LavaRand Powered Spotify Shuffle

Overview

Inspired by Cloudflare's LavaRand, this project aims to improve Spotify's shuffle feature by leveraging true randomness derived from the chaotic motion of lava lamps.

Motivation

Entirely a gimmick/fun side project. Since Spotify's shuffling algorithm often feels stale/monotonous, I wanted to create a process that would unequivocally ensure the most random possible shuffle.

Key Components

• YOLOv8: Trained model, detects and tracks lava lamp motion.
• SHA-256: Hashes keypoints to produce random numbers.
• Spotify API: Queues songs based on generated random indices.
• tkinter: Provides a user-friendly GUI for interaction.

Installation

1. Clone the repository:

   git clone https://github.com/rawcsav/LavaShuf.git
   cd LavaShuf

2. Install dependencies:

   pip install -r requirements.txt

3. Set up Spotify API credentials:

   • Create a .env file with your Spotify client ID, client secret, and redirect URI:

   CLIENT_ID=your_spotify_client_id
   CLIENT_SECRET=your_spotify_client_secret
   REDIRECT_URI=your_spotify_redirect_uri
   

Usage

1. Run the application:

   python main.py

2. Interact with the GUI:

   • Hit the Start button to begin the process, it should open a new window with the camera feed.
     • I had this setup to work with my phone's camera, so you may need to adjust the camera index in the code.
   • Once you have the camera aligned with the lava lamp(s), and you can see that it's accurately detecting with boundary boxes, press y to start the process.
     • The camera feed is strictly for visualization and getting everything aligned properly. Due to threading issues with OpenCV & MacOS, it will need to close this camera feed while actually gathering the images and calculating indices
   • Select a Spotify playlist.
   • Specify the number of songs to queue.
   • The system will then begin capturing lava lamp images and generating random picks
   • View the queued songs & associated blob data for each song queued

Technical Approach

1. Object Detection: Uses YOLOv8 to detect and track lava lamp motion.
2. Keypoint Tracking: Utilizes ORB to track chaotic motion within the detected lava lamp.
3. Random Number Generation: Hashes keypoints using SHA-256 to produce random numbers.
4. Spotify Integration: Uses the Spotify API to queue songs based on generated random indices.

Demonstration

A more exploratory and detailed write-up & demo is available on my website.

License

This project is licensed under Attribution-ShareAlike 4.0 International

Acknowledgements

The dataset used in training, validating, and testing the model was a combination of two pre-existing Roboflow datasets: Lavalamp Dataset 1 & Lavalamp Dataset 2. All props and credit to the owners/creators of these two datasets :)