docs: streamline usage guides by consolidating and decluttering

docs: add comprehensive usage manual for MoYoYo.tts

refactor(config): centralize configuration management in `project_config`

feat(tts): prevent config saving for dict-based initialization

feat(tts-voice-app): add model download setup page for first-time initialization

feat(tts-voice-app): add range request handling for audio playback

feat(cli): add `infer.py` script for GPT-SoVITS inference

feat(tts-app): redesign text-to-speech interface with voice library integration

feat(voice-library): Fix voice playback issues and enhance page functionality

feat(api/tts-voice-app): add Cantonese (yue) support and improve language-specific defaults

feat(tts-voice-app): implement voice library management and inference output features

feat(api): enhance task file management and download endpoints

feat(api): add endpoints for inference output management

feat(inference): add `_parse_list_file` to handle default values for `ref_audio_path` and `ref_text`

Update training progress messages for consistency and clarity

feat(frontend): implement Voice Clone training task management

feat(dependencies): add `aiosqlite` and `torchcodec` to `pyproject.toml`

feat(api): implement local training MVP with adapter pattern

feat(api): implement AsyncTrainingManager MVP with SQLite persistence

Use `tempfile` to handle temporary configuration files and ensure cleanup in training stages

Refactor config classes and inference pipeline to improve path handling, weight management, and modularity

Refactor config classes to generate dynamic paths and update default pretrained model references

Update all module imports to use `GPT_SoVITS` paths for consistency and clarity

Add `torchcodec` dependency and index configuration in `pyproject.toml`

Ensure target directory for weight saving exists during training

Make accelerator selection dynamic based on available hardware (CUDA, MPS, CPU) in the training pipeline

Update imports to use `GPT_SoVITS` module paths and replace `torchaudio` with `librosa` for loading reference audio

Rename project references from `GPT-SoVITS-WebUI` to `MoYoYo.tts-WebUI` across documentation

Add `pyproject.toml` with project metadata and dependencies

Add training pipeline with modular stage-based architecture

Add `.dockerignore`, `.gitignore`, and pre-commit configuration files

Upload 296 files

initial commit