From 4aab4087dc97906d0b9890035401175cdaab32d4 Mon Sep 17 00:00:00 2001 From: blackhao <13851610112@163.com> Date: Fri, 22 Aug 2025 02:51:50 -0500 Subject: 2.0 --- .../site-packages/~ip/_vendor/chardet/enums.py | 85 ++++++++++++++++++++++ 1 file changed, 85 insertions(+) create mode 100644 .venv/lib/python3.12/site-packages/~ip/_vendor/chardet/enums.py (limited to '.venv/lib/python3.12/site-packages/~ip/_vendor/chardet/enums.py') diff --git a/.venv/lib/python3.12/site-packages/~ip/_vendor/chardet/enums.py b/.venv/lib/python3.12/site-packages/~ip/_vendor/chardet/enums.py new file mode 100644 index 0000000..5e3e198 --- /dev/null +++ b/.venv/lib/python3.12/site-packages/~ip/_vendor/chardet/enums.py @@ -0,0 +1,85 @@ +""" +All of the Enums that are used throughout the chardet package. + +:author: Dan Blanchard (dan.blanchard@gmail.com) +""" + +from enum import Enum, Flag + + +class InputState: + """ + This enum represents the different states a universal detector can be in. + """ + + PURE_ASCII = 0 + ESC_ASCII = 1 + HIGH_BYTE = 2 + + +class LanguageFilter(Flag): + """ + This enum represents the different language filters we can apply to a + ``UniversalDetector``. + """ + + NONE = 0x00 + CHINESE_SIMPLIFIED = 0x01 + CHINESE_TRADITIONAL = 0x02 + JAPANESE = 0x04 + KOREAN = 0x08 + NON_CJK = 0x10 + ALL = 0x1F + CHINESE = CHINESE_SIMPLIFIED | CHINESE_TRADITIONAL + CJK = CHINESE | JAPANESE | KOREAN + + +class ProbingState(Enum): + """ + This enum represents the different states a prober can be in. + """ + + DETECTING = 0 + FOUND_IT = 1 + NOT_ME = 2 + + +class MachineState: + """ + This enum represents the different states a state machine can be in. + """ + + START = 0 + ERROR = 1 + ITS_ME = 2 + + +class SequenceLikelihood: + """ + This enum represents the likelihood of a character following the previous one. + """ + + NEGATIVE = 0 + UNLIKELY = 1 + LIKELY = 2 + POSITIVE = 3 + + @classmethod + def get_num_categories(cls) -> int: + """:returns: The number of likelihood categories in the enum.""" + return 4 + + +class CharacterCategory: + """ + This enum represents the different categories language models for + ``SingleByteCharsetProber`` put characters into. + + Anything less than CONTROL is considered a letter. + """ + + UNDEFINED = 255 + LINE_BREAK = 254 + SYMBOL = 253 + DIGIT = 252 + CONTROL = 251 -- cgit v1.2.3