ThirdParty/CharDet/chardet/escprober.py

Mon, 28 Dec 2009 16:03:33 +0000

author
Detlev Offenbach <detlev@die-offenbachs.de>
date
Mon, 28 Dec 2009 16:03:33 +0000
changeset 0
de9c2efb9d02
child 12
1d8dd9706f46
permissions
-rw-r--r--

Started porting eric4 to Python3

0
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
1 ######################## BEGIN LICENSE BLOCK ########################
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
2 # The Original Code is mozilla.org code.
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
3 #
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
4 # The Initial Developer of the Original Code is
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
5 # Netscape Communications Corporation.
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
6 # Portions created by the Initial Developer are Copyright (C) 1998
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
7 # the Initial Developer. All Rights Reserved.
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
8 #
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
9 # Contributor(s):
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
10 # Mark Pilgrim - port to Python
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
11 #
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
12 # This library is free software; you can redistribute it and/or
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
13 # modify it under the terms of the GNU Lesser General Public
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
14 # License as published by the Free Software Foundation; either
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
15 # version 2.1 of the License, or (at your option) any later version.
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
16 #
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
17 # This library is distributed in the hope that it will be useful,
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
18 # but WITHOUT ANY WARRANTY; without even the implied warranty of
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
19 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
20 # Lesser General Public License for more details.
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
21 #
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
22 # You should have received a copy of the GNU Lesser General Public
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
23 # License along with this library; if not, write to the Free Software
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
24 # Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
25 # 02110-1301 USA
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
26 ######################### END LICENSE BLOCK #########################
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
27
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
28 import constants, sys
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
29 from escsm import HZSMModel, ISO2022CNSMModel, ISO2022JPSMModel, ISO2022KRSMModel
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
30 from charsetprober import CharSetProber
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
31 from codingstatemachine import CodingStateMachine
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
32
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
33 class EscCharSetProber(CharSetProber):
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
34 def __init__(self):
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
35 CharSetProber.__init__(self)
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
36 self._mCodingSM = [ \
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
37 CodingStateMachine(HZSMModel),
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
38 CodingStateMachine(ISO2022CNSMModel),
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
39 CodingStateMachine(ISO2022JPSMModel),
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
40 CodingStateMachine(ISO2022KRSMModel)
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
41 ]
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
42 self.reset()
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
43
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
44 def reset(self):
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
45 CharSetProber.reset(self)
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
46 for codingSM in self._mCodingSM:
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
47 if not codingSM: continue
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
48 codingSM.active = constants.True
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
49 codingSM.reset()
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
50 self._mActiveSM = len(self._mCodingSM)
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
51 self._mDetectedCharset = None
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
52
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
53 def get_charset_name(self):
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
54 return self._mDetectedCharset
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
55
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
56 def get_confidence(self):
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
57 if self._mDetectedCharset:
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
58 return 0.99
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
59 else:
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
60 return 0.00
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
61
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
62 def feed(self, aBuf):
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
63 for c in aBuf:
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
64 for codingSM in self._mCodingSM:
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
65 if not codingSM: continue
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
66 if not codingSM.active: continue
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
67 codingState = codingSM.next_state(c)
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
68 if codingState == constants.eError:
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
69 codingSM.active = constants.False
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
70 self._mActiveSM -= 1
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
71 if self._mActiveSM <= 0:
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
72 self._mState = constants.eNotMe
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
73 return self.get_state()
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
74 elif codingState == constants.eItsMe:
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
75 self._mState = constants.eFoundIt
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
76 self._mDetectedCharset = codingSM.get_coding_state_machine()
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
77 return self.get_state()
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
78
de9c2efb9d02 Started porting eric4 to Python3
Detlev Offenbach <detlev@die-offenbachs.de>
parents:
diff changeset
79 return self.get_state()

eric ide

mercurial