ros-notebook is a community maintained Jupyter Docker Stack image that adds ROS to the common notebook containers such as the datascience-notebook.
Short tags are all based on the datascience-notebook where <ros-version> is the ROS version. For example, noetic is a ROS Noetic image based on the datascience-notebook.
| ROS | Jupyter |
|---|---|
 |
 |
 |
 |
 |
 |
 |
 |
All other tags are formed as follows:
<notebook-type>-<ros-version>
where <ros-version> is the ROS version and <notebook-type> is the type of notebook. For example, minimal-noetic is a ROS Noetic image based on the minimal-notebook.
An example Dockerfile for a custom notebook image:
FROM ghcr.io/mpo-web-consulting/ros-notebook:minimal-noetic
# add your customizations here
USER root
RUN mamba install -y tensorflow && \
mamba clean -y -f && \
fix-permissions "${CONDA_DIR}" && \
fix-permissions "/home/${NB_USER}"
# switch back to jovyan to avoid accidental container runs as root
USER ${NB_UID}The images and actions are generated from templates using Jinja2. The templates are located in the templates directory. The images are defined in images.yaml.
Use the following commands to generate the images and github actions files:
# generate Dockerfiles for all images
python3 scripts/generate_images.py templates/Dockerfile.j2 images images.yaml
# generate actions for all images
python3 scripts/generate_workflows.py templates/action.yml.j2 images .github/workflows