s3_path = 's3://tri-ml-sandbox-16011-us-west-2-datasets/cv_webdatasets'
dataset = 'LBMv12/allrgb_buf[24_p8]-ctx[sequence-single]_cam[ALL]'

from tqdm import tqdm
from utils import list_s3

print()
filenames = []
folders1 = list_s3(f'{s3_path}/{dataset}')
for f1 in tqdm(folders1):
    folders2 = list_s3(f'{s3_path}/{dataset}/{f1}')
    for f2 in folders2:
        folders3 = list_s3(f'{s3_path}/{dataset}/{f1}/{f2}')
        for f3 in folders3:
            folders3 = list_s3(f'{s3_path}/{dataset}/{f1}/{f2}/{f3}/indices')
            if 'filenames.txt' in folders3:
                filenames.append(f'{dataset}/{f1}/{f2}/{f3}/indices/filenames.txt')

with open('split.txt', 'w') as file:
    for filename in filenames:
        file.write(f'[\"{filename}\",1],\n')