diversify 0.1.0

Playlist generator based on artificial intelligence

Diversify

The Project

Diversify is a playlist generator based on the Spofity WEB API and the Spotipy module that aims to use concepts of AI to suggest playlists based on musical preferences between multiple people.

This project was heavily inspired by this article by @ericbonfadini and it's basically just an extra functionality in his analysis.

Goals

The goal is to use AI algorithms to generate a spotify playlist based on a user's preference and a friend of his choice. Currently the script will use a genetic algorthm to generate the playlists but this may improve in the future.

How to run

• First you need to get your spotify API key and save it to the .env file.
  • Go to spotify application web page,
  • login with your spotify account and create a new app
  • put whatever name you'd like on the project info and say no to commercial integration
  • get your client ID and client secret (by clicking show client secret)
  • put them on your .env.example file and rename it to .env
  • click on edit settings and whitelist https://edujtm.github.io/diversify/redirect

These steps are annoying but are needed because I didn't deploy this app somewhere yet, I have plans to deploy it once I make it faster.

• Clone this repo: git clone https://github.com/edujtm/diversify.git

• Create a new environment with your package manager or install the dependencies in environment.yml with pip.
  Personally I use anaconda, so it's just a matter of running conda env create --file=environmnent.yml and then conda activate diversify

• With the previous step done, you can import the interfacespfy module in an interactive prompt

from diversify.session import SpotifySession
spfy = SpotifySession()

playlists = spfy.get_user_playlists("other_username")
saved_songs = spfy.get_favorite_songs()

• You can generate playlists by running python3 diversify/main.py. (run python3 diversify/main.py --help for more info)
  Example: python3 diversify/main.py your_id -u2 your_friend_id -p my awesome playlist

Warnings

The files in the tf/ folder and splearn.py are not used, I was trying to implement the algorithm using neural networks or clustering but i didn't finish. I'll probably remove them in the future.

I didn't run much tests when making this app, so there might be errors that I didn't check for when getting data from the spofity API. You can expect some errors when running the scripts.