Add MiniMax music generation tool

[octo-patch]

Reason: Add MiniMax music generation tool wiring the /v1/music_generation endpoint for global and CN regions

Changes

• Added server/tools/music_providers/minimax_provider.py: a music generation provider that posts to the /v1/music_generation endpoint for both the global (api.minimax.io) and CN (api.minimaxi.com) regions. It supports generation models (music-3.0, music-2.6 and free variants) and cover models (music-cover, music-cover-free), the request fields prompt, lyrics, stream, output_format, audio_setting, lyrics_optimizer, is_instrumental, audio_url, audio_base64, cover_feature_id and the CN-only aigc_watermark, url/hex output formats, and response parsing via base_resp.status_code with data.status/data.audio.
• Added server/tools/generate_music_by_minimax_jaaz.py: the langchain tool generate_music_by_minimax_jaaz.
• Registered the tool in server/services/tool_service.py under the minimax provider.
• Added the minimax provider preset with its music models to server/services/config_service.py.
• Extended the model type literals in server/models/config_model.py and server/services/config_service.py to include music.
• Added server/tests/test_minimax_music_provider.py covering endpoint building, region detection, payload building, request validation and response parsing.

Checks

• python -m pytest tests/test_minimax_music_provider.py -q (run from server/) — 15 passed
• python -m ruff check on the new/changed files — clean for the new files
• python -m py_compile on all changed files — ok

[octo-patch]

I updated the MiniMax music tool to expose streaming, enforce generation and cover input constraints, and validate audio formats. I ran the focused provider tests (18 passed), Ruff, Black, compileall, and a tool-schema check.