Skip to main content

ConvertOntologyGraphToYamlWorkflow

What it is​

A workflow that:

  • Parses an ontology graph provided as a Turtle (.ttl) string into an RDFLib Graph
  • Uploads the Turtle content as a public asset to Naas storage
  • Converts the RDF graph to a YAML representation via OntologyYaml.rdf_to_yaml
  • Creates or updates an ontology in Naas using the generated YAML and uploaded asset URL
  • Returns the resulting ontology ID

Public API​

Classes​

  • ConvertOntologyGraphToYamlWorkflowConfiguration(WorkflowConfiguration)

    • Holds configuration required to initialize the workflow.
    • Field: naas_integration_config: NaasIntegrationConfiguration

  • ConvertOntologyGraphToYamlWorkflowParameters(WorkflowParameters)

    • Input parameters for a conversion run.
    • Fields:
      • graph: str (required): Graph serialized in Turtle format.
      • ontology_id: Optional[str]: Optional ontology ID to update (if not provided, workflow may resolve by label or create new).
      • label: str (default "New Ontology"): Ontology label.
      • description: str (default "New Ontology Description"): Ontology description.
      • logo_url: Optional[str] (default provided): Ontology logo URL.
      • level: str (default "USE_CASE"): Ontology level.
      • display_relations_names: bool (default True): Controls relation name display in YAML conversion.
      • class_colors_mapping: Dict (default COLORS_NODES): Class-to-color mapping used in YAML conversion.

  • ConvertOntologyGraphToYamlWorkflow(Workflow)

    • Main workflow implementation.

Methods​

  • ConvertOntologyGraphToYamlWorkflow.__init__(configuration)

    • Creates a NaasIntegration from naas_integration_config.
    • Creates an OntologyYaml converter using the module triple store service.

  • graph_to_yaml(parameters: ConvertOntologyGraphToYamlWorkflowParameters) -> str

    • Converts Turtle graph to YAML, uploads the Turtle as an asset, then creates/updates a Naas ontology.
    • Returns: ontology_id (string).
    • Raises:
      • ValueError if asset upload fails, asset URL missing, or ontology creation/update ultimately fails.
      • Re-raises exceptions from RDF-to-YAML conversion.

  • as_tools() -> list[BaseTool]

    • Exposes the workflow as a LangChain StructuredTool named convert_graph_to_yaml.

  • as_api(...) -> None

    • Present but not implemented (pass).

Configuration/Dependencies​

  • External libraries:

    • rdflib.Graph for parsing Turtle.
    • yaml for dumping YAML content (yaml.dump(..., Dumper=Dumper)).
    • pydash.get for safe response access.
    • langchain_core.tools.StructuredTool for tool exposure.

  • Internal dependencies/services:

    • NaasIntegration is used to:
      • upload_asset(...)
      • list_ontologies(workspace_id)
      • create_ontology(...)
      • update_ontology(...)
    • NaasABIModule.get_instance() is used to access:
      • configuration.workspace_id
      • configuration.storage_name
      • engine.services.triple_store (passed into OntologyYaml)
    • OntologyYaml.rdf_to_yaml(...) performs RDF graph → YAML conversion.
    • Default class_colors_mapping comes from COLORS_NODES.

Usage​

from naas_abi_marketplace.applications.naas.integrations.NaasIntegration import (
    NaasIntegrationConfiguration,
)
from naas_abi_marketplace.applications.naas.workflows.ConvertOntologyGraphToYamlWorkflow import (
    ConvertOntologyGraphToYamlWorkflow,
    ConvertOntologyGraphToYamlWorkflowConfiguration,
    ConvertOntologyGraphToYamlWorkflowParameters,
)

# Configure Naas integration (fields depend on NaasIntegrationConfiguration definition)
naas_cfg = NaasIntegrationConfiguration(...)  # fill with your integration settings

workflow = ConvertOntologyGraphToYamlWorkflow(
    ConvertOntologyGraphToYamlWorkflowConfiguration(naas_integration_config=naas_cfg)
)

ttl = """
@prefix ex: <http://example.com/> .
ex:A ex:rel ex:B .
"""

ontology_id = workflow.graph_to_yaml(
    ConvertOntologyGraphToYamlWorkflowParameters(
        graph=ttl,
        label="Example Ontology",
        description="Example description",
        level="USE_CASE",
    )
)

print(ontology_id)

As a LangChain tool​

tools = workflow.as_tools()
result = tools[0].invoke(
    {
        "graph": ttl,
        "label": "Example Ontology",
        "description": "Example description",
        "level": "USE_CASE",
    }
)
print(result)  # ontology_id

Caveats​

  • graph must be valid Turtle; invalid content will fail during Graph.parse(...).
  • Ontology selection logic:
    • If an existing ontology in the workspace has the same label, its id is used (even if ontology_id was not provided).
  • as_api(...) is not implemented; no HTTP endpoints are registered by this workflow.
  • Asset upload is required; the workflow fails if upload_asset returns no asset or no URL.