dgm_franka & dgh_franka installation

After following the installation of the prerequisites in the readme, we can now install dgm_franka and dgh_franka.

Install dgm_franka

First, clone the repo

git clone https://github.com/BolunDai0216/dgm_franka

Install lcm and the corresponding Python package

git clone https://github.com/lcm-proj/lcm.git 
cd lcm && mkdir build && cd build 
cmake .. && cmake --build .
sudo cmake --install . 

This will install the python package at /usr/local/lib/python3.<x>/site-packages. Add this to you PYTHONPATH environment variable by running

export PYTHONPATH=$PYTHONPATH:/usr/local/lib/python3.<x>/site-packages

You can test this installation by seeing if lcm can be imported in Python. Then, go back to the parent directory and build the package. But before that, we need to source the dgm-ws we created before

cd dgm_franka
source ~/dgm-ws/install/setup.bash
mkdir build && cd build
cmake ..
cmake --build .

Now that the package is built, we can install it anywhere we want

cmake --install . --prefix "<install-dir>/install"

Finally, add the path to dynamic_graph_manager_cpp_bindings to PYTHONPATH, it should be located at ~/dgm-ws/install/dynamic_graph_manager/lib/python3.8/site-packages.

Install dgh_franka

Now that dgm_franka is installed, we also need to install the dynamic graph head (DGH) to use it from Python. First, we also need to clone the repo

git clone https://github.com/BolunDai0216/dgh_franka.git
cd dgh_franka

Since this is a python package, we just need to install it

python3 -m pip install -e .

The next step is to run the executable in the install folder of dgm_franka and then control the robot from the Python API.