nao_usem_similaridade.py

def tokenizer(p, n=2):
  p = '$' + p +'#'
  tk_index = {}
  tokens = []
  for i in range(len(p)-1):
    tk = p[i:i+n]
    if tk in tk_index:
      tk_index[tk] += 1
      tokens.append(tk+str(tk_index[tk]))
    else:
      tk_index[tk] = 1
      tokens.append(tk+str(tk_index[tk]))
  return tokens

print(tokenizer('admin_pedrinho'))
print(tokenizer('arara'))

def jaccard(a,b):
  """https://www.wikiwand.com/en/Jaccard_index"""
  a = set(tokenizer(a))
  b = set(tokenizer(b))
  return len(a.intersection(b))/float(len(a.union(b)))

print(jaccard('api','apicultor'))
print(jaccard('api','_api'))
print(jaccard('api','_api_'))
print(jaccard('api','apiu'))
print(jaccard('api','api_'))
print(jaccard('api','a_p_i'))
print(jaccard('api','a_p_i____pedringo'))

import re
# regex cabulosão do tio bênis
p = re.compile(r'^[_-]a[_-]?p[_-]?i|a[_-]?p[_-]?i[_-]|^a[_-]?p[_-]?i$')

print(p.match('apicultor'))
print(p.match('_api'))
print(p.match('_api_'))
print(p.match('api_'))
print(p.match('a_p_i'))
print(p.match('a_p_i____pedringo'))