def __init__(self, data, pred, ratio=0.85):
self.data = data
self.pred = pred
self.ratio = ratio
self._training_data = None
self._test_data = None
self.logger = get_logger(self.__class__.__name__.lower())
from string import punctuation
from itertools import filterfalse
from functools import partial
from nltk.tokenize import sent_tokenize, word_tokenize
from nltk.corpus import stopwords
from buttworld.logger import get_logger
logger = get_logger(__name__)
EXTRA_PUNCTUATION = '«»—…'
TERMINATION_CHARS = '.»…!?)"\''
F_LOWERCASE = 0b00001
F_PUNCTUATION = 0b00010
F_STOPWORD = 0b00100
F_NUMBER = 0b01000
FILTER_NONE = 0b00000
FILTER_ALL = F_LOWERCASE | F_PUNCTUATION | F_STOPWORD | F_NUMBER
def _is_sent_terminator(token):
return token in TERMINATION_CHARS
def _is_number(token):
try:
def __init__(self):
self.logger = get_logger(self.__class__.__name__.lower())
def __init__(self, max_retires=None):
self.max_retries = max_retires
self.tasks = []
self.results = []
self.logger = get_logger(self.__class__.__name__.lower())
def __init__(self, classifiers: Iterable):
self.train_data = None
self.test_data = None
self.classifiers = classifiers
self.logger = get_logger(self.__class__.__name__.lower())
def __init__(self, check_url: str = DEFAULT_CHECK_URL):
self.check_url = check_url
self._instance_proxies = None
self._aiter = None
self.logger = get_logger(self.__class__.__name__.lower())
