def extract()

in muss/utils/resources.py [0:0]

• 28 lines of code
• 5 McCabe index (conditional complexity)

def extract(filepath, output_dir):
    output_dir = Path(output_dir)
    # Infer extract function based on extension
    extensions_to_functions = {
        '.tar.gz': untar,
        '.tar.bz2': untar,
        '.tgz': untar,
        '.zip': unzip,
        '.gz': ungzip,
        '.bz2': unbz2,
    }

    def get_extension(filename, extensions):
        possible_extensions = [ext for ext in extensions if filename.endswith(ext)]
        if len(possible_extensions) == 0:
            raise Exception(f'File {filename} has an unknown extension')
        # Take the longest (.tar.gz should take precedence over .gz)
        return max(possible_extensions, key=lambda ext: len(ext))

    filename = os.path.basename(filepath)
    extension = get_extension(filename, list(extensions_to_functions))
    extract_function = extensions_to_functions[extension]

    # Extract files in a temporary dir then move the extracted item back to
    # the ouput dir in order to get the details of what was extracted
    tmp_extract_dir = Path(tempfile.mkdtemp())
    # Extract
    extract_function(filepath, output_dir=tmp_extract_dir)
    extracted_items = os.listdir(tmp_extract_dir)
    output_paths = []
    for name in extracted_items:
        extracted_path = tmp_extract_dir / name
        output_path = output_dir / name
        move_with_overwrite(extracted_path, output_path)
        output_paths.append(output_path)
    return output_paths