Installation

Docker

The easiest way how to get the system running is to use published docker images of our services. There are two docker-compose files that can be taken as an example:

• fit-demo (setup with Dobot Magician, Dobot M1 and Kinect Azure)
• fanuc-demo (setup with a Fanuc robot)

PyPi

The Python packages are also published to PyPi (see robo_fit), so they can be installed e.g. by using pip install arcor2-arserver and started manually from the terminal with a respective command as arcor2_arserver.

Pants

The most suitable option for developers is to clone the (mono)repository and get Pants running. We mainly work with Ubuntu 20.04. When used for the first time (e.g., with ./pants --version), Pants initializes itself and gets all the dependencies with one exception, which is Azure Kinect SDK that is necessary for pyk4a library we use - there is an installation script for it, that needs to run in advance. However, if you don't need to work with Kinect, you may remove the arcor2_kinect_azure directory and remove pyk4a from 3rdparty/requirements.txt and from 3rdparty/constraints.txt. Please make sure to don't commit these changes.