Function processAudioGeneration

outdated/gemini/gemini.ts:1143–1215

(
msg: Api.Message, text: string, p0: string,
replyMsg?: Api.Message | null)

}

async function processAudioGeneration(
msg: Api.Message, text: string, p0: string,
replyMsg?: Api.Message | null): Promise<void> {
  const client = await getGeminiClient();
  const modelName = ConfigManager.get(CONFIG_KEYS.GEMINI_TTS_MODEL, DEFAULT_CONFIG[CONFIG_KEYS.GEMINI_TTS_MODEL]);
  const voiceName = ConfigManager.get(CONFIG_KEYS.GEMINI_TTS_VOICE, DEFAULT_CONFIG[CONFIG_KEYS.GEMINI_TTS_VOICE]);

  const response = await client.generateTTS({
    model: modelName,
    contents: [{ parts: [{ text }] }],
    voiceName
  });

  if (!response.audioData?.length) {
    throw new Error('没有收到音频数据');
  }

  const combinedAudio = Buffer.concat(response.audioData);
  if (combinedAudio.length === 0) {
    throw new Error('合并后的音频数据为空');
  }

  if (replyMsg) {
    let processedAudio: any = combinedAudio;
    
    if (response.audioMimeType && response.audioMimeType.includes('L16') && response.audioMimeType.includes('pcm')) {
      processedAudio = Utils.convertToWav(combinedAudio, response.audioMimeType) as any;
    }

    const audioFile = Object.assign(processedAudio as any, {
      name: 'gemini.ogg'
    });

    await msg.client?.sendFile(msg.peerId, {
      file: audioFile,
      caption: `<b>文本:</b> ${Utils.escapeHtml(text)}\n\n<i>Powered by Gemini TTS (${voiceName})</i>`,
      parseMode: "html",
      replyTo: replyMsg.id,
      attributes: [new Api.DocumentAttributeAudio({
        duration: 0,
        voice: true
      })]
    });

    try {
      await msg.delete();
    } catch {}
  } else {
    let processedAudio: any = combinedAudio;
    
    if (response.audioMimeType && response.audioMimeType.includes('L16') && response.audioMimeType.includes('pcm')) {
      processedAudio = Utils.convertToWav(combinedAudio, response.audioMimeType) as any;
    }

    const audioFile = Object.assign(processedAudio as any, {
      name: 'gemini.ogg'
    });