"""Grapheme to Phoneme related modeling class"""
import os
from typing import List, Optional, Union
from pororo.tasks.utils.base import PororoFactoryBase, PororoSimpleBase
[docs]class PororoG2pFactory(PororoFactoryBase):
"""
Conduct grapheme to phoneme conversion
English (`g2p.en`)
- dataset: g2pE
- metric: N/A
- ref: https://github.com/Kyubyong/g2p
Korean (`g2p.ko`)
- dataset: g2pK
- metric: N/A
- ref: https://github.com/Kyubyong/g2pK
Japanese (`g2p.ja`)
- dataset: romkan
- metric: N/A
- ref: https://github.com/soimort/python-romkan
Chinese (`g2p.zh`)
- dataset: g2pM (Park et al. 2020)
- metric: N/A
- ref: https://github.com/kakaobrain/g2pM
Args:
sent (str): input sentence to be converted to phoneme
align (bool): whether to align the result (not applied to `english` model)
Returns:
str: converted phoneme sentence
Examples:
>>> g2p = Pororo(task="g2p", lang="ko")
>>> g2p("어제는 날씨가 맑았는데, 오늘은 흐리다.")
'어제는 날씨가 말간는데, 오느른 흐리다.'
>>> g2p("어제는 날씨가 맑았는데, 오늘은 흐리다.", align=True)
[('어제는', '어제는'), ('날씨가', '날씨가'), ('맑았는데,', '말간는데,'), ('오늘은', '오느른'), ('흐리다.', '흐리다.')]
>>> g2p_en = Pororo(task="g2p", lang="en")
>>> g2p_en("I have $250 in my pocket.")
['AY1', ' ', 'HH', 'AE1', 'V', ' ', 'T', 'UW1', ' ', 'HH', 'AH1', 'N', 'D', 'R', 'AH0', 'D', ' ', 'F', 'IH1', 'F', 'T', 'IY0', ' ', 'D', 'AA1', 'L', 'ER0', 'Z', ' ', 'IH0', 'N', ' ', 'M', 'AY1', ' ', 'P', 'AA1', 'K', 'AH0', 'T', ' ', '.']
>>> g2p_zh = Pororo(task="g2p", lang="zh")
>>> g2p_zh("然而，他红了20年以后，他竟退出了大家的视线。")
'ran2 er2 , ta1 hong2 le5 2 0 nian2 yi3 hou4 , ta1 jing4 tui4 chu1 le5 da4 jia1 de5 shi4 xian4 。'
>>> g2p_zh("然而，他红了20年以后，他竟退出了大家的视线。", align=True)
[('然', 'ran2'), ('而', 'er2'), (',', ','), ('他', 'ta1'), ('红', 'hong2'), ('了', 'le5'), ('2', '2'), ('0', '0'), ('年', 'nian2'), ...]
>>> g2p_zh("然而，他红了20年以后，他竟退出了大家的视线。", align=True, tone=False)
[('然', 'ran'), ('而', 'er'), (',', ','), ('他', 'ta'), ('红', 'hong'), ('了', 'le'), ('2', '2'), ('0', '0'), ('年', 'nian'), ...]
>>> g2p_ja = Pororo(task="g2p", lang="ja")
>>> g2p_ja("pythonが大好きです")
'python ga daisuki desu'
>>> g2p_ja("pythonが大好きです", align=True)
[('python', 'python'), ('が', 'ga'), ('大好き', 'daisuki'), ('です', 'desu')]
"""
def __init__(self, task: str, lang: str, model: Optional[str]):
super().__init__(task, lang, model)
[docs] @staticmethod
def get_available_langs():
return ["en", "ko", "zh", "ja"]
[docs] @staticmethod
def get_available_models():
return {
"en": ["g2p.en"],
"ko": ["g2p.ko"],
"zh": ["g2p.zh"],
"ja": ["g2p.ja"],
}
[docs] def load(self, device: str):
"""
Load user-selected task-specific model
Args:
device (str): device information
Returns:
object: User-selected task-specific model
"""
if self.config.n_model == "g2p.ko":
try:
from g2pk import G2p as G2pK
except ModuleNotFoundError as error:
raise error.__class__(
"Please install g2pk with: `pip install g2pk`")
model = G2pK()
return PororoG2PKo(model, self.config)
if self.config.n_model == "g2p.en":
try:
from g2p_en import G2p as G2pE
except ModuleNotFoundError as error:
raise error.__class__(
"Please install g2p_en with: `pip install g2p_en`")
model = G2pE()
return PororoG2PEn(model, self.config)
if self.config.n_model == "g2p.zh":
try:
from g2pM import G2pM
except ModuleNotFoundError as error:
raise error.__class__(
"Please install g2pM with: `pip install g2pM`")
model = G2pM()
return PororoG2PZh(model, self.config)
if self.config.n_model == "g2p.ja":
try:
import fugashi
except ModuleNotFoundError as error:
raise error.__class__(
"Please install fugashi with: `pip install fugashi`")
try:
import ipadic
except ModuleNotFoundError as error:
raise error.__class__(
"Please install ipadic with: `pip install ipadic`")
try:
import romkan
except ModuleNotFoundError as error:
raise error.__class__(
"Please install ipadic with: `pip install romkan`")
dic_dir = ipadic.DICDIR
mecabrc = os.path.join(dic_dir, "mecabrc")
mecab_option = "-d {} -r {} ".format(
dic_dir,
mecabrc,
)
tagger = fugashi.GenericTagger(mecab_option)
return PororoG2PJa(tagger, romkan.to_roma, self.config)
[docs]class PororoG2PKo(PororoSimpleBase):
def __init__(self, model, config):
super().__init__(config)
self._model = model
[docs] def predict(self, text: str, **kwargs) -> Union[List, str]:
"""
Conduct grapheme to phoneme conversion
Args:
text (str): input sentence to be converted to phoneme
align (bool): whether to align the result
Returns:
str: converted phoneme sentence
"""
align = kwargs.get("align", False)
results = self._model(text)
if align:
return [(word, phoneme)
for word, phoneme in zip(text.split(), results.split())]
return results
[docs]class PororoG2PEn(PororoSimpleBase):
def __init__(self, model, config):
super().__init__(config)
self._model = model
[docs] def predict(self, text: str, **kwargs) -> str:
"""
Conduct grapheme to phoneme conversion
Args:
sent (str): input sentence to be converted to phoneme
Returns:
str: converted phoneme sentence
"""
return self._model(text)
[docs]class PororoG2PZh(PororoSimpleBase):
def __init__(self, model, config):
super().__init__(config)
self._model = model
[docs] def predict(self, text: str, **kwargs) -> Union[List, str]:
"""
Conduct grapheme to phoneme conversion
Args:
text (str): input sentence to be converted to phoneme
align (bool): whether to align the result
tone (bool): whether to show chinese tone
Returns:
str: converted phoneme sentence
"""
align = kwargs.get("align", False)
tone = kwargs.get("tone", True)
results = self._model(text, tone=tone, char_split=True)
if align:
return [(c, t) for c, t in zip(text, results)]
return " ".join(results)
[docs]class PororoG2PJa(PororoSimpleBase):
def __init__(self, tagger, romanize, config):
super().__init__(config)
self._tagger = tagger
self._romanize = romanize
[docs] def predict(self, text: str, **kwargs):
"""
Conduct grapheme to phoneme conversion
Args:
text (str): input sentence to be converted to phoneme
align (bool): whether to align the result
Returns:
List[str]: converted phoneme sentence list
"""
align = kwargs.get("align", False)
output = self._tagger.parse(text.strip())
results = list()
for line in output.splitlines():
res = line.split("\t")
if len(res) > 1:
orig = res[0]
feats = res[1].split(",")
reading = feats[-1]
if reading == "*":
reading = orig
romaji = self._romanize(reading)
results.append((orig, romaji))
if align:
return results
return " ".join([result[1] for result in results])