feat: add MiniMax as LLM provider for synthetic data generation

[octo-patch]

Summary

Add MiniMax as an alternative LLM provider for synthetic tabular data generation, complementing the existing OpenAI GPT support.

Changes

• New SingleTableMiniMaxModel: Inherits from SingleTableGPTModel, uses MiniMax's OpenAI-compatible API
• Supported models: MiniMax-M2.7 (default) and MiniMax-M2.7-highspeed
• Environment variables: MINIMAX_API_KEY and MINIMAX_API_URL (defaults to https://api.minimax.io/v1)
• Temperature clamping: Automatically clamps temperature to MiniMax's valid range (0.0, 1.0]
• 24 unit tests + 3 integration tests
• README update: Added MiniMax quick start section

Usage

import os
os.environ["MINIMAX_API_KEY"] = "your-key"

from sdgx.models.LLM.single_table.minimax import SingleTableMiniMaxModel

model = SingleTableMiniMaxModel()
model.fit(raw_data)
synthetic_data = model.sample(100)

Files Changed (4 files, ~360 additions)

File	Description
sdgx/models/LLM/single_table/minimax.py	MiniMax model implementation
sdgx/models/LLM/single_table/__init__.py	Export new model
tests/models/test_singletable_minimax.py	24 unit + 3 integration tests
README.md	MiniMax documentation section

API Reference

• Chat (OpenAI Compatible): https://platform.minimax.io/docs/api-reference/text-openai-api
• Pricing: https://platform.minimax.io/docs/guides/pricing-paygo

Test Plan

• Unit tests pass (24/24)
• Integration tests pass with real MiniMax API (3/3)
• Existing GPT model tests unaffected
• Temperature clamping works correctly
• Environment variable loading works