_data = scan_console(console, console.name)
+                    results["organized"][manufacturer.name][console.name] = console_data
+                    print(f"    {console.name}: {console_data.get("file_count", 0)} files")
+    
+    print("
+Finding duplicates...")
+    all_hashes = defaultdict(list)
+    for manufacturer, consoles in results["organized"].items():
+        for console, data in consoles.items():
+            if "roms" in data:
+                for rom in data["roms"]:
+                    if rom.get("hash"):
+                        all_hashes[rom["hash"]].append({
+                            "name": rom["name"],
+                            "console": console,
+                            "path": rom["path"]
+                        })
+    
+    results["duplicates"] = [
+        {"hash": h, "files": files} 
+        for h, files in all_hashes.items() 
+        if len(files) > 1
+    ]
+    
+    total_files = sum(
+        data.get("file_count", 0) 
+        for consoles in results["organized"].values() 
+        for data in consoles.values()
+    )
+    total_size = sum(
+        data.get("total_size_bytes", 0) 
+        for consoles in results["organized"].values() 
+        for data in consoles.values()
+    )
+    
+    results["summary"] = {
+        "total_roms": total_files,
+        "total_size_gb": round(total_size / (1024**3), 2),
+        "duplicate_count": len(results["duplicates"])
+    }
+    
+    output_file = Path(OUTPUT_DIR) / f"rom-inventory-{datetime.now().strftime('%Y%m%d-%H%M%S')}.json"
+    with open(output_file, "w") as f:
+        json.dump(results, f, indent=2)
+    
+    print(f"
+=== INVENTORY COMPLETE ===")
+    print(f"Total ROMs: {total_files:,}")
+    print(f"Total Size: {results["summary"]["total_size_gb"]:.2f} GB")
+    print(f"Duplicates Found: {len(results["duplicates"])}")
+    print(f"Output: {output_file}")
+    
+    return results
+
+if __name__ == "__main__":
+    scan_all()