christopheraburns / aws-jpl-osr-challenge Goto Github PK

After setting up the repo locally and having everything running, I noticed the robot joints would disconnect and go all over the place. After much investigation I figured out it was the reset function. I think it would be more prudent to use the model reset function in gazebo. It's exposed as a service you can invoke this with something like

from std_srvs.srv import Empty
self.reset_models = rospy.ServiceProxy('/gazebo/reset_world', Empty)


#self.gazebo_model_state_service(model_state)
#self.gazebo_model_configuration_service(model_name='rover', urdf_param_name='rover_description', joint_names=joint_names_list, joint_positions=joint_positions_list)
self.reset_models()

Using that method just causes the model to return to the state it was in from the URDF/XML/launch files.

The documentation seems to indicate this is Open Source, but there's no specific Open Source license identified in the codebase.

Line 7 - 9 of setup.sh uses sudo pip3 install for installing python packages. This installs python packages to the system-wide path and is generally considered bad practice. It is better (and more common ) practice to use a virtual environment, such as virtualenv, to manage the installation of python packages for a project. See Is it acceptable and safe to run pip install under sudo?

If it is of interest to install all python packages in a single installation script, consider specifying the minimum package requirements in an install_requires in a setup.py. See https://packaging.python.org/guides/distributing-packages-using-setuptools/#install-requires

Hi,
In the README.md there are two screenshots in the getting started section with dead links:

https://github.com/Christopheraburns/AWS-JPL-OSR-Challenge/blob/master/images/create_env.jpg

https://github.com/Christopheraburns/AWS-JPL-OSR-Challenge/blob/master/images/clear_environment.jpg

It might be worth fixing these for people that haven't used RoboMaker before.
Cheers

Was watching the intro video and it was mentioned that the simulation could be run locally. When I try to run the simulation launch file it crashes as the markov python package has many paths that rely on s3_boto_data_store.py

Is it possible for you to provide a version of this repo (or a tutorial) that runs on a local machine or do we have to figure this out ourselves?

Thanks.