#metabrainz

/

      • d4rkie has quit
      • D4RK-PH0ENiX joined the channel
      • nawcom has quit
      • nawcom joined the channel
      • nawcom has quit
      • nawcom joined the channel
      • nawcom has quit
      • nawcom joined the channel
      • supersandro2000 has quit
      • supersandro2000 joined the channel
      • c1e0 joined the channel
      • nawcom has quit
      • nawcom joined the channel
      • livingsilver94_ joined the channel
      • livingsilver94 has quit
      • c1e0 has quit
      • thomasross has quit
      • reosarevok
        Lotheric: hmm, that's an interesting idea
      • Lotheric: can you download once you've listened 9 times?
      • BrainzGit
        [musicbrainz-server] reosarevok merged pull request #1728 (beta…MBS-11147): MBS-11147: Update autocomplete serialization https://github.com/metabrainz/musicbrainz-serve...
      • BrainzBot
        MBS-11147: Beta: Composers and artists no longer show up on work search https://tickets.metabrainz.org/browse/MBS-11147
      • apiuser joined the channel
      • apiuser
        Hello Metabrainz Team
      • II'm setting up local server using docker and I ran below command to build the search index. The command is downloading files with name likes replication*.tar.bz2 and running since last 30 hours and still not completed. so just wondering how much data it is supposed to download?
      • sudo docker-compose run --rm musicbrainz fetch-dump.sh search
      • the documentation says that it would fetch around 28 GB of data. is it?
      • yvanzo
        Hi apiuser: yes, it is that large.
      • These dumps are made available in case you cannot build search indexes from your own server.
      • pristine___
      • Bitmap ^^
      • yvanzo
        Updating beta.mb.o
      • pristine___
        ishaanshah: you around?
      • apiuser
        yvanzo, any guess on approximate size of data to download?
      • and yes, I'm not able to build search index on my own server because it throws following error.
      • 2020-10-05 05:39:54,110: Checking whether the versions of the Solr cores are supported
      • HTTPError(req.get_full_url(), code, msg, hdrs, fp)
      • dseomn has quit
      • dseomn joined the channel
      • c1e0 joined the channel
      • yvanzo
        apiuser: it depends on how long ago you downloaded the data dump, there is a replication packet per hour, it takes about 1min to get and apply an hourly packet.
      • apiuser
        I just downloaded the database 2 days back.
      • yvanzo
        so that makes at most 48 packets
      • it should not be running for 30 hours
      • v6lur joined the channel
      • livingsilver94 joined the channel
      • livingsilver94_ has quit
      • ruaok
        pristine___: I hope bitmap likes steely dan.
      • but hey, it worked!
      • also, moooin!
      • Gazooo794 has quit
      • Gazooo794 joined the channel
      • ishaanshah
        pristine___: hi
      • jmp_music__
        Morning!
      • supersandro2000 has quit
      • supersandro2000 joined the channel
      • kori has quit
      • pristine___
        ishaanshah: recs were generated fir bitmap for last week, but the stat endpoint for him returns empty response, can you have a look?
      • ruaok: steely dan is his top artist so I guess he likes steely dan :)
      • apiuser
        yvanzo, thank for the clue. I need to check for any issue at my end.
      • kieto joined the channel
      • d4rkie joined the channel
      • ishaanshah
        pristine___: The full dump import has been failing because of some issue with server space for past two weeks
      • so the data in spark for last two weeks is wrong, ig something to do with that
      • pristine___
        ishaanshah: why wrong? Should be incomplete, no!
      • ?
      • D4RK-PH0ENiX has quit
      • ishaanshah
        yes incomplete sorry
      • my stats or incomplete too
      • pristine___
        ishaanshah: oooo, I was confused to see only lost frequencies in your recs.
      • Is space is the issue, I will want to have a look
      • Thanks
      • ishaanshah
        space on lemmy
      • not on spark
      • pristine___
        ah, I thought spark cluster
      • ishaanshah
        the dump is not getting generated
      • pristine___
        Ig we should fix this asap, recs and stats getting affected
      • :(
      • iliekcomputers: is this on your radar?
      • Lotheric
      • problem with that model is you end up buying lossy music
      • if I'm going to buy music, I want lossless
      • c1e0 has quit
      • MajorLurker has quit
      • _lucifer
        CatQuest: what's the difference between Bokmål, Norwegian and Norwegian ?
      • c1e0 joined the channel
      • c1e0 has quit
      • c1e0 joined the channel
      • bitmap
        pristine___: cool, thanks for the reminder!
      • I guess the similar_artist one is more useful to me since I don't listen to a lot of different artists in the span of a week
      • apiuser has quit
      • pristine___
        bitmap: yeah, rn top artist have like 200 recs, so if you have about only 5-6 tracks of your last week's top artist, will the top_artist playlist be useful for you?
      • supersandro2000 has quit
      • supersandro2000 joined the channel
      • I think you were overwhelmed by the tracks of steely dan
      • alone
      • _lucifer
        alastairp: can you check what is the latest sql schema on beta?
      • bitmap
        that might help though I'd mostly use the recs to find new music or music I've forgotten about, and top_artist has a lot of songs I've listened to in the past few weeks
      • steely dan also wasn't my top artist last week though I guess that's 'cause the stats were broken
      • last.fm says they were #3
      • pristine___
        Bitmap: hmm.... Rn you won't have any tracks in your recs which you have listened to in the last week. I think 7 days is a very small window for people to not have the taste of music they have listened to. A few other users also have the same concern, I think if we increase this window, top artist recs might make more sense.
      • bitmap
        yeah, hard to say what's best for all users. a larger window would be better for me, but that's 'cause I mostly rotate the same 2-3 albums for a few weeks and then move onto new ones
      • ruaok
        bitmap: agreed. these two recommended tracks were originally intended for a "daily mix" sort of playlist that was based on what you've recently listened to.
      • for the "jump back in" or the "we think you might like" we'll need to train more models...
      • in due time.
      • bitmap nods
      • pristine___
        feedback will help in training models :)
      • bitmap
        it seems like a good list of songs for the 'daily mix' use case, so nice work there
      • ruaok
        :)
      • case in point, we need to improve how we present these algorithms[]
      • pristine___
        Any doc of the summit?
      • c1e0 has quit
      • c1e0 joined the channel
      • livingsilver94 has quit
      • livingsilver94 joined the channel
      • d4rkie has quit
      • D4RK-PH0ENiX joined the channel
      • c1e0 has quit
      • c1e0 joined the channel
      • c1e0 has quit
      • yvanzo has quit
      • yvanzo joined the channel
      • c1e0 joined the channel
      • jwf
        bitmap: Hahah good to know I am not the only one who goes through music listening phases like that!
      • CatQuest
        [14:07] <_lucifer> CatQuest: what's the difference between Bokmål, Norwegian and Norwegian ?
      • I assume you mean nynorsk and norwegian
      • basically nynorsk and bokmål are writing systems. one is based on danish which was "norwegianified" (bokmål) the other is a constructed language based on several dialects, primarly in west and middle norway
      • there is of course many dialects, and "norwegian" is basically any of them
      • I am fro moslo so I predominantly write "bokmål" (or a easter-dialecticaly modification of such (more "a" endings and difthnogs etc)) but i do not speeak it
      • the nynorsk-bokmål thing is kidna a big issue, historical abotu independance and so on, wikipedia can describe it better thna i can in irc here :D
      • _lucifer
        CatQuest: ah that's too much info :D. i was looking at a locale error that popped in logs and traced it to that.
      • CatQuest
        in general I want both in mb, but I also want a more generic header "norwegian" for things that are written as neither (like the Vazelina Bilopphøggers' band which sing and write titles in "toten" dialenct)
      • _lucifer
        which one should we be using ?
      • oh ok
      • CatQuest
        all 3
      • (ideally)
      • bokmål is a writing system. and the most used one, but in the districts they use nynorsk. and it's a mandatory thing that all important publications be written in both forms equally
      • but for music, especially more recently, people are more and more also writing in dialect, includiong socialect and "kebabnorsk" (other language inspired youth-language)
      • they have words from polis, turk, hindi and urdu :D
      • also people are taught both forms in school
      • they are mututally intellible (mostly)
      • alastairp
        _lucifer: it's possible that there might be an issue confusing country codes and language codes here
      • CatQuest
        some words i have no idea
      • yes
      • _lucifer
        alastairp: yes, that's what i am thinking about
      • CatQuest
        for all norway is country code is no/nor, for nynorsk, it's nn and bokmål is nb
      • alastairp
        👍
      • I had no idea that norway had different languages like this! super interesting, CatQuest
      • CatQuest
        and I believe this is also the reason we once time removed the "generic" norwegian" (because als owikipedia did this I think) but for situiation where e"it's neither nn or nb i want it still