Goyamproject/React_native_app

1import os
2import shutil
3import random
4from glob import glob
5from tqdm import tqdm
6
7SOURCE_ROOT = r"C:\Users\charu\Documents\goyam\roboflow\train"
8SOURCE_IMAGES = os.path.join(SOURCE_ROOT, "images")
9SOURCE_LABELS = os.path.join(SOURCE_ROOT, "labels")
10
11
12DEST_DIR = r"C:\Users\charu\Documents\goyam\roboflow\final_split"
13
14TRAIN_RATIO = 0.8
15
16
17def split_dataset():
18  
19    if os.path.exists(DEST_DIR):
20        print(f"⚠️ Warning: Destination folder already exists: {DEST_DIR}")
21        print("   (Ideally, delete it before running this to avoid mixing old data!)")
22    
23    for split in ['train', 'val']:
24        os.makedirs(os.path.join(DEST_DIR, split, 'images'), exist_ok=True)
25        os.makedirs(os.path.join(DEST_DIR, split, 'labels'), exist_ok=True)
26
27  
28    print(f"🔍 Scanning images in: {SOURCE_IMAGES}")
29    
30
31    unique_images = set()
32    
33    # Check all extensions
34    exts = ['*.jpg', '*.jpeg', '*.png', '*.JPG', '*.JPEG', '*.PNG']
35    for ext in exts:
36        files = glob(os.path.join(SOURCE_IMAGES, ext))
37        for f in files:
38            unique_images.add(f) 
39            
40    all_images = list(unique_images)
41    
42
43    random.shuffle(all_images)
44    
45    total_count = len(all_images)
46    train_count = int(total_count * TRAIN_RATIO)
47    
48    if total_count == 0:
49        print(" Error: No images found!")
50        return
51
52    print(f"Found {total_count} unique images.")
53    print(f"   -> Training: {train_count}")
54    print(f"   -> Validation: {total_count - train_count}")
55
56    # 3. Copy Files
57    print("Organizing files...")
58    for i, img_path in enumerate(tqdm(all_images)):
59        split = 'train' if i < train_count else 'val'
60        
61        filename = os.path.basename(img_path)
62        name_no_ext = os.path.splitext(filename)[0]
63        
64        dest_img_path = os.path.join(DEST_DIR, split, 'images', filename)
65        
66        # Check label (Look for .txt)
67        src_txt_path = os.path.join(SOURCE_LABELS, name_no_ext + ".txt")
68        dest_txt_path = os.path.join(DEST_DIR, split, 'labels', name_no_ext + ".txt")
69        
70        shutil.copy(img_path, dest_img_path)
71        
72        if os.path.exists(src_txt_path):
73            shutil.copy(src_txt_path, dest_txt_path)
74
75    print(f"\nDone! Your dataset is ready at:")
76    print(f"   {DEST_DIR}")
77
78if __name__ == "__main__":
79    split_dataset()