if it's not already an official meb project it should be :)
rdswift
I think so. I would certainly prefer it to be under the MetaBrainz account rather than mine. We've started discussing it but no firm plans yet.
slriv joined the channel
samj1912
nice!
rdswift
Now that it's (mostly) out of development mode and into maintenance mode.
Still need to iron out a few details like serving different versions and translation to other languages.
slriv has quit
slriv joined the channel
MFCR_ColbyRay has quit
Gore has quit
Gore joined the channel
SomalRudra
Mr_Monkey: can you help us with a tagger icon as a dashboard icon for the app?
CallerNo6 has quit
CatQuest has quit
samj1912 has quit
CatQuest joined the channel
CallerNo6 joined the channel
CallerNo6 has quit
CallerNo6 joined the channel
nelgin joined the channel
nelgin
Hi all. I'm creating a new vm with the instructions from https://github.com/metabrainz/musicbrainz-docker. When I get to creating the indexes it gets so far then I get ""server closed the connection unexpectedly". This has happened twice now. What's going on?
kori has quit
kori joined the channel
yvanzo
hi nelgin: which server is closing the connection? how much of CPUs/RAM do you have? how long it takes to (fail to) create indexes?
nelgin
I'm not sure which server. I'm on the console so can only scroll back so far. I get Failed to import record. psycopg2.OperationalError) server closed the connection unexpectedly. This probably means the server terminated abnormally before or while processing the request. It then goes on to give a traceback for sir/indexing.py It fails after 182 minutes and 3 seconds. I have allocated 8GB
to the VM of which 4G was given to postgres via the tuning parameter.
I might be able to give 2gb more to the vm for reindexing but that'd leave the pc short, so it's be a temp measure.
yvanzo
it's usual to last a long time but the error is not
nelgin
I'll ssh into the box and rerun the index so I can scroll back further
Ok, I'll rerun the indexing and let you know if I have more information, it'll probably be tomorrow since it's midnight here and I need sleep.
it's nice that now it's technically possible to link to specific parts of the page
But realistically, how will a user know how to link a section
Should we add a copy button or something to the graphs?
shivam-kapila
You mean something like jump to a section?
iliekcomputers
i mean, i want to share my artist graph with a friend.
sumedh joined the channel
how will i know how to link that section specifically
shivam-kapila
A share button
The one with three dots
Upper right corner of graphs section
iliekcomputers
yep. makes sense to me
shivam-kapila
Ok ok
Will make a ticket and PR after the primary work
iliekcomputers
ishaanshah already has a PR, i think we can just do it in that one.
Gazooo has quit
shivam-kapila
Oh great
Gazooo joined the channel
white_shadow has quit
samj1912 joined the channel
BrainzGit
[bookbrainz-site] prabalsingh24 opened pull request #489 (UserCollection…fix-checkEntityTypeBeforeAdding): fix: throw error when trying to add collection, area, editor to a collection https://github.com/bookbrainz/bookbrainz-site/p...
ishaanshah
iliekcomputers: Ping
iliekcomputers
Pong
ishaanshah
Hi!
I was working on the backend for sitewide stats rn
It occurred to me that we have huge messages for each time range
So should I limit it to top ten per time range
or work on paginating over RMQ
iliekcomputers
I'm happy with limiting
But we should try a higher number, maybe the top 1000?
ishaanshah
Yeah sure
iliekcomputers
Or would that be too much?
ishaanshah
No that wont be the issue
iliekcomputers
I'll let you leave that configurable and we can play around with it.
But limits sound good to me.
ishaanshah
it's just that we are only gonna show the top 10
iliekcomputers
Yeah, but we can still expose more data over the API
yvanzo: finally got a bit of time to review the post
Changed a couple things, seems good to publish IMO
white_shadow
_lucifer are we using ndk in the app?
_lucifer
white_shadow: yes but not directly
white_shadow
what for?
_lucifer
the chromaprint algorithm is called from a java wrapper which is written using ndk
but the wrapper is not inside the app itself rather added as a dependency
secondly for the taglib library to read metadata
again it is added as a depenedency
so in the app we do not directly use ndk but depend on two libraries that use ndk and expose java/kotlin apis for us
white_shadow
ok but what we had to achieve by using chromaprint algorithm?
_lucifer
to obtain the fingerprint of the audio track and try to obtain a match
white_shadow
you mean determining similar audio tracks?
MFCR_ColbyRay joined the channel
_lucifer
Basically, chromaprint algorithm can generate an almost unique fingerprint for the audio file. If these audio files are already tagged with musicbrainz identifiers, the fingerprint can be submitted to acoust id database. Now, suppose someone else runs the algorithm on the same audio which does not have any metadata. Chromaprint will generate an almost equal fingerprint for it. We then query the acoust id database to check if this
fingerprint is already present. If yes, then we ask for the data they have related to it.
white_shadow
but you said we don't have the audio files of any song? Am i missing something here?
_lucifer
To be complete the rest of the workflow is as follows, AcoustID returns a small subset of the data we require. Then, we take the musicbrainz identifiers in this data and send to the musicbrainz servers to get the complete data available.
yes, we do not have any audio files white_shadow
but the users do. the users run chromaprint on their audio files