Mirrors/NonSteamLaunchers-On-Steam-Deck
2
0
Fork 0
mirror of https://github.com/moraroy/NonSteamLaunchers-On-Steam-Deck.git synced 2024-12-20 23:01:53 +01:00
Code Issues Packages Projects Releases Wiki Activity

Add files via upload

Browse Source
This commit is contained in:
Roy 2024-06-30 18:30:38 -07:00 committed by GitHub
parent 4130f1139b
commit 73e3663ce3
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
11 changed files with 4500 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

45
Modules/charset_normalizer/__init__.py
View File

@ -1 +1,46 @@
# -*- coding: utf-8 -*-
"""
Charset-Normalizer
~~~~~~~~~~~~~~
The Real First Universal Charset Detector.
A library that helps you read text from an unknown charset encoding.
Motivated by chardet, This package is trying to resolve the issue by taking a new approach.
All IANA character set names for which the Python core library provides codecs are supported.
Basic usage:
>>> from charset_normalizer import from_bytes
>>> results = from_bytes('Bсеки човек има право на образование. Oбразованието!'.encode('utf_8'))
>>> best_guess = results.best()
>>> str(best_guess)
'Bсеки човек има право на образование. Oбразованието!'
Others methods and usages are available - see the full documentation
at <https://github.com/Ousret/charset_normalizer>.
:copyright: (c) 2021 by Ahmed TAHRI
:license: MIT, see LICENSE for more details.
"""
import logging
from .api import from_bytes, from_fp, from_path, is_binary
from .legacy import detect
from .models import CharsetMatch, CharsetMatches
from .utils import set_logging_handler
from .version import VERSION, __version__
__all__ = (
"from_fp",
"from_path",
"from_bytes",
"is_binary",
"detect",
"CharsetMatch",
"CharsetMatches",
"__version__",
"VERSION",
"set_logging_handler",
)
# Attach a NullHandler to the top level logger by default
# https://docs.python.org/3.3/howto/logging.html#configuring-logging-for-a-library
logging.getLogger("charset_normalizer").addHandler(logging.NullHandler())

3
Modules/charset_normalizer/__main__.py
View File

@ -1 +1,4 @@
from .cli import cli_detect
if __name__ == "__main__":
cli_detect()

626
Modules/charset_normalizer/api.py Normal file
View File

@ -0,0 +1,626 @@
import logging
from os import PathLike
from typing import BinaryIO, List, Optional, Set, Union
from .cd import (
coherence_ratio,
encoding_languages,
mb_encoding_languages,
merge_coherence_ratios,
)
from .constant import IANA_SUPPORTED, TOO_BIG_SEQUENCE, TOO_SMALL_SEQUENCE, TRACE
from .md import mess_ratio
from .models import CharsetMatch, CharsetMatches
from .utils import (
any_specified_encoding,
cut_sequence_chunks,
iana_name,
identify_sig_or_bom,
is_cp_similar,
is_multi_byte_encoding,
should_strip_sig_or_bom,
)
# Will most likely be controversial
# logging.addLevelName(TRACE, "TRACE")
logger = logging.getLogger("charset_normalizer")
explain_handler = logging.StreamHandler()
explain_handler.setFormatter(
logging.Formatter("%(asctime)s | %(levelname)s | %(message)s")
)
def from_bytes(
sequences: Union[bytes, bytearray],
steps: int = 5,
chunk_size: int = 512,
threshold: float = 0.2,
cp_isolation: Optional[List[str]] = None,
cp_exclusion: Optional[List[str]] = None,
preemptive_behaviour: bool = True,
explain: bool = False,
language_threshold: float = 0.1,
enable_fallback: bool = True,
) -> CharsetMatches:
"""
Given a raw bytes sequence, return the best possibles charset usable to render str objects.
If there is no results, it is a strong indicator that the source is binary/not text.
By default, the process will extract 5 blocks of 512o each to assess the mess and coherence of a given sequence.
And will give up a particular code page after 20% of measured mess. Those criteria are customizable at will.
The preemptive behavior DOES NOT replace the traditional detection workflow, it prioritize a particular code page
but never take it for granted. Can improve the performance.
You may want to focus your attention to some code page or/and not others, use cp_isolation and cp_exclusion for that
purpose.
This function will strip the SIG in the payload/sequence every time except on UTF-16, UTF-32.
By default the library does not setup any handler other than the NullHandler, if you choose to set the 'explain'
toggle to True it will alter the logger configuration to add a StreamHandler that is suitable for debugging.
Custom logging format and handler can be set manually.
"""
if not isinstance(sequences, (bytearray, bytes)):
raise TypeError(
"Expected object of type bytes or bytearray, got: {0}".format(
type(sequences)
)
)
if explain:
previous_logger_level: int = logger.level
logger.addHandler(explain_handler)
logger.setLevel(TRACE)
length: int = len(sequences)
if length == 0:
logger.debug("Encoding detection on empty bytes, assuming utf_8 intention.")
if explain:
logger.removeHandler(explain_handler)
logger.setLevel(previous_logger_level or logging.WARNING)
return CharsetMatches([CharsetMatch(sequences, "utf_8", 0.0, False, [], "")])
if cp_isolation is not None:
logger.log(
TRACE,
"cp_isolation is set. use this flag for debugging purpose. "
"limited list of encoding allowed : %s.",
", ".join(cp_isolation),
)
cp_isolation = [iana_name(cp, False) for cp in cp_isolation]
else:
cp_isolation = []
if cp_exclusion is not None:
logger.log(
TRACE,
"cp_exclusion is set. use this flag for debugging purpose. "
"limited list of encoding excluded : %s.",
", ".join(cp_exclusion),
)
cp_exclusion = [iana_name(cp, False) for cp in cp_exclusion]
else:
cp_exclusion = []
if length <= (chunk_size * steps):
logger.log(
TRACE,
"override steps (%i) and chunk_size (%i) as content does not fit (%i byte(s) given) parameters.",
steps,
chunk_size,
length,
)
steps = 1
chunk_size = length
if steps > 1 and length / steps < chunk_size:
chunk_size = int(length / steps)
is_too_small_sequence: bool = len(sequences) < TOO_SMALL_SEQUENCE
is_too_large_sequence: bool = len(sequences) >= TOO_BIG_SEQUENCE
if is_too_small_sequence:
logger.log(
TRACE,
"Trying to detect encoding from a tiny portion of ({}) byte(s).".format(
length
),
)
elif is_too_large_sequence:
logger.log(
TRACE,
"Using lazy str decoding because the payload is quite large, ({}) byte(s).".format(
length
),
)
prioritized_encodings: List[str] = []
specified_encoding: Optional[str] = (
any_specified_encoding(sequences) if preemptive_behaviour else None
)
if specified_encoding is not None:
prioritized_encodings.append(specified_encoding)
logger.log(
TRACE,
"Detected declarative mark in sequence. Priority +1 given for %s.",
specified_encoding,
)
tested: Set[str] = set()
tested_but_hard_failure: List[str] = []
tested_but_soft_failure: List[str] = []
fallback_ascii: Optional[CharsetMatch] = None
fallback_u8: Optional[CharsetMatch] = None
fallback_specified: Optional[CharsetMatch] = None
results: CharsetMatches = CharsetMatches()
sig_encoding, sig_payload = identify_sig_or_bom(sequences)
if sig_encoding is not None:
prioritized_encodings.append(sig_encoding)
logger.log(
TRACE,
"Detected a SIG or BOM mark on first %i byte(s). Priority +1 given for %s.",
len(sig_payload),
sig_encoding,
)
prioritized_encodings.append("ascii")
if "utf_8" not in prioritized_encodings:
prioritized_encodings.append("utf_8")
for encoding_iana in prioritized_encodings + IANA_SUPPORTED:
if cp_isolation and encoding_iana not in cp_isolation:
continue
if cp_exclusion and encoding_iana in cp_exclusion:
continue
if encoding_iana in tested:
continue
tested.add(encoding_iana)
decoded_payload: Optional[str] = None
bom_or_sig_available: bool = sig_encoding == encoding_iana
strip_sig_or_bom: bool = bom_or_sig_available and should_strip_sig_or_bom(
encoding_iana
)
if encoding_iana in {"utf_16", "utf_32"} and not bom_or_sig_available:
logger.log(
TRACE,
"Encoding %s won't be tested as-is because it require a BOM. Will try some sub-encoder LE/BE.",
encoding_iana,
)
continue
if encoding_iana in {"utf_7"} and not bom_or_sig_available:
logger.log(
TRACE,
"Encoding %s won't be tested as-is because detection is unreliable without BOM/SIG.",
encoding_iana,
)
continue
try:
is_multi_byte_decoder: bool = is_multi_byte_encoding(encoding_iana)
except (ModuleNotFoundError, ImportError):
logger.log(
TRACE,
"Encoding %s does not provide an IncrementalDecoder",
encoding_iana,
)
continue
try:
if is_too_large_sequence and is_multi_byte_decoder is False:
str(
sequences[: int(50e4)]
if strip_sig_or_bom is False
else sequences[len(sig_payload) : int(50e4)],
encoding=encoding_iana,
)
else:
decoded_payload = str(
sequences
if strip_sig_or_bom is False
else sequences[len(sig_payload) :],
encoding=encoding_iana,
)
except (UnicodeDecodeError, LookupError) as e:
if not isinstance(e, LookupError):
logger.log(
TRACE,
"Code page %s does not fit given bytes sequence at ALL. %s",
encoding_iana,
str(e),
)
tested_but_hard_failure.append(encoding_iana)
continue
similar_soft_failure_test: bool = False
for encoding_soft_failed in tested_but_soft_failure:
if is_cp_similar(encoding_iana, encoding_soft_failed):
similar_soft_failure_test = True
break
if similar_soft_failure_test:
logger.log(
TRACE,
"%s is deemed too similar to code page %s and was consider unsuited already. Continuing!",
encoding_iana,
encoding_soft_failed,
)
continue
r_ = range(
0 if not bom_or_sig_available else len(sig_payload),
length,
int(length / steps),
)
multi_byte_bonus: bool = (
is_multi_byte_decoder
and decoded_payload is not None
and len(decoded_payload) < length
)
if multi_byte_bonus:
logger.log(
TRACE,
"Code page %s is a multi byte encoding table and it appear that at least one character "
"was encoded using n-bytes.",
encoding_iana,
)
max_chunk_gave_up: int = int(len(r_) / 4)
max_chunk_gave_up = max(max_chunk_gave_up, 2)
early_stop_count: int = 0
lazy_str_hard_failure = False
md_chunks: List[str] = []
md_ratios = []
try:
for chunk in cut_sequence_chunks(
sequences,
encoding_iana,
r_,
chunk_size,
bom_or_sig_available,
strip_sig_or_bom,
sig_payload,
is_multi_byte_decoder,
decoded_payload,
):
md_chunks.append(chunk)
md_ratios.append(
mess_ratio(
chunk,
threshold,
explain is True and 1 <= len(cp_isolation) <= 2,
)
)
if md_ratios[-1] >= threshold:
early_stop_count += 1
if (early_stop_count >= max_chunk_gave_up) or (
bom_or_sig_available and strip_sig_or_bom is False
):
break
except (
UnicodeDecodeError
) as e: # Lazy str loading may have missed something there
logger.log(
TRACE,
"LazyStr Loading: After MD chunk decode, code page %s does not fit given bytes sequence at ALL. %s",
encoding_iana,
str(e),
)
early_stop_count = max_chunk_gave_up
lazy_str_hard_failure = True
# We might want to check the sequence again with the whole content
# Only if initial MD tests passes
if (
not lazy_str_hard_failure
and is_too_large_sequence
and not is_multi_byte_decoder
):
try:
sequences[int(50e3) :].decode(encoding_iana, errors="strict")
except UnicodeDecodeError as e:
logger.log(
TRACE,
"LazyStr Loading: After final lookup, code page %s does not fit given bytes sequence at ALL. %s",
encoding_iana,
str(e),
)
tested_but_hard_failure.append(encoding_iana)
continue
mean_mess_ratio: float = sum(md_ratios) / len(md_ratios) if md_ratios else 0.0
if mean_mess_ratio >= threshold or early_stop_count >= max_chunk_gave_up:
tested_but_soft_failure.append(encoding_iana)
logger.log(
TRACE,
"%s was excluded because of initial chaos probing. Gave up %i time(s). "
"Computed mean chaos is %f %%.",
encoding_iana,
early_stop_count,
round(mean_mess_ratio * 100, ndigits=3),
)
# Preparing those fallbacks in case we got nothing.
if (
enable_fallback
and encoding_iana in ["ascii", "utf_8", specified_encoding]
and not lazy_str_hard_failure
):
fallback_entry = CharsetMatch(
sequences, encoding_iana, threshold, False, [], decoded_payload
)
if encoding_iana == specified_encoding:
fallback_specified = fallback_entry
elif encoding_iana == "ascii":
fallback_ascii = fallback_entry
else:
fallback_u8 = fallback_entry
continue
logger.log(
TRACE,
"%s passed initial chaos probing. Mean measured chaos is %f %%",
encoding_iana,
round(mean_mess_ratio * 100, ndigits=3),
)
if not is_multi_byte_decoder:
target_languages: List[str] = encoding_languages(encoding_iana)
else:
target_languages = mb_encoding_languages(encoding_iana)
if target_languages:
logger.log(
TRACE,
"{} should target any language(s) of {}".format(
encoding_iana, str(target_languages)
),
)
cd_ratios = []
# We shall skip the CD when its about ASCII
# Most of the time its not relevant to run "language-detection" on it.
if encoding_iana != "ascii":
for chunk in md_chunks:
chunk_languages = coherence_ratio(
chunk,
language_threshold,
",".join(target_languages) if target_languages else None,
)
cd_ratios.append(chunk_languages)
cd_ratios_merged = merge_coherence_ratios(cd_ratios)
if cd_ratios_merged:
logger.log(
TRACE,
"We detected language {} using {}".format(
cd_ratios_merged, encoding_iana
),
)
results.append(
CharsetMatch(
sequences,
encoding_iana,
mean_mess_ratio,
bom_or_sig_available,
cd_ratios_merged,
decoded_payload,
)
)
if (
encoding_iana in [specified_encoding, "ascii", "utf_8"]
and mean_mess_ratio < 0.1
):
logger.debug(
"Encoding detection: %s is most likely the one.", encoding_iana
)
if explain:
logger.removeHandler(explain_handler)
logger.setLevel(previous_logger_level)
return CharsetMatches([results[encoding_iana]])
if encoding_iana == sig_encoding:
logger.debug(
"Encoding detection: %s is most likely the one as we detected a BOM or SIG within "
"the beginning of the sequence.",
encoding_iana,
)
if explain:
logger.removeHandler(explain_handler)
logger.setLevel(previous_logger_level)
return CharsetMatches([results[encoding_iana]])
if len(results) == 0:
if fallback_u8 or fallback_ascii or fallback_specified:
logger.log(
TRACE,
"Nothing got out of the detection process. Using ASCII/UTF-8/Specified fallback.",
)
if fallback_specified:
logger.debug(
"Encoding detection: %s will be used as a fallback match",
fallback_specified.encoding,
)
results.append(fallback_specified)
elif (
(fallback_u8 and fallback_ascii is None)
or (
fallback_u8
and fallback_ascii
and fallback_u8.fingerprint != fallback_ascii.fingerprint
)
or (fallback_u8 is not None)
):
logger.debug("Encoding detection: utf_8 will be used as a fallback match")
results.append(fallback_u8)
elif fallback_ascii:
logger.debug("Encoding detection: ascii will be used as a fallback match")
results.append(fallback_ascii)
if results:
logger.debug(
"Encoding detection: Found %s as plausible (best-candidate) for content. With %i alternatives.",
results.best().encoding, # type: ignore
len(results) - 1,
)
else:
logger.debug("Encoding detection: Unable to determine any suitable charset.")
if explain:
logger.removeHandler(explain_handler)
logger.setLevel(previous_logger_level)
return results
def from_fp(
fp: BinaryIO,
steps: int = 5,
chunk_size: int = 512,
threshold: float = 0.20,
cp_isolation: Optional[List[str]] = None,
cp_exclusion: Optional[List[str]] = None,
preemptive_behaviour: bool = True,
explain: bool = False,
language_threshold: float = 0.1,
enable_fallback: bool = True,
) -> CharsetMatches:
"""
Same thing than the function from_bytes but using a file pointer that is already ready.
Will not close the file pointer.
"""
return from_bytes(
fp.read(),
steps,
chunk_size,
threshold,
cp_isolation,
cp_exclusion,
preemptive_behaviour,
explain,
language_threshold,
enable_fallback,
)
def from_path(
path: Union[str, bytes, PathLike], # type: ignore[type-arg]
steps: int = 5,
chunk_size: int = 512,
threshold: float = 0.20,
cp_isolation: Optional[List[str]] = None,
cp_exclusion: Optional[List[str]] = None,
preemptive_behaviour: bool = True,
explain: bool = False,
language_threshold: float = 0.1,
enable_fallback: bool = True,
) -> CharsetMatches:
"""
Same thing than the function from_bytes but with one extra step. Opening and reading given file path in binary mode.
Can raise IOError.
"""
with open(path, "rb") as fp:
return from_fp(
fp,
steps,
chunk_size,
threshold,
cp_isolation,
cp_exclusion,
preemptive_behaviour,
explain,
language_threshold,
enable_fallback,
)
def is_binary(
fp_or_path_or_payload: Union[PathLike, str, BinaryIO, bytes], # type: ignore[type-arg]
steps: int = 5,
chunk_size: int = 512,
threshold: float = 0.20,
cp_isolation: Optional[List[str]] = None,
cp_exclusion: Optional[List[str]] = None,
preemptive_behaviour: bool = True,
explain: bool = False,
language_threshold: float = 0.1,
enable_fallback: bool = False,
) -> bool:
"""
Detect if the given input (file, bytes, or path) points to a binary file. aka. not a string.
Based on the same main heuristic algorithms and default kwargs at the sole exception that fallbacks match
are disabled to be stricter around ASCII-compatible but unlikely to be a string.
"""
if isinstance(fp_or_path_or_payload, (str, PathLike)):
guesses = from_path(
fp_or_path_or_payload,
steps=steps,
chunk_size=chunk_size,
threshold=threshold,
cp_isolation=cp_isolation,
cp_exclusion=cp_exclusion,
preemptive_behaviour=preemptive_behaviour,
explain=explain,
language_threshold=language_threshold,
enable_fallback=enable_fallback,
)
elif isinstance(
fp_or_path_or_payload,
(
bytes,
bytearray,
),
):
guesses = from_bytes(
fp_or_path_or_payload,
steps=steps,
chunk_size=chunk_size,
threshold=threshold,
cp_isolation=cp_isolation,
cp_exclusion=cp_exclusion,
preemptive_behaviour=preemptive_behaviour,
explain=explain,
language_threshold=language_threshold,
enable_fallback=enable_fallback,
)
else:
guesses = from_fp(
fp_or_path_or_payload,
steps=steps,
chunk_size=chunk_size,
threshold=threshold,
cp_isolation=cp_isolation,
cp_exclusion=cp_exclusion,
preemptive_behaviour=preemptive_behaviour,
explain=explain,
language_threshold=language_threshold,
enable_fallback=enable_fallback,
)
return not guesses

395
Modules/charset_normalizer/cd.py Normal file
View File

@ -0,0 +1,395 @@
import importlib
from codecs import IncrementalDecoder
from collections import Counter
from functools import lru_cache
from typing import Counter as TypeCounter, Dict, List, Optional, Tuple
from .constant import (
FREQUENCIES,
KO_NAMES,
LANGUAGE_SUPPORTED_COUNT,
TOO_SMALL_SEQUENCE,
ZH_NAMES,
)
from .md import is_suspiciously_successive_range
from .models import CoherenceMatches
from .utils import (
is_accentuated,
is_latin,
is_multi_byte_encoding,
is_unicode_range_secondary,
unicode_range,
)
def encoding_unicode_range(iana_name: str) -> List[str]:
"""
Return associated unicode ranges in a single byte code page.
"""
if is_multi_byte_encoding(iana_name):
raise IOError("Function not supported on multi-byte code page")
decoder = importlib.import_module(
"encodings.{}".format(iana_name)
).IncrementalDecoder
p: IncrementalDecoder = decoder(errors="ignore")
seen_ranges: Dict[str, int] = {}
character_count: int = 0
for i in range(0x40, 0xFF):
chunk: str = p.decode(bytes([i]))
if chunk:
character_range: Optional[str] = unicode_range(chunk)
if character_range is None:
continue
if is_unicode_range_secondary(character_range) is False:
if character_range not in seen_ranges:
seen_ranges[character_range] = 0
seen_ranges[character_range] += 1
character_count += 1
return sorted(
[
character_range
for character_range in seen_ranges
if seen_ranges[character_range] / character_count >= 0.15
]
)
def unicode_range_languages(primary_range: str) -> List[str]:
"""
Return inferred languages used with a unicode range.
"""
languages: List[str] = []
for language, characters in FREQUENCIES.items():
for character in characters:
if unicode_range(character) == primary_range:
languages.append(language)
break
return languages
@lru_cache()
def encoding_languages(iana_name: str) -> List[str]:
"""
Single-byte encoding language association. Some code page are heavily linked to particular language(s).
This function does the correspondence.
"""
unicode_ranges: List[str] = encoding_unicode_range(iana_name)
primary_range: Optional[str] = None
for specified_range in unicode_ranges:
if "Latin" not in specified_range:
primary_range = specified_range
break
if primary_range is None:
return ["Latin Based"]
return unicode_range_languages(primary_range)
@lru_cache()
def mb_encoding_languages(iana_name: str) -> List[str]:
"""
Multi-byte encoding language association. Some code page are heavily linked to particular language(s).
This function does the correspondence.
"""
if (
iana_name.startswith("shift_")
or iana_name.startswith("iso2022_jp")
or iana_name.startswith("euc_j")
or iana_name == "cp932"
):
return ["Japanese"]
if iana_name.startswith("gb") or iana_name in ZH_NAMES:
return ["Chinese"]
if iana_name.startswith("iso2022_kr") or iana_name in KO_NAMES:
return ["Korean"]
return []
@lru_cache(maxsize=LANGUAGE_SUPPORTED_COUNT)
def get_target_features(language: str) -> Tuple[bool, bool]:
"""
Determine main aspects from a supported language if it contains accents and if is pure Latin.
"""
target_have_accents: bool = False
target_pure_latin: bool = True
for character in FREQUENCIES[language]:
if not target_have_accents and is_accentuated(character):
target_have_accents = True
if target_pure_latin and is_latin(character) is False:
target_pure_latin = False
return target_have_accents, target_pure_latin
def alphabet_languages(
characters: List[str], ignore_non_latin: bool = False
) -> List[str]:
"""
Return associated languages associated to given characters.
"""
languages: List[Tuple[str, float]] = []
source_have_accents = any(is_accentuated(character) for character in characters)
for language, language_characters in FREQUENCIES.items():
target_have_accents, target_pure_latin = get_target_features(language)
if ignore_non_latin and target_pure_latin is False:
continue
if target_have_accents is False and source_have_accents:
continue
character_count: int = len(language_characters)
character_match_count: int = len(
[c for c in language_characters if c in characters]
)
ratio: float = character_match_count / character_count
if ratio >= 0.2:
languages.append((language, ratio))
languages = sorted(languages, key=lambda x: x[1], reverse=True)
return [compatible_language[0] for compatible_language in languages]
def characters_popularity_compare(
language: str, ordered_characters: List[str]
) -> float:
"""
Determine if a ordered characters list (by occurrence from most appearance to rarest) match a particular language.
The result is a ratio between 0. (absolutely no correspondence) and 1. (near perfect fit).
Beware that is function is not strict on the match in order to ease the detection. (Meaning close match is 1.)
"""
if language not in FREQUENCIES:
raise ValueError("{} not available".format(language))
character_approved_count: int = 0
FREQUENCIES_language_set = set(FREQUENCIES[language])
ordered_characters_count: int = len(ordered_characters)
target_language_characters_count: int = len(FREQUENCIES[language])
large_alphabet: bool = target_language_characters_count > 26
for character, character_rank in zip(
ordered_characters, range(0, ordered_characters_count)
):
if character not in FREQUENCIES_language_set:
continue
character_rank_in_language: int = FREQUENCIES[language].index(character)
expected_projection_ratio: float = (
target_language_characters_count / ordered_characters_count
)
character_rank_projection: int = int(character_rank * expected_projection_ratio)
if (
large_alphabet is False
and abs(character_rank_projection - character_rank_in_language) > 4
):
continue
if (
large_alphabet is True
and abs(character_rank_projection - character_rank_in_language)
< target_language_characters_count / 3
):
character_approved_count += 1
continue
characters_before_source: List[str] = FREQUENCIES[language][
0:character_rank_in_language
]
characters_after_source: List[str] = FREQUENCIES[language][
character_rank_in_language:
]
characters_before: List[str] = ordered_characters[0:character_rank]
characters_after: List[str] = ordered_characters[character_rank:]
before_match_count: int = len(
set(characters_before) & set(characters_before_source)
)
after_match_count: int = len(
set(characters_after) & set(characters_after_source)
)
if len(characters_before_source) == 0 and before_match_count <= 4:
character_approved_count += 1
continue
if len(characters_after_source) == 0 and after_match_count <= 4:
character_approved_count += 1
continue
if (
before_match_count / len(characters_before_source) >= 0.4
or after_match_count / len(characters_after_source) >= 0.4
):
character_approved_count += 1
continue
return character_approved_count / len(ordered_characters)
def alpha_unicode_split(decoded_sequence: str) -> List[str]:
"""
Given a decoded text sequence, return a list of str. Unicode range / alphabet separation.
Ex. a text containing English/Latin with a bit a Hebrew will return two items in the resulting list;
One containing the latin letters and the other hebrew.
"""
layers: Dict[str, str] = {}
for character in decoded_sequence:
if character.isalpha() is False:
continue
character_range: Optional[str] = unicode_range(character)
if character_range is None:
continue
layer_target_range: Optional[str] = None
for discovered_range in layers:
if (
is_suspiciously_successive_range(discovered_range, character_range)
is False
):
layer_target_range = discovered_range
break
if layer_target_range is None:
layer_target_range = character_range
if layer_target_range not in layers:
layers[layer_target_range] = character.lower()
continue
layers[layer_target_range] += character.lower()
return list(layers.values())
def merge_coherence_ratios(results: List[CoherenceMatches]) -> CoherenceMatches:
"""
This function merge results previously given by the function coherence_ratio.
The return type is the same as coherence_ratio.
"""
per_language_ratios: Dict[str, List[float]] = {}
for result in results:
for sub_result in result:
language, ratio = sub_result
if language not in per_language_ratios:
per_language_ratios[language] = [ratio]
continue
per_language_ratios[language].append(ratio)
merge = [
(
language,
round(
sum(per_language_ratios[language]) / len(per_language_ratios[language]),
4,
),
)
for language in per_language_ratios
]
return sorted(merge, key=lambda x: x[1], reverse=True)
def filter_alt_coherence_matches(results: CoherenceMatches) -> CoherenceMatches:
"""
We shall NOT return "English—" in CoherenceMatches because it is an alternative
of "English". This function only keeps the best match and remove the em-dash in it.
"""
index_results: Dict[str, List[float]] = dict()
for result in results:
language, ratio = result
no_em_name: str = language.replace("—", "")
if no_em_name not in index_results:
index_results[no_em_name] = []
index_results[no_em_name].append(ratio)
if any(len(index_results[e]) > 1 for e in index_results):
filtered_results: CoherenceMatches = []
for language in index_results:
filtered_results.append((language, max(index_results[language])))
return filtered_results
return results
@lru_cache(maxsize=2048)
def coherence_ratio(
decoded_sequence: str, threshold: float = 0.1, lg_inclusion: Optional[str] = None
) -> CoherenceMatches:
"""
Detect ANY language that can be identified in given sequence. The sequence will be analysed by layers.
A layer = Character extraction by alphabets/ranges.
"""
results: List[Tuple[str, float]] = []
ignore_non_latin: bool = False
sufficient_match_count: int = 0
lg_inclusion_list = lg_inclusion.split(",") if lg_inclusion is not None else []
if "Latin Based" in lg_inclusion_list:
ignore_non_latin = True
lg_inclusion_list.remove("Latin Based")
for layer in alpha_unicode_split(decoded_sequence):
sequence_frequencies: TypeCounter[str] = Counter(layer)
most_common = sequence_frequencies.most_common()
character_count: int = sum(o for c, o in most_common)
if character_count <= TOO_SMALL_SEQUENCE:
continue
popular_character_ordered: List[str] = [c for c, o in most_common]
for language in lg_inclusion_list or alphabet_languages(
popular_character_ordered, ignore_non_latin
):
ratio: float = characters_popularity_compare(
language, popular_character_ordered
)
if ratio < threshold:
continue
elif ratio >= 0.8:
sufficient_match_count += 1
results.append((language, round(ratio, 4)))
if sufficient_match_count >= 3:
break
return sorted(
filter_alt_coherence_matches(results), key=lambda x: x[1], reverse=True
)

1995
Modules/charset_normalizer/constant.py Normal file
View File

File diff suppressed because it is too large Load Diff

54
Modules/charset_normalizer/legacy.py Normal file
View File

@ -0,0 +1,54 @@
from typing import Any, Dict, Optional, Union
from warnings import warn
from .api import from_bytes
from .constant import CHARDET_CORRESPONDENCE
def detect(
byte_str: bytes, should_rename_legacy: bool = False, **kwargs: Any
) -> Dict[str, Optional[Union[str, float]]]:
"""
chardet legacy method
Detect the encoding of the given byte string. It should be mostly backward-compatible.
Encoding name will match Chardet own writing whenever possible. (Not on encoding name unsupported by it)
This function is deprecated and should be used to migrate your project easily, consult the documentation for
further information. Not planned for removal.
:param byte_str: The byte sequence to examine.
:param should_rename_legacy: Should we rename legacy encodings
to their more modern equivalents?
"""
if len(kwargs):
warn(
f"charset-normalizer disregard arguments '{','.join(list(kwargs.keys()))}' in legacy function detect()"
)
if not isinstance(byte_str, (bytearray, bytes)):
raise TypeError( # pragma: nocover
"Expected object of type bytes or bytearray, got: "
"{0}".format(type(byte_str))
)
if isinstance(byte_str, bytearray):
byte_str = bytes(byte_str)
r = from_bytes(byte_str).best()
encoding = r.encoding if r is not None else None
language = r.language if r is not None and r.language != "Unknown" else ""
confidence = 1.0 - r.chaos if r is not None else None
# Note: CharsetNormalizer does not return 'UTF-8-SIG' as the sig get stripped in the detection/normalization process
# but chardet does return 'utf-8-sig' and it is a valid codec name.
if r is not None and encoding == "utf_8" and r.bom:
encoding += "_sig"
if should_rename_legacy is False and encoding in CHARDET_CORRESPONDENCE:
encoding = CHARDET_CORRESPONDENCE[encoding]
return {
"encoding": encoding,
"language": language,
"confidence": confidence,
}

615
Modules/charset_normalizer/md.py Normal file
View File

@ -0,0 +1,615 @@
from functools import lru_cache
from logging import getLogger
from typing import List, Optional
from .constant import (
COMMON_SAFE_ASCII_CHARACTERS,
TRACE,
UNICODE_SECONDARY_RANGE_KEYWORD,
)
from .utils import (
is_accentuated,
is_arabic,
is_arabic_isolated_form,
is_case_variable,
is_cjk,
is_emoticon,
is_hangul,
is_hiragana,
is_katakana,
is_latin,
is_punctuation,
is_separator,
is_symbol,
is_thai,
is_unprintable,
remove_accent,
unicode_range,
)
class MessDetectorPlugin:
"""
Base abstract class used for mess detection plugins.
All detectors MUST extend and implement given methods.
"""
def eligible(self, character: str) -> bool:
"""
Determine if given character should be fed in.
"""
raise NotImplementedError # pragma: nocover
def feed(self, character: str) -> None:
"""
The main routine to be executed upon character.
Insert the logic in witch the text would be considered chaotic.
"""
raise NotImplementedError # pragma: nocover
def reset(self) -> None: # pragma: no cover
"""
Permit to reset the plugin to the initial state.
"""
raise NotImplementedError
@property
def ratio(self) -> float:
"""
Compute the chaos ratio based on what your feed() has seen.
Must NOT be lower than 0.; No restriction gt 0.
"""
raise NotImplementedError # pragma: nocover
class TooManySymbolOrPunctuationPlugin(MessDetectorPlugin):
def __init__(self) -> None:
self._punctuation_count: int = 0
self._symbol_count: int = 0
self._character_count: int = 0
self._last_printable_char: Optional[str] = None
self._frenzy_symbol_in_word: bool = False
def eligible(self, character: str) -> bool:
return character.isprintable()
def feed(self, character: str) -> None:
self._character_count += 1
if (
character != self._last_printable_char
and character not in COMMON_SAFE_ASCII_CHARACTERS
):
if is_punctuation(character):
self._punctuation_count += 1
elif (
character.isdigit() is False
and is_symbol(character)
and is_emoticon(character) is False
):
self._symbol_count += 2
self._last_printable_char = character
def reset(self) -> None: # pragma: no cover
self._punctuation_count = 0
self._character_count = 0
self._symbol_count = 0
@property
def ratio(self) -> float:
if self._character_count == 0:
return 0.0
ratio_of_punctuation: float = (
self._punctuation_count + self._symbol_count
) / self._character_count
return ratio_of_punctuation if ratio_of_punctuation >= 0.3 else 0.0
class TooManyAccentuatedPlugin(MessDetectorPlugin):
def __init__(self) -> None:
self._character_count: int = 0
self._accentuated_count: int = 0
def eligible(self, character: str) -> bool:
return character.isalpha()
def feed(self, character: str) -> None:
self._character_count += 1
if is_accentuated(character):
self._accentuated_count += 1
def reset(self) -> None: # pragma: no cover
self._character_count = 0
self._accentuated_count = 0
@property
def ratio(self) -> float:
if self._character_count < 8:
return 0.0
ratio_of_accentuation: float = self._accentuated_count / self._character_count
return ratio_of_accentuation if ratio_of_accentuation >= 0.35 else 0.0
class UnprintablePlugin(MessDetectorPlugin):
def __init__(self) -> None:
self._unprintable_count: int = 0
self._character_count: int = 0
def eligible(self, character: str) -> bool:
return True
def feed(self, character: str) -> None:
if is_unprintable(character):
self._unprintable_count += 1
self._character_count += 1
def reset(self) -> None: # pragma: no cover
self._unprintable_count = 0
@property
def ratio(self) -> float:
if self._character_count == 0:
return 0.0
return (self._unprintable_count * 8) / self._character_count
class SuspiciousDuplicateAccentPlugin(MessDetectorPlugin):
def __init__(self) -> None:
self._successive_count: int = 0
self._character_count: int = 0
self._last_latin_character: Optional[str] = None
def eligible(self, character: str) -> bool:
return character.isalpha() and is_latin(character)
def feed(self, character: str) -> None:
self._character_count += 1
if (
self._last_latin_character is not None
and is_accentuated(character)
and is_accentuated(self._last_latin_character)
):
if character.isupper() and self._last_latin_character.isupper():
self._successive_count += 1
# Worse if its the same char duplicated with different accent.
if remove_accent(character) == remove_accent(self._last_latin_character):
self._successive_count += 1
self._last_latin_character = character
def reset(self) -> None: # pragma: no cover
self._successive_count = 0
self._character_count = 0
self._last_latin_character = None
@property
def ratio(self) -> float:
if self._character_count == 0:
return 0.0
return (self._successive_count * 2) / self._character_count
class SuspiciousRange(MessDetectorPlugin):
def __init__(self) -> None:
self._suspicious_successive_range_count: int = 0
self._character_count: int = 0
self._last_printable_seen: Optional[str] = None
def eligible(self, character: str) -> bool:
return character.isprintable()
def feed(self, character: str) -> None:
self._character_count += 1
if (
character.isspace()
or is_punctuation(character)
or character in COMMON_SAFE_ASCII_CHARACTERS
):
self._last_printable_seen = None
return
if self._last_printable_seen is None:
self._last_printable_seen = character
return
unicode_range_a: Optional[str] = unicode_range(self._last_printable_seen)
unicode_range_b: Optional[str] = unicode_range(character)
if is_suspiciously_successive_range(unicode_range_a, unicode_range_b):
self._suspicious_successive_range_count += 1
self._last_printable_seen = character
def reset(self) -> None: # pragma: no cover
self._character_count = 0
self._suspicious_successive_range_count = 0
self._last_printable_seen = None
@property
def ratio(self) -> float:
if self._character_count <= 24:
return 0.0
ratio_of_suspicious_range_usage: float = (
self._suspicious_successive_range_count * 2
) / self._character_count
return ratio_of_suspicious_range_usage
class SuperWeirdWordPlugin(MessDetectorPlugin):
def __init__(self) -> None:
self._word_count: int = 0
self._bad_word_count: int = 0
self._foreign_long_count: int = 0
self._is_current_word_bad: bool = False
self._foreign_long_watch: bool = False
self._character_count: int = 0
self._bad_character_count: int = 0
self._buffer: str = ""
self._buffer_accent_count: int = 0
def eligible(self, character: str) -> bool:
return True
def feed(self, character: str) -> None:
if character.isalpha():
self._buffer += character
if is_accentuated(character):
self._buffer_accent_count += 1
if (
self._foreign_long_watch is False
and (is_latin(character) is False or is_accentuated(character))
and is_cjk(character) is False
and is_hangul(character) is False
and is_katakana(character) is False
and is_hiragana(character) is False
and is_thai(character) is False
):
self._foreign_long_watch = True
return
if not self._buffer:
return
if (
character.isspace() or is_punctuation(character) or is_separator(character)
) and self._buffer:
self._word_count += 1
buffer_length: int = len(self._buffer)
self._character_count += buffer_length
if buffer_length >= 4:
if self._buffer_accent_count / buffer_length > 0.34:
self._is_current_word_bad = True
# Word/Buffer ending with an upper case accentuated letter are so rare,
# that we will consider them all as suspicious. Same weight as foreign_long suspicious.
if (
is_accentuated(self._buffer[-1])
and self._buffer[-1].isupper()
and all(_.isupper() for _ in self._buffer) is False
):
self._foreign_long_count += 1
self._is_current_word_bad = True
if buffer_length >= 24 and self._foreign_long_watch:
camel_case_dst = [
i
for c, i in zip(self._buffer, range(0, buffer_length))
if c.isupper()
]
probable_camel_cased: bool = False
if camel_case_dst and (len(camel_case_dst) / buffer_length <= 0.3):
probable_camel_cased = True
if not probable_camel_cased:
self._foreign_long_count += 1
self._is_current_word_bad = True
if self._is_current_word_bad:
self._bad_word_count += 1
self._bad_character_count += len(self._buffer)
self._is_current_word_bad = False
self._foreign_long_watch = False
self._buffer = ""
self._buffer_accent_count = 0
elif (
character not in {"<", ">", "-", "=", "~", "|", "_"}
and character.isdigit() is False
and is_symbol(character)
):
self._is_current_word_bad = True
self._buffer += character
def reset(self) -> None: # pragma: no cover
self._buffer = ""
self._is_current_word_bad = False
self._foreign_long_watch = False
self._bad_word_count = 0
self._word_count = 0
self._character_count = 0
self._bad_character_count = 0
self._foreign_long_count = 0
@property
def ratio(self) -> float:
if self._word_count <= 10 and self._foreign_long_count == 0:
return 0.0
return self._bad_character_count / self._character_count
class CjkInvalidStopPlugin(MessDetectorPlugin):
"""
GB(Chinese) based encoding often render the stop incorrectly when the content does not fit and
can be easily detected. Searching for the overuse of '丅' and '丄'.
"""
def __init__(self) -> None:
self._wrong_stop_count: int = 0
self._cjk_character_count: int = 0
def eligible(self, character: str) -> bool:
return True
def feed(self, character: str) -> None:
if character in {"丅", "丄"}:
self._wrong_stop_count += 1
return
if is_cjk(character):
self._cjk_character_count += 1
def reset(self) -> None: # pragma: no cover
self._wrong_stop_count = 0
self._cjk_character_count = 0
@property
def ratio(self) -> float:
if self._cjk_character_count < 16:
return 0.0
return self._wrong_stop_count / self._cjk_character_count
class ArchaicUpperLowerPlugin(MessDetectorPlugin):
def __init__(self) -> None:
self._buf: bool = False
self._character_count_since_last_sep: int = 0
self._successive_upper_lower_count: int = 0
self._successive_upper_lower_count_final: int = 0
self._character_count: int = 0
self._last_alpha_seen: Optional[str] = None
self._current_ascii_only: bool = True
def eligible(self, character: str) -> bool:
return True
def feed(self, character: str) -> None:
is_concerned = character.isalpha() and is_case_variable(character)
chunk_sep = is_concerned is False
if chunk_sep and self._character_count_since_last_sep > 0:
if (
self._character_count_since_last_sep <= 64
and character.isdigit() is False
and self._current_ascii_only is False
):
self._successive_upper_lower_count_final += (
self._successive_upper_lower_count
)
self._successive_upper_lower_count = 0
self._character_count_since_last_sep = 0
self._last_alpha_seen = None
self._buf = False
self._character_count += 1
self._current_ascii_only = True
return
if self._current_ascii_only is True and character.isascii() is False:
self._current_ascii_only = False
if self._last_alpha_seen is not None:
if (character.isupper() and self._last_alpha_seen.islower()) or (
character.islower() and self._last_alpha_seen.isupper()
):
if self._buf is True:
self._successive_upper_lower_count += 2
self._buf = False
else:
self._buf = True
else:
self._buf = False
self._character_count += 1
self._character_count_since_last_sep += 1
self._last_alpha_seen = character
def reset(self) -> None: # pragma: no cover
self._character_count = 0
self._character_count_since_last_sep = 0
self._successive_upper_lower_count = 0
self._successive_upper_lower_count_final = 0
self._last_alpha_seen = None
self._buf = False
self._current_ascii_only = True
@property
def ratio(self) -> float:
if self._character_count == 0:
return 0.0
return self._successive_upper_lower_count_final / self._character_count
class ArabicIsolatedFormPlugin(MessDetectorPlugin):
def __init__(self) -> None:
self._character_count: int = 0
self._isolated_form_count: int = 0
def reset(self) -> None: # pragma: no cover
self._character_count = 0
self._isolated_form_count = 0
def eligible(self, character: str) -> bool:
return is_arabic(character)
def feed(self, character: str) -> None:
self._character_count += 1
if is_arabic_isolated_form(character):
self._isolated_form_count += 1
@property
def ratio(self) -> float:
if self._character_count < 8:
return 0.0
isolated_form_usage: float = self._isolated_form_count / self._character_count
return isolated_form_usage
@lru_cache(maxsize=1024)
def is_suspiciously_successive_range(
unicode_range_a: Optional[str], unicode_range_b: Optional[str]
) -> bool:
"""
Determine if two Unicode range seen next to each other can be considered as suspicious.
"""
if unicode_range_a is None or unicode_range_b is None:
return True
if unicode_range_a == unicode_range_b:
return False
if "Latin" in unicode_range_a and "Latin" in unicode_range_b:
return False
if "Emoticons" in unicode_range_a or "Emoticons" in unicode_range_b:
return False
# Latin characters can be accompanied with a combining diacritical mark
# eg. Vietnamese.
if ("Latin" in unicode_range_a or "Latin" in unicode_range_b) and (
"Combining" in unicode_range_a or "Combining" in unicode_range_b
):
return False
keywords_range_a, keywords_range_b = unicode_range_a.split(
" "
), unicode_range_b.split(" ")
for el in keywords_range_a:
if el in UNICODE_SECONDARY_RANGE_KEYWORD:
continue
if el in keywords_range_b:
return False
# Japanese Exception
range_a_jp_chars, range_b_jp_chars = (
unicode_range_a
in (
"Hiragana",
"Katakana",
),
unicode_range_b in ("Hiragana", "Katakana"),
)
if (range_a_jp_chars or range_b_jp_chars) and (
"CJK" in unicode_range_a or "CJK" in unicode_range_b
):
return False
if range_a_jp_chars and range_b_jp_chars:
return False
if "Hangul" in unicode_range_a or "Hangul" in unicode_range_b:
if "CJK" in unicode_range_a or "CJK" in unicode_range_b:
return False
if unicode_range_a == "Basic Latin" or unicode_range_b == "Basic Latin":
return False
# Chinese/Japanese use dedicated range for punctuation and/or separators.
if ("CJK" in unicode_range_a or "CJK" in unicode_range_b) or (
unicode_range_a in ["Katakana", "Hiragana"]
and unicode_range_b in ["Katakana", "Hiragana"]
):
if "Punctuation" in unicode_range_a or "Punctuation" in unicode_range_b:
return False
if "Forms" in unicode_range_a or "Forms" in unicode_range_b:
return False
if unicode_range_a == "Basic Latin" or unicode_range_b == "Basic Latin":
return False
return True
@lru_cache(maxsize=2048)
def mess_ratio(
decoded_sequence: str, maximum_threshold: float = 0.2, debug: bool = False
) -> float:
"""
Compute a mess ratio given a decoded bytes sequence. The maximum threshold does stop the computation earlier.
"""
detectors: List[MessDetectorPlugin] = [
md_class() for md_class in MessDetectorPlugin.__subclasses__()
]
length: int = len(decoded_sequence) + 1
mean_mess_ratio: float = 0.0
if length < 512:
intermediary_mean_mess_ratio_calc: int = 32
elif length <= 1024:
intermediary_mean_mess_ratio_calc = 64
else:
intermediary_mean_mess_ratio_calc = 128
for character, index in zip(decoded_sequence + "\n", range(length)):
for detector in detectors:
if detector.eligible(character):
detector.feed(character)
if (
index > 0 and index % intermediary_mean_mess_ratio_calc == 0
) or index == length - 1:
mean_mess_ratio = sum(dt.ratio for dt in detectors)
if mean_mess_ratio >= maximum_threshold:
break
if debug:
logger = getLogger("charset_normalizer")
logger.log(
TRACE,
"Mess-detector extended-analysis start. "
f"intermediary_mean_mess_ratio_calc={intermediary_mean_mess_ratio_calc} mean_mess_ratio={mean_mess_ratio} "
f"maximum_threshold={maximum_threshold}",
)
if len(decoded_sequence) > 16:
logger.log(TRACE, f"Starting with: {decoded_sequence[:16]}")
logger.log(TRACE, f"Ending with: {decoded_sequence[-16::]}")
for dt in detectors: # pragma: nocover
logger.log(TRACE, f"{dt.__class__}: {dt.ratio}")
return round(mean_mess_ratio, 3)

340
Modules/charset_normalizer/models.py Normal file
View File

@ -0,0 +1,340 @@
from encodings.aliases import aliases
from hashlib import sha256
from json import dumps
from typing import Any, Dict, Iterator, List, Optional, Tuple, Union
from .constant import TOO_BIG_SEQUENCE
from .utils import iana_name, is_multi_byte_encoding, unicode_range
class CharsetMatch:
def __init__(
self,
payload: bytes,
guessed_encoding: str,
mean_mess_ratio: float,
has_sig_or_bom: bool,
languages: "CoherenceMatches",
decoded_payload: Optional[str] = None,
):
self._payload: bytes = payload
self._encoding: str = guessed_encoding
self._mean_mess_ratio: float = mean_mess_ratio
self._languages: CoherenceMatches = languages
self._has_sig_or_bom: bool = has_sig_or_bom
self._unicode_ranges: Optional[List[str]] = None
self._leaves: List[CharsetMatch] = []
self._mean_coherence_ratio: float = 0.0
self._output_payload: Optional[bytes] = None
self._output_encoding: Optional[str] = None
self._string: Optional[str] = decoded_payload
def __eq__(self, other: object) -> bool:
if not isinstance(other, CharsetMatch):
raise TypeError(
"__eq__ cannot be invoked on {} and {}.".format(
str(other.__class__), str(self.__class__)
)
)
return self.encoding == other.encoding and self.fingerprint == other.fingerprint
def __lt__(self, other: object) -> bool:
"""
Implemented to make sorted available upon CharsetMatches items.
"""
if not isinstance(other, CharsetMatch):
raise ValueError
chaos_difference: float = abs(self.chaos - other.chaos)
coherence_difference: float = abs(self.coherence - other.coherence)
# Below 1% difference --> Use Coherence
if chaos_difference < 0.01 and coherence_difference > 0.02:
return self.coherence > other.coherence
elif chaos_difference < 0.01 and coherence_difference <= 0.02:
# When having a difficult decision, use the result that decoded as many multi-byte as possible.
# preserve RAM usage!
if len(self._payload) >= TOO_BIG_SEQUENCE:
return self.chaos < other.chaos
return self.multi_byte_usage > other.multi_byte_usage
return self.chaos < other.chaos
@property
def multi_byte_usage(self) -> float:
return 1.0 - (len(str(self)) / len(self.raw))
def __str__(self) -> str:
# Lazy Str Loading
if self._string is None:
self._string = str(self._payload, self._encoding, "strict")
return self._string
def __repr__(self) -> str:
return "<CharsetMatch '{}' bytes({})>".format(self.encoding, self.fingerprint)
def add_submatch(self, other: "CharsetMatch") -> None:
if not isinstance(other, CharsetMatch) or other == self:
raise ValueError(
"Unable to add instance <{}> as a submatch of a CharsetMatch".format(
other.__class__
)
)
other._string = None # Unload RAM usage; dirty trick.
self._leaves.append(other)
@property
def encoding(self) -> str:
return self._encoding
@property
def encoding_aliases(self) -> List[str]:
"""
Encoding name are known by many name, using this could help when searching for IBM855 when it's listed as CP855.
"""
also_known_as: List[str] = []
for u, p in aliases.items():
if self.encoding == u:
also_known_as.append(p)
elif self.encoding == p:
also_known_as.append(u)
return also_known_as
@property
def bom(self) -> bool:
return self._has_sig_or_bom
@property
def byte_order_mark(self) -> bool:
return self._has_sig_or_bom
@property
def languages(self) -> List[str]:
"""
Return the complete list of possible languages found in decoded sequence.
Usually not really useful. Returned list may be empty even if 'language' property return something != 'Unknown'.
"""
return [e[0] for e in self._languages]
@property
def language(self) -> str:
"""
Most probable language found in decoded sequence. If none were detected or inferred, the property will return
"Unknown".
"""
if not self._languages:
# Trying to infer the language based on the given encoding
# Its either English or we should not pronounce ourselves in certain cases.
if "ascii" in self.could_be_from_charset:
return "English"
# doing it there to avoid circular import
from charset_normalizer.cd import encoding_languages, mb_encoding_languages
languages = (
mb_encoding_languages(self.encoding)
if is_multi_byte_encoding(self.encoding)
else encoding_languages(self.encoding)
)
if len(languages) == 0 or "Latin Based" in languages:
return "Unknown"
return languages[0]
return self._languages[0][0]
@property
def chaos(self) -> float:
return self._mean_mess_ratio
@property
def coherence(self) -> float:
if not self._languages:
return 0.0
return self._languages[0][1]
@property
def percent_chaos(self) -> float:
return round(self.chaos * 100, ndigits=3)
@property
def percent_coherence(self) -> float:
return round(self.coherence * 100, ndigits=3)
@property
def raw(self) -> bytes:
"""
Original untouched bytes.
"""
return self._payload
@property
def submatch(self) -> List["CharsetMatch"]:
return self._leaves
@property
def has_submatch(self) -> bool:
return len(self._leaves) > 0
@property
def alphabets(self) -> List[str]:
if self._unicode_ranges is not None:
return self._unicode_ranges
# list detected ranges
detected_ranges: List[Optional[str]] = [
unicode_range(char) for char in str(self)
]
# filter and sort
self._unicode_ranges = sorted(list({r for r in detected_ranges if r}))
return self._unicode_ranges
@property
def could_be_from_charset(self) -> List[str]:
"""
The complete list of encoding that output the exact SAME str result and therefore could be the originating
encoding.
This list does include the encoding available in property 'encoding'.
"""
return [self._encoding] + [m.encoding for m in self._leaves]
def output(self, encoding: str = "utf_8") -> bytes:
"""
Method to get re-encoded bytes payload using given target encoding. Default to UTF-8.
Any errors will be simply ignored by the encoder NOT replaced.
"""
if self._output_encoding is None or self._output_encoding != encoding:
self._output_encoding = encoding
self._output_payload = str(self).encode(encoding, "replace")
return self._output_payload # type: ignore
@property
def fingerprint(self) -> str:
"""
Retrieve the unique SHA256 computed using the transformed (re-encoded) payload. Not the original one.
"""
return sha256(self.output()).hexdigest()
class CharsetMatches:
"""
Container with every CharsetMatch items ordered by default from most probable to the less one.
Act like a list(iterable) but does not implements all related methods.
"""
def __init__(self, results: Optional[List[CharsetMatch]] = None):
self._results: List[CharsetMatch] = sorted(results) if results else []
def __iter__(self) -> Iterator[CharsetMatch]:
yield from self._results
def __getitem__(self, item: Union[int, str]) -> CharsetMatch:
"""
Retrieve a single item either by its position or encoding name (alias may be used here).
Raise KeyError upon invalid index or encoding not present in results.
"""
if isinstance(item, int):
return self._results[item]
if isinstance(item, str):
item = iana_name(item, False)
for result in self._results:
if item in result.could_be_from_charset:
return result
raise KeyError
def __len__(self) -> int:
return len(self._results)
def __bool__(self) -> bool:
return len(self._results) > 0
def append(self, item: CharsetMatch) -> None:
"""
Insert a single match. Will be inserted accordingly to preserve sort.
Can be inserted as a submatch.
"""
if not isinstance(item, CharsetMatch):
raise ValueError(
"Cannot append instance '{}' to CharsetMatches".format(
str(item.__class__)
)
)
# We should disable the submatch factoring when the input file is too heavy (conserve RAM usage)
if len(item.raw) <= TOO_BIG_SEQUENCE:
for match in self._results:
if match.fingerprint == item.fingerprint and match.chaos == item.chaos:
match.add_submatch(item)
return
self._results.append(item)
self._results = sorted(self._results)
def best(self) -> Optional["CharsetMatch"]:
"""
Simply return the first match. Strict equivalent to matches[0].
"""
if not self._results:
return None
return self._results[0]
def first(self) -> Optional["CharsetMatch"]:
"""
Redundant method, call the method best(). Kept for BC reasons.
"""
return self.best()
CoherenceMatch = Tuple[str, float]
CoherenceMatches = List[CoherenceMatch]
class CliDetectionResult:
def __init__(
self,
path: str,
encoding: Optional[str],
encoding_aliases: List[str],
alternative_encodings: List[str],
language: str,
alphabets: List[str],
has_sig_or_bom: bool,
chaos: float,
coherence: float,
unicode_path: Optional[str],
is_preferred: bool,
):
self.path: str = path
self.unicode_path: Optional[str] = unicode_path
self.encoding: Optional[str] = encoding
self.encoding_aliases: List[str] = encoding_aliases
self.alternative_encodings: List[str] = alternative_encodings
self.language: str = language
self.alphabets: List[str] = alphabets
self.has_sig_or_bom: bool = has_sig_or_bom
self.chaos: float = chaos
self.coherence: float = coherence
self.is_preferred: bool = is_preferred
@property
def __dict__(self) -> Dict[str, Any]: # type: ignore
return {
"path": self.path,
"encoding": self.encoding,
"encoding_aliases": self.encoding_aliases,
"alternative_encodings": self.alternative_encodings,
"language": self.language,
"alphabets": self.alphabets,
"has_sig_or_bom": self.has_sig_or_bom,
"chaos": self.chaos,
"coherence": self.coherence,
"unicode_path": self.unicode_path,
"is_preferred": self.is_preferred,
}
def to_json(self) -> str:
return dumps(self.__dict__, ensure_ascii=True, indent=4)

0
Modules/charset_normalizer/py.typed Normal file
View File

421
Modules/charset_normalizer/utils.py Normal file
View File

@ -0,0 +1,421 @@
import importlib
import logging
import unicodedata
from codecs import IncrementalDecoder
from encodings.aliases import aliases
from functools import lru_cache
from re import findall
from typing import Generator, List, Optional, Set, Tuple, Union
from _multibytecodec import MultibyteIncrementalDecoder
from .constant import (
ENCODING_MARKS,
IANA_SUPPORTED_SIMILAR,
RE_POSSIBLE_ENCODING_INDICATION,
UNICODE_RANGES_COMBINED,
UNICODE_SECONDARY_RANGE_KEYWORD,
UTF8_MAXIMAL_ALLOCATION,
)
@lru_cache(maxsize=UTF8_MAXIMAL_ALLOCATION)
def is_accentuated(character: str) -> bool:
try:
description: str = unicodedata.name(character)
except ValueError:
return False
return (
"WITH GRAVE" in description
or "WITH ACUTE" in description
or "WITH CEDILLA" in description
or "WITH DIAERESIS" in description
or "WITH CIRCUMFLEX" in description
or "WITH TILDE" in description
or "WITH MACRON" in description
or "WITH RING ABOVE" in description
)
@lru_cache(maxsize=UTF8_MAXIMAL_ALLOCATION)
def remove_accent(character: str) -> str:
decomposed: str = unicodedata.decomposition(character)
if not decomposed:
return character
codes: List[str] = decomposed.split(" ")
return chr(int(codes[0], 16))
@lru_cache(maxsize=UTF8_MAXIMAL_ALLOCATION)
def unicode_range(character: str) -> Optional[str]:
"""
Retrieve the Unicode range official name from a single character.
"""
character_ord: int = ord(character)
for range_name, ord_range in UNICODE_RANGES_COMBINED.items():
if character_ord in ord_range:
return range_name
return None
@lru_cache(maxsize=UTF8_MAXIMAL_ALLOCATION)
def is_latin(character: str) -> bool:
try:
description: str = unicodedata.name(character)
except ValueError:
return False
return "LATIN" in description
@lru_cache(maxsize=UTF8_MAXIMAL_ALLOCATION)
def is_punctuation(character: str) -> bool:
character_category: str = unicodedata.category(character)
if "P" in character_category:
return True
character_range: Optional[str] = unicode_range(character)
if character_range is None:
return False
return "Punctuation" in character_range
@lru_cache(maxsize=UTF8_MAXIMAL_ALLOCATION)
def is_symbol(character: str) -> bool:
character_category: str = unicodedata.category(character)
if "S" in character_category or "N" in character_category:
return True
character_range: Optional[str] = unicode_range(character)
if character_range is None:
return False
return "Forms" in character_range and character_category != "Lo"
@lru_cache(maxsize=UTF8_MAXIMAL_ALLOCATION)
def is_emoticon(character: str) -> bool:
character_range: Optional[str] = unicode_range(character)
if character_range is None:
return False
return "Emoticons" in character_range or "Pictographs" in character_range
@lru_cache(maxsize=UTF8_MAXIMAL_ALLOCATION)
def is_separator(character: str) -> bool:
if character.isspace() or character in {"｜", "+", "<", ">"}:
return True
character_category: str = unicodedata.category(character)
return "Z" in character_category or character_category in {"Po", "Pd", "Pc"}
@lru_cache(maxsize=UTF8_MAXIMAL_ALLOCATION)
def is_case_variable(character: str) -> bool:
return character.islower() != character.isupper()
@lru_cache(maxsize=UTF8_MAXIMAL_ALLOCATION)
def is_cjk(character: str) -> bool:
try:
character_name = unicodedata.name(character)
except ValueError:
return False
return "CJK" in character_name
@lru_cache(maxsize=UTF8_MAXIMAL_ALLOCATION)
def is_hiragana(character: str) -> bool:
try:
character_name = unicodedata.name(character)
except ValueError:
return False
return "HIRAGANA" in character_name
@lru_cache(maxsize=UTF8_MAXIMAL_ALLOCATION)
def is_katakana(character: str) -> bool:
try:
character_name = unicodedata.name(character)
except ValueError:
return False
return "KATAKANA" in character_name
@lru_cache(maxsize=UTF8_MAXIMAL_ALLOCATION)
def is_hangul(character: str) -> bool:
try:
character_name = unicodedata.name(character)
except ValueError:
return False
return "HANGUL" in character_name
@lru_cache(maxsize=UTF8_MAXIMAL_ALLOCATION)
def is_thai(character: str) -> bool:
try:
character_name = unicodedata.name(character)
except ValueError:
return False
return "THAI" in character_name
@lru_cache(maxsize=UTF8_MAXIMAL_ALLOCATION)
def is_arabic(character: str) -> bool:
try:
character_name = unicodedata.name(character)
except ValueError:
return False
return "ARABIC" in character_name
@lru_cache(maxsize=UTF8_MAXIMAL_ALLOCATION)
def is_arabic_isolated_form(character: str) -> bool:
try:
character_name = unicodedata.name(character)
except ValueError:
return False
return "ARABIC" in character_name and "ISOLATED FORM" in character_name
@lru_cache(maxsize=len(UNICODE_RANGES_COMBINED))
def is_unicode_range_secondary(range_name: str) -> bool:
return any(keyword in range_name for keyword in UNICODE_SECONDARY_RANGE_KEYWORD)
@lru_cache(maxsize=UTF8_MAXIMAL_ALLOCATION)
def is_unprintable(character: str) -> bool:
return (
character.isspace() is False # includes \n \t \r \v
and character.isprintable() is False
and character != "\x1A" # Why? Its the ASCII substitute character.
and character != "\ufeff" # bug discovered in Python,
# Zero Width No-Break Space located in Arabic Presentation Forms-B, Unicode 1.1 not acknowledged as space.
)
def any_specified_encoding(sequence: bytes, search_zone: int = 8192) -> Optional[str]:
"""
Extract using ASCII-only decoder any specified encoding in the first n-bytes.
"""
if not isinstance(sequence, bytes):
raise TypeError
seq_len: int = len(sequence)
results: List[str] = findall(
RE_POSSIBLE_ENCODING_INDICATION,
sequence[: min(seq_len, search_zone)].decode("ascii", errors="ignore"),
)
if len(results) == 0:
return None
for specified_encoding in results:
specified_encoding = specified_encoding.lower().replace("-", "_")
encoding_alias: str
encoding_iana: str
for encoding_alias, encoding_iana in aliases.items():
if encoding_alias == specified_encoding:
return encoding_iana
if encoding_iana == specified_encoding:
return encoding_iana
return None
@lru_cache(maxsize=128)
def is_multi_byte_encoding(name: str) -> bool:
"""
Verify is a specific encoding is a multi byte one based on it IANA name
"""
return name in {
"utf_8",
"utf_8_sig",
"utf_16",
"utf_16_be",
"utf_16_le",
"utf_32",
"utf_32_le",
"utf_32_be",
"utf_7",
} or issubclass(
importlib.import_module("encodings.{}".format(name)).IncrementalDecoder,
MultibyteIncrementalDecoder,
)
def identify_sig_or_bom(sequence: bytes) -> Tuple[Optional[str], bytes]:
"""
Identify and extract SIG/BOM in given sequence.
"""
for iana_encoding in ENCODING_MARKS:
marks: Union[bytes, List[bytes]] = ENCODING_MARKS[iana_encoding]
if isinstance(marks, bytes):
marks = [marks]
for mark in marks:
if sequence.startswith(mark):
return iana_encoding, mark
return None, b""
def should_strip_sig_or_bom(iana_encoding: str) -> bool:
return iana_encoding not in {"utf_16", "utf_32"}
def iana_name(cp_name: str, strict: bool = True) -> str:
cp_name = cp_name.lower().replace("-", "_")
encoding_alias: str
encoding_iana: str
for encoding_alias, encoding_iana in aliases.items():
if cp_name in [encoding_alias, encoding_iana]:
return encoding_iana
if strict:
raise ValueError("Unable to retrieve IANA for '{}'".format(cp_name))
return cp_name
def range_scan(decoded_sequence: str) -> List[str]:
ranges: Set[str] = set()
for character in decoded_sequence:
character_range: Optional[str] = unicode_range(character)
if character_range is None:
continue
ranges.add(character_range)
return list(ranges)
def cp_similarity(iana_name_a: str, iana_name_b: str) -> float:
if is_multi_byte_encoding(iana_name_a) or is_multi_byte_encoding(iana_name_b):
return 0.0
decoder_a = importlib.import_module(
"encodings.{}".format(iana_name_a)
).IncrementalDecoder
decoder_b = importlib.import_module(
"encodings.{}".format(iana_name_b)
).IncrementalDecoder
id_a: IncrementalDecoder = decoder_a(errors="ignore")
id_b: IncrementalDecoder = decoder_b(errors="ignore")
character_match_count: int = 0
for i in range(255):
to_be_decoded: bytes = bytes([i])
if id_a.decode(to_be_decoded) == id_b.decode(to_be_decoded):
character_match_count += 1
return character_match_count / 254
def is_cp_similar(iana_name_a: str, iana_name_b: str) -> bool:
"""
Determine if two code page are at least 80% similar. IANA_SUPPORTED_SIMILAR dict was generated using
the function cp_similarity.
"""
return (
iana_name_a in IANA_SUPPORTED_SIMILAR
and iana_name_b in IANA_SUPPORTED_SIMILAR[iana_name_a]
)
def set_logging_handler(
name: str = "charset_normalizer",
level: int = logging.INFO,
format_string: str = "%(asctime)s | %(levelname)s | %(message)s",
) -> None:
logger = logging.getLogger(name)
logger.setLevel(level)
handler = logging.StreamHandler()
handler.setFormatter(logging.Formatter(format_string))
logger.addHandler(handler)
def cut_sequence_chunks(
sequences: bytes,
encoding_iana: str,
offsets: range,
chunk_size: int,
bom_or_sig_available: bool,
strip_sig_or_bom: bool,
sig_payload: bytes,
is_multi_byte_decoder: bool,
decoded_payload: Optional[str] = None,
) -> Generator[str, None, None]:
if decoded_payload and is_multi_byte_decoder is False:
for i in offsets:
chunk = decoded_payload[i : i + chunk_size]
if not chunk:
break
yield chunk
else:
for i in offsets:
chunk_end = i + chunk_size
if chunk_end > len(sequences) + 8:
continue
cut_sequence = sequences[i : i + chunk_size]
if bom_or_sig_available and strip_sig_or_bom is False:
cut_sequence = sig_payload + cut_sequence
chunk = cut_sequence.decode(
encoding_iana,
errors="ignore" if is_multi_byte_decoder else "strict",
)
# multi-byte bad cutting detector and adjustment
# not the cleanest way to perform that fix but clever enough for now.
if is_multi_byte_decoder and i > 0:
chunk_partial_size_chk: int = min(chunk_size, 16)
if (
decoded_payload
and chunk[:chunk_partial_size_chk] not in decoded_payload
):
for j in range(i, i - 4, -1):
cut_sequence = sequences[j:chunk_end]
if bom_or_sig_available and strip_sig_or_bom is False:
cut_sequence = sig_payload + cut_sequence
chunk = cut_sequence.decode(encoding_iana, errors="ignore")
if chunk[:chunk_partial_size_chk] in decoded_payload:
break
yield chunk

6
Modules/charset_normalizer/version.py Normal file
View File

@ -0,0 +1,6 @@
"""
Expose version
"""
__version__ = "3.3.2"
VERSION = __version__.split(".")