cell_graph_validator.py 6.16 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151
from libs.cell_graph_dispatcher import CellGraphDispatcher
from libs.dummy_observation import DummyObservationBuilder

from flatland.envs.rail_env import RailEnv
from flatland.envs.rail_generators import sparse_rail_generator
from flatland.envs.schedule_generators import sparse_schedule_generator
from flatland.envs.malfunction_generators import malfunction_from_params
from flatland.envs.rail_env import RailAgentStatus

import numpy as np
import time


class CellGraphValidator:

    @staticmethod
    def multiple_tests(dispatcher_function,
                       width, height, trains, seed,
                       cities=None, rails_between_cities=None, rails_in_city=None,
                       malfunction_rate=None, prop_malfunction=None, min_prop=None, max_prop=None,
                       show_map=False):

        np.random.seed(seed)
        N = len(width)
        if cities is None:
            cities = np.random.randint(2, 35, (N,))
        if rails_between_cities is None:
            rails_between_cities = np.random.randint(2, 4, (N,))
        if rails_in_city is None:
            rails_in_city = np.random.randint(3, 6, (N,))
        if malfunction_rate is None:
            malfunction_rate = np.random.randint(500, 4000, (N,))
        if prop_malfunction is None:
            prop_malfunction = np.random.uniform(0.01, 0.01, (N,))
        if min_prop is None:
            min_prop = np.random.randint(20, 80, (N,))
        if max_prop is None:
            max_prop = np.random.randint(20, 80, (N,))
            max_prop = np.maximum(max_prop, min_prop)

        res_finished = []
        res_times = []

        for i in range(N):
            print("="*15)
            print("Test ", i)
            print("="*15)

            d = CellGraphValidator.single_test(dispatcher_function, width[i], height[i], trains[i], i, cities[i],
                                               rails_between_cities[i], rails_in_city[i],
                                               malfunction_rate[i], prop_malfunction[i],
                                               min_prop[i], max_prop[i],
                                               show_map=show_map)
            res_finished.append(d["finished"])
            res_times.append(d["time"])

        avg_finished = np.mean(res_finished)
        total_time = np.sum(res_times)

        print("="*15)
        print(f'Average finished {avg_finished}')
        print(f'Total time spent: {total_time}s')
        print("-"*15)
        print("Finished per test:", res_finished)
        print("Time per test:", res_times)

        return {'finished': avg_finished, 'time': total_time}

    @staticmethod
    def single_test(dispatcher_function,
                    width, height, trains, seed,
                    cities, rails_between_cities, rails_in_city,
                    malfunction_rate, prop_malfunction, min_prop, max_prop,
                    show_map=False):

        if show_map:
            from flatland.utils.rendertools import RenderTool, AgentRenderVariant


        start = time.time()

        speed_ration_map = {1.: 0.25,  # Fast passenger train
                            1. / 2.: 0.25,  # Fast freight train
                            1. / 3.: 0.25,  # Slow commuter train
                            1. / 4.: 0.25}  # Slow freight train

        rail_generator = sparse_rail_generator(max_num_cities=cities,
                                               seed=seed,
                                               grid_mode=False,
                                               max_rails_between_cities=rails_between_cities,
                                               max_rails_in_city=rails_in_city,
                                               )
        schedule_generator = sparse_schedule_generator(speed_ration_map)
        stochastic_data = {'malfunction_rate': malfunction_rate,  # Rate of malfunction occurence of single agent
                           'prop_malfunction': prop_malfunction,
                           'min_duration': min_prop,  # Minimal duration of malfunction
                           'max_duration': max_prop # Max duration of malfunction
                           }
        observation_builder = DummyObservationBuilder()
        env = RailEnv(width=width,
                      height=height,
                      rail_generator=rail_generator,
                      schedule_generator=schedule_generator,
                      number_of_agents=trains,
                      malfunction_generator_and_process_data=malfunction_from_params(stochastic_data),
                      # Malfunction data generator
                      obs_builder_object=observation_builder,
                      remove_agents_at_target=True
                      # Removes agents at the end of their journey to make space for others
                      )
        env.reset()

        if show_map:
            env_renderer = RenderTool(env, gl="PILSVG",
                                      agent_render_variant=AgentRenderVariant.AGENT_SHOWS_OPTIONS_AND_BOX,
                                      show_debug=False,
                                      screen_height=1920*2,  # Adjust these parameters to fit your resolution
                                      screen_width=1080*2)  # Adjust these parameters to fit your resolution


        dispatcher = dispatcher_function(env)

        max_time_steps = int(4 * 2 * (width + height + 20))

        step = 0
        while True:
            step += 1

            action_dict = dispatcher.step(step)
            next_obs, all_rewards, done, _ = env.step(action_dict)

            if show_map:
                env_renderer.render_env(show=True, show_observations=False, show_predictions=False)

            if done['__all__']:
                break

            if step == max_time_steps:
                break

        if show_map:
            env_renderer.close_window()
            del env_renderer

        finished = np.sum(
            [a.status == RailAgentStatus.DONE or a.status == RailAgentStatus.DONE_REMOVED for a in env.agents])

        finished = finished / len(env.agents)
        elapsed = time.time()-start
        print(f'Trains finished: {finished}. Time spent: {elapsed}s')
        return {'finished': finished, 'time': elapsed}