Skip to main content

CreateWorkspaceOntologyWorkflow

What it is​

A Workflow that creates or updates a Naas workspace ontology from YAML data, optionally uploading a Turtle (.ttl) graph as a public asset and linking it as the ontology download URL.

Public API​

  • CreateWorkspaceOntologyWorkflowConfiguration (dataclass)

    • Purpose: Workflow configuration container.
    • Fields:
      • naas_integration_config: NaasIntegrationConfiguration — configuration for NaasIntegration.

  • CreateWorkspaceOntologyWorkflowParameters (WorkflowParameters)

    • Purpose: Input schema for creating/updating an ontology.
    • Fields:
      • yaml_data: dict — YAML-serializable ontology source content.
      • label: str — ontology label (also used to detect existing ontology by label).
      • level: str = "USE_CASE" — ontology level.
      • description: Optional[str] = "Ontology description not provided."
      • logo_url: Optional[str] = "https://naasai-public.s3.eu-west-3.amazonaws.com/abi-demo/ontology_ULO.png"
      • ontology_id: Optional[str] = None — if not provided, the workflow searches by label.
      • graph: Optional[str] = None — Turtle string; when provided, it is parsed and uploaded as an asset.

  • CreateWorkspaceOntologyWorkflow (Workflow)

    • create_or_update_workspace_ontology(parameters) -> str
      • Purpose:
        • If parameters.ontology_id is None, lists existing ontologies in the workspace and reuses the one with matching label if found.
        • Otherwise creates a new ontology (if none exists) or updates the existing one.
        • If parameters.graph is provided, uploads it as a public asset named <label>.ttl and passes its URL as download_url.
      • Returns: the ontology id (str).
      • Raises: ValueError if creation/update fails to produce an ontology id.
    • as_tools() -> list[BaseTool]
      • Purpose: Exposes the workflow as a LangChain StructuredTool named convert_graph_to_yaml (despite performing create/update).
    • as_api(...) -> None
      • Purpose: declared but not implemented (pass).

Configuration/Dependencies​

  • Depends on:
    • NaasIntegration for:
      • list_ontologies(workspace_id)
      • create_ontology(...)
      • update_ontology(...)
      • upload_asset(...)
    • NaasABIModule.get_instance().configuration for:
      • workspace_id
      • storage_name
    • yaml.dump(..., Dumper=yaml.Dumper) to serialize yaml_data into YAML source.
    • rdflib.Graph().parse(data=..., format="turtle") to parse graph when provided.

Usage​

from naas_abi_marketplace.applications.naas.workflows.CreateWorkspaceOntologyWorkflow import (
    CreateWorkspaceOntologyWorkflow,
    CreateWorkspaceOntologyWorkflowConfiguration,
    CreateWorkspaceOntologyWorkflowParameters,
)
from naas_abi_marketplace.applications.naas.integrations.NaasIntegration import (
    NaasIntegrationConfiguration,
)

cfg = CreateWorkspaceOntologyWorkflowConfiguration(
    naas_integration_config=NaasIntegrationConfiguration(
        # populate with required integration settings
    )
)

wf = CreateWorkspaceOntologyWorkflow(cfg)

params = CreateWorkspaceOntologyWorkflowParameters(
    yaml_data={"entities": [], "relations": []},
    label="My Ontology",
    level="USE_CASE",
    # graph="... turtle content ..."  # optional
)

ontology_id = wf.create_or_update_workspace_ontology(params)
print(ontology_id)

Caveats​

  • If ontology_id is not provided, the workflow identifies an existing ontology only by exact label match.
  • When graph is provided:
    • It is parsed as Turtle; invalid Turtle will raise during rdflib parsing.
    • Upload failure is logged, but the ontology can still be created/updated with download_url=None.
  • as_tools() exposes a tool named convert_graph_to_yaml, but it actually performs create/update of the workspace ontology.