PMD-X Cross-Dataspace Demo

A fully browser-based demo showing how PMDCO (PMD Core Ontology) bridges incompatible industrial data schemas across Catena-X and Manufacturing-X dataspaces — enabling a single SPARQL query to retrieve material properties from both.

Live demo: https://materialdigital.github.io/PMD-X-Dataspaces-Demo

---

What it demonstrates

A Tension Pulley assembly has two components sourced from different dataspaces:

Component	Dataspace	Format	Material
Pulley Wheel	Catena-X	SAMM aspect model (JSON)	PA6GF30 polymer
Tensioner Bracket	Manufacturing-X	AAS 3.0 inspection document (JSON)	Steel 316/4401

Each dataset is converted once into PMDCO-conformant RDF. From that point a single SPARQL SELECT — no UNION, no bespoke adapter — retrieves material properties from both sources using the shared PMDCO predicate pattern.

Architecture

Catena-X source JSON
  → YARRRML mapping → SAMM/RDF (catx-samm.ttl)
  → SPARQL INSERT    → PMDCO graph (urn:graph:catx)

Manufacturing-X source JSON (AAS 3.0)
  → py-aas-rdf       → AAS/RDF (mfgx-aas.ttl)
  → SPARQL INSERT    → PMDCO graph (urn:graph:mfgx)

Assembly graph (assembly.ttl)  ─────────────┐
                                             ▼
                               Cross-dataspace SPARQL query
                               (FROM catx + FROM mfgx + FROM assembly)

Everything runs in-browser using Oxigraph compiled to WebAssembly — no server, no backend.

Named graphs loaded at runtime

Graph	Contents
urn:graph:catx-samm	CatX SAMM/RDF instance data + aspect model
urn:graph:mfgx-aas	MfgX AAS-ontology RDF (from py-aas-rdf)
urn:graph:assembly	Company product knowledge graph
urn:graph:pmdco-ontology	PMDCO 3.0 ontology
urn:graph:tto-ontology	Tensile Test Ontology
urn:graph:catx	PMDCO Transform output for CatX data
urn:graph:mfgx	PMDCO Transform output for MfgX data

Round-trip verification

Both input datasets are independently verified:

• Catena-X: a SPARQL SELECT over urn:graph:catx-samm reconstructs the full original SAMM JSON payload (including 547-entry PvT table) and asserts it against the reference file.
• Manufacturing-X: store.dump({format:'application/ld+json', from_graph_name:…mfgx-aas}) exports the complete named graph as expanded JSON-LD (2102 nodes); 12 key values are located by IRI and asserted against the original AAS inspection document.

Run locally

git clone https://github.com/materialdigital/PMD-X-Dataspaces-Demo.git
cd PMD-X-Dataspaces-Demo
python3 -m http.server 8000
# open http://localhost:8000

Any static file server works. A server is required because the demo fetches Turtle and JSON assets via fetch().

Source data

Dataset	Source
PA6GF30 material data (CatX)	material_data_test_pa6gf30.json
Steel 316/4401 inspection document (MfgX)	inspectiondocument_316_4401_alloy.json
Full demo dataset	PMD-X Cross-Dataspace Demo

Key technologies

• PMDCO — PMD Core Ontology (shared semantic layer)
• Oxigraph WASM — in-browser SPARQL triplestore
• SAMM — Catena-X Semantic Aspect Meta Model
• AAS 3.0 — Asset Administration Shell (Manufacturing-X)
• py-aas-rdf / AAS2KG — AAS JSON → RDF conversion
• YARRRML — declarative RDF mapping for CatX SAMM data
• Eclipse Dataspace Connector — simulated EDC data plane negotiation

License

Apache 2.0 — see LICENSE.