abstractPolicy.jl

# src/Reinforce.jl


abstract type AbstractPolicy end

"""
    a = action(policy, r, s, A)

Take in the last reward `r`, current state `s`,
and set of valid actions `A = actions(env, s)`,
then return the next action `a`.

Note that a policy could do a 'sarsa-style' update simply by saving the last state and action `(s,a)`.
"""
function action end