runicGraph schema migrations and OGM

For Cypher-based graph databases — FalkorDB, ArcadeDB, Neo4j, Memgraph, Apache AGE.

🗂️

Graph OGM

Map Python classes to nodes and edges. Typed fields, indexes, constraints, lazy/eager relationships, and a fluent query builder.

🔄

Schema migrations

Alembic-style versioned migration engine. Track index and constraint changes as replayable scripts with upgrade and downgrade paths.

🔌

Multi-backend

Pluggable driver layer supports FalkorDB, ArcadeDB, Neo4j, Memgraph, and Apache AGE (PostgreSQL). Switch backends without rewriting models.

⚡

Async first-class

AsyncSession mirrors the sync API. No hidden lazy loads — deterministic query patterns for high-throughput applications.

🧪

Testable by design

Embedded FalkorDB via redislite. No Docker required for unit tests — full CRUD, relationship, and query coverage in-process.

🔍

Vector & fulltext search

Native Vector KNN and fulltext search operations. Declare vector indexes in your model, query with .knn() — no raw Cypher needed.

Quick look

python

from runic.ogm import Field, Node, Repository, Session, create_driver

class Person(Node, labels=["Person"]):
    id: str = Field(index=True)
    name: str
    email: str = Field(index=True, unique=True)

driver = create_driver("falkordb", host="localhost", port=6379, graph="myapp")

with Session(driver) as session:
    session.add(Person(id="alice", name="Alice", email="alice@example.com"))
    session.commit()

    repo = Repository(session, Person)
    print(repo.count())   # 1

Documentation

See the OGM Quickstart or Migration Quickstart to get up and running.