External Ontologies

OWL ontologies frequently reference external vocabularies via owl:imports statements or namespace prefixes. OWL-SDA can automatically fetch, parse, and cache these external models so that the reasoner and SHACL generator have access to all imported definitions.

Configuration

extract:
  connect-timeout-ms: 2000
  read-timeout-ms: 5000
  max-retries: 1
  follow-redirects: true
  user-agent: "owlsda/1.0"
  cache-enabled: true
  cache-ttl-ms: 3600000      # 1 hour
  cache-max-entries: 100
  cache-dir: "target/cache/ontology-extract-external"
  cache-format: "TURTLE"
  mirrors: []

Key	Type	Default	Description
connect-timeout-ms	int	2000	HTTP connection timeout in milliseconds.
read-timeout-ms	int	5000	HTTP read timeout in milliseconds.
max-retries	int	1	Number of additional attempts after the first failure.
follow-redirects	boolean	true	Whether HTTP redirects are followed automatically.
user-agent	string	owlsda/1.0	User-Agent header sent with HTTP requests.
cache-enabled	boolean	true	Enable on-disk caching of resolved ontologies.
cache-ttl-ms	long	3600000	Maximum age of a cached entry in milliseconds. 0 = never expire.
cache-max-entries	int	100	Maximum number of entries in the in-memory cache.
cache-dir	string	target/cache/…	Directory for on-disk cached ontology files.
cache-format	string	TURTLE	RDF serialisation format used for the on-disk cache.
mirrors	list	[]	URI-to-mirror mappings. See Mirrors below.

How It Works

When the ontology is loaded, OWL-SDA scans for:

1. owl:imports triples — direct ontology imports declared in the file.
2. Namespace prefix URIs — all namespace prefixes defined in the ontology.

For each URI it tries to resolve a model in this order:

1. In-memory cache (within the current run).
2. On-disk file cache (cache-dir).
3. HTTP fetch (with retry and redirect support).

Successfully fetched models are written to the on-disk cache for future runs.

Mirrors

Some ontology URIs are unreliable or offline. You can map any URI to one or more mirror URLs:

extract:
  mirrors:
    - uri: "http://www.w3.org/ns/prov#"
      mirrors:
        - "https://www.w3.org/ns/prov-o"
    - uri: "http://xmlns.com/foaf/0.1/"
      mirror: "https://xmlns.com/foaf/spec/index.rdf"

Both mirrors (list) and mirror (single string) are accepted. Prefix matching is also supported, so a mirror entry for http://example.org/ applies to any URI starting with that prefix.

Supported Formats

The extractor recognises the Content-Type header returned by the server and falls back to filename extension heuristics:

Content-Type	Extension	Format
text/turtle, application/x-turtle, application/turtle, text/ttl	(default)	Turtle
application/ld+json, application/json	.jsonld, .json	JSON-LD
application/rdf+xml, application/xml, text/xml	.rdf, .xml	RDF/XML
application/n-triples, application/ntriples, text/plain	.nt, .ntriples	N-Triples
text/n3	.n3	N3

The two heuristics are independent. The Content-Type header is tried first; if it is missing or unrecognised, the extension is used instead. A reference matching no extension is parsed as Turtle. If parsing produces an empty model the extractor retries with the other formats, so a server that mislabels its content usually still works.