# The Virtual One
TheVirtualOne, 打造你的专属话痨

# Install/安装
开发时使用python版本为3.10.19

## clone/拉取仓库
选取一种方法拉取完整仓库

1. 直接拉取
```bash
git clone 仓库.git --recurse-submodules
cd 仓库
```

2. 手动拉取必要内容
```bash
git clone 仓库.git
cd 仓库
git clone -b main https://github.com/NINEMINEsigma/Convention-Python.git Convention # 或者使用仓库http://gitea.liubai.site/ninemine/Convention-Python.git
```

## requirements/依赖
安装python依赖
```bash
pip install -r requirements.txt
```

**Ollama** 准备好你要使用的ollama base url

**可选的语音合成** 如果你将使用tts那么需要准备CosyVocie2-0.5B

## config/配置
运行cli.py或web_server.py自动生成Assets/Config.json或手动生成
```java
{
    "properties": {
        "ollama_url": "http://localhost:11434",
        "ollama_model": "gemma3:4b",
        "server_port": 11451,
        "server_host": "0.0.0.0",
        "response_timeout": 300,
        "tts_host": "localhost",
        "tts_port": 43400,
        "tts_prompt_wav_path": "StreamingAssets/tts prompt.mp3",
        "tts_prompt_text": "tts prompt text",
        "tts_speaker_id": "speaker",
        "tts_timeout": 180,
        "sentence_max_length": 100,
        "tts_enable": true,//可选
        "system_prompt_path": "Assets/prompt.txt",
        "stream_enable": false,//web_server.py不要开启
        "auto_speak_wait_second": 45.0,
        "tts_server_url": "http://localhost:43400"
    }
}
```

# start/开始
如需要tts服务, 需要
```bash
python tts_server.py
```

后可启动服务
```bash
python cli.py #或web_server.py
```

# 项目引用代码
[FunAudioLLM/CosyVoice](https://github.com/FunAudioLLM/CosyVoice)