DCAT: F-UJI FAIRness regression check in CI

[sorenwacker]

Part of #24. Depends on the harvestable-exposure sub-issue.

Goal

Add an automated FAIRness regression check using F-UJI against an exported/served metaseed dataset.

Scope

• Run F-UJI (self-hostable Docker image / API) against a sample dataset's harvestable DCAT.
• Assert on the returned FsF metrics (Findability/Accessibility/Interoperability/Reusability) above a baseline; treat regressions as failures.
• Keep it opt-in / separate from the core unit suite (needs the F-UJI service), e.g. a dedicated CI job.

Why

Gives an objective, third-party measure that the DCAT export actually improves FAIRness — complements SHACL conformance.

Acceptance criteria

• A CI job runs F-UJI against a served sample dataset and reports FsF scores.
• A baseline threshold is enforced.

References

• F-UJI: https://www.f-uji.net/ · source: https://github.com/pangaea-data-publisher/fuji · metrics: https://doi.org/10.5281/zenodo.3775793

Depends on: harvestable-exposure.