An Open-source End-to-End Logic Optimization Framework for Large-scale Boolean Network with Reinforcement Learning