test.py

import environment as gym
import numpy as np
import random
import time
import logging
from pprint import pprint




logger = gym.logger.get_my_logger(__name__)

def timed_function(func):
    def decorated_func(*args, **kwargs):
        s = time.time()
        output = func(*args, **kwargs)
        e = time.time()
        logger.info("{} finished in {} seconds".format(func.__name__,e-s))
        return output
    return decorated_func


if __name__=="__main__":
    logging.disable(logging.NOTSET)
    nElevator = 3
    nFloor = 14
    spawnRates = [1/60]+[1/180]*(nFloor-1)
    avgWeight = 135
    weightLimit = 1200
    loadTime = 1


    env = gym.make(nElevator, nFloor, spawnRates, avgWeight, weightLimit, loadTime)
    s = env.reset()
    while True:
        decision_agents  = s["decision agents"]
        states           = s["states"]
        rewards          = s["rewards"]


        # A bunch of printing stuff for debugging
        parsed_states = gym.Environment.parse_states(states[0], nFloor, nElevator)

        logger.info("Rewards of decision elevators: {}".format(rewards))

        logger.info("  hall_calls_up: \n"+"".join([
            "{:8.3}".format(t) for t in parsed_states["hall_call_up_times"]
        ]))

        logger.info("  hall_calls_down: \n"+"".join([
            "{:8.3}".format(t) for t in parsed_states["hall_call_down_times"]
        ]))
        logger.info("Floor requests from within elevator {}:\n{}".format(
            decision_agents[0], parsed_states['requested_calls']
        ))

        logger.info("Number of passengers served:{}".format(env.nPassenger_served))



        # Picking action here
        actions = [random.sample(env.legal_actions(agent), 1)[0] for agent in decision_agents]# 0 calls for the legal actions of the first elevator

        # Stone Age rendering ftw.
        s = timed_function(env.step)(actions)
        env.render()
        time.sleep(0.3)