How to import large pile of songs from single directory

I’m a bit surprised at how much hand holding beets needs to import an old collection of random mp3s. I’m being asked questions about every single file which means it will take me days to import the files. I’m running beet import --group-albums /music-dump and I’ve enabled plugins: discogs fromfilename.

Is there not a way to at least move the files beets is sure about? So that I can have a better feeling about how many files are causing beets problems. Or should I just type t for every single question and let beets dump the files into an artist folder and not bother with albums?

Is this just the way it works with beets? Also I messed up a few of the early files I imported because I didn’t understand what I was doing. Is there some way to make beets rescan the files that have been imported and highlight any it thinks could be wrong?

OK so --group-albums was the wrong flag to use for my case. I don’t have albums, it’s just a big pile of singles/individual files at best I may have had 3 or 4 songs from an album. So I’ve realized that I should dump the albums approach given that’s not how I consumed or will consume music in future. I found the -s -q flags in the import documentation and they seem to be doing a much better job.

My only disappointment is that the filename plugin really doesn’t seem to be doing anything, I see files that are correctly named but the import skips them because I guess there is no metadata in them even though the filename provides all the information you’d need. Does that plugin work? The documentation says it works without configuration.

Great! Yes, singletons are what you want.

One important fact about fromfilename is that it only triggers when tracks seem to be missing embedded metadata (in which case the pipeline just uses that metadata instead). It does work, though!

@BrianDiPalma Two additional thoughts:

  1. Have you tried MusicBrainz Picard? It’s Scan feature is super useful—you can throw a big directory of singles at it and it will typically find matches for many of them. You can configure Picard to move the tracks to a different folder upon successful tagging. Then you can import the successfully tagged files into beets easily, since the metadata will be higher quality.

  2. Here are some useful beets configuration options for matching:

        countries: ['US', 'GB|UK']
        media: ['CD', 'Digital Media|File']
        original_year: yes        
    #  strong_rec_thresh: 0.50
    #  ignored: unmatched_tracks
    #  ignored: missing_tracks unmatched_tracks                                                                       
    #  max_rec:                                                                                                                                                                                                                          
    #    missing_tracks: medium

    The preferred section is nice to give beets a hint which releases you want it to prefer. It is described here in the documentation. The strong_rec_thresh, described here, tunes the threshold for when beets prompts to confirm a track versus importing it automatically as a “strong recommendation.” The ignored and max_rec options may also be useful, and are also described in the same docs.

I haven’t tried Picard, I thought the outcome would be similar as beets uses the same metadata source. I do want to try something different though because importing everything as singles with beets isn’t stable. When I reimport most of the same files beets asked me to confirm before come back up again which I find surprising. I figured once I’d selected an id and everything else in the song matched that it would be stable.

Thank you for your match suggestions, they look good. I might use them and import as albums in the hope that it makes the import process more stable/certain of what each song is.