101 lines
3.0 KiB
Python
101 lines
3.0 KiB
Python
|
|
import os
|
|
import re
|
|
from PIL import Image
|
|
|
|
TARGET_ROOT = "assets/slike"
|
|
|
|
# Regex to strip prefixes/suffixes to find "Core Name"
|
|
def get_core_name(filename):
|
|
name = filename.lower()
|
|
name = os.path.splitext(name)[0]
|
|
name = re.sub(r"assets_backup_\d+_\d+_", "", name)
|
|
name = re.sub(r"moje_slike_koncna_", "", name)
|
|
name = re.sub(r"src_assets_library_", "", name)
|
|
name = re.sub(r"assets__backup_[a-z0-9_]+_", "", name)
|
|
name = re.sub(r"_\d+$", "", name)
|
|
name = re.sub(r"_v\d+$", "", name)
|
|
name = name.replace("_copy", "").replace("_resized", "")
|
|
name = name.strip("_")
|
|
return name
|
|
|
|
def is_green_screen(filepath):
|
|
try:
|
|
with Image.open(filepath) as img:
|
|
rgb = img.convert("RGB")
|
|
# Tolerate imperfect green
|
|
p1 = rgb.getpixel((0, 0))
|
|
if p1[1] > 150 and p1[0] < 80 and p1[2] < 80: return True
|
|
p2 = rgb.getpixel((0, img.height//2))
|
|
if p2[1] > 150 and p2[0] < 80 and p2[2] < 80: return True
|
|
return False
|
|
except:
|
|
return False
|
|
|
|
def dedupe_recursive():
|
|
print(f"🚀 GLOBAL DEDUPE in {TARGET_ROOT} (Recursive)...")
|
|
|
|
# gather all files
|
|
all_files_list = []
|
|
for root, dirs, files in os.walk(TARGET_ROOT):
|
|
for f in files:
|
|
if f.lower().endswith(('.png', '.jpg')):
|
|
all_files_list.append(os.path.join(root, f))
|
|
|
|
print(f"Scanning {len(all_files_list)} files...")
|
|
|
|
groups = {}
|
|
for path in all_files_list:
|
|
fname = os.path.basename(path)
|
|
core = get_core_name(fname)
|
|
if core not in groups:
|
|
groups[core] = []
|
|
groups[core].append(path)
|
|
|
|
deleted = 0
|
|
|
|
for core, paths in groups.items():
|
|
if len(paths) < 2:
|
|
continue
|
|
|
|
# We have duplicates (by name conceptual)
|
|
# Check green
|
|
candidates = []
|
|
for p in paths:
|
|
is_g = is_green_screen(p)
|
|
candidates.append({"path": p, "green": is_g, "name": os.path.basename(p)})
|
|
|
|
greens = [c for c in candidates if c["green"]]
|
|
others = [c for c in candidates if not c["green"]]
|
|
|
|
keep = None
|
|
remove = []
|
|
|
|
if greens:
|
|
# Prefer green. Use shortest name among greens.
|
|
greens.sort(key=lambda x: len(x["name"]))
|
|
keep = greens[0]
|
|
remove = greens[1:] + others
|
|
else:
|
|
# No greens. Keep shortest name unique.
|
|
others.sort(key=lambda x: len(x["name"]))
|
|
keep = others[0]
|
|
remove = others[1:]
|
|
|
|
# Delete
|
|
for item in remove:
|
|
try:
|
|
os.remove(item["path"])
|
|
deleted += 1
|
|
# print(f"Deleted {item['name']}")
|
|
except:
|
|
pass
|
|
|
|
print("="*40)
|
|
print(f"✅ GLOBAL DEDUPE COMPLETE.")
|
|
print(f"Deleted {deleted} duplicates.")
|
|
print("="*40)
|
|
|
|
if __name__ == "__main__":
|
|
dedupe_recursive()
|