音声でChatGPTと疑似会話する環境を構築①【Python】

import requests, json
import io
import wave
import pyaudio
import time


class Voicevox:
    def __init__(self, host="127.0.0.1", port=50021):
        self.host = host
        self.port = port
        
    def speak(self, text=None, speaker=4):
        # 話し手を変える場合はspeakerのidを変更
        
        params = (
            ("text", text),
            ("speaker", speaker)
        )
        
        # VOICEVOX API ドキュメント
        # https://voicevox.github.io/voicevox_engine/api/
        # http://localhost:50021/docs
    
        init_q = requests.post(
            f"http://{self.host}:{self.port}/audio_query",
            params = params
        )
        
        res = requests.post(
            f"http://{self.host}:{self.port}/synthesis",
            headers = {"Content-Type": "application/json"},
            params = params,
            data = json.dumps(init_q.json())
        )
        
        # メモリ上で展開
        audio = io.BytesIO(res.content)
        
        with wave.open(audio, 'rb') as f:
            # 以下は再生処理
            p = pyaudio.PyAudio()
            
            def _callback(in_data, frame_count, time_info, status):
                data = f.readframes(frame_count)
                return (data, pyaudio.paContinue)
            
            stream = p.open(format = p.get_format_from_width(width = f.getsampwidth()),
            channels = f.getnchannels(),
            rate = f.getframerate(),
            output = True,
            stream_callback = _callback)
            
            # Voice再生
            stream.start_stream()
            while stream.is_active():
                time.sleep(0.1)
                
            stream.stop_stream()
            stream.close()
            p.terminate()


def main():
    vv = Voicevox()
    vv.speak(text = "こんにちは")
    

if __name__ == "__main__":
    main()