RDF URL Collector

The RDF URL Collector fetches an explicitly configured HTTP or HTTPS RDF document, parses it in the declared format, and retains its assertions with source provenance in the documentation KG. Use it when a vocabulary publisher provides an RDF download URL rather than a repository checkout or a query service.

What it documents

A downloaded RDF document can supply ontology resources, SHACL shapes, SKOS vocabularies, or EKGF use cases. The document's assertions determine what it describes; the URL's filename does not establish its semantics.

Example: collect a published ontology

Add a collector to the project in a website's publication definitions:

@prefix ga: <https://ekgf.org/ontology/grapharch#> .

<urn:example:docs> ga:collector <urn:example:commons-roles> .

<urn:example:commons-roles> a ga:CollectorUse ;
    ga:collectorType ga:RdfUrlCollector ;
    ga:source "https://www.omg.org/spec/Commons/RolesAndCompositions/" ;
    ga:format "application/rdf+xml" .

This is a collector configuration fragment, not a complete publication project. The surrounding definitions must supply the project, scope, profile and target. The configured collector runs during the website build; publication scope determines which collected terms become documentation. There is no separate publication collect --kind rdf-url CLI option.

The example URL serves the publisher's RDF/XML document. ga:source is a URL string literal; ga:format explicitly supplies the RDF media type. Supported formats include text/turtle, application/rdf+xml, application/n-triples, application/n-quads, application/trig, text/n3 and application/ld+json. An RDF-specific response content type must identify the same format. Neutral text/plain and application/octet-stream responses are parsed using the explicitly configured RDF format; this supports immutable repository downloads without guessing their syntax. HTML and conflicting RDF media types are rejected. Malformed RDF, unsupported formats and responses above the 8 MiB source limit fail explicitly. Relative IRIs resolve against the fetched URL.

An RDF download is different from an endpoint query

The SPARQL Endpoint Collector executes an explicit CONSTRUCT or DESCRIBE request and collects the returned graph. This collector instead reads the RDF document at the configured URL; it does not send a SPARQL query. It also does not automatically fetch the document's owl:imports targets. Register additional sources explicitly when their definitions are needed.

For a file maintained in Git, use a configured Git RDF source to retain repository, commit and blob provenance. For a local file, use the RDF File Collector.

Search documentation

Enter words to search this site.