SPARQL Endpoint Collector
One endpoint collector runs one explicitly configured CONSTRUCT or DESCRIBE
request and retains the returned RDF with a source-result revision. It is a
collection source, independently of where GraphArch stores its documentation KG.
A SPARQL SELECT result table is not supported as a generic record-to-document
adapter.
What it documents
Returned assertions support the same ontology resources, SHACL shapes, SKOS vocabularies, and EKGF use cases as RDF-file input. The response must actually contain their defining types and relationships.
Select a source graph
Save source.yaml:
endpoint: https://data.example.org/sparql
query: |
CONSTRUCT { ?s ?p ?o }
WHERE { GRAPH <https://example.org/vocabulary> { ?s ?p ?o } }
max_source_bytes: 8388608
http:
timeout_seconds: 30
Replace the endpoint and graph with your own source, then run:
grapharch publication collect --store output/kg \
--project urn:example:project --collector urn:example:vocabulary-query \
--kind sparql-endpoint --source source.yaml
Use distinct collector identities for different queries. Requests have explicit size/time limits; response content type identifies the RDF serialization. Failed requests and malformed responses are errors. Separate queries do not promise one atomic source-endpoint snapshot.
The adapter is SparqlEndpointSource. For authenticated requests, protocol graph
parameters, several concurrent family sources, and the separate writable-KG
contract, see SPARQL endpoints.