Reinforcement Learning Optimization for Large-Scale Learning: An Efficient and User-Friendly Scaling Library