Utilities

Utility functions shared across dataset classes.

src.utils.check_type(value: object, expected_type: type | tuple, name: str) → None

Raise TypeError if value is not an instance of expected_type.

Parameters:

• value – The value to check.

• expected_type – The expected type or tuple of types.

• name – The parameter name (used in the error message).

Raises:

TypeError – If value is not of the expected type.

src.utils.check_range(value: float, min_val: float, max_val: float, name: str) → None

Raise ValueError if value is not within [min_val, max_val].

Parameters:

• value – The numeric value to check.

• min_val – The minimum allowed value (inclusive).

• max_val – The maximum allowed value (inclusive).

• name – The parameter name (used in the error message).

Raises:

ValueError – If value is outside the allowed range.

src.utils.parse_labels_csv(labels_file: str) → dict[str, str]

Parse a CSV file mapping filenames to labels.

The CSV is expected to have no header row. Each row contains:

• column 0: filename (basename only, e.g. image.jpg)

• column 1: label (string; callers may cast to int/float as needed)

Parameters:

labels_file – Absolute or relative path to the CSV file.

Returns:

A dict mapping each filename to its label string.

Raises:

• FileNotFoundError – If labels_file does not exist.

• ValueError – If the CSV contains a row with fewer than two columns.

src.utils.load_image(path: str) → ndarray

Load an image from disk as an RGB numpy array.

Parameters:

path – Path to the image file (.jpg, .jpeg, or .png).

Returns:

A numpy array of shape (H, W, 3) with dtype uint8.

Raises:

• FileNotFoundError – If the file does not exist.

• OSError – If Pillow cannot open the file.