feat: Add MiniMax CN I2V/FL2V/S2V video pipeline
- Add MediaHubI2VSubmit node with three modes: i2v, fl2v, s2v
- I2V: first_frame_image → video (6 models, dynamic resolution/duration)
- FL2V: first_frame + last_frame → video (MiniMax-Hailuo-02 only)
- S2V: subject_image + prompt → video (S2V-01 only)
- All three image inputs are optional, validated per mode in Python
- Rename MediaHubT2VQueryStatus → MediaHubVideoQueryStatus
- Rename MediaHubT2VRetrieveVideo → MediaHubRetrieveVideo
- Legacy aliases preserved for backward-compatible workflow loading
- Add build_i2v_payload, submit_i2v_task to mediahub/minimax_cn.py
- Frontend dynamically constrains model/resolution/duration per mode
- 134 pytest tests, smoke test, JS/compile/JSON checks all pass
Van/攻城狮鹿鸣君
3d8b41af31