Aha! Thanks for sending the link. I finally understand.

Here’s the deal: beets tagged those files, but it *didn't change the `album` 
field*. The album field is left exactly how you had it.

The problem is that there’s no album information for beets to apply. In 
MusicBrainz, recordings (i.e., tracks) can appear on multiple albums. So even 
though beets found a match for a given track, it doesn’t necessarily know what 
album it was on. So it doesn’t set the fields `album`, `albumartist`, et 
cetera. You might want to clear those out with a `beet modify`.

Adrian


> On Aug 9, 2016, at 9:41 AM, Adrian Sampson <[email protected]> wrote:
> 
> Yeah, a Dropbox link would work great. And can you also include the config 
> and command we should use to reproduce the problem?
> 
>> On Aug 9, 2016, at 9:26 AM, [email protected] 
>> <mailto:[email protected]> wrote:
>> 
>> If I import a single track, it does what's expected.  beets is confused by 
>> the fact that I have 20 files that are all already tagged and file-located 
>> as though they are all from the same album, which they are not.    
>> 
>> Is there a place I can upload 85M of files for you ?  dropbox or something ?
>> 
>> On Tuesday, August 9, 2016 at 8:28:54 AM UTC-4, Adrian Sampson wrote:
>> My apologies, but I think I no longer understand what’s going on here. Is 
>> there some reproducible test case you can provide, including the file, 
>> starting from an empty library and importing a single track?
>> 
>>> On Aug 8, 2016, at 3:18 PM, cla...@ <>zzzcomputing.com 
>>> <http://zzzcomputing.com/> wrote:
>>> 
>>> The files don't have MBIDs before I start importing.  Here's what we start 
>>> with:
>>> 
>>> /home/classic/test_music/Aarti/Vaishnodevi/01 01HEY GOVIND HEY GOPAL- 
>>> NANAK.mp3
>>> /home/classic/test_music/Aarti/Vaishnodevi/02 02 HARI TUM HARO  JAN KI 
>>> PEER.mp3
>>> /home/classic/test_music/Aarti/Vaishnodevi/03 03 TUM MERI LAKHO RAJ ARI- 
>>> SU.mp3
>>> /home/classic/test_music/Aarti/Vaishnodevi/04 04 DEENAN DUKH HARAN 
>>> DEV-SURD.mp3
>>> /home/classic/test_music/Aarti/Vaishnodevi/05 05 KANHA TORI JOHAT.mp3
>>> /home/classic/test_music/Aarti/Vaishnodevi/05 05 SABSE OONCHI PREM SAGAI- 
>>> S.mp3
>>> /home/classic/test_music/Aarti/Vaishnodevi/06 06 JAI RADHA MADHAV.mp3
>>> /home/classic/test_music/Aarti/Vaishnodevi/06 06 RADHA AISI BHAI SHYAM KI 
>>> D.mp3
>>> /home/classic/test_music/Aarti/Vaishnodevi/07 07 MURLI MANOHAR GOPALA.mp3
>>> /home/classic/test_music/Aarti/Vaishnodevi/07 07 PRABHU JI TUM CHANDHAN 
>>> HUM.mp3
>>> /home/classic/test_music/Aarti/Vaishnodevi/32 Aarti Hanuman Ji Ki.mp3
>>> /home/classic/test_music/Aarti/Vaishnodevi/32 Jai Ganesh Jai Ganesh Deva.mp3
>>> /home/classic/test_music/Aarti/Vaishnodevi/32 Jai Jag Janani Jai Jai.mp3
>>> /home/classic/test_music/Aarti/Vaishnodevi/32 Jai Jagdish Hare 2.mp3
>>> /home/classic/test_music/Aarti/Vaishnodevi/32 Jai Jagdish Hare.mp3
>>> /home/classic/test_music/Aarti/Vaishnodevi/32 Jai Jai Jai Ravi Dev.mp3
>>> /home/classic/test_music/Aarti/Vaishnodevi/32 Jai Jai Jai Tulsi Mata.mp3
>>> /home/classic/test_music/Aarti/Vaishnodevi/32 Jai Laxami Mata.mp3
>>> /home/classic/test_music/Aarti/Vaishnodevi/32 Jai Parvati Mata.mp3 (19 
>>> items)
>>> 
>>> 
>>> Then Ill do an import, lets say I did about five files, many others I have 
>>> to skip because they can't be identified, and also I did "aBort" before i 
>>> got to the end just to show what it's doing.
>>> 
>>> Here's a few of the finds:
>>> 
>>> /home/classic/test_music/Aarti/Vaishnodevi/01 01HEY GOVIND HEY GOPAL- 
>>> NANAK.mp3
>>> Correcting track tags from:
>>>     Aarti - 01HEY GOVIND HEY GOPAL- NANAK
>>> To:
>>>     Jagjit Singh - Hey Gobind Hey Gopal
>>> URL:
>>>     http://musicbrainz.org/recording/f7fc29c0-d7ac-4882-a1bf-e0e2a495a192 
>>> <http://musicbrainz.org/recording/f7fc29c0-d7ac-4882-a1bf-e0e2a495a192>
>>> (Similarity: 79.5%) (artist, title)
>>> [A]pply, More candidates, Skip, Use as-is, Enter search, enter Id, aBort,
>>> eDit, edit Candidates? A
>>> 
>>> /home/classic/test_music/Aarti/Vaishnodevi/02 02 HARI TUM HARO  JAN KI 
>>> PEER.mp3
>>> Correcting track tags from:
>>>     Aarti - 02 HARI TUM HARO  JAN KI PEER
>>> To:
>>>     Jagjit Singh - Hary Tum Haro Jan Ki Peer
>>> URL:
>>>     http://musicbrainz.org/recording/e2f04a67-d8fd-4fea-818e-feb62445c453 
>>> <http://musicbrainz.org/recording/e2f04a67-d8fd-4fea-818e-feb62445c453>
>>> (Similarity: 84.5%) (artist, title)
>>> [A]pply, More candidates, Skip, Use as-is, Enter search, enter Id, aBort,
>>> eDit, edit Candidates? a
>>> 
>>> /home/classic/test_music/Aarti/Vaishnodevi/03 03 TUM MERI LAKHO RAJ ARI- 
>>> SU.mp3
>>> Correcting track tags from:
>>>     Aarti - 03 TUM MERI LAKHO RAJ ARI- SU
>>> To:
>>>     Jagjit Singh - Tum Mery Rako Laaj Hari
>>> URL:
>>>     http://musicbrainz.org/recording/778017ed-0175-4ba9-8c92-96a141badf3c 
>>> <http://musicbrainz.org/recording/778017ed-0175-4ba9-8c92-96a141badf3c>
>>> (Similarity: 76.5%) (artist, title)
>>> [A]pply, More candidates, Skip, Use as-is, Enter search, enter Id, aBort,
>>> eDit, edit Candidates? a
>>> 
>>> /home/classic/test_music/Aarti/Vaishnodevi/04 04 DEENAN DUKH HARAN 
>>> DEV-SURD.mp3
>>> Correcting track tags from:
>>>     Aarti - 04 DEENAN DUKH HARAN DEV-SURD
>>> To:
>>>     Jagjit Singh - Deenan Dukh Haran Dev
>>> URL:
>>>     http://musicbrainz.org/recording/7338031a-57d7-4f90-aab9-47d901da8151 
>>> <http://musicbrainz.org/recording/7338031a-57d7-4f90-aab9-47d901da8151>
>>> (Similarity: 81.6%) (artist, title)
>>> [A]pply, More candidates, Skip, Use as-is, Enter search, enter Id, aBort,
>>> eDit, edit Candidates? a
>>> 
>>> 
>>> these are working great but it is of course very time consuming to get 
>>> through them on track-by-track.
>>> 
>>> Then at the end, here's the end result:
>>> 
>>> [classic@zotac ~]$ .venv/bin/beet list
>>> Aarti - Vaishnodevi - 05 KANHA TORI JOHAT
>>> Aarti - Vaishnodevi - Aarti Hanuman Ji Ki
>>> Anup Jalota - Vaishnodevi - Radha Aisi Bhai
>>> Jagjit Singh - Vaishnodevi - Hey Gobind Hey Gopal
>>> Jagjit Singh - Vaishnodevi - Hary Tum Haro Jan Ki Peer
>>> Jagjit Singh - Vaishnodevi - Tum Mery Rako Laaj Hari
>>> Jagjit Singh - Vaishnodevi - Deenan Dukh Haran Dev
>>> Jagjit Singh - Vaishnodevi - Sabseoonchi Prem Sagai
>>> Jagjit Singh - Vaishnodevi - Jai Radha Madhav
>>> Jagjit Singh - Vaishnodevi - Murli Manohar Gopala
>>> Jagjit Singh & Chitra Singh - Vaishnodevi - Jai Ganesh Deva
>>> 
>>> 
>>> The problem is, *there is no such album Vaishnodevi".    It has ignored the 
>>> actual album that was located for all of those tracks from Jagjit Singh.  
>>> Only "singleton" mode has this problem.  If I run "beet import" against 
>>> each single mp3 file at a time, it doesn't go into singleton mode, and it 
>>> works great.    For example:
>>> 
>>> [classic@zotac ~]$ .venv/bin/beet import test_music/Aarti/Vaishnodevi/03\ 
>>> 03\ TUM\ MERI\ LAKHO\ RAJ\ ARI-\ SU.mp3 
>>> 
>>> /home/classic/test_music/Aarti/Vaishnodevi/03 03 TUM MERI LAKHO RAJ ARI- 
>>> SU.mp3 (1 items)
>>> Finding tags for album "Aarti - Vaishnodevi".
>>> Candidates:
>>> 1. Jagjit Singh - Kare Krishna A Live Concert (26.0%) (missing tracks, 
>>> artist, album, ...)
>>> 
>>> ...
>>> 
>>> now it found it.  beet list now shows:
>>> 
>>> Jagjit Singh - Kare Krishna A Live Concert - Tum Mery Rako Laaj Hari
>>> Jagjit Singh - Vaishnodevi - Hey Gobind Hey Gopal
>>> Jagjit Singh - Vaishnodevi - Hary Tum Haro Jan Ki Peer
>>> Jagjit Singh - Vaishnodevi - Tum Mery Rako Laaj Hari
>>> Jagjit Singh - Vaishnodevi - Deenan Dukh Haran Dev
>>> Jagjit Singh - Vaishnodevi - Sabseoonchi Prem Sagai
>>> Jagjit Singh - Vaishnodevi - Jai Radha Madhav
>>> Jagjit Singh - Vaishnodevi - Murli Manohar Gopala
>>> Jagjit Singh & Chitra Singh - Vaishnodevi - Jai Ganesh Deva
>>> 
>>> 
>>> Now we have that same track, "Tum Mery Rako Laaj Hari", *twice*. One with 
>>> the right album, one with the wrong one.
>>> 
>>> 
>>> 
>>>     I don't seem to be able to do any kind of "reimport" on these tracks 
>>> however, all the documentatinon of "beet import -L" and such doesn't seem 
>>> to do anything, even if I try pointing it at just one track.
>>> 
>>> This is something that beets is doing 99% of the work within, and there is 
>>> one weird behavior right at the end.
>>> 
>>> 
>>> 
>>> 
>>> 
>>> 
>>> 
>>> 
>>> 
>>> On Monday, August 8, 2016 at 1:51:44 PM UTC-4, Adrian Sampson wrote:
>>> Hmm, maybe I don’t quite understand—if you already have MBIDs in those 
>>> files, beets should use them when you import them as singletons. Does it 
>>> not? If not, that’s a bug we should try to reproduce.
>>> 
>>>> On Aug 8, 2016, at 12:12 PM, cla...@ <>zzzcomputing.com 
>>>> <http://zzzcomputing.com/> wrote:
>>>> 
>>>> OK so that's sounding like this is really not supported as a top-level 
>>>> task.
>>>> 
>>>> It seems quite simple:
>>>> 
>>>> 1. figure out the correct artist / album for mp3 file by searching 
>>>> musicbrainz, discogs, etc. (beets does this)
>>>> 
>>>> 2. tag the mp3 file with that artist / album that musicbrainz lists for 
>>>> the track, erasing anything currently there (beets will not do this unless 
>>>> I manually edit each by hand and re-type the name that it already has 
>>>> right there)
>>>> 
>>>> 3. place in the correct directory
>>>> 
>>>> Basically beets won't allow me to do this without manually editing the id3 
>>>> tags, even though beets has already made me go through every track 
>>>> individually and look up the musicbrainz id for each.  It's literally 
>>>> throwing away the information it's already found and which I want to keep. 
>>>>    I'm a programmer so there's no reason for me to edit things 
>>>> individually when the data is there to automate it.
>>>> 
>>>> Which means I'm back to, "hey why don't I write a script to hit 
>>>> musicbrainz for all my mp3s and force them to be cataloged correctly" .    
>>>>  I'm looking into the beets plugin process as well but that seems like I 
>>>> really have to code to the internals of Beets (like maybe flipping 
>>>> is_album=True in SingletonimportTask, not sure) - not really sure that's 
>>>> worth it here.
>>>> 
>>>> 
>>>> 
>>>> 
>>>> 
>>>> , and if I'm going to get into manual id3 tag editing, Ithen I'd just 
>>>> write a script myself to hit the musicbrainz API and write the data.  It 
>>>> refuses to ignore or erase the incorrect album / artist.
>>>> 
>>>> On Monday, August 8, 2016 at 9:25:59 AM UTC-4, Adrian Sampson wrote:
>>>> Yeah, the `edit` plugin would work, and so would the more CLI-oriented 
>>>> `beet modify album=foo`.
>>>> 
>>>> You could also import the with the flags `-ts`, which would turn on timid 
>>>> mode—i.e., avoid trusting the current metadata—and then you can enter a 
>>>> manual album and artist for the search.
>>>> 
>>>> One final option would be to turn on acoustic fingerprinting with the 
>>>> `chroma` plugin, which could do a reasonable job at guessing the original 
>>>> metadata.
>>>> 
>>>> Adrian
>>>> 
>>>> 
>>>>> On Aug 8, 2016, at 12:11 AM, cla...@ <>zzzcomputing.com 
>>>>> <http://zzzcomputing.com/> wrote:
>>>>> 
>>>>> Ok well I'm basically looking to "de-mixtape" these tracks.  That is, 
>>>>> just make them as though I just have a handful of songs that need to be 
>>>>> grouped into partial albums that they came from originally.  So I don't 
>>>>> need to group them on that "mixtape" id at all.
>>>>> 
>>>>> I'm guessing that if I totally scrubbed these of the fake artist and 
>>>>> album name, then just imported them as though they were just mixed up, 
>>>>> then it would do it?   Of course another issue is that half the songs on 
>>>>> this particular "mixtape" as well as a bunch of others aren't in the 
>>>>> musicbrainz or discogs database at all but I'll try to deal with that 
>>>>> separately.   Would I try to use "beet edit" or something to scrub them 
>>>>> out totally?  the "scrub" plugin didn't seem to really do the right thing 
>>>>> on the first try but hard to tell where beets cares about the directory 
>>>>> structure vs. the id3 tags.
>>>>> 
>>>>> thanks for the quick response.
>>>>> 
>>>>> On Sunday, August 7, 2016 at 6:55:56 PM UTC-4, Adrian Sampson wrote:
>>>>> Yeah, that’s a tricky case. When we originally introduced singletons, 
>>>>> many years ago now, the idea was for exactly this case—when you have 
>>>>> mixtapes from friends with an assortment of unrelated songs.
>>>>> 
>>>>> Really, the “beets way” would be to leave them as singletons and add an 
>>>>> extra flexible attribute that links them together. For example, `beet 
>>>>> modify mixtape=cruise2006` or something could mark all the members of a 
>>>>> mixtape. Then, you’d use a path rule to group mixtapes together in their 
>>>>> own directories.
>>>>> 
>>>>> The only trouble, of course, is that this approach doesn’t preserve 
>>>>> order. (Beets needs a “playlists” concept.) But you could consider using 
>>>>> the track field, or even another flexible attribute, to define the order.
>>>>> 
>>>>> I hope that helps. Good luck, and please ask more questions!
>>>>> 
>>>>> Adrian
>>>>> 
>>>>> 
>>>>>> On Aug 7, 2016, at 6:45 PM, cla...@ <>zzzcomputing.com 
>>>>>> <http://zzzcomputing.com/> wrote:
>>>>>> 
>>>>>> Hi there -
>>>>>> 
>>>>>> I have what I would think is one of the primary use cases for beets 
>>>>>> however it seems to be unsupported.
>>>>>> 
>>>>>> Basically I have a bunch of tracks, all underneath a particular artist 
>>>>>> and album name, where that artist/album is completely made up.  But 
>>>>>> also, each track with this album is in reality from *totally different 
>>>>>> albums*.   
>>>>>> 
>>>>>> The use case here is exactly what would happen in what I would think is 
>>>>>> the common use case of someone has a "mix tape" where every song has 
>>>>>> been tagged with someone's fake "album" name.  We want to move the 
>>>>>> tracks all underneath the actual artist/album they'd be from, given that 
>>>>>> we are generating only "partial" albums that have just a few tracks.
>>>>>> 
>>>>>> Options I'm working with to deal with this include the "scrub" plugin as 
>>>>>> well as the "group_album" option.  They are still not removing the fake 
>>>>>> album name even though a *real* album name is right in the musicbrainz 
>>>>>> listing.
>>>>>> 
>>>>>> Here is a small part of the example I'm working with, this is 
>>>>>> traditional Indian music (from my wife's collection, I don't know 
>>>>>> anything about this kind of music):
>>>>>> 
>>>>>> /Aarti/Vaishnodevi/01 01HEY GOVIND HEY GOPAL- NANAK.mp3
>>>>>> /Aarti/Vaishnodevi/02 02 HARI TUM HARO  JAN KI PEER.mp3
>>>>>> /Aarti/Vaishnodevi/04 04 DEENAN DUKH HARAN DEV-SURD.mp3
>>>>>> 
>>>>>> 
>>>>>> When I do "import" for these, beets can't find anything about an album 
>>>>>> "Aarti - Vaishnodevi" because there is no such thing.  So my only choice 
>>>>>> is to press T for "as tracks".  Whether or not I do "group albums" first 
>>>>>> does not affect the result.   Then it looks up each song individually, 
>>>>>> and slowly enough I can find close enough identifiers:
>>>>>> 
>>>>>> http://musicbrainz.org/recording/6f80178a-9fb4-4e5e-8901-59c7fe9563cf 
>>>>>> <http://musicbrainz.org/recording/6f80178a-9fb4-4e5e-8901-59c7fe9563cf>
>>>>>> http://musicbrainz.org/recording/e2f04a67-d8fd-4fea-818e-feb62445c453 
>>>>>> <http://musicbrainz.org/recording/e2f04a67-d8fd-4fea-818e-feb62445c453>
>>>>>> http://musicbrainz.org/recording/7ab29996-6032-474a-975d-1fca8f20edab 
>>>>>> <http://musicbrainz.org/recording/7ab29996-6032-474a-975d-1fca8f20edab>
>>>>>> 
>>>>>> Because I've had no choice but to select "as tracks", these are forced 
>>>>>> into "singleton" mode, which means, "there's no album", which I don't 
>>>>>> really understand since of course there's an album for each one.   I've 
>>>>>> changed my "singleton" format to read "$albumartist/$album/$track 
>>>>>> $title", rather than "No Album".  However, it leaves the ficticious 
>>>>>> album name in place (or if I use "scrub", I get a dash), and does not 
>>>>>> use the information from the musicbrainz listing, and I get:
>>>>>> 
>>>>>> /Jagjit Singh/Vaishnodevi/01 Hey Gobind Hey Gopal _ Nanak.mp3
>>>>>> /Jagjit Singh/Vaishnodevi/02 Hary Tum Haro Jan Ki Peer.mp3
>>>>>> /Jagjit Singh/Vaishnodevi/04 Deenan Dukh Haran Dev _ Surdas.mp3
>>>>>> 
>>>>>> this is wrong.  Ignoring that musicbrainz' release titles look a little 
>>>>>> off here, what it should be based on those ids, is:
>>>>>> 
>>>>>> /Jagjit Singh/Hare Krishna: A Live Concert/01 Hey Gobind Hey Gopal _ 
>>>>>> Nanak.mp3
>>>>>> /Jagjit Singh/Kare Krishna A Live Concert/02 Hary Tum Haro Jan Ki 
>>>>>> Peer.mp3
>>>>>> /Jagjit Singh/Hare Krishna: A Live Concert/04 Deenan Dukh Haran Dev _ 
>>>>>> Surdas.mp3
>>>>>> 
>>>>>> Is there some option for this behavior or is this an unsupported use 
>>>>>> case ?   (what do people organizing old mix tapes with bad album names 
>>>>>> do? )
>>>>>> 
>>>>>> 
>>>>>> 
>>>>>> 
>>>>>> 
>>>>>> 
>>>>>> 
>>>>>> -- 
>>>>>> You received this message because you are subscribed to the Google 
>>>>>> Groups "beets" group.
>>>>>> To unsubscribe from this group and stop receiving emails from it, send 
>>>>>> an email to beets-users...@ <>googlegroups.com 
>>>>>> <http://googlegroups.com/>.
>>>>>> For more options, visit https://groups.google.com/d/optout 
>>>>>> <https://groups.google.com/d/optout>.
>>>>> 
>>>>> 
>>>>> -- 
>>>>> You received this message because you are subscribed to the Google Groups 
>>>>> "beets" group.
>>>>> To unsubscribe from this group and stop receiving emails from it, send an 
>>>>> email to [email protected] <>.
>>>>> For more options, visit https://groups.google.com/d/optout 
>>>>> <https://groups.google.com/d/optout>.
>>>> 
>>>> 
>>>> -- 
>>>> You received this message because you are subscribed to the Google Groups 
>>>> "beets" group.
>>>> To unsubscribe from this group and stop receiving emails from it, send an 
>>>> email to [email protected] <>.
>>>> For more options, visit https://groups.google.com/d/optout 
>>>> <https://groups.google.com/d/optout>.
>>> 
>>> 
>>> -- 
>>> You received this message because you are subscribed to the Google Groups 
>>> "beets" group.
>>> To unsubscribe from this group and stop receiving emails from it, send an 
>>> email to [email protected] <>.
>>> For more options, visit https://groups.google.com/d/optout 
>>> <https://groups.google.com/d/optout>.
>> 
>> 
>> -- 
>> You received this message because you are subscribed to the Google Groups 
>> "beets" group.
>> To unsubscribe from this group and stop receiving emails from it, send an 
>> email to [email protected] 
>> <mailto:[email protected]>.
>> For more options, visit https://groups.google.com/d/optout 
>> <https://groups.google.com/d/optout>.
> 

-- 
You received this message because you are subscribed to the Google Groups 
"beets" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
For more options, visit https://groups.google.com/d/optout.

Reply via email to