Delft headless IIIF

You will need to install Bun.sh to run this project.

curl -fsSL https://bun.sh/install | bash

To install dependencies:

bun install

To run development server:

bun run dev

To build for production:

bun run build

Configuration

TBC

Content

• libis
• content
• manifests

Scripts

The following scripts have been created:

• delft-labels
• delft-image-source
• delft-related

These extract custom data out of the manifests, saving the results to disk. They can also add new data to the manifests.

2 main types of scripts:

• Extract
• Enrich

Extract

If you create an extract script, it will be called with every type of resource you specify (e.g. Manifest). From the extract function you can return an object with the following shape:

type Return = {
  temp?: any;
  caches?: Record<string, any>;
  meta?: any;
  indices?: Record<string, string[]>;
}

Temp will be passed to a collect() after all resources have been processed. This is useful for collecting data from multiple resources and creating an aggregation.

Caches will be persisted between runs, so you can both read and write to that and avoid running heavy operations every time.

Meta will be added to a file at the same path as the manifest, but ending in /meta.json instead of /manifest.json. This is useful for adding data to the manifest that is not part of the IIIF spec.

Indicies are used to create reverse indices, so you can "tag" a resource with a type and a subtype:

return {
  indicies: {
    Inventor: ['Some inventor'],
  }
}

Once all resources are processed, a collection will be created for each "tag" and each type of tag. So in this case, there would be an "Inventor" collection that contains a single Collection ("Some inventor") and it would contain the Manifest.

When extracting (or enriching) you get access to the entire meta.json and caches that already exist on the resource. This allows different scripts to build on each other.