Many errors in this repo

[BillMatrix]

I encountered many issues when using this repo:

• there's no 'point_mass' "body" in the given model:
  Here's the error:

Traceback (most recent call last):
  File "/Users/billzhu/.conda/envs/rep_learning/lib/python3.7/site-packages/IPython/core/interactiveshell.py", line 3325, in run_code
    exec(code_obj, self.user_global_ns, self.user_ns)
  File "<ipython-input-5-d7bebc982a78>", line 1, in <module>
    env.get_body_com('point_mass')
  File "/Users/billzhu/.conda/envs/rep_learning/lib/python3.7/site-packages/gym/envs/mujoco/mujoco_env.py", line 168, in get_body_com
    return self.data.get_body_xpos(body_name)
  File "mujoco_py/generated/wrappers.pxi", line 2604, in mujoco_py.cymj.PyMjData.get_body_xpos
  File "mujoco_py/generated/wrappers.pxi", line 1019, in mujoco_py.cymj.PyMjModel.body_name2id
ValueError: No "body" with name point_mass exists. Available "body" names = ('world', 'object1', 'object2', 'object3', 'object4', 'object5', 'r_wrist_flex_link', 'r_wrist_roll_link', 'tips_arm').

• The default XML you provide only has three objects, which causes error in the code.
• You did not define physics in the environment, where the camera uses. Also the camera_setup function does not exist on line 274 of env.py
• The mujoco environment you use I think is the deepmind's control environment (dm_control.mujoco), whose physics' qpos and qvel does not match openai gym's mujoco_env's qpos and qvel, which causes errors when taking any actions
• I don't know which mujoco_env you are using, but mujoco_env does not have the following two arguments:

max_episode_steps=maximum_episode_steps,
reward_threshold=0.,

[yidingjiang]

Hi Bill,
You are right the mujoco env that I am using is a wrapper around deepmind control environment. I am working on a fix and hope to patch the code this week.

[yidingjiang]

All these issues should be fixed now. Can you pull the latest code and confirm this?

[BillMatrix]

Thanks! I'll take a look tonight!