ThirdParty/CharDet/chardet/codingstatemachine.py

changeset 12
1d8dd9706f46
parent 0
de9c2efb9d02
child 3537
7662053c3906
equal deleted inserted replaced
11:b0996e4a289e 12:1d8dd9706f46
23 # License along with this library; if not, write to the Free Software 23 # License along with this library; if not, write to the Free Software
24 # Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 24 # Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA
25 # 02110-1301 USA 25 # 02110-1301 USA
26 ######################### END LICENSE BLOCK ######################### 26 ######################### END LICENSE BLOCK #########################
27 27
28 from constants import eStart, eError, eItsMe 28 from .constants import eStart, eError, eItsMe
29 29
30 class CodingStateMachine: 30 class CodingStateMachine:
31 def __init__(self, sm): 31 def __init__(self, sm):
32 self._mModel = sm 32 self._mModel = sm
33 self._mCurrentBytePos = 0 33 self._mCurrentBytePos = 0
38 self._mCurrentState = eStart 38 self._mCurrentState = eStart
39 39
40 def next_state(self, c): 40 def next_state(self, c):
41 # for each byte we get its class 41 # for each byte we get its class
42 # if it is first byte, we also get byte length 42 # if it is first byte, we also get byte length
43 byteCls = self._mModel['classTable'][ord(c)] 43 # PY3K: aBuf is a byte stream, so c is an int, not a byte
44 byteCls = self._mModel['classTable'][c]
44 if self._mCurrentState == eStart: 45 if self._mCurrentState == eStart:
45 self._mCurrentBytePos = 0 46 self._mCurrentBytePos = 0
46 self._mCurrentCharLen = self._mModel['charLenTable'][byteCls] 47 self._mCurrentCharLen = self._mModel['charLenTable'][byteCls]
47 # from byte's class and stateTable, we get its next state 48 # from byte's class and stateTable, we get its next state
48 self._mCurrentState = self._mModel['stateTable'][self._mCurrentState * self._mModel['classFactor'] + byteCls] 49 self._mCurrentState = self._mModel['stateTable'][self._mCurrentState * self._mModel['classFactor'] + byteCls]

eric ide

mercurial