diff options
Diffstat (limited to 'src/pip/_vendor/chardet/sbcharsetprober.py')
| -rw-r--r-- | src/pip/_vendor/chardet/sbcharsetprober.py | 73 |
1 files changed, 44 insertions, 29 deletions
diff --git a/src/pip/_vendor/chardet/sbcharsetprober.py b/src/pip/_vendor/chardet/sbcharsetprober.py index 46ba835c6..31d70e154 100644 --- a/src/pip/_vendor/chardet/sbcharsetprober.py +++ b/src/pip/_vendor/chardet/sbcharsetprober.py @@ -31,44 +31,49 @@ from collections import namedtuple from .charsetprober import CharSetProber from .enums import CharacterCategory, ProbingState, SequenceLikelihood - -SingleByteCharSetModel = namedtuple('SingleByteCharSetModel', - ['charset_name', - 'language', - 'char_to_order_map', - 'language_model', - 'typical_positive_ratio', - 'keep_ascii_letters', - 'alphabet']) +SingleByteCharSetModel = namedtuple( + "SingleByteCharSetModel", + [ + "charset_name", + "language", + "char_to_order_map", + "language_model", + "typical_positive_ratio", + "keep_ascii_letters", + "alphabet", + ], +) class SingleByteCharSetProber(CharSetProber): SAMPLE_SIZE = 64 - SB_ENOUGH_REL_THRESHOLD = 1024 # 0.25 * SAMPLE_SIZE^2 + SB_ENOUGH_REL_THRESHOLD = 1024 # 0.25 * SAMPLE_SIZE^2 POSITIVE_SHORTCUT_THRESHOLD = 0.95 NEGATIVE_SHORTCUT_THRESHOLD = 0.05 - def __init__(self, model, reversed=False, name_prober=None): - super(SingleByteCharSetProber, self).__init__() + def __init__(self, model, is_reversed=False, name_prober=None): + super().__init__() self._model = model # TRUE if we need to reverse every pair in the model lookup - self._reversed = reversed + self._reversed = is_reversed # Optional auxiliary prober for name decision self._name_prober = name_prober self._last_order = None self._seq_counters = None self._total_seqs = None self._total_char = None + self._control_char = None self._freq_char = None self.reset() def reset(self): - super(SingleByteCharSetProber, self).reset() + super().reset() # char order of last character self._last_order = 255 self._seq_counters = [0] * SequenceLikelihood.get_num_categories() self._total_seqs = 0 self._total_char = 0 + self._control_char = 0 # characters that fall in our sampling range self._freq_char = 0 @@ -76,20 +81,20 @@ class SingleByteCharSetProber(CharSetProber): def charset_name(self): if self._name_prober: return self._name_prober.charset_name - else: - return self._model.charset_name + return self._model.charset_name @property def language(self): if self._name_prober: return self._name_prober.language - else: - return self._model.language + return self._model.language def feed(self, byte_str): # TODO: Make filter_international_words keep things in self.alphabet if not self._model.keep_ascii_letters: byte_str = self.filter_international_words(byte_str) + else: + byte_str = self.remove_xml_tags(byte_str) if not byte_str: return self.state char_to_order_map = self._model.char_to_order_map @@ -103,9 +108,6 @@ class SingleByteCharSetProber(CharSetProber): # _total_char purposes. if order < CharacterCategory.CONTROL: self._total_char += 1 - # TODO: Follow uchardet's lead and discount confidence for frequent - # control characters. - # See https://github.com/BYVoid/uchardet/commit/55b4f23971db61 if order < self.SAMPLE_SIZE: self._freq_char += 1 if self._last_order < self.SAMPLE_SIZE: @@ -122,14 +124,17 @@ class SingleByteCharSetProber(CharSetProber): if self._total_seqs > self.SB_ENOUGH_REL_THRESHOLD: confidence = self.get_confidence() if confidence > self.POSITIVE_SHORTCUT_THRESHOLD: - self.logger.debug('%s confidence = %s, we have a winner', - charset_name, confidence) + self.logger.debug( + "%s confidence = %s, we have a winner", charset_name, confidence + ) self._state = ProbingState.FOUND_IT elif confidence < self.NEGATIVE_SHORTCUT_THRESHOLD: - self.logger.debug('%s confidence = %s, below negative ' - 'shortcut threshhold %s', charset_name, - confidence, - self.NEGATIVE_SHORTCUT_THRESHOLD) + self.logger.debug( + "%s confidence = %s, below negative shortcut threshold %s", + charset_name, + confidence, + self.NEGATIVE_SHORTCUT_THRESHOLD, + ) self._state = ProbingState.NOT_ME return self.state @@ -137,8 +142,18 @@ class SingleByteCharSetProber(CharSetProber): def get_confidence(self): r = 0.01 if self._total_seqs > 0: - r = ((1.0 * self._seq_counters[SequenceLikelihood.POSITIVE]) / - self._total_seqs / self._model.typical_positive_ratio) + r = ( + ( + self._seq_counters[SequenceLikelihood.POSITIVE] + + 0.25 * self._seq_counters[SequenceLikelihood.LIKELY] + ) + / self._total_seqs + / self._model.typical_positive_ratio + ) + # The more control characters (proportionnaly to the size + # of the text), the less confident we become in the current + # charset. + r = r * (self._total_char - self._control_char) / self._total_char r = r * self._freq_char / self._total_char if r >= 1.0: r = 0.99 |