oxiddb 0.8.2

Python bindings for Oxid-DB: a neurosymbolic database combining ontology reasoning with vector search

oxiddb

Python bindings for Oxid-DB: a neurosymbolic database combining ontology reasoning with vector search.

This is the embedded package: the database runs in-process inside your Python program (no server, no network). If instead you want to talk to a running Oxid-DB server over HTTP, use the separate remote client oxiddb-client — same project, two delivery models. Install whichever matches how you deploy.

Installation

pip install oxiddb

Quick Start

import oxiddb

# Create a database with 4-dimensional vectors
db = oxiddb.OxidDB(dim=4)

# Build ontology
db.add_subclass("Pilsner", "Beer")
db.add_subclass("Stout", "Beer")

# Insert individuals with vectors
db.insert("Heineken", "Pilsner", vector=[0.9, 0.1, 0.0, 0.0])
db.insert("Guinness", "Stout", vector=[0.1, 0.0, 0.8, 0.9])

# Run the OWL 2 EL reasoner
db.classify()

# Query using OxQL
results = db.query("FIND ?x WHERE ?x IS-A Beer NEAR ?x TO [0.9, 0.1, 0.0, 0.0] LIMIT 5")
for r in results:
    print(f"{r.iri}: {r.score:.4f}")

# Save to disk
db.save("my_database.oxd")