When the env is set true, then we can use an update_upstream.sh to update the upstream image.

Because we don't use user interface in docker, so we actually need apt, not aptitude.

NOTE: Since apt cannot solve some conflicts, we use aptitude in part.

skip the license key installation when installing. Because we don't need mujoco at present.

We can't install mujoco_py because of the mujoco from DockerHub, but we can add installation for local build.

• add mujoco_py installation for local
• add gym and baselines installation with mujoco_py
• add roboschool installation with mujoco_py
• add rllab installation with mujoco_py

pendulum installation failed

Building wheels for collected packages: pendulum

  Running setup.py bdist_wheel for pendulum: started

  Running setup.py bdist_wheel for pendulum: finished with status 'error'

  Complete output from command /usr/bin/python -u -c "import setuptools, tokenize;__file__='/tmp/pip-build-7AwYpN/pendulum/setup.py';exec(compile(getattr(tokenize, 'open', open)(__file__).read().replace('\r\n', '\n'), __file__, 'exec'))" bdist_wheel -d /tmp/tmpRYCK0cpip-wheel- --python-tag cp27:

  /usr/lib/python2.7/distutils/dist.py:267: UserWarning: Unknown distribution option: 'python_requires'
    warnings.warn(msg)
  usage: -c [global_opts] cmd1 [cmd1_opts] [cmd2 [cmd2_opts] ...]
     or: -c --help [cmd1 cmd2 ...]
     or: -c --help-commands
     or: -c cmd --help
  
  error: invalid command 'bdist_wheel'

Solution:

pip install --upgrade pip setuptools

Reference

• sdispater/pendulum#185

Reference: jupyter/notebook#2457

In jupyter-tensorflow-ros-base image, we open the terminal, but we get the following errors:

    Traceback (most recent call last):
      File "/opt/ros/kinetic/lib/python2.7/dist-packages/tornado/websocket.py", line 369, in _run_callback
        callback(*args, **kwargs)
      File "/opt/conda/lib/python3.6/site-packages/notebook/base/zmqhandlers.py", line 171, in open
        return super(WebSocketMixin, self).open(*args, **kwargs)
      File "/opt/conda/lib/python3.6/site-packages/terminado/websocket.py", line 48, in open
        super(TermSocket, self).open(url_component)
    TypeError: open() takes 1 positional argument but 2 were given
Websocket closed

• ffmpeg or libav installation. using the ffmpeg as the default option.
• clean installation about baselines and following installations.

add a long-term support version gazebo9 for ros image.

We will create a new docker image for this.

move data path in image which including jupyter(such as tensorflow-ros and jupyter) from /home/ros/data to /data

Run UR5 defautl as CMD, without user go into bash then run UR5 from an unfamiliar directory in container.

Refer to universal robot simulator.

• numpy python2 and python3 support. All images.

stop using the upstream image or just upgrade it.

add a docker image with ros desktop full and openai gym

add pendulum to all docker images, except UR and ubuntu-init images.

# pip3 install pendulum

# pip2 install pendulum

Reference

• osrf/docker_images
• osrf/docker_templates
• Dockerfile Generator
• Microsoft/generator-docker
• Docker starter for Cloud 66
• jboss-dockerfiles/dogen - Simple Dockerfile generator
• kaczmarj/neurodocker - Generate custom Docker images and minimize existing containers
• Red5d/docker-autocompose - Generate a docker-compose yaml definition from a running container

We will run the init process in ENTRYPOINT. To manage all process correctly, we should run the init process with root privilege, and we can change user in entrypoint script which seems setuser - phusion/baseimage-docker

This will make the docker images with the custom versions.

add nbmerge and nbdime for jupyter notebook to jupyter related image.

• add python3-tk for jupyter notebook matplotlib showing figure

%matplotlib inline 
...

%matplotlib tk

Reference

• How can I open the interactive Matplotlib window in IPython notebook?

add new image including

• dart
• pydart2
• gym
• dart_env
• tensorflow python3 latest
• ros kinetic base or ros without moveit(conflicts with dart because of libfcl)

add catkin python3 support in tensorflow-ros image.

sudo pip3 install trollius