Apply actions not definitive?

[n1els]

I'm building my library with beets at the moment, and while playing around with various plugins to get more metadata and so on I noticed that every time I run

beet import -s -L -p

There are a number of songs that keep asking me for confirmation no matter how many times I confirm that I want to apply or take any other kind of action.

/media/music/Library/House/&ME - Garden - 122 320kbps 44kHz.mp3 Tagging track: &ME - Garden URL: https://open.spotify.com/track/00cJqEccK1XmxioVYcbQPs (Similarity: 88.0%) (index, source) (Spotify) [A]pply, More candidates, Skip, Use as-is, Enter search, enter Id, aBort, Print tracks, eDit, edit Candidates?

In the example above, the match on spotify is correct. I am not sure why it only scores 88% similarity, but I would like to commit to this match and not have beets ask me again about this file, unless something about it has changed.

I'm running repeated imports and beet update commands, as well as others while getting genre identification and tagging to work correctly so I am not sure if I am inadvertently triggering beets to keep ask me about the same songs every time.

When I run exactly the same command after it finishes the first time, it starts immediately by asking me about the same song again.

[arsaboo]

I'm not sure what you are trying to do here, but your flags do not look right. -s is for singleton imports, which should only be used as a last resort (you should normally import albums); -L reimports from the library; and -p is for pretend, so nothing gets written.

If you are just trying to import your library, start with beet import. To get the confirmation prompt, add the -t flag (for timid).

[n1els]

Thanks for the quick response, I have just started out with beets so it is very likely that I am doing multiple things wrong. I have mostly single tracks as opposed to albums and beet help import shows that -p is the short form of the --resume flag (although I have seen -p used to mean pretend mode in other commands), see output and version below.

I need to use resume because gstreamer, which is used by bpmanalyser and replaygain, keeps crashing after about a hundred or so tracks. I'm using this library as a base library for DJ'ing, so right now I have just shy of 7K tracks in either MP3, Flac or AIFF format and they are almost exlusively single tracks.

My goal is, after the first large batch import is complete, to have an input directory with music. The contents of this directory will then be moved to the library during import. The genre is part of the file structure, and bpm and quality data is part of the file name. It seems to me therefore that after the initial import a second move or update or import -L cycle is needed in order to reflect the newly discovered metadata in the path and filenames.

config.yaml:

directory: /media/music/Library
library: /media/music/musiclibrary.db
plugins: duplicates embedart fetchart inline convert web chroma fromfilename info lyrics replaygain describe discogs bpmanalyser edit spotify bandcamp mbsubmit lastgenre beatport4

edit:
  itemfields: track title artist album genre
  albumfields: album albumartist

chroma:
  auto: yes

discogs:
  source_weight: 0.5
  index_tracks: yes

spotify:
    source_weight: 0.4
    mode: list
    show_failures: on
    tiebreak: first

fromfilename:
  source_weight: 0.5

beatport4:
    art: yes
    art_overwrite: no
    username: REDACTED
    password: REDACTED
bpmanalyser:
  auto: yes
  dry-run: no
  write: yes
  threads: 2
  force: yes
  quiet: no

autobpm:
  auto: yes
  overwrite: yes

bandcamp:
  include_digital_only_tracks: true
  search_max: 4
  art: yes
  comments_separator: "\n---\n"
  exclude_extra_fields: []
  genre:
    capitalize: yes
    maximum: 0
    always_include: []
    mode: progressive

lastgenre:
  auto: no
  canonical: yes
  count: 1
  fallback: ''
  force: yes
  min_weight: 10
  prefer_specific: no
  source: track
  whitelist: /media/music/genre.whitelist
  title_case: yes

duplicates:
  album: no
  move: /media/music/duplicates
  count: yes
  path: yes
  checksum: ffmpeg -i {file} -f crc -

mbsubmit:
  format: $track. $title - $artist ($length)
  threshold: medium

embedart:
  auto: yes

fetchart:
  auto: yes

lyrics:
  auto: yes

replaygain:
  backend: gstreamer
  auto: yes

web:
  host: 0.0.0.0
  port: 8337

import:
  write: yes
  move: yes
  resume: yes
  log: /media/music/beets.log
  languages: en nl
  bell: yes
  quiet_fallback: asis

match:
  strong_rec_thresh: 0.06

paths:
  default: $genre/$artist - $title - $bpm $bitrate $samplerate
  singleton: $genre/$artist - $title - $bpm $bitrate $samplerate
  comp: $genre/$artist - $title - $bpm - $bitrate $samplerate

music@music:/media/music/Library$ beet help import
Usage: beet import [options]

Options:
  -h, --help            show this help message and exit
[...]
  **-p, --resume          resume importing if interrupted**
  -P, --noresume        do not try to resume importing
  -q, --quiet           never prompt for input: skip albums instead
[...]
  -s, --singletons      import individual tracks instead of full albums

[...]
  --pretend             just print the files to import
[...]

music@music:/media/music/Library$ beet version
beets version 1.6.0
Python version 3.10.12
plugins: bandcamp, beatport4, bpmanalyser, chroma, convert, describe, discogs, duplicates, edit, embedart, fetchart, fromfilename, info, inline, lastgenre, lyrics, mbsubmit, replaygain, spotify, web

[arsaboo]

You are using a relatively old version. Please update beets - pip install --upgrade --force-reinstall --no-deps git+https://github.com/beetbox/beets.git to update to the latest version.

My recommendation would be to import the library first before you run the bpmanalyser and replaygain plugins. Once the library is imported, everything is so much easier. BTW...spotify works reasonably well with singletons. But, in general, beets works a lot better with separate folders for albums (even if each folder has only one track).

[n1els]

Ok, thanks. will try to update beets even though I installed it just a few days ago so unsure how it pulled an old version. I would not mind an album file structure for singles as the end result, but my input directory is one big chunk of unrelated files so not sure if beets would be able to make albums out of an unsorted list.

Will test with moving some of the optional plugins to a secondary step outside of the import process as well.

[arsaboo]

If you use Spotify tagger, it will update the album name. Then, you can add $album to the path for singletons, and you will end up with the right folder structure.

[n1els]

Oh that's a nice way to fix it. I have beets 1.6.1 installed now through the command you posted. I'm still not sure how I got 1.6.0 a couple of days ago by following the readthedocs instructions, and github still seems to show that version first. I currently have Spotify, beatport4, discogs and bandcamp enabled but once a song gets tagged it'll set the ID for one of these (whichever was first) and then on subsequent runs it'll send Spotify ID's to bandcamp or other bad combinations, which seem to effectively disable their purpose. Is that correct or can it still benefit from having multiple sources for Metadata? This might be fixed in the new version by the way, I'm still running the new import.

…

On Wed, Mar 13, 2024, 01:54 Alok Saboo ***@***.***> wrote: If you use Spotify tagger, it will update the album name. Then, you can add $album to the path for singletons, and you will end up with the right folder structure. — Reply to this email directly, view it on GitHub <#5143 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABU7TCC4SREU6ZUXJNQADHDYX46N3AVCNFSM6AAAAABER5OJQKVHI2DSMVQWIX3LMV43SRDJONRXK43TNFXW4Q3PNVWWK3TUHM4DONRTGQYTA> . You are receiving this because you authored the thread.Message ID: ***@***.***>