feat: add MiniMax as LLM provider (M2.7/M2.5/M2.5-highspeed)

[octo-patch]

Summary

Add MiniMax as a first-class LLM provider, supporting the latest MiniMax-M2.7, MiniMax-M2.5, and MiniMax-M2.5-highspeed models via MiniMax OpenAI-compatible API.

Changes

• model/minimax/index.ts: New MiniMax provider extending the Chat base class with streaming support, temperature clamping to [0, 1.0] range, configurable base_url/api_key/proxy, MINIMAX_API_KEY env fallback
• model/base.ts: Add MiniMaxM2_7, MiniMaxM2_5, MiniMaxM2_5_highspeed to ModelType enum; add Site.MiniMax
• model/index.ts: Register MiniMax provider in ChatModelFactory
• utils/config.ts: Add minimax config section (api_key, base_url, proxy)
• README.md / README_zh.md: Document MiniMax provider and MINIMAX_API_KEY env var
• jest.config.js + tests/minimax/: 15 unit tests + 3 integration tests

Supported Models

Model	Context Window
MiniMax-M2.7	1,048,576 tokens
MiniMax-M2.5	1,048,576 tokens
MiniMax-M2.5-highspeed	204,800 tokens

Test Plan

• 15 unit tests pass (model support, constructor, askStream, temperature clamping, error handling, env fallback)
• 3 integration tests pass (streaming M2.7, streaming M2.5-highspeed, multi-turn conversation)
• Manual test with MINIMAX_API_KEY via /ask and /ask/stream endpoints