#metabrainz

/

      • CatQuest
        sorry about that
      • 2019-06-03 15406, 2019

      • iliekcomputers
        alastairp: ok
      • 2019-06-03 15409, 2019

      • D4RK-PH0ENiX joined the channel
      • 2019-06-03 15423, 2019

      • iliekcomputers
        pushing a new image first. (although it isn't strictly needed afaik?)
      • 2019-06-03 15448, 2019

      • alastairp
        yes, there is new code in master
      • 2019-06-03 15416, 2019

      • iliekcomputers
        alright, new image pushed and site brought down.
      • 2019-06-03 15452, 2019

      • iliekcomputers
        alastairp: same command again, right?
      • 2019-06-03 15458, 2019

      • alastairp
        yes
      • 2019-06-03 15450, 2019

      • alastairp
        and then `select count(*) from lowlevel where submission_offset is null` should return 0
      • 2019-06-03 15456, 2019

      • iliekcomputers
        command run
      • 2019-06-03 15448, 2019

      • iliekcomputers
      • 2019-06-03 15458, 2019

      • alastairp
        perfect
      • 2019-06-03 15406, 2019

      • alastairp
        bring it up again
      • 2019-06-03 15422, 2019

      • alastairp
        and keep and eye on sentry :)
      • 2019-06-03 15413, 2019

      • iliekcomputers
        it is up.
      • 2019-06-03 15455, 2019

      • iliekcomputers
        restarting other two containers
      • 2019-06-03 15458, 2019

      • alastairp
        thanks!
      • 2019-06-03 15441, 2019

      • alastairp
        can you hit the stats calculator again? it's not run since March
      • 2019-06-03 15452, 2019

      • alastairp
        what are we missing here - is the cron container still not running correctly?
      • 2019-06-03 15415, 2019

      • iliekcomputers
        alastairp: we need to merge the "init refactor" pr for cron to work.
      • 2019-06-03 15421, 2019

      • iliekcomputers
        i'll run the stats commands for now
      • 2019-06-03 15426, 2019

      • alastairp
        OK, let's do that this week
      • 2019-06-03 15428, 2019

      • alastairp
        if you have time
      • 2019-06-03 15435, 2019

      • iliekcomputers
        yep.
      • 2019-06-03 15438, 2019

      • iliekcomputers
        sounds good to me.
      • 2019-06-03 15404, 2019

      • alastairp
        are you available later today too? to try and get part 3 of this PR released
      • 2019-06-03 15442, 2019

      • iliekcomputers
        an hour before meeting? (9:30 PM IST)
      • 2019-06-03 15417, 2019

      • alastairp
        yes, that seems fine. if aidanlw17 can finish rebasing this PR soon, then I can test it before that
      • 2019-06-03 15447, 2019

      • iliekcomputers
        compute_stats and cache_stats both running now.
      • 2019-06-03 15402, 2019

      • Mr_Monkey
        akhilesh: I just pushed some test helpers that you might find useful. You'll need to rebase your branch (`git fetch && git rebase master`)
      • 2019-06-03 15445, 2019

      • Mr_Monkey
        Allows me to do this in my tests:
      • 2019-06-03 15402, 2019

      • Mr_Monkey
      • 2019-06-03 15431, 2019

      • Mr_Monkey
        Makes the tests a bit more readable
      • 2019-06-03 15447, 2019

      • travis-ci joined the channel
      • 2019-06-03 15447, 2019

      • travis-ci
        Project bookbrainz-site build #2217: passed in 4 min 31 sec: https://travis-ci.org/bookbrainz/bookbrainz-site/…
      • 2019-06-03 15447, 2019

      • travis-ci has left the channel
      • 2019-06-03 15421, 2019

      • bukwurm has quit
      • 2019-06-03 15442, 2019

      • aidanlw17
        Hi alastairp sorry I didn’t see this sooner, I will rebase it right now
      • 2019-06-03 15455, 2019

      • alastairp
        it's OK, I only told you an hour ago!
      • 2019-06-03 15436, 2019

      • aidanlw17
        alastairp: an update on annoy, I have a basic implementation working. For most recordings we are getting ~60% overlap with the similar recordings from the Postgres solution!
      • 2019-06-03 15451, 2019

      • alastairp
        great
      • 2019-06-03 15402, 2019

      • travis-ci joined the channel
      • 2019-06-03 15402, 2019

      • travis-ci
        metabrainz/picard#4675 (master - 0f1c499 : Philipp Wolfer): The build passed.
      • 2019-06-03 15402, 2019

      • travis-ci
      • 2019-06-03 15402, 2019

      • travis-ci
      • 2019-06-03 15402, 2019

      • travis-ci has left the channel
      • 2019-06-03 15411, 2019

      • alastairp
        although, theoretically with the same features, I'm surprised it's so... low?
      • 2019-06-03 15422, 2019

      • alastairp
        is there a difference in the similarity metric that they're both using?
      • 2019-06-03 15435, 2019

      • alastairp
        having said that, the fact that we have at least some is great
      • 2019-06-03 15455, 2019

      • aidanlw17
        There isn’t any difference between the metrics they use, at least there shouldn’t be
      • 2019-06-03 15440, 2019

      • aidanlw17
        One thing is that annoy is approximate, and will become more precise if we build the index with a higher number of trees
      • 2019-06-03 15403, 2019

      • aidanlw17
        I will try increasing the number of trees and see if that changes anything
      • 2019-06-03 15436, 2019

      • aidanlw17
        For some mbids, I saw zero similar recordings intersect while for others I saw up to 700.
      • 2019-06-03 15408, 2019

      • alastairp
        ok, cool. is there a tradeoff with number of trees? is lookup time slower with more trees?
      • 2019-06-03 15422, 2019

      • aidanlw17
        There is a tradeoff, it increases the build time and the size of the index with more trees
      • 2019-06-03 15419, 2019

      • Slurpee joined the channel
      • 2019-06-03 15419, 2019

      • Slurpee has quit
      • 2019-06-03 15419, 2019

      • Slurpee joined the channel
      • 2019-06-03 15422, 2019

      • aidanlw17
        But the annoy readme doesn’t say anything about increased lookup time explicitly - I assume if the index is larger the lookup time may also increase?
      • 2019-06-03 15404, 2019

      • aidanlw17
        For the moment, the annoy lookup time is lightning fast compared to Postgres solution
      • 2019-06-03 15419, 2019

      • RJ2 has quit
      • 2019-06-03 15420, 2019

      • RJ2 joined the channel
      • 2019-06-03 15444, 2019

      • Slurpee has quit
      • 2019-06-03 15405, 2019

      • adhawkins has quit
      • 2019-06-03 15425, 2019

      • ruaok
        iliekcomputers: Frank keeps getting load spikes. Can. You please pick up the ratelimiting PR and deploy it soon?
      • 2019-06-03 15430, 2019

      • adhawkins joined the channel
      • 2019-06-03 15420, 2019

      • zas
        ruaok: back to home ?
      • 2019-06-03 15442, 2019

      • alastairp
        aidanlw17: ok, cool. you could do some profiling - have a list of mbids and query them, and count the lookup time
      • 2019-06-03 15429, 2019

      • BrainzGit
        [musicbrainz-server] yvanzo opened pull request #1092 (master…fix-inst-layout): Restore instrument type name in instrument page title https://github.com/metabrainz/musicbrainz-server/…
      • 2019-06-03 15433, 2019

      • alastairp
        you should check out something like https://stackoverflow.com/questions/9551838/how-t… to ensure that linux isn't caching anything that makes lookups seem faster than they are
      • 2019-06-03 15417, 2019

      • reosarevok
        yvanzo: what's the plan re: releasing beta? Today?
      • 2019-06-03 15424, 2019

      • yvanzo
        Ok!
      • 2019-06-03 15401, 2019

      • aidanlw17
        alastairp: ok thanks, I'll check for the cache and take down some numbers
      • 2019-06-03 15416, 2019

      • pristine__
        ruaok: Hey! Have a look at #PR21 whenever you are free.
      • 2019-06-03 15403, 2019

      • akhilesh
        Mr_Monkey: Are you tried test-helper that you pushed?
      • 2019-06-03 15449, 2019

      • ruaok
        hi pristine__, will do.
      • 2019-06-03 15407, 2019

      • bitmap
        Cyna: what's the issue you're having with the form?
      • 2019-06-03 15446, 2019

      • reosarevok
        bitmap: IIRC it was how to pass it
      • 2019-06-03 15451, 2019

      • reosarevok
        To Node
      • 2019-06-03 15448, 2019

      • bitmap
        do you remember which form this is? we've always just pass it as a prop, unless this is something other than FormHandler
      • 2019-06-03 15429, 2019

      • akhilesh
        Mr_Monkey: when truncateEntities function is called, it gives <entity> table not found.
      • 2019-06-03 15436, 2019

      • akhilesh
      • 2019-06-03 15427, 2019

      • ruaok submits a review on #PR21
      • 2019-06-03 15450, 2019

      • yvanzo
        bitmap, reosarevok: blog post up for review
      • 2019-06-03 15449, 2019

      • aidanlw17
        alastairp: AB #341 should be rebased now, with the offset script removed
      • 2019-06-03 15457, 2019

      • alastairp
        I just saw it update, thanks
      • 2019-06-03 15456, 2019

      • aidanlw17
        no problem 👍
      • 2019-06-03 15440, 2019

      • aidanlw17
        alastairp: have you ever encountered an issue where running `./test.sh` takes a very long time and eventually times out with the following error: `An HTTP request took too long to complete.`
      • 2019-06-03 15458, 2019

      • alastairp
        mmm, no
      • 2019-06-03 15414, 2019

      • alastairp
        is it clear what is printing that error?
      • 2019-06-03 15417, 2019

      • alastairp
        is it docker?
      • 2019-06-03 15423, 2019

      • pristine__
        ruaok: yeah. thanks :)
      • 2019-06-03 15439, 2019

      • Mr_Monkey
        akhilesh: My bad, I pushed a correction
      • 2019-06-03 15445, 2019

      • aidanlw17
        I'm not sure. After the error, it says `Retry with --verbose to obtain debug information. If you encounter this issue regularly because of slow network conditions, consider setting COMPOSE_HTTP_TIMEOUT to a higher value (current value: 60).`
      • 2019-06-03 15452, 2019

      • aidanlw17
        If that gives any hint.
      • 2019-06-03 15433, 2019

      • yvanzo
        zas, reosarevok, bitmap: updating mb containers
      • 2019-06-03 15435, 2019

      • aidanlw17
        I tried stopping and deleting the containers with `./test.sh -s` and `./test.sh -d`, and also rebuilding with `./develop.sh build` to no avail.
      • 2019-06-03 15449, 2019

      • alastairp
        yeah, that's definitely a docker error
      • 2019-06-03 15456, 2019

      • CatQuest
        COMPOSE_HTTP_TIMEOUT
      • 2019-06-03 15456, 2019

      • CatQuest
        COMPOSE_EDGY_ETUDE
      • 2019-06-03 15456, 2019

      • CatQuest
        COMPOSE_ARIA
      • 2019-06-03 15456, 2019

      • CatQuest
        COMPOSE_ALLEGRETTO
      • 2019-06-03 15456, 2019

      • CatQuest
        oh no the AI is taking over
      • 2019-06-03 15407, 2019

      • alastairp
        next time it happens, try and restart docker
      • 2019-06-03 15412, 2019

      • CatQuest
        ok
      • 2019-06-03 15440, 2019

      • CatQuest
        restarting docker. the AI overloards weekness!
      • 2019-06-03 15449, 2019

      • CatQuest
        sorry i'l lbe quiet
      • 2019-06-03 15411, 2019

      • alastairp
        the AI systems aren't doing too well if we can subjugate them by just shutting down the process
      • 2019-06-03 15411, 2019

      • akhilesh
        Mr_Monkey: Is it require again rebase?
      • 2019-06-03 15418, 2019

      • Mr_Monkey
        Yes
      • 2019-06-03 15423, 2019

      • alastairp
        it's OK, just run the AI on google cloud, and it'll go down
      • 2019-06-03 15424, 2019

      • akhilesh
        ok
      • 2019-06-03 15431, 2019

      • CatQuest
        alastairp: well, to be fair, that's how you overthrow *human* overloards to :/
      • 2019-06-03 15451, 2019

      • CatQuest
        lords not loards. loard.
      • 2019-06-03 15455, 2019

      • aidanlw17
        aha! restarting docker fixed the issue :)
      • 2019-06-03 15457, 2019

      • alastairp
        this discussion suddenly got really dark
      • 2019-06-03 15402, 2019

      • CatQuest
        yea :/ sorry
      • 2019-06-03 15404, 2019

      • reosarevok
        bitmap: add alias IIRC, which doesn't seem to have a $form
      • 2019-06-03 15413, 2019

      • reosarevok
        yvanzo, bitmap: checking
      • 2019-06-03 15415, 2019

      • reosarevok
        The blog post
      • 2019-06-03 15424, 2019

      • travis-ci joined the channel
      • 2019-06-03 15424, 2019

      • travis-ci
        Project bookbrainz-site build #2219: passed in 5 min 9 sec: https://travis-ci.org/bookbrainz/bookbrainz-site/…
      • 2019-06-03 15424, 2019

      • travis-ci has left the channel
      • 2019-06-03 15426, 2019

      • CatQuest
        huh 🤔 "just run the AI on google cloud"
      • 2019-06-03 15426, 2019

      • CatQuest
        soudns a bit like something
      • 2019-06-03 15431, 2019

      • CatQuest
        aI on a cloud....
      • 2019-06-03 15442, 2019

      • CatQuest
        something fro ma film. where machines took over
      • 2019-06-03 15448, 2019

      • CatQuest
        even had time machines and everything
      • 2019-06-03 15411, 2019

      • CatQuest
        🤔 😽
      • 2019-06-03 15412, 2019

      • alastairp
        wait, are you talking about terminator or the matrix?
      • 2019-06-03 15419, 2019

      • aidanlw17
        the matrix?
      • 2019-06-03 15426, 2019

      • alastairp
        no time machines in the matrix?
      • 2019-06-03 15436, 2019

      • CatQuest
        ohohohoh am i talknig about terminator or the matrix indeed? ô_ô
      • 2019-06-03 15445, 2019

      • reosarevok
        yvanzo: made a couple changes
      • 2019-06-03 15447, 2019

      • CatQuest
        yes, terminator
      • 2019-06-03 15454, 2019

      • alastairp
        the CAT-2000
      • 2019-06-03 15407, 2019

      • CatQuest
        "running ai on google cloud" made me thing "skynet" immediately
      • 2019-06-03 15421, 2019

      • CatQuest
        vi skal være *litt* forsikting med det der
      • 2019-06-03 15431, 2019

      • CatQuest
        I mean it's bogus right now ffs
      • 2019-06-03 15433, 2019

      • CatQuest
        but yea
      • 2019-06-03 15449, 2019

      • CatQuest
        we're dependant on computers and machiens enouh as it is
      • 2019-06-03 15459, 2019

      • CatQuest
        no need ot paint uselves in a corner
      • 2019-06-03 15456, 2019

      • alastairp
        I think we should migrate musicbrainz to a paper spreadsheet
      • 2019-06-03 15403, 2019

      • CatQuest
        eh what I mean is: it's cool at computers, machines and aI are tools, that's awesome
      • 2019-06-03 15407, 2019

      • alastairp
        if you want to look up some data, you can make a phone call, and someone will look it up for you
      • 2019-06-03 15414, 2019

      • CatQuest
        but they sohudl remain *tools*
      • 2019-06-03 15416, 2019

      • CatQuest
        for *humans*
      • 2019-06-03 15427, 2019

      • reosarevok
        Humans are tools for humans too!
      • 2019-06-03 15433, 2019

      • CatQuest
        reosarevok: uff da!
      • 2019-06-03 15402, 2019

      • CatQuest
        I'm digressing awfully with the impending "robots taking over jobs" thing. some stuff is kinda really not ok