Installation

The GEO Knowledge Hub is built on top of InvenioRDM, a turn-key research data management repository developed by CERN and its community.

Thanks to this deep integration with InvenioRDM, the installation workflow benefits from all mature tooling and best practices developed by the InvenioRDM community. In the general, there are three main installation paths. They are presented below and detailed in the following sub-pages:

Local (Python)

Run the application directly on your machine in a Python virtual environment. Best for developers extending or customizing the platform.

Containerized (Docker)

Build and run everything in Docker for a quick local preview. Best for evaluating the platform without installing Python dependencies.

Kubernetes (Helm)

Deployment on Kubernetes using the helm-invenio chart.

Which path should I use?

Path	Use when…	Iteration speed	Production-ready
Local (Python)	You want to change Python or React code and see results immediately.	Fastest, with code reload in seconds.	No (development server only).
Containerized (Docker)	You want a quick, throwaway preview without installing Python toolchains.	Slow, every code change requires an image rebuild.	Yes (but requires hardening).
Kubernetes (Helm)	You are running a real instance for users, with HA, scaling, persistence and TLS.	Production deploy cadence.	Yes (but requires hardening).

Common prerequisites

Regardless of the path you choose you will need:

Git to clone the geo-knowledge-hub repository.
Docker (Engine 20+) for the local services (i.e., Postgres, Redis, RabbitMQ and OpenSearch) that the application talks to. The Kubernetes path uses Docker only to build and push images.
A working knowledge of the InvenioRDM CLI for the local and Docker paths, and of Helm 3 for the Kubernetes path.

Each sub-page lists its additional prerequisites.