Supaero robotics, 2023

This repository contains the exercices for the robotics class at Supaero, 2023. The exercices are organized by notebook. Each notebook corresponds to one chapter of the class. The notebooks are in Python and based on the software Pinocchio.

Set up

Linux, Python 3, PyPI

On a Linux system with Python 3, you can get the dependencies directly with +pip (see installation procedure and update below):

python3 -m pip install -r requirements.txt

NB: you should consider using a virtualenv

Once you have the dependencies, you can start the server with jupyter notebook

Docker

On other systems, use use the virtualization provided by +Docker. A Docker image is provided, and can be started with:

sudo docker run --rm -p 7000:7000 -p 7001:7001 -p 7002:7002 -p 7003:7003 -p 7004:7004 -p 8888:8888 -v data:/home/user/tp -it gepetto/supaero

On Linux host systems, you may simply start the Docker with:

sudo docker run --rm --net host -v data:/home/user/tp -it gepetto/supaero

In case of big update, you must update the docker:

sudo docker pull gepetto/supaero

Update the notebooks

If the repository changes (for example when new tutorials are pushes), you need to update your local version by "pulling" it from the repository. On a native installation, just go in the folder containing the tutorials and execute git pull

With a docker, execute the following:

sudo docker run --rm -v data:/home/user/tp -it gepetto/supaero  git remote set-url origin https://github.com/gepetto/supaero2023

Then

sudo docker run --rm -v data:/home/user/tp -it gepetto/supaero  git pull --rebase origin main

To avoid conflict when pulling a new version, you should better to your modifications in copy of the original files, not directly in the original files itself.

Side notes

Installing pip

Pip is a tool for installing and managing Python packages. You can install it with

sudo apt install python3-pip

The default version of +pip installed by +apt is not up to date, so upgrade it with

python3 -m pip install --upgrade --user

In general, running +pip is likely to run an alias on +pip in /usr, so either run it through python3 as explained above, or make sure your path select the right pip executable in your ~/.local. The option --user is kind of optional for recent +pip version, but removing it should work with a warning.

Installing docker

On linux, install docker with +apt (or see https://docs.docker.com/engine/install/ubuntu/).

sudo apt install docker.io

On other OS, see how to get Docker.

Join me on [Matrix]

[Matrix] is a distributed chat system that will be used during the class. Consider creating an account and join the classroom channel.

Set up of the system ... for admin only

Compile the code snipets

run python utils/generate.py. The directories to be generated are listed in plain code (line 54). This script take the tp*/py and produces tp/generated/* files following the "# %jupyter_snippet" / "# %end_jupyter_snippet" hashtag in the code.

Build and push the docker

Build the docker as specified in Dockerfile. You need the buildkit for that. The following file +/etc/docker/deamon_json+ should contain:

{ "features" : { "buildkit" : true } }

Add DOCKER_BUILDKIT=1 in your environment and restart docker.

sudo systemctl restart docker
export DOCKER_BUILDKIT=1

Then build:

docker build . -t gepetto/supaero

Log to Docker-hub using your credentials.

docker login

Use the login corresponding to +https://hub.docker.com/+. And push it to the Gepetto repository of Docker-hub.

docker build -t gepetto/supaero

Docker as normal user

sudo usermod -aG docker nmansard
newgrp docker