Skip to content

README.md

Latest commit

 

History

History
165 lines (122 loc) · 9.05 KB

README.md

File metadata and controls

165 lines (122 loc) · 9.05 KB
Raw

MelodyMind: Hybrid Music Recommendation System 🎧

Welcome to the MelodyMind project repository!
This repository showcases my work in building music recommendation systems, evolving from a real-time API-based recommender to a sophisticated hybrid model.

🚀 Project 1: Last.fm API-Based Music Recommender (Live Demo)

👉 Live Demo

Powered by Last.fm Streamlit Python License: MIT

This application is your personal guide to discovering new music, powered by the extensive Last.fm API. It offers intelligent and diverse recommendations through an intuitive Streamlit interface.

✨ Project Overview

MelodyMind isn't just another music recommender; it's a hybrid system that intelligently blends various data points from Last.fm to unearth unique music tailored to your taste. Whether you provide an artist, a specific track, or both, MelodyMind digs deep to find your next favorite song. Its hybrid approach ensures you get a broad spectrum of suggestions, moving beyond simple direct similarities to explore connected artists and genres.

🌟 Key Features

  • Intelligent Input Handling: Smartly processes your artist and track inputs, with built-in auto-correction for common typos, ensuring you always find what you're looking for.
  • Diverse Recommendation Strategies:
    • Track-Focused: Discover songs similar to a given track, expanding your playlist effortlessly.
    • Artist-Centric: Explore the top hits from your favorite artists or discover new tracks from artists you'll love.
    • Contextual Hybridization: When direct matches are scarce, the system intelligently broadens its search to include top tracks from similar artists and songs associated with relevant genres, providing richer recommendations.
  • Comprehensive Insights: Dive deeper into the music with detailed information about recommended artists and tracks, including their top tags and concise summaries.
  • Sleek User Interface: Enjoy a smooth and responsive experience thanks to the interactive Streamlit web application.
  • Instant Access: Direct links to Last.fm enable you to listen to recommended tracks immediately.

💻 Technology Stack

  • Python: The robust foundation for all logic and data processing.
  • Streamlit: For creating a beautiful, interactive, and easy-to-use web application.
  • Requests: Handles all HTTP communications with the Last.fm API.
  • Last.fm API: The rich data source powering the music recommendations.

⚙️ Getting Started (Last.fm API App)

To run this version of MelodyMind on your local machine, follow these simple steps!

Prerequisites

  • Python 3.12 (highly recommended for best compatibility)
  • Git
  • A Last.fm API Key – essential for accessing music data.

Local Installation

  1. Clone the Repository: 
    git clone [https://github.com/indranil143/Hybrid-Music-Recommendation-System.git](https://github.com/indranil143/Hybrid-Music-Recommendation-System.git)
cd Hybrid-Music-Recommendation-System
  2. Create a Virtual Environment (Highly Recommended): 
    python -m venv venv
# On Windows:
.\venv\Scripts\activate
# On macOS/Linux:
source venv/bin/activate
  3. Install Dependencies: Ensure your requirements.txt file is present in the current directory. 
    pip install -r requirements.txt
  4. Set Your Last.fm API Key: The application needs your Last.fm API Key to function. Set it as an environment variable:
    • For Windows (Command Prompt): 
      set LASTFM_API_KEY="YOUR_ACTUAL_LASTFM_API_KEY"
    • For Windows (PowerShell): 
      $env:LASTFM_API_KEY="YOUR_ACTUAL_LASTFM_API_KEY"
    • For macOS/Linux (Bash/Zsh): 
      export LASTFM_API_KEY="YOUR_ACTUAL_LASTFM_API_KEY"
    • Remember to replace YOUR_ACTUAL_LASTFM_API_KEY with the key you obtained from Last.fm. For persistent local development, consider using a .env file with the python-dotenv package.
  5. Run the Streamlit App: 
    streamlit run music_app_api_only.py
    Your web browser should automatically open the MelodyMind application.

☁️ Deployment to Streamlit Community Cloud

Deploying MelodyMind to Streamlit Community Cloud is straightforward:

  1. GitHub Repository: Your project must be hosted on a public GitHub repository.
  2. requirements.txt: Ensure this file is located in the same directory as music_app_api_only.py. It should accurately list all project dependencies and avoid platform-specific packages (like pywin32) or problematic +cpu PyTorch/Torchaudio versions, as these can cause deployment issues.
  3. API Key as Secret: Securely store your LASTFM_API_KEY in the Streamlit Community Cloud dashboard's secrets. Go to your app settings, then "Advanced settings," and finally "Secrets," adding the following: 
    LASTFM_API_KEY="YOUR_ACTUAL_LASTFM_API_KEY_HERE"

🎮 How to Use

  • Simply enter an artist name, a track name, or both in the sidebar on the left.
  • Adjust the "Number of Recommendations" slider to control the quantity of suggestions.
  • Then click "Get Recommendations."
  • Explore the results under the "🎵 Recommendations" tab, and dive into deeper context with the "📊 Insights" tab.

🎶 Project 2: Hybrid Music Recommendation System (LightFM)

This project implements a Hybrid Music Recommendation System combining Content-Based Filtering and Collaborative Filtering using the LightFM library. It provides personalized song recommendations by analyzing audio features and simulating user interactions (artist-track associations).

Python LightFM Streamlit License: MIT

✨ Features

  • Hybrid Approach: Combines Content-Based and Collaborative Filtering.
  • Content-Based Filtering: Recommends songs based on audio features (danceability, tempo, energy, etc.) using cosine similarity.
  • Collaborative Filtering: Utilizes the LightFM library to learn user (artist) and item (track) embeddings from interaction data.
  • Simulated User Interactions: Demonstrates collaborative filtering using artist-track associations as implicit feedback.
  • Streamlit Web App: Provides an interactive web interface to get recommendations.
  • Modular Design: Separated data processing and deployment (music_app.py).

💡 Concepts Explained

  • Content-Based Filtering: Imagine recommending songs that sound similar to what you already like. This method analyzes the characteristics (features) of items (songs) and suggests others with similar characteristics.
  • Collaborative Filtering: This method looks at the behavior of many users. If User A and User B like similar songs, and User A likes a song that User B hasn't heard, Collaborative Filtering might recommend that song to User B. It finds patterns in user-item interactions.
  • Hybrid System: Combines both methods to recommend new, niche songs (Content-Based) while leveraging user patterns (Collaborative Filtering) for enhanced accuracy.

Dataset

Uses the Spotify Dataset 1921-2020, 160k+ Tracks. Source: https://www.kaggle.com/datasets/fcpercival/160k-spotify-songs-sorted

⚙️ Getting Started (LightFM Hybrid App)

Running the Notebook

Execute the cells in Hybrid_Music_Recommendation_System.ipynb using Kaggle or Google Colab. This trains the model and saves music_recommender_components.pkl.

Running the Streamlit App

  1. Ensure music_recommender_components.pkl is in the root directory.
  2. Activate your virtual environment (from the previous setup steps, or create a new one).
  3. Run the app: 
    streamlit run music_app.py
    The app will open in your browser.

🤝 Contributing

Feel free to contribute to this project by opening issues or submitting pull requests!

📄 License

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