jtalkを用いた読み上げコマンド

jtalk.py

#!/usr/bin/env python3

# 参考
# https://qiita.com/kkoba84/items/b828229c374a249965a9
# https://mekou.com/linux-magazine/open-jtalklinux-%E6%97%A5%E6%9C%AC%E8%AA%9E%E9%9F%B3%E5%A3%B0%E5%90%88%E6%88%90%E3%82%A8%E3%83%B3%E3%82%B8%E3%83%B3/

# Ubuntu 20.04で動作検証

SETUP = """
# (1) Docoptのインストール
sudo python3 -m pip install docopt

# (2) Open JTalkのセットアップ
sudo apt-get install open-jtalk open-jtalk-mecab-naist-jdic hts-voice-nitech-jp-atr503-m001

# (3) 音声ファイルのセットアップ
wget https://sourceforge.net/projects/mmdagent/files/MMDAgent_Example/MMDAgent_Example-1.8/MMDAgent_Example-1.8.zip/download -O MMDAgent_Example-1.8.zip
unzip MMDAgent_Example-1.8.zip
sudo cp -r MMDAgent_Example-1.8/Voice/mei/ /usr/share/hts-voice
"""

import re
import sys
import subprocess

from docopt import docopt


wav_file_template = "/tmp/open_jtalk_%d.wav"


def speech_lines(lines, show_text=False):
    open_jtalk=['open_jtalk']
    mech=['-x','/var/lib/mecab/dic/open-jtalk/naist-jdic']
    htsvoice=['-m','/usr/share/hts-voice/mei/mei_normal.htsvoice']
    speed=['-r','1.0']
    outwav=['-ow']
    wav_gen_cmd = open_jtalk + mech + htsvoice + speed + outwav
    play_cmd = ['aplay','-q']

    speech_process = None
    for i, L in enumerate(lines):
        wav_file = wav_file_template % (i % 2)
        c = subprocess.Popen(wav_gen_cmd + [wav_file], stdin=subprocess.PIPE)
        c.stdin.write(L.encode())
        c.stdin.close()
        c.wait()
        if speech_process is not None:
            speech_process.wait()
        if show_text:
            print(L, file=sys.stderr)
        speech_process= subprocess.Popen(play_cmd + [wav_file])
    if speech_process is not None:
        speech_process.wait()


def now_text():
    import locale
    from datetime import datetime

    locale.setlocale(locale.LC_TIME, 'ja_JP.UTF-8')
    d = datetime.now()
    wd = d.strftime('%A')  # 曜日
    text = '%s月%s日%s、%s時%s分%s秒' % (d.month, d.day, wd, d.hour, d.minute, d.second)
    return text


__doc__ = """日本語テキストを読み上げます。

Usage:
  jtalk [options] [<textfile>]

Options:
  -N        改行で文を区切らないようにする
  -t        発声されているテキストを表示する
"""


def main():
    args = docopt(__doc__)
    if not args['<textfile>']:
        text = now_text()
    elif args['<textfile>'] == '-':
        text = sys.stdin.read()
    else:
        with open(args['<textfile>']) as inp:
            text = inp.read()
    
    if args['-N']:
        text = text.replace('\n\n', '。')
        text = text.replace('\n', '')
        lines = re.split(r"。", text)
    else:
        lines = re.split(r"[。\n]", text)
    lines = [L for L in lines if L]

    speech_lines(lines, show_text=args['-t'])


if __name__ == '__main__':
    main()