Fri, 25 Apr 2014 22:07:19 +0200
updated CharDet to 2.2.1, updated changelog
3537
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
1 | ######################## BEGIN LICENSE BLOCK ######################## |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
2 | # The Original Code is Mozilla Universal charset detector code. |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
3 | # |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
4 | # The Initial Developer of the Original Code is |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
5 | # Netscape Communications Corporation. |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
6 | # Portions created by the Initial Developer are Copyright (C) 2001 |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
7 | # the Initial Developer. All Rights Reserved. |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
8 | # |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
9 | # Contributor(s): |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
10 | # Mark Pilgrim - port to Python |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
11 | # Shy Shalom - original C code |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
12 | # |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
13 | # This library is free software; you can redistribute it and/or |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
14 | # modify it under the terms of the GNU Lesser General Public |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
15 | # License as published by the Free Software Foundation; either |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
16 | # version 2.1 of the License, or (at your option) any later version. |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
17 | # |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
18 | # This library is distributed in the hope that it will be useful, |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
19 | # but WITHOUT ANY WARRANTY; without even the implied warranty of |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
20 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
21 | # Lesser General Public License for more details. |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
22 | # |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
23 | # You should have received a copy of the GNU Lesser General Public |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
24 | # License along with this library; if not, write to the Free Software |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
25 | # Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
26 | # 02110-1301 USA |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
27 | ######################### END LICENSE BLOCK ######################### |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
28 | |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
29 | from .charsetgroupprober import CharSetGroupProber |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
30 | from .sbcharsetprober import SingleByteCharSetProber |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
31 | from .langcyrillicmodel import (Win1251CyrillicModel, Koi8rModel, |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
32 | Latin5CyrillicModel, MacCyrillicModel, |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
33 | Ibm866Model, Ibm855Model) |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
34 | from .langgreekmodel import Latin7GreekModel, Win1253GreekModel |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
35 | from .langbulgarianmodel import Latin5BulgarianModel, Win1251BulgarianModel |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
36 | from .langhungarianmodel import Latin2HungarianModel, Win1250HungarianModel |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
37 | from .langthaimodel import TIS620ThaiModel |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
38 | from .langhebrewmodel import Win1255HebrewModel |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
39 | from .hebrewprober import HebrewProber |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
40 | |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
41 | |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
42 | class SBCSGroupProber(CharSetGroupProber): |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
43 | def __init__(self): |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
44 | CharSetGroupProber.__init__(self) |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
45 | self._mProbers = [ |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
46 | SingleByteCharSetProber(Win1251CyrillicModel), |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
47 | SingleByteCharSetProber(Koi8rModel), |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
48 | SingleByteCharSetProber(Latin5CyrillicModel), |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
49 | SingleByteCharSetProber(MacCyrillicModel), |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
50 | SingleByteCharSetProber(Ibm866Model), |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
51 | SingleByteCharSetProber(Ibm855Model), |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
52 | SingleByteCharSetProber(Latin7GreekModel), |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
53 | SingleByteCharSetProber(Win1253GreekModel), |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
54 | SingleByteCharSetProber(Latin5BulgarianModel), |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
55 | SingleByteCharSetProber(Win1251BulgarianModel), |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
56 | SingleByteCharSetProber(Latin2HungarianModel), |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
57 | SingleByteCharSetProber(Win1250HungarianModel), |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
58 | SingleByteCharSetProber(TIS620ThaiModel), |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
59 | ] |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
60 | hebrewProber = HebrewProber() |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
61 | logicalHebrewProber = SingleByteCharSetProber(Win1255HebrewModel, |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
62 | False, hebrewProber) |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
63 | visualHebrewProber = SingleByteCharSetProber(Win1255HebrewModel, True, |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
64 | hebrewProber) |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
65 | hebrewProber.set_model_probers(logicalHebrewProber, visualHebrewProber) |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
66 | self._mProbers.extend([hebrewProber, logicalHebrewProber, |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
67 | visualHebrewProber]) |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
68 | |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
69 | self.reset() |