Implement auto-ontology generation

[krapfj23]

Summary

Build the system that automatically generates and maintains an ontology from ingested documents.

Blocked by

• Research: auto-generated ontology from ingested documents #74 (auto-ontology research)

Blocks

• Expose ontology editing endpoint #76 (ontology editing endpoint)

Steps

1. Based on the research recommendation, implement the chosen approach
2. Create backend/app/services/ontology.py with function generate_ontology(dataset_name) -> str that returns an OWL/TTL string
3. After cognee.cognify(), analyze extracted entities/relationships and generate/update the ontology file
4. Store the generated ontology at a known path (e.g., data/ontologies/<dataset_name>.ttl)
5. On subsequent ingestions to the same dataset, pass the existing ontology to cognee.cognify(ontology_file_path=...)

Acceptance Criteria

• Ontology is auto-generated after document ingestion
• Ontology file is persisted and reused on subsequent ingestions
• Re-ingestion updates (not overwrites) the ontology
• Extraction quality improves with the ontology applied
• PR body includes Closes #75