def test_split_simple(self):
assert simple_word_tokenize('Мама мыла раму') == [
'Мама', 'мыла', 'раму'
]
assert simple_word_tokenize('Постой, паровоз!') == [
'Постой', ',', 'паровоз', '!'
]
def general_data():
path_opencorpora = "." + os.sep + "!data" + os.sep + \
"newcorpus" + os.sep + "OpenCorpora_txt_clean"
files = [
item for item in os.listdir(path_opencorpora) if item.endswith(".txt")
]
total_pars = []
total_avs = []
total_pars_count = 0
for file in files:
with open(path_opencorpora + os.sep + file, "r",
encoding="utf-8") as f:
pars = f.readlines()
total_pars_count += len(pars)
total_pars.append(len(pars))
try:
av = []
for par in pars:
av.append(len(tokenizers.simple_word_tokenize(par)))
total_avs.append(mean(av))
except:
# это пустые файлы
pass
print("В среднем слов в абзаце: {:.2f}\n".format(mean(total_avs)) +
"Всего абзацев: {:,}\n".format(total_pars_count).replace(",", " ") +
"В среднем абзацев в тексте: {:.2f}".format(mean(total_pars)))
def test_exctract_words(self):
text = '''Это отразилось: на количественном,и на качествен_ном
- росте карельско-финляндского сотрудничества - офигеть! кони+лошади=масло.
-сказал кто-то --нет--'''
assert simple_word_tokenize(text) == [
'Это',
'отразилось',
':',
'на',
'количественном',
',',
'и',
'на',
'качествен_ном',
'-',
'росте',
'карельско-финляндского',
'сотрудничества',
'-',
'офигеть',
'!',
'кони',
'+',
'лошади',
'=',
'масло',
'.',
'-сказал',
'кто-то',
'--нет--',
]
def lemmatize(s):
l = [
morph.parse(w)[0].normal_form for w in simple_word_tokenize(s)
if not is_punctuation(w)
]
l = [w for w in l if w not in stop_words]
return " ".join(l)
def api_morphy(request):
query = request.json_body
if query["all"]:
words = simple_word_tokenize(query["phrase"])
new_phrase = []
for word in words:
new_phrase.append(lean(word, case=query["case"]))
return {"phrase": " ".join(new_phrase)}
def pymorphy_simple(text):
"""
Разбивает текст на слова функцией из pymorphy2
:param text: Входной текст
:return: список слов
"""
words = simple_word_tokenize(text)
return words
def lemmatize_text(text):
text = re.sub("\s+", " ", text)
split_text = tokenizers.simple_word_tokenize(text)
lemmatized_text = [lemmatize(t) for t in split_text]
lemmas = [l[0] for l in lemmatized_text]
t_tags = [l[1] for l in lemmatized_text]
tags = [t_tags.count(gm) for gm in gram]
words_known = sum([word_is_known(t) for t in split_text])
return ' '.join(lemmas), words_known, tags
def tokenize(definition: str) -> List[str]:
"""
разбивает строку на токены
:param definition: строка - определение
:return: токены (без знаков препинания)
"""
return [
x for x in simple_word_tokenize(definition) if x not in punctuation
]
def test_exctract_words(self):
text = '''Это отразилось: на количественном,и на качествен_ном
- росте карельско-финляндского сотрудничества - офигеть! кони+лошади=масло.
-сказал кто-то --нет--'''
assert simple_word_tokenize(text) == [
'Это', 'отразилось', ':', 'на', 'количественном', ',', 'и', 'на',
'качествен_ном', '-', 'росте', 'карельско-финляндского',
'сотрудничества', '-', 'офигеть', '!', 'кони', '+', 'лошади',
'=', 'масло', '.', '-сказал', 'кто-то', '--нет--',
]
def combine_sent(sent, lemma_seq, pos_seq):
tokens = simple_word_tokenize(sent)
norm_tokens = [token for token in tokens if token[0].isalpha()]
if type(lemma_seq) != str:
return []
lemmas = lemma_seq.split()
pos = pos_seq.split()
combined_list = []
for i in range(len(pos)):
combined_list.append(' ' + norm_tokens[i] + '|' + lemmas[i] + '|' +
pos[i])
return combined_list
def str_handler(in_string):
"""Обработчик строк. Удаление лишних символов, детранслитерация, нормализация(приведение к инфинитиву).
Keyword arguments:
in_string -- строка для обработки
"""
tokens = tokenizers.simple_word_tokenize(re.sub('[!?,.%]', '', in_string))
new_string = ''
for word_in in tokens:
if re.search(latin_pattern, word_in):
word_in = translit.detranslify(word_in)
new_string += morph.parse(word_in)[0].normal_form + ' '
return new_string
def pos(self, line):
self.pos_dict = {}
# parsing
if self.language == "rus":
play_pos = [
self.analyzer.parse(token)[0].tag.POS
for token in simple_word_tokenize(line)
]
elif self.language in self.spacy_analyzers.keys():
play_pos = [token.pos_ for token in self.analyzer(line)]
elif self.language in self.cltk_analyzers.keys():
print(self.language)
play_pos = self.analyzer.analyze(text=line).pos
return play_pos
def tokenize(text_data):
text_data = str(text_data).lower() # к нижнему регистру
text_data = re.sub(
'\s+', ' ',
text_data) # нормализовать все переносы и табы к просто пробелу
text_data = re.sub('\[id\d*\|\w*\]', 'username',
text_data) # замена всех упоминаний пользователей
text_data = re.sub('"', '', text_data) # удаление двойных кавычек
text_data = re.sub("'", '', text_data) # удаление апострофов
text_data = re.sub('ё', 'е', text_data) # замена Ё на Е
text_data = re.sub(r'(?<=\w)\*', 'ё',
text_data) # замена * на Ё, если перед ним буква
tokens = simple_word_tokenize(text_data) # токенизация
return ' '.join(tokens)
def tokenizeSingleText(text, configurations):
minimal_words_in_sentence = 1
if (configurations != None):
minimal_words_in_sentence = configurations.get(
"minimal_words_in_sentence", 1)
remove_index_list = []
for index, sentence in enumerate(text.original_sentences):
if (len(sentence) > 1):
tokenized_sentence = tokenizers.simple_word_tokenize(sentence)
updated_tokenized_sentence = []
for word in tokenized_sentence:
if word.isalpha() and len(word) > 1:
updated_tokenized_sentence.append(word)
tokenized_sentence = updated_tokenized_sentence
if (len(tokenized_sentence) >= minimal_words_in_sentence):
text.tokenized_sentences.append(tokenized_sentence)
else:
remove_index_list.append(index)
else:
remove_index_list.append(index)
# Печать предложений перед вырезкой предложений
# string_for_print = ''
# for index, sentence in enumerate(text.original_sentences):
# string_for_print = string_for_print + str(index) + ')' + sentence + '\n'
# writeStringToFile(string_for_print, 'output_files/preProcessing_before_cut.txt')
#
need_agresive_filtration = False
if (configurations != None):
need_agresive_filtration = configurations.get(
"need_agresive_filtration", False)
sorted_remove_index_list = sorted(remove_index_list,
key=lambda x: x,
reverse=True)
if (need_agresive_filtration):
for index in sorted_remove_index_list:
text.original_sentences.pop(index)
# Печать предложений после вырезки
# string_for_print = ''
# for index, sentence in enumerate(text.original_sentences):
# string_for_print = string_for_print + str(index) + ')' + sentence + '\n'
# writeStringToFile(string_for_print, 'output_files/preProcessing_after_cut.txt')
return text.tokenized_sentences
def lemmatize(self, line):
self.lemmas = []
play_lemmas = []
if self.language == "rus":
play_lemmas = [
self.analyzer.parse(token)[0].normal_form
for token in simple_word_tokenize(line)
]
elif self.language in self.spacy_analyzers.keys():
play_lemmas = [token.lemma_ for token in self.analyzer(line)]
elif self.language in self.cltk_analyzers.keys():
print(self.language)
play_lemmas = self.analyzer.analyze(text=line).lemmata
self.lemmas += play_lemmas
return " ".join(play_lemmas)
def main():
morph = MorphAnalyzer()
texts_path = '..' + os.sep + '..' + os.sep + '..' + os.sep + '..' + os.sep + 'RuCoref' + os.sep + 'rucoref_texts'
for folder in os.listdir(texts_path):
text_folder = texts_path + os.sep + folder
for filename in os.listdir(text_folder):
if filename.endswith('.txt'):
with open(text_folder + os.sep + filename,
'r',
encoding='utf-8') as source_file:
source_text = source_file.read()
# get tokens
tokens = tokenizers.simple_word_tokenize(source_text)
# parse tokens
tokens_with_tags = morphology_features(tokens, morph)
# write tokens to new file
write_info(tokens_with_tags, text_folder, filename)
def tokenize_me(file_text):
file_text = file_text.lower()
tokens = tokenizers.simple_word_tokenize(file_text)
tokens = [morph.parse(w)[0].normal_form for w in tokens]
#deleting punctuation symbols
tokens = [i for i in tokens if (i not in string.punctuation)]
#deleting stop_words
stop_words = stopwords.words('russian')
stop_words.extend([
'что', 'это', 'так', 'вот', 'быть', 'как', 'в', '—', 'к', 'на', '...'
])
tokens = [i for i in tokens if (i not in stop_words)]
return ' '.join(tokens)
def _get_gender(self, profession: str) -> GENDER:
"""
Predict gender, without using cache
"""
if not profession.strip():
# Empty string
return GENDER.unknown
toks = simple_word_tokenize(profession)
observed_genders = [self.get_word_gender(tok) for tok in toks]
if not observed_genders:
# No observed gendered words - return neutral
return GENDER.neutral
# Return the most commonly observed gender
return Counter(observed_genders).most_common()[0][0]
def normalized(twit):
twit = re.sub(r'#\S+', '', #вычищение хештегов
re.sub(r'@\S+', '', #вычищение @name
re.sub(r'http\S+', '', #вычищение ссылок
re.sub(r'RT ', '', #убрать RT
re.sub("\d+", "", twit)))) #убираем все цифры
).strip()
tokens = simple_word_tokenize(twit)
parses = [
morph.parse(w)[0]
for w in tokens
if w.lower() not in stopwords
]
parses = [
p for p in parses
if p.tag.POS not in {'PNCT', 'UNKN', 'NUMB', 'CONJ', 'LATN'}
]
return [p.normal_form.lower() for p in parses]
def api_morphy(request):
query = request.json_body
command = query["command"]
phrase = query["phrase"]
if command == "all":
words = simple_word_tokenize(phrase)
gender = russian.gender(words, "M", "F", "_")
if gender == "_":
new_phrase = []
for word in words:
new_phrase.append(lean(word, case=query["case"]))
else:
case_ = query["case"]
c = TR.get(case_, case_)
new_phrase = russian.make_human_case(words, c)
return {"phrase": " ".join(new_phrase)}
elif command == "gender":
p = russian.gender(phrase, query["M"], query["F"], ";-)")
return {"phrase": p}
def preprocessing(sentence, clist):
s = re.sub('[^а-яА-Яa-zA-Z]+', ' ', sentence).strip().lower()
s = re.sub('ё', 'е', s)
result = []
for word in tokenizers.simple_word_tokenize(s):
if word not in clist:
if not morph.word_is_known(word):
new_words = split_word(word)
else:
new_words = word,
for new_word in new_words:
parse = morph.parse(new_word)[0]
pos = parse.tag.POS
if pos is not None and pos not in [
'NPRO', 'PREP', 'CONJ', 'PRCL', 'INTJ'
]:
result.append(parse.normal_form)
else:
result.append(word)
return ' '.join(result)
def normalize_text(text):
lemmas = []
for t in simple_word_tokenize(text):
if t not in stops:
lemmas.append(m.parse(t)[0].normal_form)
return ' '.join(lemmas)
def tokenize_if_needed(tokens):
if not isinstance(tokens, (list, tuple)):
return simple_word_tokenize(tokens)
return tokens
def predict(self, tokens):
if not isinstance(tokens, (list, tuple)):
tokens = simple_word_tokenize(tokens)
return [self.morph.parse(token)[0] for token in tokens]
def tokenize(self, text):
return simple_word_tokenize(text)
def test_split_signs(self):
assert simple_word_tokenize('a+b=c_1') == ['a', '+', 'b', '=', 'c_1']
def read_text_lemmas(fileobj):
m = MorphAnalyzer()
for line in fileobj:
yield ' '.join((m.parse(t)[0].normal_form for t in simple_word_tokenize(line.decode('utf-8'))))
def tokenize(text):
return tokenizers.simple_word_tokenize(text)
def test_split_simple(self):
assert simple_word_tokenize('Мама мыла раму') == ['Мама', 'мыла', 'раму']
assert simple_word_tokenize('Постой, паровоз!') == ['Постой', ',', 'паровоз', '!']
def pos_tag(text):
return [f"{m.parse(word)[0].normal_form}_{m.parse(word)[0].tag.POS}"
for word in simple_word_tokenize(text)]
def test_split_signs(self):
assert simple_word_tokenize('a+b=c_1') == ['a','+','b','=','c_1']
import pickle
from pymorphy2 import MorphAnalyzer, tokenizers
from constants import PAGES_PATH
morph = MorphAnalyzer()
inverted_index = {}
page_occurrences = {}
pages = os.listdir(path=PAGES_PATH)
for index, page in enumerate(pages):
file = open(PAGES_PATH + page, 'r', encoding="utf-8")
text = file.read()
tokens = tokenizers.simple_word_tokenize(text)
page_occurrences[index] = len(tokens)
for token in tokens:
lemma = morph.parse(token)[0].normal_form.lower()
value = inverted_index.get(lemma)
if value is None:
inverted_index[lemma] = {index: 1}
elif inverted_index[lemma].get(index) is None:
inverted_index[lemma][index] = 1
else:
inverted_index[lemma][index] += 1
inverted_index_file = open("inverted_index.pkl", "wb")
def test_split_hyphen(self):
assert simple_word_tokenize('Ростов-на-Дону') == ['Ростов-на-Дону']
assert simple_word_tokenize('Ура - победа') == ['Ура', '-', 'победа']
def get_sents(input_data):
return [normalized(simple_word_tokenize(sent)) for sent in input_data]
def _iter_tokens_tokenize(fp):
""" Return an iterator of input tokens; each line is tokenized """
return (token for line in fp for token in simple_word_tokenize(line))
def lemmatize(text):
return [
m.parse(word)[0].normal_form for word in simple_word_tokenize(text)
]
def _iter_tokens_tokenize(fp):
""" Return an iterator of input tokens; each line is tokenized """
return (token for line in fp for token in simple_word_tokenize(line))
def test_split_hyphen(self):
assert simple_word_tokenize('Ростов-на-Дону') == ['Ростов-на-Дону']
assert simple_word_tokenize('Ура - победа') == ['Ура', '-', 'победа']
def tokenize(line):
return [token for token in simple_word_tokenize(line) if token not in string.punctuation]
