refactor(core): Architecturally decouple Audio, ASR, and TTS modules

添加获取TTS模型参考音频的API路由，支持通过模型ID获取音频文件。

调整prompt的逻辑，确保展示的提示词中不包含/no_think指令

添加设置相关的API路由

Add echo cancellation and VAD toggle support in service factories and routes

Enhance WebSocket handling for connection management and reliability

Refactor WebSocket handling with connection manager

Replace `logging` with centralized `loguru`-based logger across all modules.

Add robust lifecycle management for `audio_player` service in system routes

Standardize service lifecycle management by replacing `stop` with `exit` and introducing `is_exited` check

Refactor speech recognizer, audio capture, and system routes for improved clarity and functionality

Add pause and resume functionality to voice dialogue system

Rename 'src/VoiceDialogue' to 'src/voice_dialogue'.