extract_images.py

#!/usr/bin/python

import argparse
import os
import shutil
import sys


def parse_arguments():
    parser = argparse.ArgumentParser(description='Process a directory for images')
    parser.add_argument('--input_dir')
    parser.add_argument('--output_dir')
    args = parser.parse_args()

    if not args.input_dir or not args.output_dir:
        parser.print_usage()
        sys.exit(-1)

    if not os.path.exists(args.input_dir):
        print('Input directory {0} does not exist'.format(args.input_dir))
        sys.exit(-1)

    if not os.path.exists(args.output_dir):
        print('Output directory {0} does not exist'.format(args.output_dir))
        sys.exit(-1)

    return args.input_dir, args.output_dir


def find_duplicates(files):
    result = []
    files_with_names = [(f, os.path.basename(f)) for f in files]
    work_list = files_with_names.copy()
    for pair in files_with_names:
        work_list = [x for x in work_list if x != pair]
        duplicates = [d for d in work_list if d[0] != pair[0] and d[1] == pair[1]]
        if duplicates:
            work_list = [x for x in work_list if x not in duplicates]
            duplicates.append(pair)
            duplicates = [x[0] for x in duplicates]
            result.append(duplicates)
    return result


def find_conflicting(files):
    for file in find_duplicates(files):
        temp = set([os.path.getsize(x) for x in file])
        if len(temp) > 1:
            print(file)
    return True

def main():
    input_dir, output_dir = parse_arguments()
    to_copy = []

    for root, dir, files in os.walk(input_dir):
        images = [f for f in files if os.path.splitext(f)[1] in ['.jpg']]
        to_copy += [os.path.join(root, f) for f in images]
    

    find_conflicting(to_copy)


if __name__ == '__main__':
    main()