import os
import hashlib
import shutil

TARGET_DIR = os.path.abspath("assets/slike/glavna_referenca")
SAMPLE_DIR = os.path.abspath("_TRASH_SAMPLES")
TRASH_DIR = os.path.abspath("_TRASH_BIN/targeted_cleanup")

def get_file_hash(filepath):
    hasher = hashlib.md5()
    try:
        with open(filepath, 'rb') as f:
            buf = f.read(65536)
            while len(buf) > 0:
                hasher.update(buf)
                buf = f.read(65536)
        return hasher.hexdigest()
    except:
        return None

def main():
    if not os.path.exists(SAMPLE_DIR):
        print("❌ Sample directory missing.")
        return
        
    if not os.path.exists(TRASH_DIR):
        os.makedirs(TRASH_DIR)

    # 1. Get hashes of samples
    print(f"🎯 analyzing samples in {SAMPLE_DIR}...")
    target_hashes = set()
    for f in os.listdir(SAMPLE_DIR):
        if f.startswith("."): continue
        h = get_file_hash(os.path.join(SAMPLE_DIR, f))
        if h:
            target_hashes.add(h)
    
    print(f"🎯 Loaded {len(target_hashes)} unique target signatures.")

    # 2. Scan and destroy
    print(f"🔍 Scanning {TARGET_DIR} for matches...")
    count = 0
    matches = 0
    
    for root, dirs, files in os.walk(TARGET_DIR):
        for filename in files:
            if filename.startswith("."): continue
            
            filepath = os.path.join(root, filename)
            file_hash = get_file_hash(filepath)
            
            if file_hash in target_hashes:
                # MATCH FOUND!
                # print(f"💥 Destroying: {filename}")
                shutil.move(filepath, os.path.join(TRASH_DIR, filename))
                matches += 1
            
            count += 1
            if count % 1000 == 0:
                print(f"   Scanned {count} files...")

    print(f"✨ DONE! Removed {matches} matching files to {TRASH_DIR}")

if __name__ == "__main__":
    main()