snd-sweden/data-storage-information-interface

Ro-crate json-schema and documentation for file description metadata to SND DORIS

About the storage manifest

The purpose of this manifest is to integrate a local research storage with SND DORIS so that the researchers can easily describe, refer to and publish data without data having to leave the organisation's own storage.

The goal is to have as few required fields as possible and focus on using a RO-Crate manifest to describe a set of files connected to an organisation (e.g. a university) and a list of researchers (identified by Swamid/eduGAIN identifiers).

The manifest for describing the storage manifest and files is based on RO-Crate 1.1.

JSON schema for the manifest

JSON schema - documentation

Example implementations

Node.js command line tool

C# command line tool

Java command line tool

PHP Nextcloud extension example

Properties overview

This is an overview of the properties supported by SND's RO-Crate profile.

• @type CreativeWork
• @id ro-crate-metadata.json reference to the manifest, must be ro-crate-metadata.json
• conformsTo
  • @id https://w3id.org/ro/crate/1.1
• identifier 04679b46-964c-11ec-b909-0242ac120002 UUID for the manifest
• publisher organisation responsible for the dataset
  • @type Organization
  • identifier
    • @type PropertyValue
    • propertyID domain
    • value example.com organisation's top domain
• creator one or more "owners" of the files described in the manifest
  • @type Person
  • identifier
    • @type PropertyValue
    • propertyID eduPersonPrincipalName
    • value localid@example.com person Swamid/eduGAIN id
• about
  • @type Dataset
  • name optional: provide a name for the dataset
  • hasPart
    • @type File
    • @id example-file.csv relative path from dataset root
    • contentSize optional: size in bytes
    • dateCreated optional: created date (ISO 8601)
    • dateModified optional: modified date (ISO 8601)
    • encodingFormat optional: MIME format
    • sha256 optional: file checksum (SHA256)
    • url https://example.com/04679b46-964c-11ec-b909-0242ac120002/example-file.csv optional: public URL to file

Example manifest

Simple manifest with two files

Note: RO-Crate manifests must be expressed as a JSON-LD flattened graph, nested properties are not allowed. This means references must be expressed with local id:s, for example Person.identifier to PropertyValue to reference the value for eduPersonPrincipalName (see example manifest and example scripts).

Example for posting a manifest

To get access to the staging environment for handling the server contact SND.

1. Generate ro-crate-metadata.json or use ro-crate-metadata.example.json
2. curl -X POST https://example-index-server.se/ro-crate -H 'Content-Type: application/json' -d @'ro-crate-metadata.json'
3. For updates post the updated manifest (the identifier in the manifest needs to be identical)
4. Your files should be listed in DORIS (staging)