quickstart.py

"""Minimal, runnable example — a tiny generic ontology + a small spec table.

    python examples/minimal/quickstart.py

Requires ``rdflib`` (``pip install -e ".[dev]"`` installs it). Demonstrates:
declare a spec table, parse an extension TTL that imports a core, and generate
typed JSON configs holding the extracted named subclasses + individuals.
"""

import json
import tempfile
from pathlib import Path

from ontology_config_generator import ConfigSpec, generate_configs

HERE = Path(__file__).resolve().parent
EXTENSION = HERE / "ontology" / "extension.ttl"

# 1) Declare YOUR spec table: which file binds to which class, under which key.
#    (Plain dicts work too — see spec_table.json — these are the typed form.)
spec_table = [
    ConfigSpec(filename="regions.json", extract={"regions": "Region"}),
    ConfigSpec(filename="categories.json", extract={"categories": "Category"}),
    # A "static-only" config the TTL has nothing to extract for keeps its shape.
    ConfigSpec(filename="glossary.json", static={"terms": [], "synonyms": []}),
]

with tempfile.TemporaryDirectory() as d:
    out = Path(d)

    # 2) Generate. The namespace defaults to http://example.org/ontology/core/
    #    (the IRI the bound class local-names — "Region", "Category" — resolve to).
    status = generate_configs(spec_table, EXTENSION, out)
    print("generated:", status)

    # 3) Inspect what was extracted.
    regions = json.loads((out / "regions.json").read_text())
    print("\nregions.json:")
    for r in regions["regions"]:
        print(f"  - {r['id']}: {r['label']} — {r['description']}")
    print("  provenance:", regions["_generated"])

    categories = json.loads((out / "categories.json").read_text())
    print("\ncategories.json:")
    for c in categories["categories"]:
        print(f"  - {c['id']}: {c['label']}")

    glossary = json.loads((out / "glossary.json").read_text())
    print("\nglossary.json (static-only):", {k: v for k, v in glossary.items() if k != "_generated"})