fork download
copy 
  1. # your code goes here
  2.  
  3. import sre_parse
  4. import pprint
  5. import random
  6.  
  7.  
  8. class RandomRegexStringGenerator:
  9.     def __init__(self):
  10.         self._words = 'abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789_'
  11.         self._digits = '0123456789'
  12.         self._whitespaces = ' \t'  #  ' \t\n\r\v\f' 에서 조정
  13.         self._max_repeat = 5 # 무한 수량자가 나왔을때 최대 반복 리밋
  14.  
  15.         self._result = ''
  16.  
  17.     def _gen_char(self, value):
  18.         return chr(value)
  19.  
  20.     def _gen_range(self, min_value, max_value):
  21.         # random 선택
  22.         idx = random.randint(min_value, max_value)
  23.         return chr(idx)
  24.  
  25.     def _gen_in_word(self):
  26.         return self._words[random.randint(0, len(self._words)-1)]
  27.  
  28.     def _gen_in_digit(self):
  29.         return self._digits[random.randint(0, len(self._digits)-1)]
  30.  
  31.     def _gen_in_whitespace(self):
  32.         return self._whitespaces[random.randint(0, len(self._whitespaces)-1)]
  33.  
  34.  
  35.     def _proc_cmd(self, itm):
  36.         # 단문자
  37.         if itm[0] == sre_parse.LITERAL:
  38.             self._result += self._gen_char(itm[1])
  39.  
  40.         # 문자클래스
  41.         elif itm[0] == sre_parse.IN:
  42.             start_num = 0
  43.             body = itm[1]
  44.             if body[0][0] == sre_parse.NEGATE: # 제외조건 TODO:
  45.                 start_num = 1
  46.  
  47.             # random 선택
  48.             idx = random.randint(start_num, len(body)-1)
  49.             cls_itm = body[idx]
  50.  
  51.             if cls_itm[0] == sre_parse.LITERAL:
  52.                 self._result += self._gen_char(cls_itm[1])
  53.  
  54.             elif cls_itm[0] == sre_parse.RANGE:
  55.                 self._result += self._gen_range(cls_itm[1][0], cls_itm[1][1])
  56.  
  57.             elif cls_itm[0] == sre_parse.CATEGORY:
  58.                 if cls_itm[1] == sre_parse.CATEGORY_DIGIT:
  59.                     self._result += self._gen_in_digit()
  60.  
  61.                 elif cls_itm[1] == sre_parse.CATEGORY_WORD:
  62.                     self._result += self._gen_in_word()
  63.  
  64.                 elif cls_itm[1] == sre_parse.CATEGORY_SPACE:
  65.                     self._result += self._gen_in_whitespace()
  66.                 # TODO: NOT CATEGORY 처리해야함..
  67.  
  68.         # 수량자
  69.         elif itm[0] in (sre_parse.MAX_REPEAT, sre_parse.MIN_REPEAT):
  70.             # 패턴 => XXX_REPEAT, (min, max, p)                 
  71.             min_repeat = itm[1][0] 
  72.             max_repeat = itm[1][1]
  73.             # set limit
  74.             if itm[1][1] == sre_parse.MAXREPEAT:
  75.                 max_repeat = self._max_repeat  # 지정한 max값으로 한정
  76.  
  77.             # random 선택
  78.             rand_repeat = random.randint(min_repeat, max_repeat)
  79.  
  80.             # random 만큼 반복
  81.             for i in range(rand_repeat):
  82.                 self._read_body(itm[1][2])
  83.  
  84.         # Union
  85.         elif itm[0] == sre_parse.BRANCH:
  86.             body = itm[1][1]
  87.             # random 선택
  88.             idx = random.randint(0, len(body)-1)
  89.             self._read_body(body[idx])
  90.  
  91.         else:            
  92.             if isinstance(itm[1], tuple):
  93.                 for titm in itm[1]:
  94.                     if isinstance(titm, sre_parse.SubPattern):
  95.                         #print('CMD: %s in body '%(itm[0]))
  96.                         self._read_body(titm)
  97.  
  98.     # [ (cmd), (cmd), ...] 이런 형태에 대해 cmd 마다 처리
  99.     def _read_body(self, body):
  100.         for itm in body:
  101.             self._proc_cmd(itm)
  102.  
  103.  
  104.     def generate(self, regex_pattern):
  105.  
  106.         p = sre_parse.parse( regex_pattern )
  107.         #pprint.pprint(p)
  108.  
  109.         self._result = ''
  110.         self._read_body( p )
  111.  
  112.         return self._result
  113.  
  114.  
  115.  
  116.  
  117.  
  118. pat = r'01[069]-\d\d\d\d?(-\d\d\d\d){2}'
  119.  
  120. g = RandomRegexStringGenerator()
  121.  
  122. for i in range(10):
  123.     rstr = g.generate( pat )
  124.     print( rstr )
  125.  
  126.  
Success #stdin #stdout 0.03s 11848KB
comments (?)
stdin
copy 
Standard input is empty
stdout
copy 
010-631-4279-4705
010-6518-5752-3683
016-124-0882-4711
019-583-3439-1319
016-072-6866-4919
019-4837-2730-3732
019-9947-2561-3228
019-7507-0317-3081
019-563-9371-9878
016-8216-5173-5505
https://ideone.com/NNoIax
language:
Python 3 (python  3.12)
created:
4 years ago
visibility:
secret

Share or Embed source code

Discover > Sphere Engine API

The brand new service which powers Ideone!

Discover > IDE Widget

Widget for compiling and running the source code in a web browser!