fabric-ontology-owl

Import, export, design, and synchronize OWL ontologies with Microsoft Fabric OneLake and Fabric IQ.

4|1|Updated May 22, 2026
One-click install
npx skills add https://github.com/ThomazRossito/ai-data-agents --skill fabric-ontology-owl-thomazrossito
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: fabric-ontology-owl
Source: https://github.com/ThomazRossito/ai-data-agents/tree/main/plugins/ai-data-agents/skills/fabric-ontology-owl
Command: npx skills add https://github.com/ThomazRossito/ai-data-agents --skill fabric-ontology-owl-thomazrossito

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires rdflib, owlready2.

What problem does it solve? Managing OWL ontologies across local files and Microsoft Fabric is error-prone: formats vary, triples must be ingested into Delta tables, and schema changes must be propagated to Fabric IQ Ontology without recreating everything from scratch. This Skill provides structured playbooks for the full ontology lifecycle in Fabric. ## Core Features & Use Cases - Import & Export Playbooks: Validate external OWL/RDF files with rdflib, normalize to Turtle, upload to OneLake, ingest triples into a Delta table via Spark, and export back to Turtle, RDF/XML, N-Triples, or JSON-LD. - Domain Ontology Design: Step-by-step guidance for creating T-Box ontologies from scratch with namespace conventions, labeling checklists, and structural validation. - Schema Sync with Fabric IQ: Diff local OWL classes and properties against Fabric IQ entity and relationship types, then apply surgical add/update/remove operations via MCP tools in dependency-safe order. - Use Case: After adding a new class to your local Turtle ontology, run the Schema Sync playbook to compute the diff against Fabric IQ, present it for approval, and apply only the required entity type and relationship type changes. ## Quick Start Ask the agent to import an external OWL ontology file into the Fabric ontology_lh lakehouse and synchronize its classes with the Fabric IQ Ontology.

Frequently Asked Questions about fabric-ontology-owl

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I import an OWL ontology into Microsoft Fabric?▼

Load and validate the file with rdflib, normalize it to Turtle, upload it to the ontology_lh lakehouse via the OneLake MCP upload tool, then ingest the triples into the ontology_triples Delta table using a Spark notebook with an ABFSS path.

How to sync OWL ontology changes with Fabric IQ Ontology?▼

List existing entity and relationship types via the fabric-ontology MCP tools, diff them against local OWL classes and object properties, present the diff for approval, then apply add, update, or remove operations in dependency-safe order.

What export formats does rdflib support for OWL ontologies?▼

The Skill recommends Turtle for Git review, RDF/XML for Protégé and reasoners, N-Triples for Spark pipeline ingestion, and JSON-LD for REST APIs. rdflib serializes the same graph to any of these formats.

Can I use owlready2 reasoning on Spark executors?▼

No. The Skill explicitly warns against running owlready2 reasoning on Spark executors because it is not distributed. Run reasoning on the driver node or outside the cluster, and use Spark only for triple storage and querying.

Why does a Fabric IQ relationship return zero rows?▼

A documented bug shows the factsales_has_dimproducts edge returned zero rows because its contextualization used StoreId instead of ProductId as the destination key. Destination node key columns must contain the primary key of the destination table.

When should I separate T-Box from A-Box in an ontology?▼

Separate the stable T-Box (classes and properties) from the volatile A-Box (individuals) when the instance volume exceeds 10,000 individuals. Store them in different OneLake folders such as ontologies/domain and ontologies/instances.