feat: add MiniMax as LLM provider for AI script generation

[octo-patch]

Summary

• Add MiniMax M2.7 and MiniMax-M2.5-highspeed as LLM provider options for the AI script generation feature in the role-play tab
• MiniMax provides an OpenAI-compatible API at https://api.minimax.io/v1, making integration seamless with the existing openai SDK usage
• Handle MiniMax-specific quirks: temperature default (0.7) and <think> tag stripping from responses

Changes

File	Change
webui_mix.py	Add MiniMax models to llm_change() URL mapping and Radio selector; add MiniMax platform link to free API channels
llm_utils.py	Set temperature=0.7 for MiniMax models; strip <think> tags from MiniMax responses
README.md	Add MiniMax to the list of supported LLM providers
tests/test_llm_minimax.py	22 unit tests covering URL mapping, temperature handling, think-tag stripping, JSON extraction
tests/test_minimax_integration.py	3 integration tests calling real MiniMax API (script generation, basic request, JSON output)

How to use

1. Get a free API key from MiniMax Platform
2. In the role-play tab, select MiniMax-M2.7 or MiniMax-M2.5-highspeed from the model selector
3. Enter your MiniMax API key and click AI script generation

Test plan

• 22 unit tests pass (pytest tests/test_llm_minimax.py)
• 3 integration tests pass with real MiniMax API (pytest tests/test_minimax_integration.py -m integration)
• Manual test: select MiniMax-M2.7 in the role-play tab and generate a script