import os
import difflib

# Target directory
NPC_DIR = os.path.abspath("assets/slike/glavna_referenca/_FOLDER_NPC")

def find_fuzzy_duplicates():
    print(f"🕵️‍♂️ Analyzing similar files in {os.path.basename(NPC_DIR)}...")
    
    files = {} # filename -> filepath
    for f in os.listdir(NPC_DIR):
        if f.startswith("."): continue
        files[f] = os.path.join(NPC_DIR, f)
        
    filenames = list(files.keys())
    files_to_remove = set()
    
    # Compare names
    count = 0
    total = len(filenames)
    # Simple check: Name containment (e.g. "gronk.png" vs "gronk_01.png")
    # sorted by length so we check shorter names against longer ones
    sorted_names = sorted(filenames, key=len)
    
    for i in range(len(sorted_names)):
        name1 = sorted_names[i]
        stem1, ext1 = os.path.splitext(name1)
        path1 = files[name1]
        
        # Skip if already marked
        if path1 in files_to_remove: continue

        for j in range(i + 1, len(sorted_names)):
            name2 = sorted_names[j]
            stem2, ext2 = os.path.splitext(name2)
            path2 = files[name2]
            
            if path2 in files_to_remove: continue
            
            # RULE 1: If stem1 is contained in stem2 (e.g. "image" in "image_copy")
            # AND extensions match or are compatible image types
            if stem1 in stem2:
                # Check similarity ratio to avoid false positives like "man" in "woman"
                if difflib.SequenceMatcher(None, stem1, stem2).ratio() > 0.8 or stem2.startswith(stem1):
                     # Potential duplicate! 
                     # Strategy: Keep the one with clearer name or larger size?
                     # Let's keep LARGER file usually (better quality)
                     size1 = os.path.getsize(path1)
                     size2 = os.path.getsize(path2)
                     
                     print(f"   ⚠️ Potential dupe: '{name1}' ({size1}b) vs '{name2}' ({size2}b)")
                     
                     # REMOVE THE SMALLER ONE
                     if size1 >= size2:
                         print(f"      🗑️ Deleting smaller/same: {name2}")
                         files_to_remove.add(path2)
                     else:
                         print(f"      🗑️ Deleting smaller: {name1}")
                         files_to_remove.add(path1)
                         break # name1 is gone, stop checking it

    # DO THE DELETION
    print(f"\n🗑️ Deleting {len(files_to_remove)} files...")
    for p in files_to_remove:
        try:
            os.remove(p)
        except OSError as e:
            print(f"Error removing {p}: {e}")
            
    print("✨ Fuzzy cleanup done.")

if __name__ == "__main__":
    find_fuzzy_duplicates()