Drake contribution test environment

---

I use this environment to keep track of various docker images for testing drake, manipulation and underactuated. Usually this requires quickly mounting a fresh branch from that deviates from the master, therefore I dont clone the repositories. I manage individual repositories through submodules.

Setting up the repo

To setup the repository, clone it into your local folder and then inside the parent folder of the repo run the following

git submodule update --init --recursive

Workflow

---

This section describes how I ususally fwork with this repository, some common commands etc.

Git workflow

Always remember to update the submodules

git submodule update --remote

Common Docker commands

The following are some of the standard docker commands that are useful for developing using docker The following command attaches to absh shell within a container that goes by CONTAINER_NAME

docker attach CONTAINER_NAME

Drake

To just load up the drake container first build it with the following

docker-compose -f .docker/drake.nightly.yaml build

Next, run the container using

docker-compose -f .docker/drake.nightly.yaml up

Open a shell into the container either using portainer or one of the above docker commands. I use portainer if I am only running a script/binary from within a container. For build jobs it is better to attach to a container's terminal. Next, Set up the container for the build

./setup/ubuntu/install_prereqs.sh

This should prompt you for a bunch of permissions, reply by typing Y/n. In case you want to go ahead and build the repository, the following commands are useful

cd path/to/drake_workspace
bazel build //... # build the entire project
bazel test //... # build and test the entire project

# just build relevant binaries and run the planar scenegrpah test
bazel test -j 2 //bindings/pydrake/systems:py/planar_scenegraph_visualizer_test
# or
bazel test -j HOST_CPUS*0.5 //bindings/pydrake/systems:py/planar_scenegraph_visualizer_test

bazel test //bindings/pydrake/systems:py/planar_scenegraph_visualizer_py_pycodestyle

Manipulation

Run the docker container according to the following command (build it if you haven't already)

docker-compose -f .docker/manipulation.nightly.yaml up

shell into the running container and activate pyenv with the following command

eval "$(/pyenv/bin/pyenv init -)"

Once this is done activate the manipulation environment

# from the manipulation folder
source .venv/bin/activate