$ show KRR-Oxford/DeepOnto
DeepOnto is a Python package for ontology engineering with deep learning, centered on ontology alignment (BERTMap) and subsumption inference (BERTSubs, OntoLAMA). It wraps the OWLAPI Java library through JPype to load, query, reason over, and prune OWL ontologies, and uses PyTorch/Hugging Face Transformers to train or apply BERT-based classifiers for mapping prediction and refinement. The package exposes config-driven pipelines (BERTMapPipeline, BERTSubs pipelines, OntoLAMA inference) plus evaluation utilities for OAEI-style benchmarks.
The LLM identified 4 files worth highlighting for reuse:
Facade over the OWLAPI Java library via JPype. Centralizes loading, querying, axiom manipulation, annotation extraction, and saving of OWL ontologies. Reusable across any OWL-based ontology engineering project; demonstrates non-obvious JVM interop and ontology API design.
Hybrid sub-word/rule-based tokenizer with an InvertedIndex for fast candidate generation. Self-contained text processing utility that can be dropped into any ontology matching or entity linking project.
Implements precision/recall/F1, Hits@K, and MRR for ontology alignment outputs. Isolated metric computation pattern reusable for any matching or retrieval evaluation pipeline.
Defines EntityMapping and ReferenceMapping and generates equivalence/subsumption candidates including negative sampling strategies. The candidate generation logic is reusable for ontology matching datasets and alignment benchmarks.