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()
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()
def filter_without_english_letters(self, aBuf):
aBuf = re.sub(_b(r'([A-Za-z])+'), _b(' '), aBuf)
return aBuf
def filter_high_bit_only(self, aBuf):
aBuf = re.sub(_b(r'([\x00-\x7F])+'), _b(' '), aBuf)
return aBuf
def filter_without_english_letters(self, aBuf):
aBuf = re.sub(_b(r'([A-Za-z])+'), _b(' '), aBuf)
return aBuf
def filter_high_bit_only(self, aBuf):
aBuf = re.sub(_b(r'([\x00-\x7F])+'), _b(' '), aBuf)
return aBuf
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()
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()