inlabgeo/inlab repository overview

⁠InLab Container Stacks

Run InLab CoFI projects without setting up local scientific Python environments.

• CoFI installed
• CoFI examples ready to run
• Full example dependency stack included in the inlab image
• Published for linux/amd64 and linux/arm64

⁠Getting Started

Make sure you have Docker⁠ or Podman⁠ installed and running.

⁠Pull and Run

$ docker run -p 8888:8888 inlabgeo/inlab:latest

Then open Jupyter Lab in your browser at http://127.0.0.1:8888. Enter the token shown in the terminal when prompted.

With Podman:

$ podman run -it -p 8888:8888 inlabgeo/inlab:latest

The inlab image also serves the cofi-paper marimo notebooks. Add -p 2718:2718 to either command above to reach them at http://127.0.0.1:2718.

⁠Build Locally

The full inlab target is adapted from the validated Apptainer stack in inlab-geo/inlab-apptainer/inlab.py313.def, ported to Ubuntu 24.04 LTS (supported until April 2029) with native Python 3.12, NumPy 2-compatible packages, and PyGIMLi installed as a pinned prebuilt pip wheel on amd64 / built from source on arm64 (see docs/how-it-works.md).

$ docker build --target inlab --file image/Containerfile --tag inlabgeo/inlab:local .
$ docker run -p 8888:8888 inlabgeo/inlab:local

Build the lighter CoFI-only image with:

$ docker build --target cofi --file image/Containerfile --tag inlabgeo/cofi:local .

Build a specific platform locally:

$ docker buildx build --load --platform linux/amd64 --target inlab --file image/Containerfile --tag inlabgeo/inlab:amd64 .

Build a multi-platform image manifest for publishing:

$ docker buildx build --push --platform linux/amd64,linux/arm64 --target inlab --file image/Containerfile --tag inlabgeo/inlab:latest .

⁠Images

PyGIMLi installs differently per architecture (prebuilt wheel on amd64, from source on arm64, both pinned to the same version), so the two inlab image sizes differ substantially even though their package versions match.

⁠More References

The InLab containers are built based on the Jupyter Docker Stacks, and we inherit the entry point created by the Jupyter team. If you have any further questions about running the containers, we kindly recommend referring to the Getting started guide provided by the Jupyter team.

• inlab-geo/inlab-apptainer⁠
• Apptainer⁠
• Jupyter Docker Stacks quick start⁠

⁠Acknowledgement

This repository was originally generated from the Jupyter Docker Stacks cookiecutter⁠.