Load Fibo Ontology

---
name: load-fibo-ontology
description: Load and verify Financial Industry Business Ontology content in Neo4j with the Onto2AI loader. Use for FIBO presets, explicit ontology IRIs, namespace failures, and source-ontology preparation.
---
# FIBO Loading Expert Instructions

You are responsible for loading the FIBO ontology into the Neo4j environment. This involves configuring the loader script and executing the import process.

## Loading Process

### Configure the Loader
The core loading logic resides in `neo4j_onto2ai_toolset/onto2ai_loader.py`.

Use the package CLI instead of editing the `if __name__ == "__main__":` block. The loader now expects either an explicit ontology URI or a known preset.

Common commands:

```bash
python -m neo4j_onto2ai_toolset.onto2ai_loader load --uri <ontology_iri> --no-reset
python -m neo4j_onto2ai_toolset.onto2ai_loader load --preset default-domains --no-reset \
  --source-version <fibo_release> --source-commit <upstream_commit>
```

`load` defaults to `--reset`. Use `--no-reset` unless the user explicitly requested a clean replacement of the target database and the configured database name has been confirmed.

Use the default FIBO domain slice as the normal starting point for industry-ontology exploration. Expand domains only when the user needs a broader source landscape.

For reproducible loads, prefer a published FIBO release URI or supply both the
upstream release identifier and commit. If the selected RDF uses
`master/latest`, the loader records it as a rolling version and computes a
SHA-256 fingerprint of every source document; do not describe that load as a
specific quarterly release unless `--source-version` identifies one.

For a checked-out FIBO release tag, set `ONTOLOGY_CATALOG_PATH` to its
`catalog-v001.xml` and set `ONTOLOGY_ROOT_PATH` to an isolated dependency cache.
Use `--local-files-only` only when the catalog and cache contain every imported
dependency; otherwise allow missing non-FIBO dependencies to be fetched into
the isolated cache.

### Execute the Load
Run from the root of the workspace so package imports and prefixes are correctly resolved:

```bash
python -m neo4j_onto2ai_toolset.onto2ai_loader load --preset default-domains --no-reset
```

### Post-Load Verification
After loading, the script automatically:
- Materializes OWL restrictions into Neo4j relationships and properties.
- Cleans up duplicate relationships.
- Resolves namespaces using the controlled list in `neo4j_onto2ai_toolset/onto2ai_core/prefixes.py`.
- Writes an `Onto2AILoadProvenance` node with source version, optional commit,
  timestamps, document counts, root IRIs, and exact-content fingerprint.

## Troubleshooting
- **ShortenStrictException**: If this occurs, a new namespace was found. Add it to `neo4j_onto2ai_toolset/onto2ai_core/prefixes.py`.
- **Database Reset**: The `load` command resets by default. Confirm the configured target database before omitting `--no-reset` or passing `--reset`; a reset deletes the existing graph before loading.
- **Source vs Target**: FIBO is a source ontology. Use MCP/Modeller Source Ontology to search and extract a focused subset, then stage and curate the target ontology separately.
- **Post-load Check**: After loading, use `search_ontology_concepts` or `list_model_classes` against the loaded database to confirm expected concepts are present.