Modes

The same string can mean different things depending on where it came from. Blink-182-All The Small Things is clearly artist and title on YouTube. In a music library, a tight dash like that is more likely part of a name. Modes tell trackparse how much guessing the source can take.

Mode	For	What it adds
clean	store metadata, ID3 tags	Only a spaced dash separates artist and title
youtube	video titles, scrobbles	Pipe segments, "Title" by Artist, Artist: "Title", unspaced dashes, trailing junk, the uploader fallback
filename	files on disk	Strips the extension, reads _ as a space, accepts 01-Track numbering
tracklist	one line of a tracklist	The first spaced dash always separates, and the words after it are never read as a version or genre suffix
auto	anything (the default)	Picks filename, youtube or clean from the string. It never picks tracklist

The resolved mode is in track.mode, so you can always see which one ran.

How auto decides

1. A known audio or video extension at the end (.mp3, .flac, .wav, .mp4…) means filename.
2. Junk, a |, or a platform suffix like - YouTube means youtube.
3. Anything else is clean.

JavaScript

parse("07_Burial_-_Archangel.mp3");
// mode 'filename', position 7, artist 'Burial', title 'Archangel'

parse("Noisia - Stigma | Official Video");
// mode 'youtube', title 'Stigma', junk [ { raw: 'Official Video', kind: 'video' } ]

Python

parse("07_Burial_-_Archangel.mp3")
# mode 'filename', position 7, artist 'Burial', title 'Archangel'

parse("Noisia - Stigma | Official Video")
# mode 'youtube', title 'Stigma', junk (Junk(raw='Official Video', kind='video'),)

youtube mode

Turn it on explicitly when you know the strings are video titles. It takes guesses that clean won't, and every guess shows up in warnings:

JavaScript

parse('"Numb" by Linkin Park', { mode: "youtube" });
// artists [ 'Linkin Park' ], title 'Numb', warnings [ 'bySplit' ]

parse('Linkin Park: "Numb"', { mode: "youtube" });
// artists [ 'Linkin Park' ], title 'Numb', warnings [ 'quotedTitleSplit' ]

parse("Blink-182-All The Small Things", { mode: "youtube" });
// artists [ 'Blink-182' ], title 'All The Small Things', warnings [ 'unspacedDashSplit' ]

Python

parse('"Numb" by Linkin Park', mode="youtube")
# artists ['Linkin Park'], title 'Numb', warnings ('bySplit',)

parse('Linkin Park: "Numb"', mode="youtube")
# artists ['Linkin Park'], title 'Numb', warnings ('quotedTitleSplit',)

parse("Blink-182-All The Small Things", mode="youtube")
# artists ['Blink-182'], title 'All The Small Things', warnings ('unspacedDashSplit',)

In clean mode the last one comes back with no artist, the whole string as the title and a noSeparator warning. That's the safe answer when you can't be sure.

The uploader fallback

Many music videos are titled with just the song name, and the artist is the channel. Pass the channel as uploader and youtube mode uses it when there's no separator. Auto-generated suffixes like - Topic and VEVO are dropped:

JavaScript

parse("Around the World", { mode: "youtube", uploader: "Daft Punk - Topic" });
// artists [ 'Daft Punk' ], title 'Around the World'

Python

parse("Around the World", mode="youtube", uploader="Daft Punk - Topic")
# artists ['Daft Punk'], title 'Around the World'

filename mode

JavaScript

parse("07_Burial_-_Archangel.mp3");                    // position 7, Burial / Archangel
parse("311 - Amber.mp3").position;                     // { raw: '311', number: 311 }
parse("311 - Amber").artists[0].name;                  // '311'

Python

parse("07_Burial_-_Archangel.mp3")                     # position 7, Burial / Archangel
parse("311 - Amber.mp3").position                      # Position(raw='311', number=311)
parse("311 - Amber").artists[0].name                   # '311'

The last two show why the mode matters. In a file name a leading number with a dash is a track number. Anywhere else, 311 could be a band.

tracklist mode

Inside a tracklist almost every line is Artist - Title, so the first spaced dash can be trusted. A single string doesn't get that benefit of the doubt:

JavaScript

parse("Electrosoul System - Liquid Metal");
// title 'Electrosoul System', junk [ { raw: 'Liquid Metal', kind: 'genre' } ]

parse("Electrosoul System - Liquid Metal", { mode: "tracklist" });
// artists [ 'Electrosoul System' ], title 'Liquid Metal'

Python

parse("Electrosoul System - Liquid Metal")
# title 'Electrosoul System', junk (Junk(raw='Liquid Metal', kind='genre'),)

parse("Electrosoul System - Liquid Metal", mode="tracklist")
# artists ['Electrosoul System'], title 'Liquid Metal'

parseTracklist (parse_tracklist) uses this mode for every line. You only need to set it yourself when you parse tracklist lines one at a time.