The Definitive Guide to Policy Gradients in Deep Reinforcement Learning: Theory, Algorithms and Implementations