Notes from #MetaBrainz Meeting 2023-03-27

[aerozol]

MetaBrainz Meeting 2023-03-27

Meeting start: IRC Logs for #metabrainz | MetaBrainz Chatlogs
Meeting wiki page: MetaBrainz Meeting - MusicBrainz Wiki

Agenda

• Reviews
• April

Last meeting: 2023-03-20
Next meeting: 2023-04-03

Today is International Heirloom Chili day!

Reviews

@akshaaatt

Last week I added some 3D stuff to LB android YIM
lso, reviewing a lot of stuff from prospective gsoc candidates
We are planning to freeze the apps for a month or two now
After a final release today/tomorrow
Because I feel our releases should be more focused and offer more to users through our blogs and stuff
We don’t have any bug fixing to do, hence we should now only release stuff once a month or so

@zas

last week, I worked mainly with atj about testing borg/borgmatic move to Ansible
the idea is to ease backup configuration, and to ease deployment of this stuff
apart that, usual maintenance, upgrades, user support, MB edits, etc…

@reosarevok

Last week I tried to figure out some schema change tickets, and we ended up deciding to drop one of them (about spammer ratings/tags) because it seemed like a patch over a very messy spammer system
We’ll try to figure out what we should do with spammers instead once we can get to a proper review of the whole thing + spambrainz
Other than that, I am on holiday, and I spent the day walking around, feeling hot (came from snow straight to 22C) and looking at things like funny ducks with blue beaks

@CatQuest

Worked with reo to add instrument images to several instruments (like mouth harps and so)
also added INST-1067 (tulum) and am now working on INST-879 (Biniou)

@snaekboi

last week was again super productive!

1. Except for the exams, I was able to do some more work on my PR on the MusicBrainz-Bot with the help of reosarevok & bitmap, where I experimented with Mechanize and added some features to enable addition of new areas with ISO Codes, External Links, etc. (Hoping to get that reviewed soon :))
2. Apart from that, I also worked on some experimental stuff like fetching new unseen areas from Wikidata, and comparing them with existing areas from the MusicBrainz database.
   I went the extra mile and tried to separate out spam areas from Wikidata from the good ones, since there still seems to be quite a lot of it on Wikidata!

I have also been working hard on my GSoC proposal for which the first draft should be posted very soon.
:))
Given the super-tight deadlines, I hope I’d be able to get some feedback from you guys on the proposal, and tailor it to perfection!

@bitmap

last week I worked on setting up a musicbrainz-server mirror with old data to test the dbmirror v2 packets once more (replicating up to the most recent packet numbers). going to leave that running and make sure nothing breaks
the team discussed entity naming in our codebase re [MBS-12552 (pre): Refine most Flow.js types] again, and I pushed some more changes there (this PR is more internal refactoring)
we also chatted about MBS-12794 (Don’t use spammer tags/ratings when calculating tag counts / rating averages) and decided to leave it out of the schema change release because we aren’t happy with the implementation and feel it needs more discussion
also, reosarevok noticed that the upgrade script for MBS-12256 (Keep rating and rating_count column on *_meta tables up-to-date with triggers) (from last schema change) didn’t fix existing rating counts, but I only found one that was inconsistent in the whole DB, so I just fixed it by hand
other than that I just worked a bit more on writing the design document for the json dumps improvements

@kellnerd

Last week I finally finished the BB Wikipedia extract integration and implemented caching for it.
Now we can already include cached extracts in the static page and don’t have to load them asynchronously.
As a “side effect” this also prevents abuse of the Wikimedia APIs.
In the same spirit I’m also dynamically generating a user-agent for BB, which required refactoring the way in which the current git commit is passed into the Docker container.
Over the weekend I finally started working on my GSoC proposal to import MARC records into BB.
I’ve studied lots of existing but unused code from a previous GSoC project and read old chatlogs to prepare my notes for a very constructive (and long) discussion with monkey today!
Now I’m well prepared to finish the first draft of my GSoC proposal, which I will hopefully post soon.

@mr_monkey

Last week I continued work on the redesign. It has now gotten to a good enough state to invite you to try it: test server [this server is not static/may look different by the time you click this link, forum users]
Word of warning: it’s still WIP, if you find bugs/ things that don’t look right, please report them to me
I also continued and finished LB#2419 ( Distinguish “no cover art” and “not mapped” listens) that improves the actions users can take on LB when data is missing in MB (link listen, add cover art, and catching potential errors loading cover art)
And I continued wrestling with the LB frontend test suite, we still have a Schroedinger’s test that fails only sometimes.
On the BookBrainz side I reviewed PRs and GSOC proposals and had some discussions around that.
Also read up on the multiple ongoing guidelines proposals
You’re doing good work, community!

@yvanzo

Last week, I helped with releasing MBS v-2023-03-20 and releasing its hotfix the next day.
Then prepared services on trille for zas to reboot the server and run Docker with systemd cgroup driver.
Started testing ionice (unsuccessful) and block i/o weight (hopefully on Wed).
Had a Thursday MB dev meeting here about schema change, search, and entity type categories.
Reviewed the list of genre-specific other databases linked to MB.
Videogam.in is gone. 277 (over 1350) Operabase URLs were not normalized.
Thanks a lot to derat for taking care of this data with their bot!
Plus usual PR reviews, support, tickets triage, maintenance.

@atj

last week i finished working on the borgmatic ansible role
performed a successful deploy on a test borg server
started working on adding the backup server in the office to ansible

@jasjeet

last week was very average
did some work
fixed listens service

@rob

monday was: Eat all the street tacos in Oaxaca. Tuesday was a mexican mole cooking class.
and then on wednesday/thursday I worked to make sure that the the taco glut in mexico was under control.
thursday I flew home.
not very worky, but man was it delicious.
I did some background work, a bit of PR stuff, but really not a whole lot other than some email.
we now understand why SoundCloud does not give out API keys!
and hopefully we can convice them to give us one – I have a meeting with them on April 11.
mostly about the fact the people ended up monitizing outside of the soundcloud platform and it was hurting their business and the artists.
fair, I can understand that.

@lucifer

last week, i worked on testing the listens table schema stuff, then hit an issue with vlan. will restart testing after that’s resolved.
also worked on adding stats for like top listeners and total listen count for any artist and album.
still a WIP.
helped some SoC applicants in getting started and multiple conversations about SoC projects/ideas.

@aerozol

I’m just in time to say… I worked on non MeB stuff last week! Looking forward to getting back into it this week. Done!

Quite a few people were missing or late, and it was agreed that it was because there was a location-specific timezone change that not everybody would be aware of, and that timezones and daylight savings are stupid.

GSoC Proposals

@rob

for the UX UI focused GSoC applications, we now have created a bit of a problems for ourselves
take a look at for instance: [forum post]
the proposal is mostly comprised of images that aerozol made. I think there are two in there that are original.
which is not bad at all, we want people to implement the UI that aerozol designs.
but I feel that we’ve lost some key tools for being able to understand and see the thought processes that students are undergoing.
previously if the design has horribly bad, you’d get a clue that this student might not be up to snuff.
anyone else see this as well?

After some discusson it is agreed:

@rob

ok, then lets put up a soft requirement for this year to ask students to implement a small portion of the UI.
and next year we can have a more coordinated approach to dealing with this.

@yvanzo shares GSoC links to: guide for applicants, guide for mentors

Absent regulars: Moot because of the weird daylight savings thing

[UltimateRiff]

@rob

we now understand why SoundCloud does not give out API keys!
and hopefully we can convice them to give us one – I have a meeting with them on April 11.

ooh, this could be exciting… for ListenBrainz, perhaps? or something else?

@lucifer

also worked on adding stats for like top listeners and total listen count for any artist and album.
still a WIP.

also very exciting~ I’d love to see if I’m the top listener to some of my favorite artists~

(probably the only listener for some, lol)

I am in general looking forward to ListenBrainz pages for artists, releases, etc. tho~ that alone will make it feel more like a cohesive product to others, I feel~

[rob]

@UltimateRiff

ooh, this could be exciting… for ListenBrainz, perhaps? or something else?

Yes, for ListenBrainz. We would love to work with them to be able to play and recommend SC content on LB.