Mercurial > repos > guerler > springsuite
annotate planemo/lib/python3.7/site-packages/chardet/charsetgroupprober.py @ 0:d30785e31577 draft
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
| author | guerler | 
|---|---|
| date | Fri, 31 Jul 2020 00:18:57 -0400 | 
| parents | |
| children | 
| rev | line source | 
|---|---|
| 
0
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
1 ######################## BEGIN LICENSE BLOCK ######################## | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
2 # The Original Code is Mozilla Communicator client code. | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
3 # | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
4 # The Initial Developer of the Original Code is | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
5 # Netscape Communications Corporation. | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
6 # Portions created by the Initial Developer are Copyright (C) 1998 | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
7 # the Initial Developer. All Rights Reserved. | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
8 # | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
9 # Contributor(s): | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
10 # Mark Pilgrim - port to Python | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
11 # | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
12 # This library is free software; you can redistribute it and/or | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
13 # modify it under the terms of the GNU Lesser General Public | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
14 # License as published by the Free Software Foundation; either | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
15 # version 2.1 of the License, or (at your option) any later version. | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
16 # | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
17 # This library is distributed in the hope that it will be useful, | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
18 # but WITHOUT ANY WARRANTY; without even the implied warranty of | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
19 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
20 # Lesser General Public License for more details. | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
21 # | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
22 # You should have received a copy of the GNU Lesser General Public | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
23 # License along with this library; if not, write to the Free Software | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
24 # Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
25 # 02110-1301 USA | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
26 ######################### END LICENSE BLOCK ######################### | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
27 | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
28 from .enums import ProbingState | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
29 from .charsetprober import CharSetProber | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
30 | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
31 | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
32 class CharSetGroupProber(CharSetProber): | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
33 def __init__(self, lang_filter=None): | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
34 super(CharSetGroupProber, self).__init__(lang_filter=lang_filter) | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
35 self._active_num = 0 | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
36 self.probers = [] | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
37 self._best_guess_prober = None | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
38 | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
39 def reset(self): | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
40 super(CharSetGroupProber, self).reset() | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
41 self._active_num = 0 | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
42 for prober in self.probers: | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
43 if prober: | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
44 prober.reset() | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
45 prober.active = True | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
46 self._active_num += 1 | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
47 self._best_guess_prober = None | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
48 | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
49 @property | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
50 def charset_name(self): | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
51 if not self._best_guess_prober: | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
52 self.get_confidence() | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
53 if not self._best_guess_prober: | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
54 return None | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
55 return self._best_guess_prober.charset_name | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
56 | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
57 @property | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
58 def language(self): | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
59 if not self._best_guess_prober: | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
60 self.get_confidence() | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
61 if not self._best_guess_prober: | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
62 return None | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
63 return self._best_guess_prober.language | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
64 | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
65 def feed(self, byte_str): | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
66 for prober in self.probers: | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
67 if not prober: | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
68 continue | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
69 if not prober.active: | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
70 continue | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
71 state = prober.feed(byte_str) | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
72 if not state: | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
73 continue | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
74 if state == ProbingState.FOUND_IT: | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
75 self._best_guess_prober = prober | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
76 return self.state | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
77 elif state == ProbingState.NOT_ME: | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
78 prober.active = False | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
79 self._active_num -= 1 | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
80 if self._active_num <= 0: | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
81 self._state = ProbingState.NOT_ME | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
82 return self.state | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
83 return self.state | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
84 | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
85 def get_confidence(self): | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
86 state = self.state | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
87 if state == ProbingState.FOUND_IT: | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
88 return 0.99 | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
89 elif state == ProbingState.NOT_ME: | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
90 return 0.01 | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
91 best_conf = 0.0 | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
92 self._best_guess_prober = None | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
93 for prober in self.probers: | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
94 if not prober: | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
95 continue | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
96 if not prober.active: | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
97 self.logger.debug('%s not active', prober.charset_name) | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
98 continue | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
99 conf = prober.get_confidence() | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
100 self.logger.debug('%s %s confidence = %s', prober.charset_name, prober.language, conf) | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
101 if best_conf < conf: | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
102 best_conf = conf | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
103 self._best_guess_prober = prober | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
104 if not self._best_guess_prober: | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
105 return 0.0 | 
| 
 
d30785e31577
"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
 
guerler 
parents:  
diff
changeset
 | 
106 return best_conf | 
