For additional details, please see the SUMO challenge website
This API is written using a combination of C++, Python, and Cython. The user-facing interface is Python only.
This library was tested with Python 3.5.2 on Ubuntu 16.04.
sudo apt-get install git cmake python3-dev python3.5-venv \
libeigen3-dev libboost-python-dev libopencv-dev python-opencv \
libgmp-dev libcgal-qt5-dev swig
sudo yum install -y python35 python3-pip python3-devel cmake git swig \
opencv-python opencv-devel boost-devel gmp-devel qt5-qtbase-devel \
freetype-devel libpng-devel libtiff-devel
Note: on the Amazon EC2 instance where we tested, the Eigen package was not available. As a workaround, you can manually install:
wget http://springdale.math.ias.edu/data/puias/computational/7/x86_64//eigen3-devel-3.2.3-2.sdl7.noarch.rpm
sudo rpm -Uvh eigen3-devel-3.2.3-2.sdl7.noarch.rpm
git clone https://github.com/facebookresearch/sumo-challenge.git
Create a virtual evironment where the required packages will be installed in isolation.
cd sumo-challenge/sumo-api
python3 -m venv venv
source venv/bin/activate
pip install --upgrade pip
pip install -r requirements.txt
python setup.py build_ext --inplace
To run all unit tests:
python -m unittest discover -p "*tests.py" -v
To run all unit tests under a specific directory:
python -m unittest discover -s <path> -p "*tests.py" -v
For example:
python -m unittest discover -s sumo/threedee -p "*tests.py" -v
python setup.py cleanall