Python-Based Reinforcement Learning on Simulink Models