Ontologies and RDF resources
The RDF File Collector reads RDF files and retains
their assertions and named-graph identity. The ontology publication pipelines recognise explicit
owl:Ontology declarations and document named resources associated with the
selected vocabulary.
Example: a customer vocabulary
@prefix ex: <https://example.org/customer#> .
@prefix owl: <http://www.w3.org/2002/07/owl#> .
@prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> .
@prefix skos: <http://www.w3.org/2004/02/skos/core#> .
<https://example.org/customer> a owl:Ontology ;
rdfs:label "Customer vocabulary"@en .
ex:Customer a owl:Class ;
rdfs:isDefinedBy <https://example.org/customer> ;
rdfs:label "Customer"@en ;
skos:definition "A person or organisation receiving our products."@en .
When this vocabulary is selected for publication, Customer gets a class reference entry with its definition and IRI. Adding labelled properties and named individuals supplies entries in their respective sections. This is a source example; the family workflow adds the namespace, membership, and publication configuration.
For a real vocabulary, compare DPROD's sources with its generated specification. Look at the DataProduct class and its links to related resources.
Supporting collectors
| Source | Collector or discovery path |
|---|---|
| An RDF file on disk | RDF File Collector |
| RDF files in a local or remote Git repository | Repository scanners, which select the specialised RDF-file collectors |
| RDF returned by a configured query | SPARQL Endpoint Collector |
These collectors retain RDF. The ontology documentation stage recognises the declared resource types and turns the selected vocabulary into reference entries. Reading RDF and generating those entries are separate steps.
Recognised resources
The family reference recognises these resource categories:
- Classes declared as
owl:Classorrdfs:Class, and RDFS datatypes. - RDF properties and OWL object, datatype, and annotation properties.
owl:NamedIndividualinstances, with their declared classes and other roles.- Named SHACL shapes and SKOS resources, with their own reference categories.
- Other explicitly typed resources belonging to a selected vocabulary.
This describes source declarations; it does not infer a resource's type by its spelling or by executing ontology reasoning. A named individual that also has another declared role retains those assertions without acquiring duplicate reference anchors.
Source descriptions and relationships
Reference fragments retain identifiers, labels, definitions, annotations, and
relationships to other resources. English or untagged source descriptions are
used by the current reference pipeline. The EKGF markdownDescription
annotation supplies richer authored explanations alongside the RDF.
Anonymous restrictions and RDF collections are retained as part of the named resource's description. They are subject to explicit depth and size budgets. External imports remain source assertions; they are not fetched implicitly. Missing or competing annotations can be exposed in a quality report.
Ontology identity and namespace are separate metadata. A single source's ontology IRI need not be identical to its vocabulary namespace. Publisher, profile, scope, and section hints determine the publication rather than changing the ontology's identity.
Publish a reference
The ontology family workflow works with selected vocabularies and explicit family membership. Choose the specification renderer for a standards reference, or bookshelves for module-sized books.