Skip to main content

OntologyEngineerAgent

What it is​

An Agent implementation configured to act as a BFO (Basic Formal Ontology) expert and ontology engineering specialist, with a system prompt that guides:

  • answering educational BFO questions, and
  • delegating text-to-ontology mapping and triplestore operations to subordinate agents.

Public API​

  • create_agent(agent_shared_state: Optional[AgentSharedState] = None, agent_configuration: Optional[AgentConfiguration] = None) -> Optional[Agent]

    • Factory that builds and returns an OntologyEngineerAgent preconfigured with:
      • default chat model (naas_abi.models.default.get_model()),
      • AgentConfiguration(system_prompt=SYSTEM_PROMPT) if none provided,
      • AgentSharedState() if none provided,
      • sub-agents:
        • EntitytoSPARQLAgent.create_agent()
        • KnowledgeGraphBuilderAgent.create_agent()

  • class OntologyEngineerAgent(Agent)

    • Concrete agent type; no additional methods beyond the base Agent (inherits behavior).

Configuration/Dependencies​

  • Depends on core agent types:
    • naas_abi_core.services.agent.Agent.Agent
    • AgentConfiguration, AgentSharedState
  • Loads model via:
    • naas_abi.models.default.get_model()
  • Composes two other agents:
    • naas_abi.agents.EntitytoSPARQLAgent.create_agent
    • naas_abi.agents.KnowledgeGraphBuilderAgent.create_agent
  • Key constants:
    • NAME = "Ontology_Engineer"
    • DESCRIPTION = "A agent that helps users understand BFO Ontology and transform text into ontologies."
    • SYSTEM_PROMPT (defines operating guidelines and constraints)

Usage​

from naas_abi.agents.OntologyEngineerAgent import create_agent

agent = create_agent()
# The returned object is an OntologyEngineerAgent (subclass of Agent),
# configured with the SYSTEM_PROMPT and two sub-agents.

Caveats​

  • The module’s behavior is primarily defined by SYSTEM_PROMPT; the OntologyEngineerAgent class itself does not add custom logic (it is pass).
  • create_agent() instantiates sub-agents immediately (entity_to_sparql_agent() and knowledge_graph_builder_agent()), which may have side effects depending on those implementations.