Add MiniMax async and WebSocket TTS operations

[octo-patch]

Reason: Extend the MiniMax TTS adapter with async task and WebSocket synthesis operations.

• Add async create and query methods that use the configured API base and validate API errors.
• Add WebSocket synthesis with protocol event validation and audio frame assembly.
• Normalize the default voice setting payload and filter operation-specific request fields.
• Add focused tests for HTTP paths, query payloads, regional WebSocket URLs, and audio frames.

Checks:

• uv run --no-project --python 3.11 python tests/models_provider/test_minimax_tts.py
• uv run --no-project --python 3.11 python -m py_compile apps/models_provider/impl/minimax_model_provider/model/tts.py tests/models_provider/test_minimax_tts.py
• uvx --from ruff==0.15.12 ruff check apps/models_provider/impl/minimax_model_provider/model/tts.py tests/models_provider/test_minimax_tts.py
• uvx --from ruff==0.15.12 ruff format --check apps/models_provider/impl/minimax_model_provider/model/tts.py tests/models_provider/test_minimax_tts.py
• git diff --check

[kapelame]

Updated the MiniMax WebSocket synthesis flow to process task_continued frames, send task_finish, and validate task_finished. I ran the focused unit tests, Python compilation, Ruff check/format, and git diff --check.

[liuruibin]

no plan