#metabrainz

mayhem, reosarevok et al: I spent some time this weekend testing and verifying a few things regarding publishing two documents on ReadTheDocs from the same GitHub repository (and branch). It takes a bit more effort to set it up initially, but it seems to work pretty well (although I haven't yet figured out how to make common pages available to both documents from a single set of files). The test repo that I'm using is

https://github.com/rdswift/rtd_testing and the two documents generated from it are https://rtd-testing-1.readthedocs.io/en/latest/ and https://rtd-testing-2.readthedocs.io/en/latest/ using two projects on ReadTheDocs, "rtd_testing_1" https://app.readthedocs.org/projects/rtd-testin... and "rtd_testing_2" https://app.readthedocs.org/projects/rtd-testin... (under my personal ReadTheDocs account), both pointing to the same GitHub repo

and branch. The GitHub repo is a bit messy because I'm still in development / testing mode. Anyway, it looks like we can make that configuration work.

[bookbrainz-site] 14shalini-bhandari closed pull request #1148 (03master…bb-791): BB-791 https://github.com/metabrainz/bookbrainz-site/p...

kellnerd: Hi, if you have some time, could you check what the newest timestamp is for Tidal request snapshots in Harmony? I have the v2 Tidal API basically ready, just need to decide when to use the old API for snapshots.

<GautamShorewala[> "It took forever to upload 😅" <- Should I make any changes to it?

jasje:

<GautamShorewala[> "jasje: currently in LB when we..." <- Gautam Shorewala: need some time to test

outsidecontext: We have the last two Tidal API requests at 1737454946 (which is on 21 January), so any timestamp between then and today should be fine.

Ok, thanks

[listenbrainz-server] 14mayhem merged pull request #3162 (03master…fix/top-artists-feed-template): Fix: Add missing <ul> element to the Top Artists feed template. https://github.com/metabrainz/listenbrainz-serv...

[@mayhem:chatbrainz.org](https://matrix.to/#/@mayhem:chatbrainz.org) it seems that listening activity stats are working but everything else is not according to that thread, either way I'll look at it in a couple of hours

<jasje[m]> "Gautam Shorewala: need some time..." <- Gautam Shorewala: PR?

jasje[m]: Looks good in video

jasje[m]: But it depends on implementation on how good it turns out

jasje[m]: How maintainable it is

<jasje[m]> "Gautam Shorewala: PR?" <- Will raise in some time

mayhem: can you review my pr

quite busy today, will try. no promises.

@fettuccinae:matrix.org: ok, reviewed -- looking pretty good, but there are a few things to tidy up.

[listenbrainz-server] 14amCap1712 merged pull request #3148 (03master…popularity-incremental-stats): Popularity incremental stats https://github.com/metabrainz/listenbrainz-serv...

[listenbrainz-server] 14amCap1712 opened pull request #3164 (03master…popularity-incremental-granular): More granular incremental stats https://github.com/metabrainz/listenbrainz-serv...

[listenbrainz-android] 1407jasjeet opened pull request #547 (03main…main): Bug fixes https://github.com/metabrainz/listenbrainz-andr...

mayhem: i think the stats are outdated because no listens since 31st jan have been imported into the cluster.

[listenbrainz-android] 1407jasjeet merged pull request #547 (03main…main): Bug fixes https://github.com/metabrainz/listenbrainz-andr...

oh. I guess everyone stopped listening... ;)

full dumps taking almost 2 days and that blocks incremental dumps too.

a total of 54 hours to produce and upload the dump.

we're going to need to think up something.... "better".

what we could possibly do is split the dumps for spark listens to be seprarate from rest of the stuff.

something more scalable, but so far I've not got any ideas.

we could stream listens inserts and deletes to spark cluster using rabbitmq.

yes, I think that makes lots of sense.

but also processing ALL the listen data for each dump seem so wasteful.

if it were not for deleting listens... this would be easy

hdfs is not optimized for realtime writes though so i would probably do the inserts hourly.

yeah true. supporting deletes is hard.

that sounds like an impedance mismatch with reabbitmq.

you'll have to receive the listens and store them somewhere?

sure we can do that, write to redis or a local file for temp storage. upload the file at the end of the hour to hdfs.

or we could just leave it to accumulate in rabbitmq and have a job to connect after an hour, process, upload and sleep.

is rabbitmq ok with that?

hour is of course configurable and we can make it more granular or coarse depending on listens load.

that would be the easiest.

should be methinks.

ok, sounds good.

will have to check how much memory usage it takes before i can be sure. but let's start at that and then do something else if need be.

the other thing that happened in the last few weeks is that parquet dumps have become 3x in size so i'll debug why that is too.

<mayhem[m]> "@fettuccinae:matrix.org: ok..." <- thanks for the detailed review, ill tidy it up.

[musicbrainz-server] 14reosarevok merged pull request #3462 (03master…genie-sidebar): MBS-13921: Add Genie links to sidebar https://github.com/metabrainz/musicbrainz-serve...

.

.... ..

-.-.--

..._..._

..-. ..- -.-. -.- / - .-. ..- -- .--.

-.-.

⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⢠⣾⣷⣄⠀⠀⠀⣀⣤⣤⣤⡀⠀⠀⠀⠀⠀⠀⠀... (full message at <https://matrix.chatbrainz.org/_matrix/media/v3/...>)

[listenbrainz-android] 14GautamCoder4019k opened pull request #548 (03main…adaptive_ui_for_orientation_changes): feat: made the ui adaptive for orientation changes https://github.com/metabrainz/listenbrainz-andr...

<-... .- -. --.>

Eh

<BANG>

lol

Welcome to another MetaBrainz Monday Meeting!

On my list for today: aerozol, julian45, monkey, yvanzo, kellnerd, zas, Gautam Shorewala, mayhem, bitmap, ansh, jasje, reosarevok, lucifer, Shreyas, MonkeyPython

And I have two mailed in reviews:

outsidecontext says: "Not much to say. My day job kept me busy after I had been absent the week before. That left little time for MB. But zas and I discussed a Picard release and plan to do one this week"

And derat says: "I filed MBS-13920 and will try to spend some time fixing it later this week, with an eye toward also supporting MBS-13711."

MBS-13920: Make it harder to skip warning about feat. artists in track titles https://tickets.metabrainz.org/browse/MBS-13920

MBS-13711: Prompt beginner users about likely capitalization mistakes https://tickets.metabrainz.org/browse/MBS-13711

More help to point users towards the right thing would be lovely, thanks derat

monkey: as the first from today's list who is clearly around, wanna go?

I will !

Hello everyone

🌊

Last week I did a lot of reviewing PRs

Had conversatioons with ansh and mayhem about potential GSOC projects

I worked with LupinIII (IRC) / MonkeyPython (IRC) on revamping the BookBrainz user guide and style guidelines

Also worked on separating user listens from the ListenBrainz feed (new page with streams of listens from users you follow, and similar users) which you can preview at https://test.listenbrainz.org/feed/follows/

And a bunch of ticket triage/creation and sentriage

That's most of my week

yvanzo: How about you?

HI!

Last week, we had two/three meetings about MusicBrainz search development.

Fixed SIR development setup and improved it to allow using Python 3.

Also reviewed collecting metrics for mirror/search performance analysis.

Published a security advisory for MusicBrainz mirrors, after CVE-2025-24814.

Fin, zas?

Still on my list: aerozol, julian45, kellnerd, Gautam Shorewala, mayhem, bitmap, ansh, jasje, reosarevok, lucifer, Shreyas, MonkeyPython

Hey

Last week, I ordered 2 new servers in order to replace hendrix/jimmy, in relation to the mainboard issue Hetzner announced.

monkey[m]: Looks like web is catching up with android now 😜

I also investigated few abuses on our websites, blocked few IPs (and also unblocked few users).

Apart that, usual PRs reviews, MB edits, upgrades, etc. fin. aerozol ?

zas[m]: Oh right, there was that too!

Not sure he's around yet

kellnerd maybe?

hi

Hey!

Last week I continued working on Harmony provider tests.

Almost everything is working as I want it to now, I only have to think about reducing the amount of testdata which is stored in the repo (for offline tests.)

Then I can actually start to collect eaningful example releases for the tests and commit them.

Other than that a bit of maintenance for my related other open-source projects like MB and LB TS clients.

That's it, ansh please :)

Hi Everyone!

Last week, I had couple of discussions with monkey regarding the GSoC projects for LB.

Also had another discussion on revamping the feed and the dashboard page to make it more personalized.

Did a few POCs for adding react suspense in LB.