garmentiq.utils.check_unzipped_dir

 1"""Validating the structure of an extracted dataset."""
 2import os
 3
 4
 5def check_unzipped_dir(folder):
 6    """
 7    Verifies the structure and contents of an unzipped dataset directory.
 8
 9    This function checks that the specified folder contains:
10    - A subdirectory named 'images' with at least one `.jpg`, `.png`, `.jpeg` file.
11    - A file named 'metadata.csv'.
12
13    If any of these components are missing, a `FileNotFoundError` is raised.
14
15    Args:
16        folder (str): Path to the root folder of the unzipped dataset.
17
18    Raises:
19        FileNotFoundError: If the 'images' folder is missing,
20                           if no `.jpg`, `.png`, `.jpeg` files are found in the 'images' folder,
21                           or if 'metadata.csv' is missing.
22
23    Returns:
24        None
25    """
26    image_dir = os.path.join(folder, "images")
27    metadata_path = os.path.join(folder, "metadata.csv")
28
29    if not os.path.isdir(image_dir):
30        raise FileNotFoundError(f"Missing 'images' folder in: {folder}")
31
32    jpg_files = [f for f in os.listdir(image_dir) if f.lower().endswith(".jpg") or f.lower().endswith(".png") or f.lower().endswith(".jpeg")]
33    if not jpg_files:
34        raise FileNotFoundError(f"No .jpg, .png, or .jpeg file found in: {image_dir}")
35
36    if not os.path.isfile(metadata_path):
37        raise FileNotFoundError(f"Missing 'metadata.csv' in: {folder}")