Get a list of number of files and size of files for all file types in a directory. Example: "python file_types.py ./dir_path"

file_types.py

import os
from os.path import join, getsize
import argparse
import operator


def parse_args():
    parser = argparse.ArgumentParser(description='Count the number of files for each file type in a directory.')
    parser.add_argument('path', metavar='p', type=str, help='path of the directory')
    parser.add_argument('-threshold', metavar='t', type=int, default=10,
                        help='file count threshold (default: 10)')
    args = parser.parse_args()
    return args


def count_files(path):
    extension_dict = {}
    all_files = 0
    for root, dirs, files in os.walk(path):
        for file in files:
            all_files = all_files + 1
            chunks = file.split(".")
            if len(chunks) > 1:
                size = getsize(join(root, file))
                if chunks[1] in extension_dict:
                    (num, total_sizes) = extension_dict[chunks[1]]
                    num = num + 1
                    total_sizes = total_sizes + size
                    extension_dict[chunks[1]] = (num, total_sizes)
                else:
                    extension_dict[chunks[1]] = (1, size)
    print("total files: %d" % all_files)
    return extension_dict


def trim_dict(extension_dict, threshold):
    trimmed_dict = {}
    for k, v in extension_dict.items():
        if v[0] > threshold:
            trimmed_dict[k] = v
    return trimmed_dict


def print_dict(unsorted_dict):
    sorted_list = sorted(unsorted_dict.items(), key=operator.itemgetter(1), reverse=True)
    print("file type - number of files - size of files(bytes)\n")
    for i in sorted_list:
        # tuples like ('c', (991, 21108147))
        print("%s - %d - %d" % (i[0], i[1][0], i[1][1]))


def main():
    path = parse_args().path
    threshold = parse_args().threshold
    output_dict = count_files(path)
    output_dict = trim_dict(output_dict, threshold)
    print_dict(output_dict)


if __name__ == "__main__":
    main()