wolfang/anki-hsk-creator
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

change from chaterbox to gwen-tts

Browse Source
This commit is contained in:
Wolfang Torres
2026-06-25 20:58:44 +08:00
parent df64a5fc15
commit d153a441ed
5 changed files with 34 additions and 27 deletions
Download Patch File Download Diff File Expand all files Collapse all files

20
src/anki_hsk_creator/proccessor.py
View File

@@ -8,7 +8,7 @@ import argostranslate.translate
import torchaudio
# Local
from .constants import LANGUAGES, WORD, COMMA
from .constants import LANGUAGES
from .utility import CCCEDICT, TTS, DictionaryResult, ProcessFile, TranslationResult
# Constants
@@ -31,10 +31,8 @@ def dictation_process(
audio_line = audio_line.replace("，", "，。。。]")
audio_path = process_file.resources / f"N{n:03n}.wav"
if not audio_path.exists():
audio = TTS.MODEL.generate(
f"{audio_line}。", language_id=LANGUAGES.CN, **TTS.DEFAULTS
)
torchaudio.save(audio_path, audio, TTS.MODEL.sr)
wavs, sr = TTS.generate(f"{audio_line}。")
torchaudio.save(audio_path, wavs[0], sr)
translated = argostranslate.translate.translate(
line, LANGUAGES.CN, process_file.language_id
)
@@ -56,10 +54,8 @@ def translator_process(
audio_line = audio_line.replace("，", "，。。。]")
audio_path = process_file.resources / f"N{n:03n}.wav"
if not audio_path.exists():
audio = TTS.MODEL.generate(
f"{audio_line}。", language_id=LANGUAGES.CN, **TTS.DEFAULTS
)
torchaudio.save(audio_path, audio, TTS.MODEL.sr)
wavs, sr = TTS.generate(f"{audio_line}。")
torchaudio.save(audio_path, wavs[0], sr)
translated = argostranslate.translate.translate(
line, LANGUAGES.CN, process_file.language_id
)
@@ -121,10 +117,8 @@ def dictionary_process(process_file: ProcessFile) -> list[DictionaryResult]:
for line in reader:
audio_path = process_file.resources / f"{line['pinyin']}.wav"
if not audio_path.exists():
audio = TTS.MODEL.generate(
f"{line['simplified']}。", language_id=LANGUAGES.CN, **TTS.DEFAULTS
)
torchaudio.save(audio_path, audio, TTS.MODEL.sr)
wavs, sr = TTS.generate(f"{line['simplified']}。")
torchaudio.save(audio_path, wavs[0], sr)
print(line)
result = DictionaryResult(
**line, audio_path=audio_path, language_id=process_file.language_id