-
Notifications
You must be signed in to change notification settings - Fork 2
/
robosuite_env.py
103 lines (80 loc) · 2.87 KB
/
robosuite_env.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
#!/usr/bin/env python
"""
This uses env in https://github.com/ARISE-Initiative/robosuite
# Example:
## normal run:
python robosuite_env.py
## run as server and client
python robosuite_env.py --server
python robosuite_env.py --client
NOTE: make sure https://github.com/ARISE-Initiative/robosuite/pull/497 is merged.
else type: git pull origin refs/pull/497/head to use the pending PR branch.
"""
import robosuite as suite
from robosuite.wrappers.gym_wrapper import GymWrapper
import cv2
from agentlace.gym_env import GymEnvServerWrapper, GymEnvClient
import argparse
def make_env():
# create environment instance
env = suite.make(
env_name="Lift", # try with other tasks like "Stack" and "Door"
robots="Panda", # try with other robots like "Sawyer" and "Jaco"
has_renderer=True,
has_offscreen_renderer=True,
use_camera_obs=True,
)
# convert to gym environment
env = GymWrapper(
env,
keys=["robot0_proprio-state", "object-state", "agentview_image"],
flatten_obs=False
)
print(" -> action space:", env.action_space)
print(" -> observation space:", env.observation_space)
env.reset()
return env
def run_env(env):
print("Running environment")
# reset the environment
obs, info = env.reset()
for i in range(1000):
print("step", i)
# action = np.random.randn(env.robots[0].dof) # sample random action
action = env.action_space.sample()
print(action)
obs, reward, done, trunc, info = env.step(action) # take action in the environment
print(obs.keys(), obs["agentview_image"].shape)
# Convert the image from RGB to BGR (OpenCV uses BGR by default)
image = cv2.cvtColor(obs["agentview_image"], cv2.COLOR_RGB2BGR)
# Display the image
cv2.imshow("Agent View", image)
# Press 'q' to quit the display window
if cv2.waitKey(1) & 0xFF == ord('q'):
break
env.render() # render on display
# Release the display window
cv2.destroyAllWindows()
if __name__ == '__main__':
parser = argparse.ArgumentParser()
parser.add_argument("--host", type=str, default="localhost")
parser.add_argument("--port", type=int, default=5556)
parser.add_argument("--client", action="store_true")
parser.add_argument("--server", action="store_true")
args = parser.parse_args()
if args.server:
print("Running Action Env server")
env = make_env()
env = GymEnvServerWrapper(env, port=args.port)
env.start()
env.stop()
print("Server stopped")
elif args.client:
print("Running Action Env client")
env = GymEnvClient(host=args.host, port=args.port, timeout_ms=1500)
run_env(env)
else:
print("Running default cartpole env")
env = make_env()
run_env(env)
print("Done")