Knowledge Graph Engine

Multi-tenant knowledge graph engine for the SilvaEngine AI management platform. The module is shipped as a GraphQL plugin for the silvaengine_base gateway; the gateway owns authentication, request routing, and connection-pool lifecycle, while this module supplies the domain logic, Graphene resolvers, and the canonical schema (Postgresql

• Neo4j).

Boundaries

The module is reached only through the gateway's GraphQL dispatch. It does not:

• Bind an HTTP server.
• Start a background loop, thread pool, or scheduled job.
• Read environment variables or perform its own secret lookups.
• Maintain an HTTP-facing JWT contract.

Authentication, authorization, rate limiting, routing, and connection pool creation are the gateway's responsibility. See docs/requirements_realization_statement.md for the full boundary declaration.

Module layout

knowledge_graph_engine/
├── __init__.py
├── main.py                       # Engine plugin class + deploy() metadata
├── schema.py                     # Graphene Query / Mutation roots
├── config.json                   # SilvaEngine module declaration
├── exceptions.py                 # Business-module error taxonomy
├── _compat.py                    # Stub modules for partial environments
├── docs/                         # Requirements Realization Statement
├── handlers/
│   ├── config.py                 # Thread-safe Config singleton
│   ├── neo4j_connection_manager.py
│   ├── partition_manager.py
│   ├── schema_resolver.py
│   ├── extractor.py
│   ├── search_handler.py
│   └── rag_handler.py
├── models/postgresql/            # SQLAlchemy models + session helpers
├── mutations/                    # Graphene mutations
├── queries/                      # Graphene queries
├── types/                        # Graphene output / input types
├── utils/                        # Repository helpers + GraphRAG util
├── migrations/                   # SQL migration scripts
└── tests/                        # Unit + integration tests

Local development

# Install the module dependencies (Python 3.10+).
pip install -e .[dev]

# Run the unit tests.
python -m unittest discover knowledge_graph_engine/tests -v

Configuration

The module reads runtime configuration through the gateway's ConfigModel.find(setting_id="knowledge_graph_engine") call. The shape of the configuration dict is declared in config.json under settings[knowledge_graph_engine_engine].

License

Internal — see your platform's licensing terms.