Your First Ontology

This tutorial walks through building a real ontology, loading it, inserting data, and querying with subsumption.

The Domain: A Bookstore

We’ll model a bookstore with books, authors, genres, and publishers. The ontology will let us:

Query all “Fiction” and get both “ScienceFiction” and “Fantasy” results
Insert a book with an author and have inverse relationships created automatically
Use domain/range inference to auto-type entities

Step 1: Define Classes

@prefix : <http://example.org/bookstore#> .
@prefix owl: <http://www.w3.org/2002/07/owl#> .
@prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> .
@prefix xsd: <http://www.w3.org/2001/XMLSchema#> .

# Top-level classes
:CreativeWork a owl:Class .
:Person a owl:Class .
:Organization a owl:Class .
:Genre a owl:Class .

# CreativeWork hierarchy
:Book a owl:Class ; rdfs:subClassOf :CreativeWork .
:Article a owl:Class ; rdfs:subClassOf :CreativeWork .

# Person specializations
:Author a owl:Class ; rdfs:subClassOf :Person .
:Editor a owl:Class ; rdfs:subClassOf :Person .

# Organization specializations
:Publisher a owl:Class ; rdfs:subClassOf :Organization .

# Genre hierarchy
:Fiction a owl:Class ; rdfs:subClassOf :Genre .
:NonFiction a owl:Class ; rdfs:subClassOf :Genre .
:ScienceFiction a owl:Class ; rdfs:subClassOf :Fiction .
:Fantasy a owl:Class ; rdfs:subClassOf :Fiction .
:Biography a owl:Class ; rdfs:subClassOf :NonFiction .
:History a owl:Class ; rdfs:subClassOf :NonFiction .

# Disjointness
:Fiction owl:disjointWith :NonFiction .

Step 2: Define Properties

# Object properties (link entities)
:writtenBy a owl:ObjectProperty ;
    rdfs:domain :CreativeWork ;
    rdfs:range :Author .

:authorOf a owl:ObjectProperty ;
    owl:inverseOf :writtenBy .

:publishedBy a owl:ObjectProperty ;
    rdfs:domain :Book ;
    rdfs:range :Publisher .

:hasGenre a owl:ObjectProperty ;
    rdfs:domain :CreativeWork ;
    rdfs:range :Genre .

# Data properties (link entity to literal)
:title a owl:DatatypeProperty ;
    rdfs:domain :CreativeWork ;
    rdfs:range xsd:string ;
    a owl:FunctionalProperty .

:name a owl:DatatypeProperty ;
    rdfs:range xsd:string ;
    a owl:FunctionalProperty .

:isbn a owl:DatatypeProperty ;
    rdfs:domain :Book ;
    rdfs:range xsd:string .

:publicationYear a owl:DatatypeProperty ;
    rdfs:domain :CreativeWork ;
    rdfs:range xsd:integer .

Step 3: Load the Ontology

Save the complete file as bookstore.ttl and load it:

curl -X POST http://YOUR-INSTANCE/ontology \
  -H "Content-Type: text/turtle" \
  --data-binary @bookstore.ttl

Response:

{
  "classes": 14,
  "objectProperties": 3,
  "dataProperties": 4,
  "status": "success",
  "message": "Ontology loaded and materializer activated"
}

Step 4: Insert Data

curl -X POST http://YOUR-INSTANCE/mutate?commitNow=true \
  -H "Content-Type: application/json" \
  -d '{
    "set": [
      {
        "uid": "_:dune",
        "dgraph.type": "ScienceFiction",
        "title": "Dune",
        "isbn": "978-0441172719",
        "publicationYear": 1965,
        "writtenBy": { "uid": "_:herbert" }
      },
      {
        "uid": "_:herbert",
        "dgraph.type": "Author",
        "name": "Frank Herbert"
      }
    ]
  }'

What OWLGraph inferred automatically:

For the book node (_:dune):

Added types: Fiction, Genre, Book, CreativeWork (from ScienceFiction hierarchy and domain inference)

For the author node (_:herbert):

Added types: Person (from Author hierarchy)
Created edge: authorOf → _:dune (from inverseOf on writtenBy)

Step 5: Query with Subsumption

# All fiction (returns Dune even though it was typed as ScienceFiction)
curl -X POST http://YOUR-INSTANCE/query \
  -H "Content-Type: application/dql" \
  -d '{ fiction(func: type(Fiction)) { title dgraph.type } }'

# All creative works
curl -X POST http://YOUR-INSTANCE/query \
  -H "Content-Type: application/dql" \
  -d '{ works(func: type(CreativeWork)) { title writtenBy { name } } }'

# What did Herbert author? (inverse property)
curl -X POST http://YOUR-INSTANCE/query \
  -H "Content-Type: application/dql" \
  -d '{ herbert(func: eq(name, "Frank Herbert")) { name authorOf { title } } }'

Step 6: Introspect

# View the Fiction class hierarchy
curl "http://YOUR-INSTANCE/ontology/introspect?class=Fiction"

{
  "iri": "Fiction",
  "superClasses": ["Genre"],
  "subClasses": ["ScienceFiction", "Fantasy"],
  "disjointWith": ["NonFiction"],
  "properties": [...]
}

What You Learned

Class hierarchies drive subsumption queries — query a parent, get all descendants
Inverse properties create bidirectional relationships automatically
Domain/range inference auto-types entities based on the properties they use
Disjointness prevents invalid data at write time
Introspection lets you explore the ontology at runtime