lexigram-graph

Graph database support for the Lexigram Framework (Neo4j, in-memory).

Overview

lexigram-graph provides graph storage backends with DI wiring for in-memory and Neo4j implementations behind the graph contracts. It supports node and edge creation, graph traversal queries, Cypher compilation for Neo4j, and lazy graph creation.

Full documentation: docs.lexigram.dev

Install

uv add lexigram-graph
# With Neo4j support
uv add "lexigram-graph[neo4j]"

Quick Start

from lexigram import Application
from lexigram.di.module import Module, module
from lexigram.graph import GraphConfig, GraphModule
from lexigram.contracts.data.graph import (
    GraphStoreProtocol,
    TraversalQuery,
    StartSpec,
    TraversalStep,
)


@module(imports=[GraphModule.configure(GraphConfig(backend="memory"))])
class AppModule(Module):
    pass


async def main() -> None:
    async with Application.boot(modules=[AppModule]) as app:
        store = await app.container.resolve(GraphStoreProtocol)
        graph = await store.get_graph()

        await graph.create_node(["Person"], {"name": "Alice"}, node_id="alice")
        await graph.create_node(["Person"], {"name": "Bob"}, node_id="bob")
        await graph.create_edge("alice", "bob", "KNOWS")

        paths = await graph.traverse(
            TraversalQuery(
                start=StartSpec(node_ids=("alice",)),
                steps=(TraversalStep(edge_types=("KNOWS",)),),
            )
        )
        assert paths


if __name__ == "__main__":
    import asyncio

    asyncio.run(main())

Configuration

Zero-config usage: Call GraphModule.configure() with no arguments to use all defaults (in-memory backend).

Option 1 — YAML file

# application.yaml
graph:
  enabled: true
  backend: neo4j
  default_traversal_max_depth: 10
  neo4j:
    uri: bolt://localhost:7687
    password: "${NEO4J_PASSWORD}"

Option 2 — Profiles + Environment Variables (recommended)

export LEX_GRAPH__BACKEND=neo4j
export LEX_GRAPH__NEO4J__URI=bolt://localhost:7687

Option 3 — Python

from lexigram.graph import GraphConfig, GraphModule
from lexigram.graph.config import Neo4jConfig

GraphModule.configure(
    GraphConfig(
        backend="neo4j",
        neo4j=Neo4jConfig(
            uri="bolt://localhost:7687",
            password="${NEO4J_PASSWORD}",
        ),
    )
)

Config reference

Module Factory Methods

Key Features

• In-memory backend — no external service needed; for development and tests
• Neo4j backend — async Neo4j driver with Cypher query compilation
• Graph traversal — TraversalQuery, StartSpec, TraversalStep for graph walks
• Named graphs — lazy graph creation per name
• Connection pooling — configurable pool size for Neo4j driver

Testing

async with Application.boot(modules=[GraphModule.stub()]) as app:
    store = await app.container.resolve(GraphStoreProtocol)
    graph = await store.get_graph()
    # Test with in-memory backend

Key Source Files

Multi-Tenancy

lexigram-graph supports two isolation strategies:

Strategies

Configuration

from lexigram.graph import GraphModule
from lexigram.graph.config import GraphConfig, GraphTenancyConfig

config = GraphConfig(
    backend="neo4j",
    tenancy=GraphTenancyConfig(
        enabled=True,
        strategy="node_property",
        template="{logical}_t_{tenant}",
    ),
)
GraphModule.configure(config)

Components