└── handler.py


/handler.py:
--------------------------------------------------------------------------------
  1 | import os
  2 | import sys
  3 | import hashlib
  4 | # import argparse
  5 | 
  6 | 
  7 | # if __name__ == '__main__':
  8 | #     parser = argparse.ArgumentParser()
  9 | #     parser.add_argument("directory")
 10 | #     args = parser.parse_args()
 11 | #
 12 | #     if args.directory:
 13 | #         os.chdir(args.directory)
 14 | #         for root, dirs, files in os.walk('.'):
 15 | #             for name in files:
 16 | #                 print(os.path.join(root, name))
 17 | #             for name in dirs:
 18 | #                 print(os.path.join(root, name))
 19 | #################################################
 20 | def file_type():
 21 |     f_type = input('\nEnter file format:\n')
 22 |     if f_type:
 23 |         return f".{f_type}"
 24 |     return None
 25 | 
 26 | 
 27 | def sort_options():
 28 |     print("\nSize sorting options:\n"
 29 |           "1. Descending\n"
 30 |           "2. Ascending\n")
 31 |     while True:
 32 |         s_option = input('Enter a sorting option:\n')
 33 |         if s_option in ['1', '2']:
 34 |             return s_option
 35 |         else:
 36 |             print('\nWrong option\n')
 37 | 
 38 | 
 39 | def read_file(path_):
 40 |     with open(path_, 'rb') as file:
 41 |         return file.read()
 42 | 
 43 | 
 44 | def find_in_list_of_list(list_, index_):
 45 |     for sub_list in list_:
 46 |         if index_ in sub_list:
 47 |             return list_.index(sub_list)
 48 | 
 49 | 
 50 | def delete_files(file_list_):
 51 |     answer = input("\nDelete files?\n")
 52 |     if answer.lower() == 'no':
 53 |         return True
 54 |     elif answer.lower() == 'yes':
 55 |         while True:
 56 |             print("\nEnter file numbers to delete:")
 57 |             # files_delete_list = [int(x) for x in input().split()]
 58 |             try:
 59 |                 files_delete_list = list(map(int, input().strip().split()))
 60 |             except ValueError:
 61 |                 print("\nWrong format")
 62 |             else:
 63 |                 if not len(files_delete_list) or (max(files_delete_list) > len(file_list_)):
 64 |                     print("\nWrong format")
 65 |                 else:
 66 |                     space_sum = 0
 67 |                     for num in files_delete_list:
 68 |                         index = find_in_list_of_list(file_list_, num)
 69 |                         os.remove(file_list_[index][1])
 70 |                         space_sum += file_list_[index][2]
 71 |                     print(f'\nTotal freed up space: {space_sum} bytes')
 72 |                     break
 73 |         return True
 74 |     else:
 75 |         return False
 76 | 
 77 | 
 78 | def check_duplicates(files_dict_, s_type_):
 79 |     answer = input("\nCheck for duplicates?\n")
 80 |     if answer.lower() == 'no':
 81 |         return True
 82 |     elif answer.lower() == 'yes':
 83 |         i = 1
 84 |         file_list = []
 85 |         for key, val in sorted(files_dict_.items(), reverse=s_type_):
 86 |             hash_dict = {}
 87 |             if len(val) > 1:
 88 |                 print(f'\n{key} bytes')
 89 |                 for v in val:
 90 |                     md5 = hashlib.md5()
 91 |                     md5.update(read_file(v))
 92 |                     if md5.hexdigest() in hash_dict:
 93 |                         hash_dict[md5.hexdigest()].append(v)
 94 |                     else:
 95 |                         hash_dict[md5.hexdigest()] = [v]
 96 |             for key_, val_ in hash_dict.items():
 97 |                 if len(val_) > 1:
 98 |                     print(f'Hash: {key_}')
 99 |                     for v in val_:
100 |                         print(f'{i}. {v}')
101 |                         file_list.append([i, v, key])
102 |                         i += 1
103 |         while True:
104 |             if delete_files(file_list):
105 |                 break
106 |             else:
107 |                 print("\nWrong choice")
108 |         return True
109 |     else:
110 |         return False
111 | 
112 | 
113 | def main():
114 |     files_dict = {}
115 |     args = sys.argv
116 |     if len(args) < 2:
117 |         print('Directory is not specified')
118 |     else:
119 |         f_type = file_type()
120 |         s_type = sort_options()
121 |         os.chdir(args[1])
122 |         full_path = os.getcwd()
123 |         for root, dirs, files in os.walk(full_path):
124 |             for name in files:
125 |                 path = os.path.join(root, name)
126 |                 size = os.path.getsize(path)
127 |                 ext = os.path.splitext(path)
128 |                 if size in files_dict and (ext[1] == f_type or f_type is None):
129 |                     files_dict[size].append(path)
130 |                 elif ext[1] == f_type or f_type is None:
131 |                     files_dict[size] = [path]
132 |         if s_type == '1':
133 |             s_type = True
134 |         elif s_type == '2':
135 |             s_type = False
136 |         for key, val in sorted(files_dict.items(), reverse=s_type):
137 |             if len(val) > 1:
138 |                 print(f'\n{key} bytes')
139 |                 print(*val, sep='\n')
140 |         while True:
141 |             if check_duplicates(files_dict, s_type):
142 |                 break
143 | 
144 | 
145 | if __name__ == '__main__':
146 |     main()
147 | 


--------------------------------------------------------------------------------