forked from HumanCompatibleAI/overcooked_ai
-
Notifications
You must be signed in to change notification settings - Fork 2
/
run_complex.py
185 lines (148 loc) · 6 KB
/
run_complex.py
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
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
import pygame
import matplotlib
matplotlib.use('TkAgg')
import matplotlib.pyplot as plt
from argparse import ArgumentParser
from overcooked_ai_py.mdp.overcooked_mdp import OvercookedGridworld, Direction, Action
from overcooked_ai_py.mdp.overcooked_env import OvercookedEnv
from overcooked_ai_py.agents.agent import StayAgent, RandomAgent, AgentFromPolicy, GreedyHumanModel
from overcooked_ai_py.planning.planners import MediumLevelPlanner
from overcooked_ai_py.mdp.layout_generator import LayoutGenerator
from overcooked_ai_py.utils import load_dict_from_file
from human_aware_rl.baselines_utils import load_baselines_model, get_agent_from_saved_model
from human_aware_rl.ppo.ppo import get_ppo_agent
from human_aware_rl.imitation.behavioural_cloning import get_bc_agent_from_saved
UP = 273
RIGHT = 275
DOWN = 274
LEFT = 276
SPACEBAR = 32
no_counters_params = {
'start_orientations': False,
'wait_allowed': False,
'counter_goals': [],
'counter_drop': [],
'counter_pickup': [],
'same_motion_goals': True
}
valid_counters = [(5, 3)]
one_counter_params = {
'start_orientations': False,
'wait_allowed': False,
'counter_goals': valid_counters,
'counter_drop': valid_counters,
'counter_pickup': [],
'same_motion_goals': True
}
class App:
"""Class to run an Overcooked Gridworld game, leaving one of the agents as fixed.
Useful for debugging. Most of the code from http://pygametutorials.wikidot.com/tutorials-basic."""
def __init__(self, env, agent, player_idx, slow_time):
self._running = True
self._display_surf = None
self.env = env
self.agent = agent
self.agent_idx = player_idx
self.slow_time = slow_time
print("Human agent index:", player_idx)
def on_init(self):
pygame.init()
# Adding pre-trained agent as teammate
self.agent.set_agent_index(self.agent_idx)
self.agent.set_mdp(self.env.mdp)
print(self.env)
self._running = True
def on_event(self, event):
done = False
if event.type == pygame.KEYDOWN:
pressed_key = event.dict['key']
action = None
if pressed_key == UP:
action = Direction.NORTH
elif pressed_key == RIGHT:
action = Direction.EAST
elif pressed_key == DOWN:
action = Direction.SOUTH
elif pressed_key == LEFT:
action = Direction.WEST
elif pressed_key == SPACEBAR:
action = Action.INTERACT
if action in Action.ALL_ACTIONS:
done = self.step_env(action)
if self.slow_time and not done:
for _ in range(2):
action = Action.STAY
done = self.step_env(action)
if done:
break
if event.type == pygame.QUIT or done:
print("TOT rew", self.env.cumulative_sparse_rewards)
self._running = False
def step_env(self, my_action):
agent_action = self.agent.action(self.env.state)
if self.agent_idx == 0:
joint_action = (agent_action, my_action)
else:
joint_action = (my_action, agent_action)
s_t, r_t, done, info = self.env.step(joint_action)
print(self.env)
print("Curr reward: (sparse)", r_t, "\t(dense)", info["shaped_r"])
print(self.env.t)
return done
def on_loop(self):
pass
def on_render(self):
pass
def on_cleanup(self):
pygame.quit()
def on_execute(self):
if self.on_init() == False:
self._running = False
while (self._running):
for event in pygame.event.get():
self.on_event(event)
self.on_loop()
self.on_render()
self.on_cleanup()
def setup_game(run_type, run_dir, run_seed, agent_num, player_idx):
if run_type == "ppo":
print("Seed", run_seed)
agent, config = get_ppo_agent(run_dir, run_seed, best=True)
elif run_type == "pbt":
run_path = "data/" + run_type + "_runs/" + run_dir + "/seed_{}".format(run_seed)
config = load_dict_from_file(run_path + "/config.txt")
agent_path = run_path + '/agent' + str(agent_num) + "/best"
agent = get_agent_from_saved_model(agent_path, config["sim_threads"])
elif run_type == "bc":
agent, config = get_bc_agent_from_saved(run_dir)
else:
raise ValueError("Unrecognized run type")
env = OvercookedEnv(OvercookedGridworld.from_layout_name(**config["mdp_params"]), **config["env_params"])
return env, agent, player_idx
if __name__ == "__main__":
"""
Sample commands
-> pbt
python overcooked_interactive.py -t pbt -r pbt_simple -a 0 -s 8015
->
python overcooked_interactive.py -t ppo -r ppo_sp_simple -s 386
-> BC
python overcooked_interactive.py -t bc -r simple_bc_test_seed4
"""
parser = ArgumentParser()
parser.add_argument("-t", "--type", dest="type",
help="type of run, (i.e. pbt, bc, ppo, etc)", required=True)
parser.add_argument("-r", "--run_dir", dest="run",
help="tag of run dir in data/*_runs/", required=True)
parser.add_argument("-no_slowed", "--no_slowed_down", dest="slow",
help="Slow down time for human to simulate actual test time", action='store_false')
parser.add_argument("-s", "--seed", dest="seed", required=False, default=0)
parser.add_argument("-a", "--agent_num", dest="agent_num", default=0)
parser.add_argument("-i", "--idx", dest="idx", default=0)
args = parser.parse_args()
run_type, run_dir, slow_time, run_seed, agent_num, player_idx = args.type, args.run, bool(args.slow), int(
args.seed), int(args.agent_num), int(args.idx)
env, agent, player_idx = setup_game(run_type, run_dir, run_seed, agent_num, player_idx)
theApp = App(env, agent, player_idx, slow_time)
print("Slowed time:", slow_time)
theApp.on_execute()