Collect native RDF statements
RDF collection retains resource identities, annotations, relationships, and named graph context. Documentation generators can then recognise documentable subjects and build readable explanations from the supplied statements.
File formats
Directory and repository discovery currently selects:
| Serialization | Extension |
|---|---|
| Turtle | .ttl |
| RDF/XML | .rdf |
| N-Triples | .nt |
| N-Quads | .nq |
| TriG | .trig |
| Notation3 | .n3 |
The explicit RDF-file collector also accepts JSON-LD through .jsonld input.
That does not make arbitrary JSON a documentable-record format, and automatic
family/repository discovery currently uses the six extensions above. XML
catalogue files are not treated as RDF source files.
Describe what the source actually says
GraphArch preserves supplied labels, definitions, rich descriptions, and relationships. Source-aware reference generation exposes missing annotations and competing values as quality findings. It does not invent definitions or implicitly download external ontology imports.
Named classes, properties, individuals, shapes, schemes, and concepts retain their RDF identities. Supported anonymous expressions and RDF collections are rendered from retained statements, with explicit limits and cycle checks. Unsupported expressions or malformed input are errors.
See ontology-family publication for source selection and SPARQL collection for RDF supplied by a service.