|
Version |
Author |
Changelog |
|---|---|---|
|
20260818 |
Matthias Liffers |
First version |
|
20261112 |
Matthias Liffers |
Core schema blocks corrected from 11 to 12 (added metadata) to match the RAiD metadata schema |
These criteria define what the correct operation of a RAiD Service looks like. The intent is to verify that a Registration Agency's deployment of the RAiD Service Software Stack behaves equivalently to the ARDC RAiD Service, within the allowances for local customisation described below.
Allowances for local customisation: a Registration Agency may differ from ARDC in authentication providers, branding, and the use of Extended and Local metadata components. All other behaviour is expected to be identical.
Validation approach: to be agreed between ARDC and the Registration Agency. Version currency requirements are covered in the RAiD Registration Agency Agreement.
1. RAiD minting
-
The service can mint new RAiDs through both the web app and the Member API.
-
Minted RAiDs are assigned DOIs under the Registration Agency's organisation within the RAiD DataCite Consortium.
-
Each minted RAiD is immediately resolvable via the global RAiD resolver (raid.org).
-
RAiD records include correct provenance metadata identifying the Registration Agency.
2. Core metadata schema
-
All 12 Core schema blocks are implemented and enforced: identifier, metadata, date, title, description, contributor, organisation, relatedObject, alternateIdentifier, alternateUrl, relatedRaid, and access.
-
Mandatory fields are validated on create and update. The service rejects records that do not satisfy Core requirements.
-
Controlled vocabularies referenced in the schema (for example contributor roles, organisation roles, related object types) resolve correctly.
-
Date ranges on contributor positions and organisation roles are stored and returned correctly.
-
The access block correctly enforces the 18-month embargo cap and requires an embargo expiry date when access type is embargoed.
3. Member API
-
The Registration Agency's Member API is functionally equivalent to the ARDC Member API (as documented at https://api.prod.raid.org.au/swagger-ui/index.html) for all Core operations: mint, read, update, and list.
-
The API returns correct HTTP status codes and error responses.
-
API responses conform to the RAiD schema.
4. Authentication and authorisation
-
Keycloak is deployed and operational.
-
The Registration Agency's chosen identity provider(s) (for example AAF, Google, ORCID) allow researchers to log in successfully.
-
OAuth2/OIDC authorisation code token flow works correctly.
-
OAuth2/OIDC client credential flow works correctly.
-
Service Point admins can create Client Credentials scoped to their Service Point.
-
Users can created short-lived API tokens.
-
Service point membership controls access correctly: users can only create and edit RAiDs within service points they belong to.
5. Service points
-
The Registration Agency can create and manage service points within their deployment.
-
Users can be assigned to service points.
-
RAiDs created through a service point are correctly attributed to that service point in the identifier block.
6. Web app
-
The web app allows authenticated users to create, view, edit, and search RAiD records.
-
All Core metadata blocks can be created and edited through the UI.
-
The Registration Agency's branding is applied. ARDC branding is not present.
7. Landing pages
-
The Registration Agency deploys and hosts landing pages for all RAiDs minted by their service, using the static site generator from the au-research/raid-au repository.
-
Landing pages are publicly accessible without authentication.
-
Resolving any RAiD minted by the Registration Agency via raid.org correctly redirects to the Registration Agency's landing page for that record.
-
Open Access RAiDs have a landing page displaying all metadata in a human-readable format.
-
Embargoed RAiDs have a stub landing page.
-
The Registration Agency's branding is applied. ARDC branding is not present.
8. Data integrity and persistence
-
RAiD records are persisted correctly and can be retrieved after creation.
-
Updates to RAiD records are reflected immediately in subsequent reads.
-
The version field increments correctly on each update.
9. Public metadata archive
-
The Registration Agency publishes a regular archive of all open access RAiD records from their service to a stable, publicly accessible location with a persistent identifier (for example a DOI-identified deposit on Zenodo or an equivalent repository).
-
The archive includes public information about the Registration Agency's service points: Service Point ID, name, and ROR.
-
The archive is updated on a regular cadence (for example monthly, as the ARDC RAiD Service does).
Out of scope for these criteria
-
Service levels and uptime: the Registration Agency's own Service Level Specifications apply.
-
Version currency: covered in the RAiD Registration Agency Agreement.
-
Federation API (api.raid.org) participation: planned implementation by end 2026.