Collection is smaller after import into beets

[vext01]

Hi,

Loving the idea of beets, but there is one thing that stops me using it for real.

I ran beets over my collection in copy mode (thereby keeping my old collection as-is). I then diffed the beets collection with my existing collection and found that the beets collection was a few GB smaller.

Why would this be? Is there a way to find out?

Cheers

[sampsyo]

Perhaps you had other files (images, etc., or other media types like videos) that were left behind? Or you skipped some albums, not copying the data?

You could consider using something like diff -r to see which files are present in one and not the other. Let us know if you figure out what's being left behind and you think that's in error.

[vext01]

Well diffing wont work since beets will rename the files ;)

[sampsyo]

Indeed. Not sure what to recommend. Maybe you could list out the metadata using beets and some other tool, diff that, and manually remove cases where beets corrected the metadata?

Or if your directory structure is similar before and after the import, you could use du -s * to get the sizes of each directory and look for large discrepancies.

[vext01]

What does beets currently do with anything it can't process? I am guessing, nothing?

If so, can I recommend that it be possible to copy these items into a path configured in the config file? Like you say, I reckon there is some junk that beets can't figure out what to do with.

[sampsyo]

Yes, it ignores non-music files. See #111 for discussion of attachments, a hypothetical feature where we handle those.

[vext01]

I notice that it imports mp4. Is that expected?

A decent stop gap would be an entry in the log file saying what could not be imported?

[sampsyo]

Yes. MPEG-4 is a supported format (see the docs).

If you want details about which files were detected as non-music, you might consider enabling verbose mode (beet -v import ...). That can lead to a lot of output, of course.

[vext01]

OK, I think I see what has happened. There are several GBs of non-music in my music dir. All kinds of stuff from album art to PDF freebies and LSDJ SAV files. There are also some zip files which I forgot to remove after extracting.

I found these with:

find . ! -name '*.mp3' ! -name '*.ogg' ! -name '*.wav' ! -name '*.flac' ! -type d

I think this is therefore an extension of bug #111. It would be nice to just copy unprocessed files into the same directory as the other music files they were with. In the case of my collection, that would be the album's directory.

[sampsyo]

Yep! That's exactly attachments. Stay tuned; this is an oft-requested feature that we should address before long.

[vext01]

Thanks for your time.