opengeohub/geocomp-ide repository overview

⁠opengeohub/geocomp-ide

A multi-language geocomputation IDE: a single Docker image bundling a JupyterHub-ready JupyterLab with kernels for Python, R, Julia, JavaScript and Rust, all backed by a shared GDAL/PROJ/GEOS stack, plus Quarto (knitr engine) for reproducible publishing with cross-language variable sharing.

⁠What is it?

One image, five languages, one geospatial foundation. Each language ships its own Jupyter kernel and GDAL-compatible raster/vector libraries, so you can read, analyse and publish raster and vector data in whichever language fits the task — and a Quarto document can pass objects between them from R chunks via reticulate (R ↔ Python) and JuliaCall (R ↔ Julia).

⁠What's inside

Base — Ubuntu 24.04 with GDAL 3.11.4, GEOS and PROJ (ubuntugis-unstable).

Kernels & geospatial stacks

Publishing — Quarto CLI 1.10.18 (bundles its own pandoc) with the knitr engine: knitr + rmarkdown, reticulate and JuliaCall. RETICULATE_PYTHON points at the Python venv, so R chunks reach the full scikit-map stack.

JupyterLab — ipywidgets, geemap, ipyleaflet, ipympl, jupyter_bokeh, spell-checker and the jupyterlab-quarto extension (preview/render .qmd).

⁠Usage

All the available options of jupyter/datascience-notebook⁠ works in this image

1. Download the appropiate BLAS/LAPACK implementation for you chipset (image tags⁠):

docker pull opengeohub/geocomp-ide:v3.12.3-gdal3114

1. Create the container

docker run -d --restart=always --name geocomp_ide_312 \
  -v /home:/home/host -p 8888:8888 --tmpfs /tmp:mode=1777 \
  opengeohub/geocomp-ide:v3.12.3-gdal3114 jupyter lab \
  --allow-root --LabApp.token='opengeohub' --ServerApp.root_dir='/' --no-browser --ip=0.0.0.0

1. Access the JupyterLab through http://localhost:8888⁠ using the credentials:

• Password or token: opengeohub