Mercurial > repos > guerler > hhblits

annotate lib/python3.8/site-packages/pip/_vendor/chardet/utf8prober.py @ 0:9e54283cc701 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
"planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
author guerler
date Mon, 27 Jul 2020 03:47:31 -0400
parents
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
0
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
1 ######################## BEGIN LICENSE BLOCK ########################
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
2 # The Original Code is mozilla.org code.
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
3 #
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
4 # The Initial Developer of the Original Code is
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
5 # Netscape Communications Corporation.
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
6 # Portions created by the Initial Developer are Copyright (C) 1998
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
7 # the Initial Developer. All Rights Reserved.
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
8 #
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
9 # Contributor(s):
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
10 # Mark Pilgrim - port to Python
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
11 #
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
12 # This library is free software; you can redistribute it and/or
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
13 # modify it under the terms of the GNU Lesser General Public
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
14 # License as published by the Free Software Foundation; either
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
15 # version 2.1 of the License, or (at your option) any later version.
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
16 #
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
17 # This library is distributed in the hope that it will be useful,
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
18 # but WITHOUT ANY WARRANTY; without even the implied warranty of
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
19 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
20 # Lesser General Public License for more details.
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
21 #
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
22 # You should have received a copy of the GNU Lesser General Public
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
23 # License along with this library; if not, write to the Free Software
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
24 # Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
25 # 02110-1301 USA
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
26 ######################### END LICENSE BLOCK #########################
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
27
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
28 from .charsetprober import CharSetProber
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
29 from .enums import ProbingState, MachineState
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
30 from .codingstatemachine import CodingStateMachine
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
31 from .mbcssm import UTF8_SM_MODEL
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
32
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
33
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
34
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
35 class UTF8Prober(CharSetProber):
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
36 ONE_CHAR_PROB = 0.5
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
37
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
38 def __init__(self):
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
39 super(UTF8Prober, self).__init__()
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
40 self.coding_sm = CodingStateMachine(UTF8_SM_MODEL)
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
41 self._num_mb_chars = None
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
42 self.reset()
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
43
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
44 def reset(self):
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
45 super(UTF8Prober, self).reset()
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
46 self.coding_sm.reset()
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
47 self._num_mb_chars = 0
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
48
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
49 @property
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
50 def charset_name(self):
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
51 return "utf-8"
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
52
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
53 @property
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
54 def language(self):
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
55 return ""
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
56
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
57 def feed(self, byte_str):
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
58 for c in byte_str:
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
59 coding_state = self.coding_sm.next_state(c)
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
60 if coding_state == MachineState.ERROR:
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
61 self._state = ProbingState.NOT_ME
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
62 break
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
63 elif coding_state == MachineState.ITS_ME:
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
64 self._state = ProbingState.FOUND_IT
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
65 break
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
66 elif coding_state == MachineState.START:
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
67 if self.coding_sm.get_current_charlen() >= 2:
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
68 self._num_mb_chars += 1
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
69
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
70 if self.state == ProbingState.DETECTING:
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
71 if self.get_confidence() > self.SHORTCUT_THRESHOLD:
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
72 self._state = ProbingState.FOUND_IT
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
73
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
74 return self.state
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
75
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
76 def get_confidence(self):
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
77 unlike = 0.99
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
78 if self._num_mb_chars < 6:
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
79 unlike *= self.ONE_CHAR_PROB ** self._num_mb_chars
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
80 return 1.0 - unlike
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
81 else:
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
82 return unlike