Installing Plugins¶
OXFORDIA's extensibility model is a two-axis plugin system:
- Data Plugins describe a kind of data — a domain-specific shape, schema, and vocabulary.
- Statistic Plugins describe a kind of computation — mean, Kaplan–Meier, regression, and so on.
The two axes compose: a Mean Statistic Plugin can be used against any compatible field exposed by any Data Plugin. Adding a new disease's data model does not require touching any statistic; adding a new statistic does not require touching any data model.
Plugin installation¶
Plugins ship as standard installable packages. Install them on your node alongside the core oxfordia-pod package.
# Example: install the nemaline data plugin
sudo apt install oxfordia-plugin-data-nemaline
# Example: install the Kaplan-Meier statistic plugin
sudo apt install oxfordia-plugin-stat-kaplan-meier
# Restart the node to load new plugins
sudo systemctl restart oxfordia-pod
For Kubernetes deployments, add the plugin image references to your Helm values file:
plugins:
data:
- name: nemaline
image: ghcr.io/oxfordia-project/plugin-data-nemaline:latest
statistic:
- name: kaplan-meier
image: ghcr.io/oxfordia-project/plugin-stat-kaplan-meier:latest
Operational note¶
New statistic plugins must be installed on every node in the federation where they will be available. This is a known operational cost of the plugin model. Plugins ship as standard installable packages to minimize that cost, and the long-term roadmap includes a more general query layer that reduces the per-node install requirement.
Available plugins¶
Data Plugins¶
| Plugin | Disease / Domain | Package |
|---|---|---|
| Nemaline | Nemaline myopathy clinical trials | oxfordia-plugin-data-nemaline |
See Data Plugins for details on what each plugin provides.
Statistic Plugins¶
| Plugin | Computation | Package |
|---|---|---|
| Mean | Arithmetic mean | oxfordia-plugin-stat-mean |
| Median | Median | oxfordia-plugin-stat-median |
| Kaplan–Meier | Survival analysis | oxfordia-plugin-stat-kaplan-meier |
See Statistic Plugins for details on query parameters and access rule schemas.