use std::collections::HashMap;

/// A step in the environment, containing the new state, reward, and termination info
#[derive(Debug, Clone)]
pub struct Step<State, Reward> {
    pub state: State,
    pub reward: Reward,
    pub done: bool,
    pub info: HashMap<String, String>,
}

/// Environment trait for reinforcement learning
pub trait Environment {
    type State: Clone;
    type Action: Clone;
    type Reward: Clone;

    /// Reset the environment to initial state
    fn reset(&mut self) -> Self::State;

    /// Take one step in the environment
    fn step(&mut self, action: &Self::Action) -> Step<Self::State, Self::Reward>;

    /// Get the action space bounds (low, high)
    fn action_space(&self) -> (Self::Action, Self::Action);

    /// Get the observation space bounds (low, high)
    fn observation_space(&self) -> (Self::State, Self::State);
}