Vector Sets Browser

A modern web-based visualization and interaction tool for Redis Vector Sets. This application provides an intuitive interface for exploring and analyzing vector embeddings stored in Redis vector-sets.

Overview

Vector Sets Browser is a Next.js application that provides real-time visualization of vector embeddings and their relationships. It features multiple visualization layouts including:

• Force-directed graph layout
• UMAP dimensionality reduction
• PCA (Principal Component Analysis) visualization
• 3D vector space visualization

Prerequisites

• Node.js (Latest LTS version recommended)
• A Redis server with vector sets. ((Available in Beta with the latest Redis 8))
• (optional) OpenAI API key (for AI-assisted template generation)
• (optional) Ollama for embedding generation

Installation

Clone the repository:

git clone https://github.com/redis/vector-sets-browser.git
cd vector-sets-browser

Running with Docker

You can run this project using the provided Dockerfile. This allows you to avoid building the project manually. To do so, follow these steps:

1. Build the Docker image:

   docker build -t vector-sets-browser .

2. Run the Docker container:

   docker run -p 3000:3000 vector-sets-browser

3. Open your browser and navigate to http://localhost:3000. If Redis is running at localhost, you'll need to configure the application to connecto to host.docker.internal:6379.

Using a .env File

If you need to configure environment variables, you can create a .env file based on the provided .env.example. This file can be used to set variables such as NEXT_PUBLIC_OLLAMA_URL. It is important if Ollama cannot be reached directly from docker using the default localhost address.

Running locally

1. Install dependencies:

npm install

2. Configure environment variables:

   • Create a .env file in the root directory
   • Add your OpenAI API key: OPENAI_API_KEY=your_api_key_here

3. Start the development server:

npm run dev

4. Open your browser and navigate to http://localhost:3000

Features

• Interactive Visualization: Real-time visualization of vector embeddings with multiple layout algorithms
• Multiple Projection Methods:
  • UMAP for preserving both local and global structure
  • PCA for linear dimensionality reduction
  • Force-directed layout for graph visualization
• AI-Assisted CSV Import: Automatically generate optimal templates for CSV imports using OpenAI
• Modern UI: Built with modern React components and Tailwind CSS
• Real-time Updates: Live visualization updates as vector sets change
• Flexible Integration: Works with any vector embeddings stored in Redis vector-sets

Technology Stack

• Next.js 14
• React 18
• Three.js for visualization
• Transformers.js (for built-in embedding models)
• Redis vector sets
• Tailwind CSS for styling
• Various data visualization libraries (D3.js, UMAP, PCA)

Redis Vector Sets Integration

This browser requires a Redis instance running with vector sets. Vector sets provide high-performance vector similarity search capabilities. Make sure you have the latest version installed as it includes important features like:

• Proper node deletion with relinking
• 8-bit and binary quantization
• Threaded queries
• Filtered search with predicate callbacks

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

License

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