PitchRuby

A Ruby script for parsing Pitchfork's best new albums RSS feed and adding songs to an Rdio playlist.

Originally inspired by bawish's Rdiofork.

Data structures

Since this was in part an excuse for me to learn redis, I tried to plan out storing data based not only on what information I want to keep, but on how I wanted to query it later.

Key => value to keep track of unique ids for each album: albums:nextID

We can get the next available ID with redis.incr('albums:nextID')
Example: albums:nextID = 1

Most importantly, I'm storing a redis hash of albums: albums:#{id} where id is from albums:nextID

We want to store the id, the artist, and the album title
Example: albums:1 = { "id" => 1, "artist" => "Purity Ring", "title" => "Shrines" }

Storage based on queries

What was the most recently parsed RSS post?

This is just a simple key => value which stores the unix timestamp of the latest post as the key albums:latest_post.

Example: albums:latest_post = 1361822312

Do we already have an entry for this artist + album?

Redis key => value, where the key is albums:artist:title and the value is the id of the album in the albums:id hash

Example: albums:Purity Ring:Shrines = 1

Is this album available for streaming on rdio?

Redis "set" of unavailable albums: albums:unavailable