feat: add MiniMax as LLM provider

[octo-patch]

Summary

Adds MiniMax AI as a fourth LLM provider option alongside OpenAI, Azure OpenAI, and Anthropic.

• MINIMAX_API_KEY env var in CodeInterpreterAPISettings for auto-detection
• MiniMax branch in _choose_llm() using ChatOpenAI with base_url=https://api.minimax.io/v1 (OpenAI-compatible API)
• Default model: MiniMax-M2.5 when no MiniMax-specific model is configured; also supports MiniMax-M2.5-highspeed for faster responses
• Temperature clamping: [0.01, 1.0] range required by MiniMax API
• Agent compatibility: Returns ChatOpenAI instance, so OpenAIFunctionsAgent is used automatically (function calling works)
• README updated with MiniMax configuration instructions

Usage

export MINIMAX_API_KEY=your-minimax-api-key
export MODEL=MiniMax-M2.5  # or MiniMax-M2.5-highspeed

from codeinterpreterapi import CodeInterpreterSession

with CodeInterpreterSession() as session:
    response = session.generate_response("Plot the bitcoin chart of 2023")
    response.show()

Changes

File	Change
src/codeinterpreterapi/config.py	Add MINIMAX_API_KEY setting
src/codeinterpreterapi/session.py	Add MiniMax branch in _choose_llm()
README.md	Add MiniMax configuration docs
tests/test_minimax_provider.py	12 unit tests (config, model selection, temp clamping, priority)
tests/test_minimax_integration.py	3 integration tests (LLM creation, completion, session)

Test plan

• 12 unit tests pass (pytest tests/test_minimax_provider.py)
• 3 integration tests pass (MINIMAX_API_KEY=... pytest tests/test_minimax_integration.py)
• Existing tests unaffected
• MiniMax returns ChatOpenAI instance for OpenAIFunctionsAgent compatibility
• Temperature clamped to valid range
• Provider priority maintained: Azure > OpenAI > Anthropic > MiniMax