Skip to main content

Ontology Building

Overview​

The ABI project uses a hierarchical ontology structure organized into four distinct layers, providing a flexible and extensible knowledge representation system. The ontology is built using RDF/OWL and follows the Basic Formal Ontology (BFO) and Common Core Ontologies (CCO) as upper-level ontologies.

Ontology Hierarchy​

1. Top-Level Ontologies​

  • Foundation layer using BFO (Basic Formal Ontology)
  • Provides fundamental concepts and relationships
  • Example: bfo-core.ttl
<http://purl.obolibrary.org/obo/bfo.owl> rdf:type owl:Ontology ;
                                      owl:versionIRI <http://purl.obolibrary.org/obo/bfo/2020/bfo-core.ttl> ;
                                      dc11:contributor "Alan Ruttenberg" ,
                                                       "Albert Goldfain" ,
                                                       "Barry Smith" ,
                                                       "Bill Duncan" ,
                                                       "Bjoern Peters" ,
                                                       "Chris Mungall" ,
                                                       "David Osumi-Sutherland" ,
                                                       "Fabian Neuhaus" ,
                                                       "James A. Overton" ,
                                                       "Janna Hastings" ,
                                                       "Jie Zheng" ,
                                                       "John Beverley" ,
                                                       "Jonathan Bona" ,
                                                       "Larry Hunter" ,
                                                       "Leonard Jacuzzo" ,
                                                       "Ludger Jansen" ,
                                                       "Mark Jensen" ,
                                                       "Mark Ressler" ,
                                                       "Mathias Brochhausen" ,
                                                       "Mauricio Almeida" ,
                                                       "Melanie Courtot" ,
                                                       "Neil Otte" ,

2. Mid-Level Ontologies​

  • Built on top of BFO
  • Uses Common Core Ontologies (CCO)
  • Includes general concepts shared across domains
  • Examples:
    • Agent Ontology
    • Artifact Ontology
    • Facility Ontology
    • Geospatial Ontology
    • Information Entity Ontology
    • Quality Ontology
    • Units of Measure Ontology

3. Domain-Level Ontologies​

Domain-specific concepts and relationships. Examples include:

  • Person Ontology
abi:Person a owl:Class ;
    rdfs:label "Person"@en ;
    rdfs:subClassOf bfo:BFO_0000040 ;  # Material entity
    skos:definition "An individual person with associated professional and interaction data"@en .
  • Organization Ontology
abi:Organization a owl:Class ;
    rdfs:label "Organization"@en ;
    rdfs:subClassOf bfo:BFO_0000027 ;  # Object aggregate
    skos:definition "A business or company entity with associated metadata"@en .
  • Task Ontology
abi:Task rdf:type owl:Class ;
    rdfs:subClassOf cco:ont00000228 ; # Planned Act
    rdfs:label "Task"@en ;
    rdfs:comment "A defined unit of work that needs to be accomplished within a certain period of time."@en ;
    skos:definition "A Task is a planned activity with a specific objective, timeline, and measurable outcome."@en ;
    skos:example "Writing documentation, fixing a bug, or implementing a new feature."@en .

4. Application-Level Ontologies​

Specific implementations for different platforms and services:

  • GitHub
  • LinkedIn
  • HubSpot
  • Perplexity

Creating New Ontologies​

1. File Structure​

Create new ontology files in the appropriate directory under src/ontologies/:

  • Top-level: src/ontologies/top-level/
  • Mid-level: src/ontologies/mid-level/
  • Domain-level: src/ontologies/domain-level/
  • Application-level: src/ontologies/application-level/

2. Basic Ontology Template​

@prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> .
@prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> .
@prefix owl: <http://www.w3.org/2002/07/owl#> .
@prefix xsd: <http://www.w3.org/2001/XMLSchema#> .
@prefix bfo: <http://purl.obolibrary.org/obo/bfo/> .
@prefix cco: <http://www.ontologyrepository.com/CommonCoreOntologies/> .
@prefix skos: <http://www.w3.org/2004/02/skos/core#> .
@prefix abi: <http://ontology.naas.ai/abi/> .

<http://ontology.naas.ai/abi/your-ontology.ttl> rdf:type owl:Ontology ;
    dc11:contributor "Your Name" ;
    dc:description "Description of your ontology"@en ;
    dc:title "Your Ontology Title" .

3. Adding Classes and Properties​

Example of defining a new class:

abi:YourClass a owl:Class ;
    rdfs:label "Your Class"@en ;
    rdfs:subClassOf bfo:BFO_0000001 ;  # Appropriate BFO or CCO parent class
    skos:definition "Definition of your class"@en ;
    skos:example "Example of your class usage"@en .

Integration with ABI​

1. Registering Ontologies​

Use the Naas Integration API to register new ontologies:

    class CreateOntologySchema(BaseModel): 
        workspace_id: str = Field(..., description="Workspace ID to create an ontology in")
        label: str = Field(..., description="Label for the ontology")
        ontology: str = Field(..., description="Ontology source/content")
        level: str = Field(..., description="Level of the ontology - one of: USE_CASE, DOMAIN, MID, TOP")
    class GetOntologySchema(BaseModel):
        workspace_id: str = Field(..., description="Workspace ID to get an ontology from")
        ontology_id: Optional[str] = Field("", description="Optional ontology ID to get a specific ontology. If not provided, lists all ontologies")

    class GetOntologiesSchema(BaseModel):
        workspace_id: str = Field(..., description="Workspace ID to get ontologies from")

    class UpdateOntologySchema(BaseModel):
        workspace_id: str = Field(..., description="Workspace ID to update an ontology in")
        ontology_id: str = Field(..., description="ID of the ontology to update")
        ontology_source: Optional[str] = Field(None, description="Updated ontology source")
        level: Optional[str] = Field(None, description="Updated ontology level")

    class DeleteOntologySchema(BaseModel):
        workspace_id: str = Field(..., description="Workspace ID to delete an ontology from")
        ontology_id: str = Field(..., description="ID of the ontology to delete")

2. Visualization​

The system includes an ontology visualizer for exploring relationships:

class OntologyVisualizer:
    def __init__(self, ontology_path: str = "ontology/consolidated_ontology.ttl"):
        self.graph = rdflib.Graph()
        self.logger = logging.getLogger(__name__)
        self.ontology_path = Path(ontology_path)
        
        # Initialize pyvis network
        self.net = Network(
            height="750px",
            width="100%",
            bgcolor="#ffffff",
            font_color="#000000"
        )
        
        # Configure physics
        self.net.force_atlas_2based()
        self.net.show_buttons(filter_=['physics'])
        
        # Define color scheme for different namespaces
        self.namespace_colors = {
            'bfo': '#4287f5',    # Blue
            'cco': '#f5d142',    # Yellow
            'domain': '#f58442', # Orange
            'application': '#42f554'  # Green
        }

Best Practices​

  1. Alignment with Upper Ontologies

    • Always align new concepts with BFO and CCO where possible
    • Use existing relationships from upper ontologies when available

  2. Documentation

    • Include clear labels and definitions using rdfs:label and skos:definition
    • Provide examples using skos:example
    • Document the purpose and scope of each ontology

  3. Naming Conventions

    • Use CamelCase for class names
    • Use lowerCamelCase for properties
    • Maintain consistent prefixes across related concepts

  4. Version Control

    • Include version information in ontology metadata
    • Document major changes in ontology evolution
    • Maintain backward compatibility when possible

  5. Testing

    • Validate ontology syntax and structure
    • Test integration with existing ontologies
    • Verify alignment with upper-level ontologies

Tools and Resources​

  1. Visualization: Use the built-in OntologyVisualizer class for exploring ontology relationships
  2. Validation: RDFlib for parsing and validating ontologies
  3. Integration: Naas Integration API for managing ontologies in the workspace

For more detailed information about specific ontology components, refer to the individual ontology files in the src/ontologies/ directory.