openGauss Vector Database Adapter

This is a community-contributed adapter for integrating openGauss DataVec with Cognee.

About openGauss

openGauss is an open-source relational database with ACID transactions and built-in vector capabilities via DataVec (6.0.0+). It provides native VECTOR(n) type, HNSW / IVFFLAT / IVFPQ indexes, and hybrid SQL + vector similarity search — all in a single database.

Installation

1. Install the required dependencies:

   # Option 1: Install dependencies directly
   pip install psycopg2-binary>=2.9.9
   
   # Option 2: Install as a package (if published)
   pip install cognee-community-vector-adapter-opengauss
   
   # Option 3: Install from source
   cd packages/vector/opengauss
   pip install .
   

2. Import and register the adapter in your code:

   from cognee_community_vector_adapter_opengauss import register
   

Setup

Start an openGauss container:

docker run -d \
    --name opengauss \
    --privileged=true \
    --restart=always \
    -p 5432:5432 \
    -e GS_PASSWORD=openGauss@123 \
    -v ./opengauss:/var/lib/opengauss/data \
    opengauss/opengauss:7.0.0-RC1

No additional extension required — DataVec is built in.

Configuration

Create a .env file from the template:

cp .env.example .env
# Edit .env with your LLM / embedding / database settings

Configure Cognee to use openGauss:

from dotenv import load_dotenv
from cognee import config
from cognee_community_vector_adapter_opengauss import register

load_dotenv()

config.set_vector_db_config(
    {
        "vector_db_provider": "opengauss",
        "vector_db_url": "postgresql://gaussdb:openGauss%40123@localhost:5432/postgres",
    }
)

Required environment variables:

OPENGAUSS_URL=postgresql://gaussdb:OpenGauss%40123@localhost:5432/postgres
ENABLE_BACKEND_ACCESS_CONTROL=false

LLM and embedding providers must also be configured — see the Cognee docs.

Optional openGauss settings (all have defaults):

OPENGAUSS_SCHEMA_NAME=cognee           # Schema for vector tables
OPENGAUSS_INDEX_TYPE=HNSW              # HNSW | IVFFLAT | IVFPQ | HNSW-PQ
OPENGAUSS_DISTANCE_STRATEGY=COSINE     # COSINE | EUCLIDEAN | MANHATTAN | INNER_PROD
OPENGAUSS_CREATE_INDEX=false           # Enable when index implementation is stable

Usage Example

import asyncio
import os
import pathlib
from dotenv import load_dotenv
from cognee import add, cognify, config, search, SearchType
from cognee_community_vector_adapter_opengauss import register

load_dotenv()


async def main():
    # Set working directories
    root = pathlib.Path(__file__).parent
    config.system_root_directory(str(root / ".cognee_system"))
    config.data_root_directory(str(root / ".cognee_data"))

    # Configure openGauss
    config.set_vector_db_config(
        {
            "vector_db_provider": "opengauss",
            "vector_db_url": os.getenv(
                "OPENGAUSS_URL",
                "postgresql://gaussdb:openGauss%40123@localhost:5432/postgres",
            ),
        }
    )

    # Add and process data
    await add("Natural language processing is a subfield of AI.")
    await cognify()

    # Search
    results = await search("Tell me about NLP", query_type=SearchType.GRAPH_COMPLETION)
    print(results)


asyncio.run(main())

Features

• ACID transactions: Full transactional guarantees for vector writes
• Multiple index types: HNSW, IVFFLAT, IVFPQ, and HNSW-PQ index algorithms
• Hybrid search: Combine SQL filtering with vector similarity search in a single query
• Schema isolation: All vector tables live in a configurable cognee schema
• Self-contained: No separate vector database or extension required — DataVec is built in

Testing

Run the tests to verify the adapter works correctly:

python tests/test_opengauss.py

Dependencies

• psycopg2-binary>=2.9.9: PostgreSQL database adapter for Python

Production Deployment

Deploy openGauss on bare metal or Kubernetes for high availability. It supports active-standby replication and automatic failover. See the openGauss installation guide for details.

Support

For issues specific to this adapter:

1. Check the openGauss documentation
2. Create an issue in the main Cognee repository with the "community-adapter" label
3. Refer to the examples/ and tests/ directories for usage patterns

License

This adapter is licensed under the MIT license, same as the main Cognee project.