From 3c86f24419c6c902a3f9659b8d79965e698b6245 Mon Sep 17 00:00:00 2001
From: spiglerg <spiglerg@gmail.com>
Date: Wed, 3 Apr 2019 11:24:09 +0200
Subject: [PATCH] initial draft interface for env.py, derived from
rllib.env.MultiAgentEnv
---
flatland/core/env.py | 92 ++++++++++++++++++++++++++++++++++++++++++++
1 file changed, 92 insertions(+)
diff --git a/flatland/core/env.py b/flatland/core/env.py
index e69de29..98e5778 100644
--- a/flatland/core/env.py
+++ b/flatland/core/env.py
@@ -0,0 +1,92 @@
+"""
+The env module defines the base Environment class.
+The base Environment class is adapted from rllib.env.MultiAgentEnv
+(https://github.com/ray-project/ray).
+"""
+
+# TODO: add derived classes and environments, e.g., GridEnvironment and
+# RailEnvironment.
+
+class Environment:
+ """Base interface for multi-agent environments in Flatland.
+
+ Agents are identified by agent ids (handles).
+ Examples:
+ >>> obs = env.reset()
+ >>> print(obs)
+ {
+ "train_0": [2.4, 1.6],
+ "train_1": [3.4, -3.2],
+ }
+ >>> obs, rewards, dones, infos = env.step(
+ action_dict={
+ "train_0": 1, "train_1": 0})
+ >>> print(rewards)
+ {
+ "train_0": 3,
+ "train_1": -1,
+ }
+ >>> print(dones)
+ {
+ "train_0": False, # train_0 is still running
+ "train_1": True, # train_1 is done
+ "__all__": False, # the env is not done
+ }
+ >>> print(infos)
+ {
+ "train_0": {}, # info for train_0
+ "train_1": {}, # info for train_1
+ }
+ """
+
+ def __init__(self):
+ pass
+
+ def reset(self):
+ """
+ Resets the env and returns observations from agents in the environment.
+
+ Returns:
+ obs : dict
+ New observations for each agent.
+ """
+ raise NotImplementedError()
+
+ def step(self, action_dict):
+ """
+ Performs an environment step with simultaneous execution of actions for
+ agents in action_dict.
+ Returns observations from agents in the environment.
+ The returns are dicts mapping from agent_id strings to values.
+
+ Parameters
+ -------
+ action_dict : dict
+ Dictionary of actions to execute, indexed by agent id.
+
+ Returns
+ -------
+ obs : dict
+ New observations for each ready agent.
+ rewards: dict
+ Reward values for each ready agent.
+ dones : dict
+ Done values for each ready agent. The special key "__all__"
+ (required) is used to indicate env termination.
+ infos : dict
+ Optional info values for each agent id.
+ """
+ raise NotImplementedError()
+
+ def render(self):
+ """
+ Perform rendering of the environment.
+ """
+ raise NotImplementedError()
+
+ def get_agent_handles(self):
+ """
+ Returns a list of agents' handles to be used as keys in the step()
+ function.
+ """
+ raise NotImplementedError()
--
GitLab