A workspace around Neuroglancer, extended by plugins. Live at https://alleninstitute.github.io/ngplatform/.

ngplatform hosts domain-specific tools for Zarr, OME-Zarr and Zarr Vectors datasets (track filtering, registration, annotation and segmentation, filters, data formats) as plugins around an unchanged Neuroglancer. Tools that are too specialised or too large to add to Neuroglancer itself live here instead.

The platform provides:

• A workspace: datasets, scenes, and a link that holds them all.
• Local data: files on your computer read in place as virtual Zarr stores.
• An SDK (@ngplatform/sdk): plugins add tabs, commands, views, overlays, jobs and data formats through it.

Docs: https://ngplatform.readthedocs.io/ has the user guide, a page for each plugin and how to contribute (sources in docs/).

Add a plugin

Each plugin is a folder in plugins/:

plugins/
  index.ts             the list of plugins the app ships: add yours here
  workers.ts           worker modules of data-type plugins
  my-plugin/
    manifest.ts        what the platform shows before loading any code
    plugin.ts          export const plugin: PluginModule = { activate(ctx) { … } }
    tests/             *.test.ts, run by npm test
    README.md

1. Copy a plugin folder close to what you need: track-filtering or volume-view for a tool, or zarr-vectors for a data format.
2. Give it a manifest with a unique id, and import the platform only as @ngplatform/sdk.
3. Add one line for it to plugins/index.ts (and, for a data format, its worker modules to plugins/workers.ts).
4. npm test checks the manifest and the entries; the Plugins tab switches it on and off.

To try one without building, the Plugins tab's Add loads a JavaScript module or a Python file from your computer, and reloads it after each edit.

The full guide, with the manifest, the ctx API and examples: plugins/README.md.