Mercurial > repos > shellac > sam_consensus_v3

annotate env/lib/python3.9/site-packages/chardet/sbcsgroupprober.py @ 0:4f3585e2f14b draft default tip

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
"planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
author shellac
date Mon, 22 Mar 2021 18:12:50 +0000
parents
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
0
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
1 ######################## BEGIN LICENSE BLOCK ########################
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
2 # The Original Code is Mozilla Universal charset detector code.
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
3 #
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
4 # The Initial Developer of the Original Code is
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
5 # Netscape Communications Corporation.
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
6 # Portions created by the Initial Developer are Copyright (C) 2001
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
7 # the Initial Developer. All Rights Reserved.
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
8 #
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
9 # Contributor(s):
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
10 # Mark Pilgrim - port to Python
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
11 # Shy Shalom - original C code
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
12 #
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
13 # This library is free software; you can redistribute it and/or
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
14 # modify it under the terms of the GNU Lesser General Public
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
15 # License as published by the Free Software Foundation; either
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
16 # version 2.1 of the License, or (at your option) any later version.
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
17 #
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
18 # This library is distributed in the hope that it will be useful,
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
19 # but WITHOUT ANY WARRANTY; without even the implied warranty of
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
20 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
21 # Lesser General Public License for more details.
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
22 #
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
23 # You should have received a copy of the GNU Lesser General Public
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
24 # License along with this library; if not, write to the Free Software
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
25 # Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
26 # 02110-1301 USA
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
27 ######################### END LICENSE BLOCK #########################
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
28
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
29 from .charsetgroupprober import CharSetGroupProber
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
30 from .hebrewprober import HebrewProber
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
31 from .langbulgarianmodel import (ISO_8859_5_BULGARIAN_MODEL,
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
32 WINDOWS_1251_BULGARIAN_MODEL)
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
33 from .langgreekmodel import ISO_8859_7_GREEK_MODEL, WINDOWS_1253_GREEK_MODEL
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
34 from .langhebrewmodel import WINDOWS_1255_HEBREW_MODEL
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
35 # from .langhungarianmodel import (ISO_8859_2_HUNGARIAN_MODEL,
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
36 # WINDOWS_1250_HUNGARIAN_MODEL)
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
37 from .langrussianmodel import (IBM855_RUSSIAN_MODEL, IBM866_RUSSIAN_MODEL,
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
38 ISO_8859_5_RUSSIAN_MODEL, KOI8_R_RUSSIAN_MODEL,
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
39 MACCYRILLIC_RUSSIAN_MODEL,
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
40 WINDOWS_1251_RUSSIAN_MODEL)
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
41 from .langthaimodel import TIS_620_THAI_MODEL
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
42 from .langturkishmodel import ISO_8859_9_TURKISH_MODEL
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
43 from .sbcharsetprober import SingleByteCharSetProber
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
44
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
45
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
46 class SBCSGroupProber(CharSetGroupProber):
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
47 def __init__(self):
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
48 super(SBCSGroupProber, self).__init__()
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
49 hebrew_prober = HebrewProber()
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
50 logical_hebrew_prober = SingleByteCharSetProber(WINDOWS_1255_HEBREW_MODEL,
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
51 False, hebrew_prober)
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
52 # TODO: See if using ISO-8859-8 Hebrew model works better here, since
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
53 # it's actually the visual one
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
54 visual_hebrew_prober = SingleByteCharSetProber(WINDOWS_1255_HEBREW_MODEL,
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
55 True, hebrew_prober)
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
56 hebrew_prober.set_model_probers(logical_hebrew_prober,
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
57 visual_hebrew_prober)
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
58 # TODO: ORDER MATTERS HERE. I changed the order vs what was in master
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
59 # and several tests failed that did not before. Some thought
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
60 # should be put into the ordering, and we should consider making
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
61 # order not matter here, because that is very counter-intuitive.
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
62 self.probers = [
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
63 SingleByteCharSetProber(WINDOWS_1251_RUSSIAN_MODEL),
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
64 SingleByteCharSetProber(KOI8_R_RUSSIAN_MODEL),
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
65 SingleByteCharSetProber(ISO_8859_5_RUSSIAN_MODEL),
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
66 SingleByteCharSetProber(MACCYRILLIC_RUSSIAN_MODEL),
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
67 SingleByteCharSetProber(IBM866_RUSSIAN_MODEL),
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
68 SingleByteCharSetProber(IBM855_RUSSIAN_MODEL),
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
69 SingleByteCharSetProber(ISO_8859_7_GREEK_MODEL),
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
70 SingleByteCharSetProber(WINDOWS_1253_GREEK_MODEL),
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
71 SingleByteCharSetProber(ISO_8859_5_BULGARIAN_MODEL),
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
72 SingleByteCharSetProber(WINDOWS_1251_BULGARIAN_MODEL),
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
73 # TODO: Restore Hungarian encodings (iso-8859-2 and windows-1250)
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
74 # after we retrain model.
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
75 # SingleByteCharSetProber(ISO_8859_2_HUNGARIAN_MODEL),
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
76 # SingleByteCharSetProber(WINDOWS_1250_HUNGARIAN_MODEL),
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
77 SingleByteCharSetProber(TIS_620_THAI_MODEL),
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
78 SingleByteCharSetProber(ISO_8859_9_TURKISH_MODEL),
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
79 hebrew_prober,
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
80 logical_hebrew_prober,
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
81 visual_hebrew_prober,
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
82 ]
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
83 self.reset()