Exemplo n.º 1
0
 def reset(self):
     self.result = {'encoding': None, 'confidence': 0.0}
     self.done = False
     self._mStart = True
     self._mGotData = False
     self._mInputState = ePureAscii
     self._mLastChar = _b('')
     if self._mEscCharSetProber:
         self._mEscCharSetProber.reset()
     for prober in self._mCharSetProbers:
         prober.reset()
Exemplo n.º 2
0
 def reset(self):
     self.result = {'encoding': None, 'confidence': 0.0}
     self.done = False
     self._mStart = True
     self._mGotData = False
     self._mInputState = ePureAscii
     self._mLastChar = _b('')
     if self._mEscCharSetProber:
         self._mEscCharSetProber.reset()
     for prober in self._mCharSetProbers:
         prober.reset()
Exemplo n.º 3
0
 def filter_without_english_letters(self, aBuf):
     aBuf = re.sub(_b(r'([A-Za-z])+'), _b(' '), aBuf)
     return aBuf
Exemplo n.º 4
0
 def filter_high_bit_only(self, aBuf):
     aBuf = re.sub(_b(r'([\x00-\x7F])+'), _b(' '), aBuf)
     return aBuf
Exemplo n.º 5
0
 def filter_without_english_letters(self, aBuf):
     aBuf = re.sub(_b(r'([A-Za-z])+'), _b(' '), aBuf)
     return aBuf
Exemplo n.º 6
0
 def filter_high_bit_only(self, aBuf):
     aBuf = re.sub(_b(r'([\x00-\x7F])+'), _b(' '), aBuf)
     return aBuf
Exemplo n.º 7
0
 def __init__(self):
     self._highBitDetector = re.compile(_b(r'[\x80-\xFF]'))
     self._escDetector = re.compile(_b(r'(\033|~{)'))
     self._mEscCharSetProber = None
     self._mCharSetProbers = []
     self.reset()
Exemplo n.º 8
0
 def __init__(self):
     self._highBitDetector = re.compile(_b(r'[\x80-\xFF]'))
     self._escDetector = re.compile(_b(r'(\033|~{)'))
     self._mEscCharSetProber = None
     self._mCharSetProbers = []
     self.reset()