README

About

The project allows you to create playlists in Spotify from a CSV file. It can be used with this frontend app.

Installation

Execute make docker-build in order to build a Docker image with the binary. Running make docker-run will create a Docker container and the application can be reached at port 8000 by default. If the client application is a website which will run on a different port make sure that ALLOWED_ORIGINS environment variable is updated with the website origin (for CORS reasons).

Settings

The complete list of possible environment variables can be found in .env.example file. The file should be renamed to .env so that all the variables are automatically loaded in the application environment.
You should set MARKET environment variable to the country tied to your Spotify account.
TEST_REFRESH_TOKEN is only required for running tests.
SPOTIFY_CLIENT_ID_SECRET_BASE64 is of form <base64 encoded client_id:client_secret>. You can read more about Spotify authorization here.

Project Overview

The original inspiration for the project was to copy Apple Music playlists to Spotify. This can be done by exporting Itunes playlist as a CSV file (Itunes 12.8) as follows: Itunes -> File -> Library -> Export Playlist. As a result any CSV file formatted to the project-specific format can be uploaded to become a Spotify playlist. I'm aware that there're some solutions on the Internet which can perform the task but most of them are paid and I wasn't satisfied with the one I tried.

The CSV format is:

The header row with the first column containing song name, the second column containing artist name.
The file must comma-separated.

The application requires some Spotify user data, most importantly refresh token in order to perform track lookups and add them to user playlist. The application doesn't collect user email.

Due to Spotify API rate limiting all playlist tracks can't be looked up at once, instead they're looked up in batches and there's TRACK_LOOKUP_INTERVAL seconds in between each batch lookup (TRACK_LOOKUP_INTERVAL is an environment variable, by default it's 5 seconds).

The application also has a websocket server which updates client websockets with lookup progress: how many tracks have been found/not found.

The application uses Kafka for messaging between application modules. Inside kafka folder there's a docker-compose.yml in order to start a kafka service locally. Personally, I was using cloudkarafka managed Kafka service which has a free tier.

Roadmap