tudorjnu/cathsim

mj_loadXML error: b"Error: could not open STL file

Closed this issue · 4 comments

I try to run cathsim.(just run python3 tests/test_env.py) There are some issues.
First, scene.xml file (in cathsim/cathsim_env/envs/) includes catheter.xml, but there is cathter_template.xml in this directory, I don't understand why this file names are different.

Second, there are no walls.stl file. scene directory should be in meshes directory but there are only phantom_1, phantom_2 and slave. How can I solve this problem?

Hello @SangbaekYoo!

Indeed, I tried to re-run it and it seemed to throw an error. I have investigated it and now it should be fixed.

Let me know if it is fixed for you or if there is any other issue.

Thank you for response my question. Now I can run basic example. I also want to render simulation with obs_type 'image', but there is some error. Is there need anothor modification in code? My current code is below.

import cathsim_env
import gym

env = gym.make('cathsim_env/CathSim-v0',
scene=1, # 1 or 2 for Type-I Aortic Arch and Type-II Aortic Arch
target="bca", # "bca" or "lcca"
obs_type="image", # image or internal
image_size=128,
delta=0.008, # the distance threshold between catheter head and target
success_reward=10.0, # the reward for reaching the target
compute_force=False, # whether to compute the force
dense_reward=True, # whether to use a dense reward or a sparse reward,
)

obs = env.reset()
for _ in range(2000):
action = env.action_space.sample()
obs, reward, done, info = env.step(action)
env.render()

Ahh, this is a more complicated issue that has to do with glfw (see issue 2989 on openai/gym).

However, as a wrap-around, you can try to the cv2 images by rendering the single frames.

I am closing this issue. Please let me know if there are more issues.