Combining ideas from graph search and graph of convex sets
This repo uses Poetry for dependency management. To setup this project, first install Poetry and, make sure to have Python3.12 installed on your system.
(Optional) Configure poetry to create virtual environment in project
poetry config virtualenvs.in-project true
Then, configure poetry to setup a virtual environment that uses >= Python 3.10:
poetry env use python3.10
Next, install all the required dependencies to the virtual environment with the following command:
poetry install -vvv
(the -vvv flag adds verbose output).
Clone this fork of pypolycontain locally. https://github.com/shaoyuancc/pypolycontain
Next, add the following lines to .venv/bin/activate replacing the paths with your actual paths:
PYPOLYCONTAIN="/path/to/pypolycontain"
export PROJECT_ROOT="/path/to/project/folder/large_gcs"
export PYTHONPATH="$PROJECT_ROOT:$PYPOLYCONTAIN:$PYTHONPATH"
export MOSEKLM_LICENSE_FILE="/path/to/mosek/license/mosek/mosek.lic"
export GRB_LICENSE_FILE="/path/to/gurobi/license/gurobi.lic"
On Ubuntu
sudo apt-get install python3-tk # For interactive plotting in graph generator script
sudo apt install ffmpeg # For saving animation videos
On MacOS, run the following command:
brew install python-tk
brew install ffmpeg
Additional packages that I was not able to install using poetry, and had to use pip instead
pip install kaleido
If you have problems with poetry, when in the virtual env shell, run the following:
pip install numpy matplotlib ipykernel scipy black tqdm pytest wandb hydra-core omegaconf autoflake isort pdbpp plotly docformatter drake kaleido pre-commit
The repo is setup to do automatic linting and code checking on every commit through the use of pre-commits. To run all the pre-commit hooks (which will clean up all files in the repo), run the following command:
poetry shell
pre-commit install
In typical development, regularly run
pytest -m "not slow_test"
which runs all the tests that take shorter than 5 seconds each.
To run the slow tests alone use
pytest -m slow_test
Or to run all the tests simply use
pytest
To run a specific test(s) by referring to some part of the test's name, use
pytest -k "shortcut_edge_cg_simple_2_inc"
To make tests verbose and allow print statements to be shown use -v -s flags.
Create a config file specifying the experiment in config and run it using the following command:
python3 experiments/run_contact_graph_experiment.py --config-name quickstart
where quickstart should be replaced with your config name.
After running quickstart you can compare your results (which should appear in the multirun/date/time folder) to the contents of the quickstart_output folder
python3 experiments/run_contact_graph_experiment.py --config-path ../config/WAFR_experiments --config-dir config --config-name trajectory_figures
Configs for the GGCS experiments are located in config/ggcs_experiments
To run an experiment:
python3 experiments/run_ggcs_se3_maze_experiment.py --config-path ../config/ggcs_experiments --config-dir config --config-name reaches_new_sampling_ggcs
This will produce output files in the folder multirun/{current_date}/{current_time}
To aggregate multirun data run:
python3 scripts/process_multirun_data.py --dir multirun/{date}/{time}
This will save a json file of the data from all the runs (aggregated_run_data.json) to the same multirun folder.
Create a bash script in scripts and make it executable with chmod +x run_multiple_experiments.sh.
Then run it with run_multiple_experiments.sh
On Ubuntu, 'Times' font can be installed via
sudo apt update
sudo apt install ttf-mscorefonts-installer
After installation, you may need to update the font cache:
fc-cache -f -v
On Mac, it already comes in the system
Additionally, need LaTex on your system For Ubuntu,
sudo apt install texlive-latex-extra
sudo apt-get install texlive-fonts-recommended
sudo apt-get install cm-super
sudo apt-get install dvipng
This repo references and contains code from: Bernhard Paus Græsdal https://github.com/bernhardpg/planning-through-contact and Tobia Marcucci https://github.com/TobiaMarcucci/shortest-paths-in-graphs-of-convex-sets.