Skip to content

Installation

Jump to bottom
Rika edited this page Aug 28, 2021 · 7 revisions

Prerequisites

DEDO is a lightweight environment, so all dependencies, including the physics engine, can be installed through pip.

Though we still recommend using a conda environment or virtualenv to compartmentalize your python projects Here, a conda example is provided:

conda create --name DEDO python=3.7
conda activate DEDO

Supported operating systems: Linux/BSD,MacOS, Windows* (windows support has not been tested)

Installing DEDO

git clone https://github.com/contactrika/dedo
cd dedo
pip install numpy  # important: necessary for compiling numpy-enabled PyBullet
pip install -e .

To enable recording/logging videos install ffmpeg:

sudo apt-get install ffmpeg

Troubleshooting

  • Compilation errors while installing PyBullet

Please ensure that GCC and G++ are both installed. In conda + linux, one can install the latest version of gxx

  • Why do I get numpy and Pybullet Assertion Error when running Pybullet

Although DEDO works with most modern version of Python, we use a special version of PyBullet that is compiled with NumPy. We have encountered that on some systems (AMD + Ubuntu 16.04) where PyBullet was having trouble recognizing numpy in Python 3.8 due to (we think) a packaging issue with PyBullet repo. However, we do not see this problem with Python 3.7.

Clone this wiki locally