def _latinize_internal(text, ascii=False):
if ascii:
if not hasattr(latinize_text, '_ascii'):
# Transform to latin, separate accents, decompose, remove
# symbols, compose, push to ASCII
latinize_text._ascii = Transliterator.createInstance(
'Any-Latin; NFKD; [:Symbol:] Remove; [:Nonspacing Mark:] Remove; NFKC; Accents-Any; Latin-ASCII'
) # noqa
return latinize_text._ascii.transliterate(text)
if not hasattr(latinize_text, '_tr'):
latinize_text._tr = Transliterator.createInstance('Any-Latin')
return latinize_text._tr.transliterate(text)
def main(argv):
inputfile = ''
outputfile = ''
try:
opts, args = getopt.getopt(argv,"hi:o:",["ifile=","ofile="])
except getopt.GetoptError:
print 'test.py -i <inputfile> -o <outputfile>'
sys.exit(2)
for opt, arg in opts:
if opt == '-h':
print 'zg-my.py -i <inputfile> -o <outputfile>'
sys.exit()
elif opt in ("-i", "--ifile"):
inputfile = arg
outputfile = "converted_" + inputfile
elif opt in ("-o", "--ofile"):
outputfile = arg
print 'Input file is ', inputfile
print 'Output file is ', outputfile
uni = Transliterator.createInstance('Zawgyi-my')
f = open(inputfile, "r")
converted = uni.transliterate(f.read())
f.close()
fo = open(outputfile, "w")
fo.write(converted.encode('utf8'))
fo.close()
def compose_nfc(text):
"""Perform unicode composition."""
if text is None:
return None
if not hasattr(compose_nfc, '_tr'):
compose_nfc._tr = Transliterator.createInstance('Any-NFC')
return compose_nfc._tr.transliterate(text)
def decompose_nfkd(text):
"""Perform unicode compatibility decomposition.
This will replace some non-standard value representations in unicode and
normalise them, while also separating characters and their diacritics into
two separate codepoints.
"""
if text is None:
return None
if not hasattr(decompose_nfkd, '_tr'):
decompose_nfkd._tr = Transliterator.createInstance('Any-NFKD')
return decompose_nfkd._tr.transliterate(text)
def latinize_text(text, ascii=False):
"""Transliterate the given text to the latin script.
This attempts to convert a given text to latin script using the
closest match of characters vis a vis the original script.
"""
if text is None or not isinstance(text, six.string_types) or not len(text):
return text
if ascii:
if not hasattr(latinize_text, '_ascii'):
# Transform to latin, separate accents, decompose, remove
# symbols, compose, push to ASCII
latinize_text._ascii = Transliterator.createInstance(
'Any-Latin; NFKD; [:Symbol:] Remove; [:Nonspacing Mark:] Remove; NFKC; Accents-Any; Latin-ASCII'
) # noqa
return latinize_text._ascii.transliterate(text)
if not hasattr(latinize_text, '_tr'):
latinize_text._tr = Transliterator.createInstance('Any-Latin')
return latinize_text._tr.transliterate(text)
def make_transliterator(script):
try:
from icu import Transliterator
inst = Transliterator.createInstance(script)
return inst.transliterate
except ImportError:
from text_unidecode import unidecode
warnings.warn("Install 'pyicu' for better text transliteration.",
ICUWarning,
stacklevel=4) # noqa
def transliterate(text):
text = compose_nfkc(text)
return unidecode(text)
return transliterate
def make_trans(script: str) -> Callable[[str], Optional[str]]:
try:
from icu import Transliterator # type: ignore
inst = Transliterator.createInstance(script)
return cast(Callable[[str], str], inst.transliterate)
except ImportError:
from text_unidecode import unidecode # type: ignore
warnings.warn("Install 'pyicu' for better text transliteration.",
ICUWarning,
stacklevel=4) # noqa
def transliterate(text: str) -> Optional[str]:
clean = compose_nfkc(text)
if clean is None:
return None
return cast(Optional[str], unidecode(clean))
return transliterate
def _getTransliterator(self, name):
return Transliterator.createInstance(name, UTransDirection.FORWARD)
# -*- coding: utf-8 -*-
"""
Transliterating text to International Phonetic Alphabet (IPA)
Using International Components for Unicode (ICU)
https://github.com/ovalhub/pyicu
"""
from icu import Transliterator
_ICU_THAI_TO_LATIN = Transliterator.createInstance("Thai-Latin")
# ถอดเสียงภาษาไทยเป็นอักษรละติน
def transliterate(text: str) -> str:
"""
Use ICU (International Components for Unicode) for transliteration
ถอดเสียงภาษาไทยเป็นอักษรละติน รับค่า ''str'' ข้อความ คืนค่า ''str'' อักษรละติน
:param str text: Thai text to be transliterated.
:return: A string of Internaitonal Phonetic Alphabets indicating how the text should be pronounced.
"""
return _ICU_THAI_TO_LATIN.transliterate(text)
import os
import re
import six
import yaml
from icu import Transliterator
DATA_PAGE = 10000
WS_PATTERN = re.compile('\s+')
tr = Transliterator.createInstance('Any-Latin')
def resolve_includes(file_path, data):
"""Handle include statements in the configuration file."""
if isinstance(data, (list, tuple, set)):
data = [resolve_includes(file_path, i) for i in data]
elif isinstance(data, dict):
include_paths = data.pop('include', [])
if not isinstance(include_paths, (list, tuple, set)):
include_paths = [include_paths]
for include_path in include_paths:
dir_prefix = os.path.dirname(file_path)
include_path = os.path.join(dir_prefix, include_path)
data.update(load_config_file(include_path))
for key, value in data.items():
data[key] = resolve_includes(file_path, value)
return data
def load_config_file(file_path):
"""Load a YAML (or JSON) model configuration file."""
def _getTransliterator(self, name):
return Transliterator.createInstance(name, UTransDirection.FORWARD)
def _decompose_nfkd(text):
if not hasattr(_decompose_nfkd, '_tr'):
_decompose_nfkd._tr = Transliterator.createInstance('Any-NFKD')
return _decompose_nfkd._tr.transliterate(text)
def _compose_nfc(text):
if not hasattr(_compose_nfc, '_tr'):
_compose_nfc._tr = Transliterator.createInstance('Any-NFC')
return _compose_nfc._tr.transliterate(text)
flags.DEFINE_bool("build_fasttext", False, "build fasttext features")
flags.DEFINE_bool("build_tfrecord", False,
"build tensorflow record input files")
flags.DEFINE_integer("nrows", 100, "The TOP number of rows to query")
prog = re.compile("[\\W\\d]", re.UNICODE)
prog_with_digits = re.compile("[\\W]", re.UNICODE)
stemmer = SnowballStemmer("russian", ignore_stopwords=True)
float_prog = re.compile(r"[-+]?\d*\.\d+|\d+", re.UNICODE)
dot_prog = re.compile(r'[xх*]', re.UNICODE)
TransTable = str.maketrans(dict.fromkeys(r'~/-\[\]()|{}:^+', ' '))
wt = WordTokenizer()
trans = Transliterator.createInstance('Latin-Cyrillic')
unit_lookup = {
'г': 'грамм',
'грам': 'грамм',
'гр': 'грамм',
'грамм': 'грамм',
'gr': 'грамм',
'ml': 'мл',
'милл': 'мл',
'млитр': 'мл',
'млтр': 'мл',
'мл': 'мл',
'ш': 'шт',
'шт': 'шт',
'тон': 'тонна',