py-rocket-geospatial-2: a Docker image for Python-R Geospatial

This release provides a cloud-ready Docker image for Python–R geospatial and earth-science workflows, designed for use in shared JupyterHub and similar multi-user environments.

py-rocket-geospatial-2 integrates modern Python and R geospatial stacks with tools commonly used across oceanography, climate science, hydrology, and remote sensing. The image follows design patterns from the Pangeo ecosystem and emphasizes scalable, array-based analysis of large spatiotemporal datasets.

What this image provides

• Python geospatial stack optimized for large datasets, distributed computing, and cloud-native object storage
• R + RStudio with geospatial packages installed using Rocker Project scripts
• JupyterLab with both Python and R kernel support
• Desktop environment (VNC) for GUI-based tools such as QGIS and Panoply
• VS Code OSS configured for scientific notebooks and Quarto
• Publishing toolchain including Quarto, JupyterBook, MyST, Pandoc, and TeX Live
• Helper scripts (pyrocket_scripts, rocker_scripts) to support customization and extension

Intended use

This image is intended to lower barriers for reproducible, cloud-ready analysis of large earth-system datasets, including workflows that authenticate and access data from NASA Earthdata, NOAA and other major earth-observation archives.

Licensing & attribution

The resulting container image is released under the Apache-2.0 License.
Rocker installation scripts retain their original GPL-2.0-or-later licensing.

Attribution is appreciated when this container image is used, adapted, or redistributed.

Whats Changed

• Update py-rocket-base image version in Dockerfile to fix LD_LIBRARY_PATHS bug
• Add automated Python tests via Jupyter notebooks
• Add pinned package versions
• Add actions to create draft GitHub release