ThirdParty/CharDet/chardet/sbcsgroupprober.py

Sat, 02 Jan 2010 15:11:35 +0000

author
Detlev Offenbach <detlev@die-offenbachs.de>
date
Sat, 02 Jan 2010 15:11:35 +0000
changeset 12
1d8dd9706f46
parent 0
de9c2efb9d02
child 3537
7662053c3906
permissions
-rw-r--r--

First commit after changing to Python 3.1.

12
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
1 ######################## BEGIN LICENSE BLOCK ########################
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
2 # The Original Code is Mozilla Universal charset detector code.
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
3 #
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
4 # The Initial Developer of the Original Code is
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
5 # Netscape Communications Corporation.
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
6 # Portions created by the Initial Developer are Copyright (C) 2001
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
7 # the Initial Developer. All Rights Reserved.
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
8 #
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
9 # Contributor(s):
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
10 # Mark Pilgrim - port to Python
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
11 # Shy Shalom - original C code
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
12 #
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
13 # This library is free software; you can redistribute it and/or
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
14 # modify it under the terms of the GNU Lesser General Public
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
15 # License as published by the Free Software Foundation; either
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
16 # version 2.1 of the License, or (at your option) any later version.
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
17 #
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
18 # This library is distributed in the hope that it will be useful,
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
19 # but WITHOUT ANY WARRANTY; without even the implied warranty of
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
20 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
21 # Lesser General Public License for more details.
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
22 #
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
23 # You should have received a copy of the GNU Lesser General Public
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
24 # License along with this library; if not, write to the Free Software
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
25 # Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
26 # 02110-1301 USA
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
27 ######################### END LICENSE BLOCK #########################
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
28
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
29 from . import constants
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
30 import sys
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
31 from .charsetgroupprober import CharSetGroupProber
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
32 from .sbcharsetprober import SingleByteCharSetProber
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
33 from .langcyrillicmodel import Win1251CyrillicModel, Koi8rModel, Latin5CyrillicModel, MacCyrillicModel, Ibm866Model, Ibm855Model
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
34 from .langgreekmodel import Latin7GreekModel, Win1253GreekModel
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
35 from .langbulgarianmodel import Latin5BulgarianModel, Win1251BulgarianModel
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
36 from .langhungarianmodel import Latin2HungarianModel, Win1250HungarianModel
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
37 from .langthaimodel import TIS620ThaiModel
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
38 from .langhebrewmodel import Win1255HebrewModel
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
39 from .hebrewprober import HebrewProber
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
40
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
41 class SBCSGroupProber(CharSetGroupProber):
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
42 def __init__(self):
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
43 CharSetGroupProber.__init__(self)
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
44 self._mProbers = [ \
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
45 SingleByteCharSetProber(Win1251CyrillicModel),
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
46 SingleByteCharSetProber(Koi8rModel),
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
47 SingleByteCharSetProber(Latin5CyrillicModel),
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
48 SingleByteCharSetProber(MacCyrillicModel),
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
49 SingleByteCharSetProber(Ibm866Model),
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
50 SingleByteCharSetProber(Ibm855Model),
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
51 SingleByteCharSetProber(Latin7GreekModel),
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
52 SingleByteCharSetProber(Win1253GreekModel),
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
53 SingleByteCharSetProber(Latin5BulgarianModel),
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
54 SingleByteCharSetProber(Win1251BulgarianModel),
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
55 SingleByteCharSetProber(Latin2HungarianModel),
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
56 SingleByteCharSetProber(Win1250HungarianModel),
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
57 SingleByteCharSetProber(TIS620ThaiModel),
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
58 ]
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
59 hebrewProber = HebrewProber()
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
60 logicalHebrewProber = SingleByteCharSetProber(Win1255HebrewModel, False, hebrewProber)
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
61 visualHebrewProber = SingleByteCharSetProber(Win1255HebrewModel, True, hebrewProber)
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
62 hebrewProber.set_model_probers(logicalHebrewProber, visualHebrewProber)
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
63 self._mProbers.extend([hebrewProber, logicalHebrewProber, visualHebrewProber])
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
64
1d8dd9706f46 First commit after changing to Python 3.1.
Detlev Offenbach <detlev@die-offenbachs.de>
parents: 0
diff changeset
65 self.reset()

eric ide

mercurial