Installation

The code was tested for Python >=3.8 and this is what we recommend.

Note: If you want to work with a version Python<3.8, that is mostly possible at the moment. Only the moteus python package needs to be commented in the setup.py and it will not be possible to operate mjbots hardware.

We recommend using a virtual python environment.

  1. Install dependencies:

    # ubuntu20
sudo apt-get install git unzip libyaml-cpp-dev libeigen3-dev libpython3.8 libx11-6 libsm6 libxt6 libglib2.0-0 python3-pip python3-sphinx python3-numpydoc python3-sphinx-rtd-theme

# ubuntu22
sudo apt-get install git unzip libyaml-cpp-dev libeigen3-dev libx11-6 libsm6 libglib2.0-0 libboost-dev python3-pip python3-sphinx python3-numpydoc python3-sphinx-rtd-theme

  2. The header only eigen library can be installed via (For more information visit `eigen<https://eigen.tuxfamily.org/index.php?title=Main_Page>`__):

    wget -O Eigen.zip https://gitlab.com/libeigen/eigen/-/archive/3.4.0/eigen-3.4.0.zip
unzip Eigen.zip
sudo mv eigen-*/Eigen /usr/local/include/

  3. Clone the repository:

    git clone git@github.com:dfki-ric-underactuated-lab/double_pendulum.git

  4. Install the double pendulum package:

    cd double_pendulum
make install

With this you are done and you can use all functionalities and some controllers in this repository. If you want install all requirements (this includes also larger depoendencies such as tensorflow, drake etc.) for all controllers do

  1. (Optional) Full installation

    make pythonfull

The following two bullet points may be useful when only the C++/Python code is needed, if you work on the code and need to recompile often, or if errors appear during the installation call in 4.

  1. If you want to install only the Python package, you can do:

    make python
# make pythonfull

# or

cd src/python
pip install .
# pip install .[all]

  2. If you want to build C++ code and install python bindings:

    make cpp

# or

cd src/cpp/python
make