🔬 EMBL · AI4Access · Case study

Euro-BioImaging
Technology Registry

Submission framework that lets imaging-facility staff describe their technologies in a structured, validated way and exposes them via a typed REST API + linked-data projections.

Choose a version to explore
Core submission · 5 hours
Main version
As per the task brief evaluation criteria

The base framework: FastAPI · Pydantic v2 · PostgreSQL 16 · Alembic · controlled SKOS vocabularies · three-layer validation · RDF + SHACL · audit log · bundled vanilla-JS UI · full Swagger.

  • POST / GET / PATCH / DELETE on facilities + technologies
  • Server-minted PIDs via w3id.org/eurobioimaging
  • RDF projection (Turtle + JSON-LD) + small SPARQL endpoint
  • Dry-run validation endpoints
  • Bundled UI: Catalogue · Submit · Validate · Linked-data · Schemas
Open main
Extended features · +4 hours
Extended version
Everything in main, plus the AI / RAG / ops layer

Builds on the same data substrate to demonstrate how a real Research-Navigator-style consumer would integrate against it.

  • Hybrid RAG over the catalogue (pgvector + FTS, streamed grounded answers with citations)
  • AI auto-fill of submissions from a 1-page PDF or pasted text
  • Modern Cmd-K palette with model badge, country flags, follow-up suggestions
  • Stats dashboard: KPIs · hero charts · click-to-filter
  • World map with marker clustering · RSS / Atom feeds · world coords on every facility
  • ?dry_run=1 on every write endpoint · pytest isolation via dedicated DB
Open extended