JNRH'2023: Pinocchio Tutorial

Support chat: https://matrix.to/#/#jnrh2023-tuto:laas.fr

Slides: https://gepettoweb.laas.fr/talks/jnrh2023/

⚠️ as you can guess from the name, this repo is from 2023 and will stay there. Some of the content from here is available in https://github.com/gepetto/supaero2025 (or ulterior versions), or the upstream projects directly.

Getting started

Start by cloning this repository:

git clone https://github.com/nmansard/jnrh2023
cd jnrh2023

The only maintained use of this tutorial is with nix:

nix run

Check your installation

Jupyter-lab should show you a couple links in https://localhost:8888 / https://127.0.0.1:8000: you should be able to open them, and in your webbrowser open and run 0_setup.ipynb to say hi to Talos.

Obsolete and not maintaind instructions

You have to choose one of the three following supported methods:

pip

⚠️ this is obsolete, not tested, and probably broken

1. create an environment:

   python -m venv .venv

2. activate it:

   source .venv/bin/activate

3. update pip:

   pip install -U pip

4. install dependencies:

   pip install example-robot-data-jnrh2023 jupyterlab meshcat scipy ipywidgets matplotlib

5. start PYTHONPATH=. jupyter-lab

conda (Linux, macOS and Windows)

⚠️ this is obsolete, not tested, and probably broken

1. update conda:

   conda update -n base -c defaults conda

2. create an environment:

   conda create -n jnrh-2023 python=3.9

3. activate it :

   conda activate jnrh-2023

4. add conda-forge channel:

   conda config --add channels conda-forge

   Note: If you already have the conda-forge channel set-up in your config, this will have no effect and you will likely get the warning message Warning: 'conda-forge' already in 'channels' list, moving to the top. You can simply discard it.

5. install specific dependencies:

   conda install -c olivier.roussel hpp-fcl example-robot-data pinocchio=2.99.0

   Note: Forcing the Pinocchio package version to 2.99.0 is not mandatory depending on your channel priority configuration, which can vary depending on your conda installation source.

6. check your Pinocchio install:

   conda list

   Check for correct packages versions and source channel (4th column) in your output:

   example-robot-data    4.1.0     <some_hash>    olivier.roussel
   hpp-fcl               2.99.0    <some_hash>    olivier.roussel
   pinocchio             2.99.0    <some_hash>    olivier.roussel
   

   If versions or source channel does not match this, then something gone wrong in your installation process.

7. install regular dependencies:

   conda install -c conda-forge jupyterlab meshcat-python scipy ipywidgets matplotlib

8. start JupyterLab:

   Be sure that you are still in the jnrh2023 root directory of the repository that you cloned at the Getting started step.

   Linux / macOS

   PYTHONPATH=. jupyter-lab

   Windows

   jupyter-lab

docker

⚠️ this might still work, but it use old versions of everything

1. start docker run -v ./:/tuto --net=host -it nim65s/jnrh2023