Last updated: 2023-10-04 16:15:38 +0100
Upstream URL: git clone http://chriswarbo.net/git/music-scripts.git
Contents of README follows
This repo contains a bunch of scripts for fetching, tagging and organising a music collection. It makes lots of assumptions about the directory structure, mount points, etc. so it’s probably more useful for copy/paste than for download/execute.
Each script might be useful on its own, but a few useful “workflows” have emerged.
NOTE: You should never need to run a script directly from this repo.
Instead, use nix-build to create a “bin” directory and run
them from there; that will ensure their dependencies are wired-in.
Most scripts deal with “fixing” one aspect or another of a music collection, e.g. finding potential duplicates, moving tracks into album-appropriate directories, spotting non-music cruft, etc.
The main entry point for coordinating all of this is
fix-music. This should be run with ~/Public as
the current working directory (i.e. the /opt/shared
directory of the raspberry pi). The only “actions” this script will take
are to download metadata and cache CRCs and AcoustIDs. Everything else
is simply printed to stdout; it is up to us whether we want to follow
those recommendations.
The main script for this is get_album, which will
download and tag the content of a YouTube playlist. This performs a lot
of I/O, to download, then extract audio, then add artist tags, then add
album tags; so it’s recommended to do it locally, in
~/Downloads/Music, rather than over SSHFS or such.
Alternatively, to get a single track from YouTube or other site, we
can use yt-audio, which is a simple wrapper around
youtube-dl with sensible defaults.
Once we’ve got a bunch of music, we should tag it using something
like Picard; the artist and album tags added by get_album
are often sufficient to make the MusicBrainz lookup work.
Tags should be saved in files (Options -> Save Tags), and files should be renamed (Options -> Rename Files) to the following format:
$if($gt(%totaldiscs%,1),%discnumber%-,)$num(%tracknumber%,2) %title%For example 1-04 My Favourite Song.opus where
1- is the disc number (iff there is more than one),
04 is the track number, My Favourite Song is
the track title and opus is the audio format.
We do not currently use Picard to move files (i.e. we do not use Options -> Move Files).
Once we have fixed up these files locally, we can copy them over the
the raspberry pi using the move_downloaded_music
script.