Mercurial > repos > shellac > sam_consensus_v3

annotate env/lib/python3.9/site-packages/chardet/sbcharsetprober.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 collections import namedtuple
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
30
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
31 from .charsetprober import CharSetProber
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
32 from .enums import CharacterCategory, ProbingState, SequenceLikelihood
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
33
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
34
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
35 SingleByteCharSetModel = namedtuple('SingleByteCharSetModel',
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
36 ['charset_name',
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
37 'language',
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
38 'char_to_order_map',
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
39 'language_model',
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
40 'typical_positive_ratio',
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
41 'keep_ascii_letters',
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
42 'alphabet'])
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
43
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
44
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
45 class SingleByteCharSetProber(CharSetProber):
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
46 SAMPLE_SIZE = 64
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
47 SB_ENOUGH_REL_THRESHOLD = 1024 # 0.25 * SAMPLE_SIZE^2
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
48 POSITIVE_SHORTCUT_THRESHOLD = 0.95
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
49 NEGATIVE_SHORTCUT_THRESHOLD = 0.05
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
50
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
51 def __init__(self, model, reversed=False, name_prober=None):
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
52 super(SingleByteCharSetProber, self).__init__()
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
53 self._model = model
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
54 # TRUE if we need to reverse every pair in the model lookup
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
55 self._reversed = reversed
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
56 # Optional auxiliary prober for name decision
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
57 self._name_prober = name_prober
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
58 self._last_order = None
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
59 self._seq_counters = None
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
60 self._total_seqs = None
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
61 self._total_char = None
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
62 self._freq_char = None
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
63 self.reset()
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
64
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
65 def reset(self):
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
66 super(SingleByteCharSetProber, self).reset()
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
67 # char order of last character
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
68 self._last_order = 255
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
69 self._seq_counters = [0] * SequenceLikelihood.get_num_categories()
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
70 self._total_seqs = 0
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
71 self._total_char = 0
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
72 # characters that fall in our sampling range
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
73 self._freq_char = 0
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
74
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
75 @property
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
76 def charset_name(self):
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
77 if self._name_prober:
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
78 return self._name_prober.charset_name
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
79 else:
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
80 return self._model.charset_name
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
81
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
82 @property
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
83 def language(self):
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
84 if self._name_prober:
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
85 return self._name_prober.language
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
86 else:
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
87 return self._model.language
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
88
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
89 def feed(self, byte_str):
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
90 # TODO: Make filter_international_words keep things in self.alphabet
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
91 if not self._model.keep_ascii_letters:
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
92 byte_str = self.filter_international_words(byte_str)
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
93 if not byte_str:
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
94 return self.state
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
95 char_to_order_map = self._model.char_to_order_map
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
96 language_model = self._model.language_model
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
97 for char in byte_str:
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
98 order = char_to_order_map.get(char, CharacterCategory.UNDEFINED)
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
99 # XXX: This was SYMBOL_CAT_ORDER before, with a value of 250, but
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
100 # CharacterCategory.SYMBOL is actually 253, so we use CONTROL
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
101 # to make it closer to the original intent. The only difference
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
102 # is whether or not we count digits and control characters for
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
103 # _total_char purposes.
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
104 if order < CharacterCategory.CONTROL:
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
105 self._total_char += 1
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
106 # TODO: Follow uchardet's lead and discount confidence for frequent
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
107 # control characters.
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
108 # See https://github.com/BYVoid/uchardet/commit/55b4f23971db61
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
109 if order < self.SAMPLE_SIZE:
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
110 self._freq_char += 1
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
111 if self._last_order < self.SAMPLE_SIZE:
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
112 self._total_seqs += 1
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
113 if not self._reversed:
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
114 lm_cat = language_model[self._last_order][order]
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
115 else:
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
116 lm_cat = language_model[order][self._last_order]
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
117 self._seq_counters[lm_cat] += 1
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
118 self._last_order = order
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
119
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
120 charset_name = self._model.charset_name
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
121 if self.state == ProbingState.DETECTING:
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
122 if self._total_seqs > self.SB_ENOUGH_REL_THRESHOLD:
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
123 confidence = self.get_confidence()
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
124 if confidence > self.POSITIVE_SHORTCUT_THRESHOLD:
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
125 self.logger.debug('%s confidence = %s, we have a winner',
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
126 charset_name, confidence)
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
127 self._state = ProbingState.FOUND_IT
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
128 elif confidence < self.NEGATIVE_SHORTCUT_THRESHOLD:
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
129 self.logger.debug('%s confidence = %s, below negative '
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
130 'shortcut threshhold %s', charset_name,
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
131 confidence,
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
132 self.NEGATIVE_SHORTCUT_THRESHOLD)
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
133 self._state = ProbingState.NOT_ME
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
134
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
135 return self.state
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
136
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
137 def get_confidence(self):
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
138 r = 0.01
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
139 if self._total_seqs > 0:
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
140 r = ((1.0 * self._seq_counters[SequenceLikelihood.POSITIVE]) /
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
141 self._total_seqs / self._model.typical_positive_ratio)
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
142 r = r * self._freq_char / self._total_char
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
143 if r >= 1.0:
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
144 r = 0.99
4f3585e2f14b "planemo upload commit 60cee0fc7c0cda8592644e1aad72851dec82c959"
shellac
parents:
diff changeset
145 return r