Song Similarity App(AI/ML)

Images of the App 📝

References

• Chat GPT
• V0(by vercel)
• YouTube
• Shadcn documentation
• GeekForGeeks
• MillionSongDataset
• Google Developer Console

Overview

This project is an Ai/Ml - web based project where one can find similar songs based on the track/song they uploaded. File can be of the format of .mp3/.wav. Based on that top 5 results will be shown, along with their youtube link if available(api taken from YouTube Data API v3).

Check the video for the brief of the project without running here -> Link ⭐

Table of Contents

• Features
• Technologies Used
• Installation
• Usage
• How to Run

Features

• Upload: Users can uplod their track(.mp3/.wav) format.
• View: Users can see the top 5 similar track along with the youtube link(if available).
• Responsive Design: Optimized for various screen sizes, providing a seamless experience on both mobile and desktop devices.

Technologies Used

• Frontend: React.js+vite
• Backend: python
• Api: Google's (for youtube api)
• Styling: Shadcn Ui and custom Tailwind CSS for responsive design
• Version Control: Git & GitHub

Installation

Prerequisites

Ensure you have the following installed on your machine:

• Node.js (v14 or later)
• npm (Node Package Manager)
• Python(v3)

Step 1: Clone the Repository

git clone https://github.com/Neel-max-cpu/Aiml-project-Finding-similar-song.git

Step 2: Navigate to the Project Directory

Change into the project directory:

cd frontend

cd backend

Step 3: Install Dependencies

Run the following command to install the necessary dependencies for the frontend:

npm install

For the backend you need to check and install the dependencies directly or can do it from requirements.txt

pip install -r requirements.txt

Step 4. Download Dataset

Important to, download the dataset you can go to MillionSongData's website Link here. I have used the million song subset(1.8gb) the full data set if of 280gb. Download it and keep the file in the public folder of the backend.

Step 5: Set Up the Environment Variables

Create the .env file, just paste your api(can be fetched from here Link). Create an account and enable YouTube Data API v3 and create an api key. create an .env file and write your youtube api key

VITE_API_KEY="your api key"

Links to the gpt prompt if you would like to read

• Here Link1
• Here Link2

Usage

After setting up the environment variables, you can start the application.

How to Run

Step 1: Start the Backend Server

Before running the backend build/precompute your faiss index. Navigate to the corefolder then build_faiss_index.py run that command or directly run the command from the backend folder, wait for it as it can take sometime!

cd backend\app\core
python build_faiss_index.py

or

python app/core/build_faiss_index.py

Navigate to the server directory and start the backend server, set it to virtual environment and start the server:

cd backend
.\venv\Scripts\activate
uvicorn app.main:app --reload

Step 2: Start the Frontend

In a new terminal window, navigate to the client directory and start the React application:

cd frontend
npm run dev

Step 3: Access the Application

Open your web browser and go to http://localhost:5173/ to access the App.