#metabrainz

/

      • MajorLurker joined the channel
      • 2020-08-28 24155, 2020

      • MajorLurker has quit
      • 2020-08-28 24117, 2020

      • MajorLurker joined the channel
      • 2020-08-28 24108, 2020

      • thomasross joined the channel
      • 2020-08-28 24139, 2020

      • MajorLurker has quit
      • 2020-08-28 24135, 2020

      • MajorLurker joined the channel
      • 2020-08-28 24159, 2020

      • MajorLurker has quit
      • 2020-08-28 24125, 2020

      • Gore has quit
      • 2020-08-28 24139, 2020

      • Gore joined the channel
      • 2020-08-28 24145, 2020

      • thomasross has quit
      • 2020-08-28 24142, 2020

      • MajorLurker joined the channel
      • 2020-08-28 24159, 2020

      • MajorLurker has quit
      • 2020-08-28 24114, 2020

      • MajorLurker joined the channel
      • 2020-08-28 24137, 2020

      • ishaanshah
        Morning
      • 2020-08-28 24154, 2020

      • dseomn1 joined the channel
      • 2020-08-28 24152, 2020

      • dseomn has quit
      • 2020-08-28 24115, 2020

      • sumedh joined the channel
      • 2020-08-28 24144, 2020

      • ishaanshah
      • 2020-08-28 24109, 2020

      • ishaanshah
        almost all of the entries are present in the MB database
      • 2020-08-28 24107, 2020

      • Higilopochtli has quit
      • 2020-08-28 24136, 2020

      • pristine___
        ishaanshah: how did you check?
      • 2020-08-28 24111, 2020

      • ishaanshah
        all of them are pretty famous
      • 2020-08-28 24150, 2020

      • ishaanshah
      • 2020-08-28 24117, 2020

      • MajorLurker has quit
      • 2020-08-28 24112, 2020

      • pristine___
        ishaanshah: umm... I think there should be some other means to verify this other than the names since they can be misleading
      • 2020-08-28 24137, 2020

      • ishaanshah
        meaning?
      • 2020-08-28 24114, 2020

      • ishaanshah
        essentially I should find a corr. MBID in mb right
      • 2020-08-28 24131, 2020

      • ishaanshah
        we can also use labs api
      • 2020-08-28 24136, 2020

      • ishaanshah
      • 2020-08-28 24115, 2020

      • pristine___
        Do we have *Love Me Like You Do - From \"Fifty Shades of Grey\* track of Ellie Goulding in MB?
      • 2020-08-28 24138, 2020

      • pristine___
        It's not just about the arist, we are also checking if a recording is in the MB or not? Is there a way to check for a recording MBID from recording MSID ishaanshah ?
      • 2020-08-28 24159, 2020

      • sumedh has quit
      • 2020-08-28 24114, 2020

      • ishaanshah
      • 2020-08-28 24124, 2020

      • ishaanshah
        the first one is not present
      • 2020-08-28 24129, 2020

      • ishaanshah
        the second entry is
      • 2020-08-28 24159, 2020

      • sumedh joined the channel
      • 2020-08-28 24102, 2020

      • ishaanshah
        Hmm, matches are not found for lot of msids
      • 2020-08-28 24135, 2020

      • pristine___
        Right.
      • 2020-08-28 24136, 2020

      • v6lur joined the channel
      • 2020-08-28 24145, 2020

      • pristine___
        But it is there for the second one?
      • 2020-08-28 24145, 2020

      • ishaanshah
        but the issue is, that these tracks are famous and do exist in the MB database
      • 2020-08-28 24148, 2020

      • pristine___
        It's a bug
      • 2020-08-28 24101, 2020

      • ishaanshah
      • 2020-08-28 24106, 2020

      • pristine___
        Right. But we really cannot use MSIDs in recs
      • 2020-08-28 24117, 2020

      • pristine___
        And MBIDs limit them.
      • 2020-08-28 24144, 2020

      • ishaanshah
        hmm, so right now, we are getting limited by the mapping available?
      • 2020-08-28 24150, 2020

      • pristine___
      • 2020-08-28 24159, 2020

      • pristine___
        Does this makes sense to you?
      • 2020-08-28 24109, 2020

      • pristine___
        > hmm, so right now, we are getting limited by the mapping available?
      • 2020-08-28 24113, 2020

      • pristine___
        ?
      • 2020-08-28 24117, 2020

      • pristine___
        Didn't get you
      • 2020-08-28 24145, 2020

      • ishaanshah
        I mean, those tracks should have been mapped to a valid MBID, but they didn't
      • 2020-08-28 24100, 2020

      • ishaanshah
        and thats why its showing up as missing MB data
      • 2020-08-28 24109, 2020

      • pristine___
        Yes. So the plan it to show recs on site and request users to fill in those missing MBIDs for better recs in future.
      • 2020-08-28 24132, 2020

      • ishaanshah
        oh, so crowd source msid-mbid mapping?
      • 2020-08-28 24114, 2020

      • pristine___
        Two or more MSIDs can map to single MBID. MSIDs are noisy, we initially used MSIDs for recs but the results were horrifying. Also, MSIDs are case sensitive which was leading to repetitive recs.
      • 2020-08-28 24124, 2020

      • pristine___
        > oh, so crowd source msid-mbid mapping?
      • 2020-08-28 24126, 2020

      • pristine___
        Kinda
      • 2020-08-28 24124, 2020

      • pristine___
        Did you red the query ishaanshah ?
      • 2020-08-28 24139, 2020

      • ishaanshah
        yep reading
      • 2020-08-28 24144, 2020

      • pristine___
        Repeating*
      • 2020-08-28 24148, 2020

      • pristine___
        Cool :)
      • 2020-08-28 24134, 2020

      • ishaanshah
        the query looks correct to me
      • 2020-08-28 24141, 2020

      • pristine___
        But can you open a ticket for the bug?
      • 2020-08-28 24149, 2020

      • ishaanshah
        yep will do
      • 2020-08-28 24150, 2020

      • pristine___
        The existing MBID thing?
      • 2020-08-28 24157, 2020

      • pristine___
        And assign it to me
      • 2020-08-28 24100, 2020

      • pristine___
        Thank you :)
      • 2020-08-28 24117, 2020

      • ishaanshah
        we should check if the mapping version is same on bono and our cluster too
      • 2020-08-28 24139, 2020

      • ishaanshah
        Maybe it improved in a later version
      • 2020-08-28 24147, 2020

      • pristine___
        Umm.... Right. I will check when was the mapping last updated on FTP.
      • 2020-08-28 24159, 2020

      • Higilopochtli joined the channel
      • 2020-08-28 24106, 2020

      • pristine___
        I will ask. ruaok once he returns
      • 2020-08-28 24117, 2020

      • ishaanshah
        I'll still open a ticket, just so that we dont forget
      • 2020-08-28 24137, 2020

      • pristine___
        Meanwhile if you wish to you submit your missing musicbrainz data :p
      • 2020-08-28 24145, 2020

      • pristine___
        > I'll still open a ticket, just so that we dont forget
      • 2020-08-28 24150, 2020

      • pristine___
        Right. Thanka
      • 2020-08-28 24156, 2020

      • pristine___
        Thanks*
      • 2020-08-28 24117, 2020

      • ishaanshah
        The data already exists in MB
      • 2020-08-28 24129, 2020

      • ishaanshah
        we just have to link it correctly ig
      • 2020-08-28 24105, 2020

      • pristine___
        Which data? I thought you said most of the data you have fetched from the API endpoint doesn't exist
      • 2020-08-28 24120, 2020

      • pristine___
        > Hmm, matches are not found for lot of msids
      • 2020-08-28 24121, 2020

      • pristine___
        Here
      • 2020-08-28 24111, 2020

      • ishaanshah
        I mean the recording "Payphone" exists in MB database
      • 2020-08-28 24143, 2020

      • ishaanshah
        but the msid "aa780803-0c00-44c7-b965-48644a49fe81" doesnt map to it
      • 2020-08-28 24108, 2020

      • pristine___
        Oh. Right.
      • 2020-08-28 24104, 2020

      • pristine___
        Is there a check in MB when a user submits data? Like if they are submitting correct MBIDs or something like that?
      • 2020-08-28 24145, 2020

      • pristine___
        ishaanshah: I think in most of the cases data exists in MBID, but it isn't mapped to the MSID. Gut feeling.
      • 2020-08-28 24105, 2020

      • pristine___
        Can you include the Payphone example in the ticket? It's a nice one.
      • 2020-08-28 24123, 2020

      • ishaanshah
        yep thats what I think too
      • 2020-08-28 24126, 2020

      • ishaanshah
        yep will do
      • 2020-08-28 24147, 2020

      • pristine___
        Hey. I think there should be a check in Lemmy too if the data has already been submitted to MB? It may happen that some other user submits the data that is shown to you.
      • 2020-08-28 24156, 2020

      • sumedh has quit
      • 2020-08-28 24101, 2020

      • BrainzGit
        [listenbrainz-server] vansika opened pull request #1059 (master…dont-send-last-week-listens-in-rec): Recommended recordings should not include recordings the user listened to recently. https://github.com/metabrainz/listenbrainz-server…
      • 2020-08-28 24102, 2020

      • MajorLurker joined the channel
      • 2020-08-28 24132, 2020

      • MajorLurker has quit
      • 2020-08-28 24114, 2020

      • mckean
        pristine___: are you also using the new performance monitoring from sentry?
      • 2020-08-28 24122, 2020

      • pristine___
        mckean: I am not sure what is that so I guess not using
      • 2020-08-28 24116, 2020

      • mckean
        basically APM, newrelic started with thtat, datadog is doing it, now also sentry provides it. Indepth application performance monitoring... the nice things is you can provide an id with the frontend request and it will link the two.
      • 2020-08-28 24138, 2020

      • mckean
        but yeah we're still waiting, proper php support is not there yet.
      • 2020-08-28 24142, 2020

      • pristine___
        Hmm... Sounds interesting.
      • 2020-08-28 24139, 2020

      • BrainzGit
        [listenbrainz-server] vansika opened pull request #1060 (master…remove-unused-return): don't return playcounts_df which is not used later in the code https://github.com/metabrainz/listenbrainz-server…
      • 2020-08-28 24133, 2020

      • pristine___
        iliekcomputers: did you see the `Connection Closed` error in sentry. I think the recs weren't pushed to Lemmy.
      • 2020-08-28 24118, 2020

      • diru1100
        Morning!!
      • 2020-08-28 24122, 2020

      • yvanzo
        mo’’in’
      • 2020-08-28 24105, 2020

      • mckean
        good morning
      • 2020-08-28 24142, 2020

      • BrainzGit
        [listenbrainz-server] vansika opened pull request #1061 (master…make-release-fields-in-model-optional): Make release msid and release name field optional in the pydantic model https://github.com/metabrainz/listenbrainz-server…
      • 2020-08-28 24150, 2020

      • _lucifer
        pristine___: which field is used from the input dataset to compute recs, i mean recording name, recording msid or something else?
      • 2020-08-28 24149, 2020

      • pristine___
      • 2020-08-28 24114, 2020

      • pristine___
        each distinct user_name is assigned a user_id
      • 2020-08-28 24134, 2020

      • pristine___
        each distinct recording_mbid is assigned a recording_id
      • 2020-08-28 24141, 2020

      • pristine___
        count is the listen count
      • 2020-08-28 24103, 2020

      • _lucifer
        ok, thanks
      • 2020-08-28 24117, 2020

      • pristine___
        This is used as the input to train the model as well as generate recs from trained model
      • 2020-08-28 24123, 2020

      • pristine___
        no prob :)
      • 2020-08-28 24135, 2020

      • _lucifer
        another ques, i had is whether we want to assign equal weights to the user?
      • 2020-08-28 24153, 2020

      • pristine___
        as in?
      • 2020-08-28 24115, 2020

      • pristine___
        we are basically using user-user similarity to get recs
      • 2020-08-28 24131, 2020

      • mckean
        alastairp: just shouting out, I'm here whenever you have some spare time.
      • 2020-08-28 24146, 2020

      • pristine___
        iliekcomputers: if a dump fails, is there any message on LB site stating that we weren't able to update stats for that week or something like that?
      • 2020-08-28 24158, 2020

      • pristine___
        Dump fails for that week*
      • 2020-08-28 24120, 2020

      • iliekcomputers
        Not really
      • 2020-08-28 24159, 2020

      • pristine___
        I think it will nice thing to have.
      • 2020-08-28 24104, 2020

      • pristine___
        Be*
      • 2020-08-28 24149, 2020

      • pristine___
        iliekcomputers: connection error in sentry means recs weren't pushed for all or some users?
      • 2020-08-28 24120, 2020

      • iliekcomputers
        I'd have to look into it
      • 2020-08-28 24121, 2020

      • pristine___
        Cool. What do they mean in general though? I have seen that error many times before this?
      • 2020-08-28 24105, 2020

      • _lucifer
        i meant normalization each users ratings pristine___, do we want to do that?
      • 2020-08-28 24140, 2020

      • pristine___
        I think we should do that. I remember we chatted about it. There are a couple of bugs I want to address before taking this one in hand.
      • 2020-08-28 24151, 2020

      • pristine___
        Can you open a ticket and assign to me?
      • 2020-08-28 24154, 2020

      • antlarr2 is now known as antlarr
      • 2020-08-28 24100, 2020

      • shivam-kapila
        Morning
      • 2020-08-28 24125, 2020

      • shivam-kapila
        iliekcomputers: All PRs are ready
      • 2020-08-28 24102, 2020

      • Gazooo7 has quit
      • 2020-08-28 24110, 2020

      • Gazooo7 joined the channel
      • 2020-08-28 24116, 2020

      • iliekcomputers
        shivam-kapila: thanks!
      • 2020-08-28 24128, 2020

      • iliekcomputers
        shivam-kapila: i'll take a look.
      • 2020-08-28 24146, 2020

      • iliekcomputers
        shivam-kapila: can you start looking at the LB-686 today? I think we can get it finished by tomorrow, if you get an initial version up for review today
      • 2020-08-28 24147, 2020

      • BrainzBot
        LB-686: Separate, obvious page for ways to scrobble and import listens https://tickets.metabrainz.org/browse/LB-686
      • 2020-08-28 24136, 2020

      • iliekcomputers
        the ticket is pretty specific in its suggestions, but feel free to play around with what the best UX should be
      • 2020-08-28 24148, 2020

      • iliekcomputers
        however, do keep the timeline in mind.
      • 2020-08-28 24118, 2020

      • MajorLurker joined the channel
      • 2020-08-28 24101, 2020

      • MajorLurker has quit
      • 2020-08-28 24141, 2020

      • pristine___
        iliekcomputers: can you have a look at #1059, #1060, #1061 today if you get the time? They are small PRs. I want to merge them today if possible.
      • 2020-08-28 24144, 2020

      • iliekcomputers
        sure
      • 2020-08-28 24106, 2020

      • pristine___
        Thank you so much 🌹
      • 2020-08-28 24146, 2020

      • shivam-kapila
        iliekcomputers: I have roughly started but can I get the work product done first
      • 2020-08-28 24155, 2020

      • shivam-kapila
        Its raw and incomplete
      • 2020-08-28 24159, 2020

      • iliekcomputers
        sure.
      • 2020-08-28 24140, 2020

      • _lucifer
        sure, will do pristine___. i'd like to work on it as well in the upcoming weeks