Bitnami PostgreSQL with pgvector

This project provides a Docker image that combines Bitnami's PostgreSQL with the pgvector extension for vector similarity search capabilities.

Quick Install

For specific architectures:

# For linux/amd64
docker pull ghcr.io/4xxi/bitnami-pgvector:latest@sha256:1065402f43c9384a0b34a64d6cab0839f9b332b5cb4d75c97161fb12ad25fc92

# For linux/arm64
docker pull ghcr.io/4xxi/bitnami-pgvector:latest@sha256:4df7853e68c428460c1b529f41e9b5b7dbc7052925c8077980c92c27343dd84f

For multi-arch support (automatically selects the right architecture):

docker pull ghcr.io/4xxi/bitnami-pgvector:latest

Available tags:

• latest - Latest build
• pg17 - Latest PostgreSQL 17 build
• pg17-233f3-67470 - Specific version combination

Features

• Based on the official Bitnami PostgreSQL image
• Includes pgvector extension for vector similarity search
• Multi-architecture support (amd64 and arm64)
• Automated builds via GitHub Actions
• Version tracking based on both PostgreSQL and pgvector versions

Prerequisites

• Docker with BuildKit and multi-architecture support enabled
• GitHub CLI (optional, for authentication)
• curl and wget (for fetching version information)

Usage

Pull the Image

docker pull ghcr.io/${YOUR_GITHUB_USERNAME}/bitnami-pgvector:latest
# or use a specific version
docker pull ghcr.io/${YOUR_GITHUB_USERNAME}/bitnami-pgvector:pg17-xxxxx-xxxxx

Run the Container

docker run -d \
  --name postgres-vector \
  -e POSTGRESQL_PASSWORD=your_password \
  -p 5432:5432 \
  ghcr.io/${YOUR_GITHUB_USERNAME}/bitnami-pgvector:latest

Enable pgvector Extension

After connecting to your database:

CREATE EXTENSION vector;

Building

Automated Builds

The image is automatically built:

• Every day at 20:00 UTC
• On every push to the main branch
• Manually through GitHub Actions interface

To trigger a manual build:

1. Go to the "Actions" tab in your GitHub repository
2. Select the "Build Postgres" workflow
3. Click "Run workflow"
4. (Optional) Change the PostgreSQL version
5. Click "Run workflow"

Building Locally

1. Clone the repository:

git clone https://github.com/${YOUR_GITHUB_USERNAME}/bitnami-pgvector.git
cd bitnami-pgvector

2. Build the image:

PG_MAJOR_VERSION=17 ./build.sh

The script will:

• Check for existing images in the registry
• Download the latest compatible versions of PostgreSQL and pgvector
• Build multi-architecture images (amd64 and arm64)
• Push the images to GitHub Container Registry (if authenticated)

Environment Variables

• PG_MAJOR_VERSION: PostgreSQL major version (required for building)
• GITHUB_TOKEN: GitHub token for authentication (optional, required for registry checks)
• REPO_NAME: Override the default repository name (optional)

Tags

The images are tagged using the following format:

• latest: Latest successful build
• pg{VERSION}: Latest build for a specific PostgreSQL major version
• pg{VERSION}-{BITNAMI_HASH}-{PGVECTOR_HASH}: Specific version combination

Contributing

1. Fork the repository
2. Create your feature branch
3. Commit your changes
4. Push to the branch
5. Create a new Pull Request

License

This project is licensed under the MIT License - see the LICENSE file for details.