Do some caching to speed up subsequent runs

[artyuum]

The following steps can take very long depending on the library's size:

1. fetching every saved track (long)
2. fetching every saved album along with their tracks (longer)

We can imagine caching these data into a small & simple SQLite database. Regarding the second step, we will still need to query the Spotify API to get a list of the saved albums, but only their tracks will be cached (this should drastically speed up this step).

The end-user will be able to pass the --no-cache option to disable caching.