Emacs用読み上げ拡張のIvisSpeech対応

#! /usr/bin/env python
"""
say_ivisspeech_engine.py

Reads text from standard input, performs speech synthesis, and plays
back speech.

* LICENSE

  This software is free software: you can redistribute it and/or modify
  it under the terms of the GNU General Public License as published by
  the Free Software Foundation, either version 3 of the License, or
  (at your option) any later version.

  This software is distributed in the hope that it will be useful,
  but WITHOUT ANY WARRANTY; without even the implied warranty of
  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  GNU General Public License for more details.

  You should have received a copy of the GNU General Public License
  along with GNU Emacs.  If not, see <https://www.gnu.org/licenses/>.

"""
import os.path
import pathlib
import sys
import time
import wave

import numpy as np
import sounddevice as sd

from voicevox_engine.aivm_manager import AivmManager
from voicevox_engine.core.core_initializer import MOCK_VER
from voicevox_engine.model import AudioQuery
from voicevox_engine.tts_pipeline.style_bert_vits2_tts_engine import StyleBertVITS2TTSEngine
from voicevox_engine.tts_pipeline.tts_engine import TTSEngineManager

__author__ = "TakesxiSximada"
__version__ = "1"

aivm_dir_path = os.path.expanduser("~/ng/symdon/articles/posts/1732150183")  # AVIMファイルを配置したディレクトリへのパス
aivm_manager = AivmManager(pathlib.PosixPath())
tts_engines = TTSEngineManager()
tts_engines.register_engine(StyleBertVITS2TTSEngine(aivm_manager, False, True), MOCK_VER)
engine = tts_engines.get_engine("0.0.0")
style_id = engine.aivm_manager.get_speakers()[0].styles[0].id

while True:
    text = sys.stdin.read().strip().replace("\n\n", "。")  # 見出しと本文を続けて読まないために、空行を句点で置き換える。

    if len(text) > 30:
        sentences = text.split("。")
    else:
        sentences = [text]

    for sentence in sentences:
        sentence = sentence.strip()
        if not sentence:  # 雑音が入ってしまう事になるため、空文字列の場合は無視する。
            continue

        accent_phrases = engine.create_accent_phrases(sentence, style_id)
        query = AudioQuery(
                accent_phrases=accent_phrases,
                speedScale=1.0,
                intonationScale=1.0,
                tempoDynamicsScale=1.0,
                pitchScale=0.0,
                volumeScale=1.0,
                prePhonemeLength=0.1,
                postPhonemeLength=0.1,
                pauseLength=None,
                pauseLengthScale=1,
                outputSamplingRate=engine.default_sampling_rate,
                outputStereo=False,
                kana=sentence)

        wave_data = engine.synthesize_wave(query, style_id, enable_interrogative_upspeak=True)

        sd.wait()
        sd.play(wave_data, samplerate=44100)  # , device=1)
        print(repr(sentence))

        output_file_path = f"voice-{str(time.time())}.wav"
        with wave.open(output_file_path, 'wb') as wf:
            wf.setnchannels(1)  # モノラル
            wf.setsampwidth(2)  # 16ビット
            wf.setframerate(engine.default_sampling_rate)
            wf.writeframes((wave_data * 32767).astype(np.int16).tobytes())  # WAVファイルに書き出す