def setUpClass(cls):
cls.stress_predictor = CombinedStressPredictor(
stress_model_path=RU_STRESS_DEFAULT_MODEL,
zalyzniak_dict=ZALYZNYAK_DICT,
cmu_dict=CMU_DICT,
raw_stress_dict_path=RU_GRAPHEME_STRESS_PATH,
stress_trie_path=RU_GRAPHEME_STRESS_TRIE_PATH)
def inflate_vocab(self, top_n=None) -> None:
"""
Получение словаря с ударениями по этому словарю.
:param top_n: сколько первых записей взять?
"""
vocab = Vocabulary(GENERATOR_VOCAB_PATH)
stress_predictor = CombinedStressPredictor()
forms = self.word_forms
if top_n is not None:
forms = forms[:top_n]
for index, word_form in tqdm(enumerate(forms), desc="Accenting words"):
text = word_form.text
stresses = stress_predictor.predict(text)
word = Word(-1, -1, text, Graphemes.get_syllables(text))
word.set_stresses(stresses)
vocab.add_word(word, index)
vocab.save()
def setUpClass(cls):
cls.stress_predictor = CombinedStressPredictor(
stress_model_path=RU_STRESS_DEFAULT_MODEL,
zalyzniak_dict=ZALYZNYAK_DICT,
cmu_dict=CMU_DICT,
raw_stress_dict_path=RU_GRAPHEME_STRESS_PATH,
stress_trie_path=RU_GRAPHEME_STRESS_TRIE_PATH)
logging.basicConfig(stream=sys.stdout, level=logging.DEBUG)
def setUpClass(cls):
cls.stress_predictor = CombinedStressPredictor(
stress_model_path=RU_STRESS_DEFAULT_MODEL,
g2p_model_path=RU_G2P_DEFAULT_MODEL,
zalyzniak_dict=ZALYZNYAK_DICT,
ru_wiki_dict=RU_WIKI_DICT,
cmu_dict=CMU_DICT,
raw_stress_dict_path=RU_GRAPHEME_STRESS_PATH,
stress_trie_path=RU_GRAPHEME_STRESS_TRIE_PATH,
aligner_dump_path=RU_ALIGNER_DEFAULT_PATH,
g2p_dict_path=RU_G2P_DICT_PATH)
def inflate_vocab(self, dump_path, top_n=None) -> None:
"""
Получение словаря с ударениями по этому словарю.
:param top_n: сколько первых записей взять?
:param dump_path: путь, куда сохранить словарь.
"""
from rupo.main.vocabulary import Vocabulary
from rupo.stress.predictor import CombinedStressPredictor
vocab = Vocabulary(dump_path)
stress_predictor = CombinedStressPredictor()
forms = self.word_forms
if top_n is not None:
forms = forms[:top_n]
for index, word_form in tqdm(enumerate(forms), desc="Accenting words"):
text = word_form.text
stresses = stress_predictor.predict(text)
word = Word(-1, -1, text, Graphemes.get_syllables(text))
word.set_stresses(stresses)
vocab.add_word(word, index)
vocab.save()
def get_stress_predictor(self,
language="ru",
stress_model_path: str = None,
raw_stress_dict_path=None,
stress_trie_path=None,
zalyzniak_dict=ZALYZNYAK_DICT,
cmu_dict=CMU_DICT):
if self.stress_predictors.get(language) is None:
self.stress_predictors[language] = CombinedStressPredictor(
language, stress_model_path, raw_stress_dict_path,
stress_trie_path, zalyzniak_dict, cmu_dict)
return self.stress_predictors[language]
def inflate_vocab(self, dump_path, top_n=None) -> None:
"""
Получение словаря с ударениями по этому словарю.
:param top_n: сколько первых записей взять?
:param dump_path: путь, куда сохранить словарь.
"""
from rupo.main.vocabulary import StressVocabulary
from rupo.stress.word import StressedWord, Stress
from rupo.stress.predictor import CombinedStressPredictor
vocab = StressVocabulary(dump_path)
stress_predictor = CombinedStressPredictor()
forms = self.word_forms
if top_n is not None:
forms = forms[:top_n]
for index, word_form in tqdm(enumerate(forms), desc="Accenting words"):
text = word_form.text
stresses = [Stress(pos, Stress.Type.PRIMARY) for pos in stress_predictor.predict(text)]
word = StressedWord(text, set(stresses))
vocab.add_word(word, index)
vocab.save()
def get_stress_predictor(self,
language="ru",
stress_model_path: str = None,
g2p_model_path: str = None,
grapheme_set=RU_GRAPHEME_SET,
g2p_dict_path=None,
aligner_dump_path=None,
raw_stress_dict_path=None,
stress_trie_path=None,
zalyzniak_dict=ZALYZNYAK_DICT,
ru_wiki_dict=RU_WIKI_DICT,
cmu_dict=CMU_DICT):
if self.stress_predictors.get(language) is None:
self.stress_predictors[language] = CombinedStressPredictor(
language, stress_model_path, g2p_model_path, grapheme_set,
g2p_dict_path, aligner_dump_path, raw_stress_dict_path,
stress_trie_path, zalyzniak_dict, cmu_dict, ru_wiki_dict)
return self.stress_predictors[language]
def setUpClass(cls):
cls.stress_predictor = CombinedStressPredictor()
