#metabrainz

/

      • Gentlecat
        oh, that's neat
      • 2015-09-21 26414, 2015

      • zas
        it looks like it was a fun night ruaok ;)
      • 2015-09-21 26424, 2015

      • ruaok
        zas: massively fun. :)
      • 2015-09-21 26434, 2015

      • ruaok
        not the best for a sunday. :)
      • 2015-09-21 26439, 2015

      • zas
        ahah
      • 2015-09-21 26400, 2015

      • zas
        what is the history behind this event ?
      • 2015-09-21 26402, 2015

      • Gentlecat
        possibly the most fun I ever had :P
      • 2015-09-21 26424, 2015

      • alastairp
      • 2015-09-21 26424, 2015

      • ruaok
        Gentlecat: wow. :)
      • 2015-09-21 26446, 2015

      • alastairp
        though I don't really trust english wikipedia to explain everything
      • 2015-09-21 26410, 2015

      • alastairp
      • 2015-09-21 26404, 2015

      • zas
        the french page tells it exists back to middle age, around 11th century, and it also happens in the very south of France
      • 2015-09-21 26410, 2015

      • reosarevok has quit
      • 2015-09-21 26425, 2015

      • alastairp
        yeah, which is also historically catalunya
      • 2015-09-21 26450, 2015

      • zas
        yup
      • 2015-09-21 26419, 2015

      • samphippen has quit
      • 2015-09-21 26408, 2015

      • samphippen joined the channel
      • 2015-09-21 26437, 2015

      • zas
        Hell seems to become more organized : "Federació de Diables i Dimonis de Catalunya" ;)
      • 2015-09-21 26414, 2015

      • ruaok
        this is very serious business here.
      • 2015-09-21 26423, 2015

      • ruaok
        every neighborhood has a troupe
      • 2015-09-21 26421, 2015

      • zas
        similar to carnavals in Brazil
      • 2015-09-21 26401, 2015

      • ruaok nods
      • 2015-09-21 26424, 2015

      • Mineo joined the channel
      • 2015-09-21 26433, 2015

      • UmkaDK_ joined the channel
      • 2015-09-21 26403, 2015

      • UmkaDK has quit
      • 2015-09-21 26411, 2015

      • CatCat has quit
      • 2015-09-21 26417, 2015

      • alastairp
        ocharles__: strange. your other listens still haven't come up yet. it looks like it's doing a lot of catching up...
      • 2015-09-21 26425, 2015

      • ocharles__
        indeed
      • 2015-09-21 26442, 2015

      • alastairp
        (unless it crashed again... *checks*)
      • 2015-09-21 26450, 2015

      • Freso
        alastairp: Did you deploy the new code?
      • 2015-09-21 26410, 2015

      • alastairp
        no
      • 2015-09-21 26415, 2015

      • Gentlecat
        ruaok: I assume this one would work http://uk.rs-online.com/web/p/wireless-adapters/7…, right?
      • 2015-09-21 26433, 2015

      • CatCat joined the channel
      • 2015-09-21 26458, 2015

      • alastairp
        ocharles__: still running, I guess. It's caught up 2 million (!) since I started it
      • 2015-09-21 26409, 2015

      • ruaok
        alastairp: did it crash?
      • 2015-09-21 26436, 2015

      • alastairp
        ruaok: yeah, LB-47
      • 2015-09-21 26422, 2015

      • ruaok
        ah, yes. I've seen this one before.
      • 2015-09-21 26414, 2015

      • alastairp
        but it's different to no response from MessyBrainz, because in that case it loads but validate returns false
      • 2015-09-21 26412, 2015

      • ruaok
        http://tickets.musicbrainz.org/browse/LB-48 as per our discussion on saturday
      • 2015-09-21 26415, 2015

      • reosarevok joined the channel
      • 2015-09-21 26417, 2015

      • reosarevok has quit
      • 2015-09-21 26429, 2015

      • ruaok
        alastairp: yes, periodically we get a response from messybrainz that lacks all the msids.
      • 2015-09-21 26439, 2015

      • ruaok
        those get tossed by the listenstore importer.
      • 2015-09-21 26458, 2015

      • alastairp
        yep, that problem is LB-42
      • 2015-09-21 26459, 2015

      • ruaok
        but the root of the problem lies at msb, since the response from msb is valid, but empty
      • 2015-09-21 26404, 2015

      • alastairp
        I'm looking into it now
      • 2015-09-21 26418, 2015

      • ruaok
        k
      • 2015-09-21 26441, 2015

      • samphippen has quit
      • 2015-09-21 26418, 2015

      • samphippen joined the channel
      • 2015-09-21 26436, 2015

      • xps2 joined the channel
      • 2015-09-21 26446, 2015

      • alastairp
        ruaok: do we want to start looking at how we are going to counting? Open a ticket, get some people giving feedback?
      • 2015-09-21 26416, 2015

      • ruaok
        to document our redis idea?
      • 2015-09-21 26416, 2015

      • samphippen has quit
      • 2015-09-21 26428, 2015

      • alastairp
        yeah
      • 2015-09-21 26432, 2015

      • ruaok
        sure
      • 2015-09-21 26416, 2015

      • samphippen joined the channel
      • 2015-09-21 26442, 2015

      • ocharles__
        considered any large data set approximations like hyperloglog?
      • 2015-09-21 26402, 2015

      • ruaok
        what does that do?
      • 2015-09-21 26440, 2015

      • alastairp
        perhaps not until real counting becomes too slow for us
      • 2015-09-21 26451, 2015

      • alastairp
        we only have 200 users (at the moment...)
      • 2015-09-21 26437, 2015

      • alastairp
        our plan is to have a redis counter per user, which shouldn't cause many problems for now
      • 2015-09-21 26450, 2015

      • ocharles__
        It's a probabilistic data structure that gives you an approximate count of elements in a multiset
      • 2015-09-21 26450, 2015

      • alastairp
        I think hyperloglog is too complex too quickly
      • 2015-09-21 26401, 2015

      • ocharles__
        fair enough
      • 2015-09-21 26412, 2015

      • alastairp
        but it's cool
      • 2015-09-21 26430, 2015

      • alastairp
        and if we want to count per-track and per-artist too, it might be something worth looking at
      • 2015-09-21 26455, 2015

      • ocharles__
        are there plans to have LB have those kind of metrics like last.fm?
      • 2015-09-21 26459, 2015

      • ocharles__
        or is that mostly to be decided?
      • 2015-09-21 26413, 2015

      • alastairp
        mostly to be decided :)
      • 2015-09-21 26420, 2015

      • alastairp
        what do you want? we'll add it!
      • 2015-09-21 26441, 2015

      • ocharles__
        you mean "what do you want? you add it!", right? :)
      • 2015-09-21 26448, 2015

      • alastairp
        :(
      • 2015-09-21 26453, 2015

      • ocharles__
        though maybe I shouldn't argue with your willingness to do work for me
      • 2015-09-21 26459, 2015

      • ocharles__
        so... a pony?
      • 2015-09-21 26404, 2015

      • alastairp
        oh, right. /you'll/ do it
      • 2015-09-21 26406, 2015

      • alastairp
        go for it
      • 2015-09-21 26409, 2015

      • github joined the channel
      • 2015-09-21 26409, 2015

      • github
        [messybrainz-server] alastair created ujson (+2 new commits): https://github.com/metabrainz/messybrainz-server/…
      • 2015-09-21 26409, 2015

      • github
        messybrainz-server/ujson f3d0775 Alastair Porter: Add some missing indexes for lookup
      • 2015-09-21 26409, 2015

      • github
        messybrainz-server/ujson 394719f Alastair Porter: Use ujson to read requests and make responses
      • 2015-09-21 26409, 2015

      • github has left the channel
      • 2015-09-21 26432, 2015

      • ruaok
      • 2015-09-21 26454, 2015

      • github joined the channel
      • 2015-09-21 26454, 2015

      • github
        [messybrainz-server] alastair opened pull request #10: Ujson (master...ujson) https://github.com/metabrainz/messybrainz-server/…
      • 2015-09-21 26454, 2015

      • github has left the channel
      • 2015-09-21 26444, 2015

      • alastairp
        ruaok: today I want to do LB-21, LB-40, LB-42, LB-43. Let me know if you want to pick up one of them
      • 2015-09-21 26418, 2015

      • alastairp
        --> lunch
      • 2015-09-21 26418, 2015

      • ruaok
        alastairp: the big question for me is if we still want to move the LB sanity checking to MSB in order to avoid duplicate parsing.
      • 2015-09-21 26408, 2015

      • xps2 has quit
      • 2015-09-21 26444, 2015

      • xps2 joined the channel
      • 2015-09-21 26454, 2015

      • samphippen has quit
      • 2015-09-21 26457, 2015

      • alastairp
        ruaok: I agree. I like that it could make things faster, with less parsing, but it still feels like the wrong place to do it, to me
      • 2015-09-21 26449, 2015

      • ruaok
        no doubt. logically it is a mess. :(
      • 2015-09-21 26446, 2015

      • Mogmi joined the channel
      • 2015-09-21 26458, 2015

      • CatQuest has quit
      • 2015-09-21 26401, 2015

      • alastairp
        so many jira tabs opened
      • 2015-09-21 26415, 2015

      • Freso has quit
      • 2015-09-21 26422, 2015

      • alastairp
        Gentlecat: http://tickets.musicbrainz.org/browse/AB-61 the PR is merged. can this be closed?
      • 2015-09-21 26403, 2015

      • Freso joined the channel
      • 2015-09-21 26441, 2015

      • Gentlecat
        alastairp: yeah, let's close it
      • 2015-09-21 26446, 2015

      • alastairp
        cool
      • 2015-09-21 26456, 2015

      • alastairp
        I'm doing AB dev full time now!
      • 2015-09-21 26411, 2015

      • alastairp
        well, maybe not full time, 60% or so
      • 2015-09-21 26416, 2015

      • alastairp
        not that you are, traitor
      • 2015-09-21 26409, 2015

      • Gentlecat
        I'm there if you need any help
      • 2015-09-21 26418, 2015

      • alastairp
        awww
      • 2015-09-21 26425, 2015

      • alastairp
        ruaok: you pay the right people
      • 2015-09-21 26436, 2015

      • ruaok
        wait, I did something right?
      • 2015-09-21 26437, 2015

      • Gentlecat
        actually hope that I get some project which would basically allow me to continue working on it
      • 2015-09-21 26445, 2015

      • Gentlecat
        on "university time"
      • 2015-09-21 26435, 2015

      • ruaok
        alastairp: I've got a pile of non-coding stuff to do and a skype call this afternoon.
      • 2015-09-21 26445, 2015

      • alastairp
        ruaok: no problem
      • 2015-09-21 26400, 2015

      • ruaok
        around 17:30h I can be free to help code the optimizations, unless you're done with them by then.
      • 2015-09-21 26401, 2015

      • Freso
        alastairp: \o/
      • 2015-09-21 26404, 2015

      • alastairp
        I'll see if I can get some more msb stuff done, since it kind of fits in with real work
      • 2015-09-21 26404, 2015

      • ruaok
        chances are...
      • 2015-09-21 26429, 2015

      • alastairp
        Gentlecat: right, let's talk
      • 2015-09-21 26453, 2015

      • alastairp
        it'd be a good project if we can find a end-to-end project with a research output
      • 2015-09-21 26405, 2015

      • ruaok
        mozartkugeln in the demo room.
      • 2015-09-21 26416, 2015

      • alastairp
        yes
      • 2015-09-21 26417, 2015

      • ruaok
        alastairp: you should go steal some for the team. :)
      • 2015-09-21 26419, 2015

      • alastairp
        already eaten
      • 2015-09-21 26423, 2015

      • ruaok
        damn.
      • 2015-09-21 26424, 2015

      • alastairp
        all gone, I think
      • 2015-09-21 26429, 2015

      • alastairp
        at least I had one
      • 2015-09-21 26445, 2015

      • alastairp
        ruaok: http://tickets.musicbrainz.org/browse/AB-50 last sentence of this
      • 2015-09-21 26414, 2015

      • alastairp
        1) if you're able to give an opinion now, please do; 2) if you think it requires a discussion in person, can we do that this week?
      • 2015-09-21 26452, 2015

      • ruaok
        mirror what MB does, I would say.
      • 2015-09-21 26456, 2015

      • ruaok
        that is to say that we should:
      • 2015-09-21 26412, 2015

      • ruaok
        dump a last incremental (and stop accepting new submissions during this)
      • 2015-09-21 26422, 2015

      • ruaok
        do the schema change
      • 2015-09-21 26428, 2015

      • ruaok
        do a full dump
      • 2015-09-21 26435, 2015

      • ruaok
        continue to do incrementals
      • 2015-09-21 26421, 2015

      • alastairp
        thanks. Added as a comment. I'll work towards that
      • 2015-09-21 26425, 2015

      • ruaok
        k
      • 2015-09-21 26444, 2015

      • ruaok
        the stopping submissions is kinda hard in this case.
      • 2015-09-21 26451, 2015

      • ruaok
        not hard, sucky
      • 2015-09-21 26457, 2015

      • ruaok
        because even an incremental will take a while.
      • 2015-09-21 26404, 2015

      • ruaok
        maybe consider doing two incrementals.
      • 2015-09-21 26413, 2015

      • ruaok
        one to get the bulk out of the way.
      • 2015-09-21 26418, 2015

      • alastairp
        right
      • 2015-09-21 26421, 2015

      • ruaok
        and another one while the site is in read only mode
      • 2015-09-21 26431, 2015

      • alastairp
        man, we should put kafka in acousticbrainz!
      • 2015-09-21 26439, 2015

      • ruaok
        and MB!
      • 2015-09-21 26445, 2015

      • alastairp
        allthebrainz
      • 2015-09-21 26447, 2015

      • ruaok
        it is clearly the right tool for that job.
      • 2015-09-21 26458, 2015

      • alastairp
        actually, this is interesting: the schema change will only be on the high level tables
      • 2015-09-21 26403, 2015

      • alastairp
        can we use that to our advantage?
      • 2015-09-21 26419, 2015

      • alastairp
        I'll think about it
      • 2015-09-21 26434, 2015

      • ruaok
        it allows zero downtime.