#metabrainz

Some last questions remain, and possibly the need to rewrite some parts to turn a hack into a more proper feature.

reosarevok: definitely a "coming soon" feature

reosarevok: yes. the taglookup feature should be replaced with that.

Neat!

:o

Also, Mr_Monkey: neat! Where do I find time to *also* send data to BB?

Also prepared for the hack weekend with good discussions, and had a grand ol’ time hacking away at the new playlist feature this weekend with ruaok alastairp shivam-kapila and _lucifer . We're a god 75% of the way there, I'd say.

Can someone donate me some hours?

reosarevok: i do week/week thing. idk

god 75%? what happens at 100% god?

Going to work this week to finish it up and wire all the separate parts together.

Good Omens

ruaok: bad things, I think. And then PeopleOtherThanUs go to heaven eternal?

Something like that :)

Also some continued forum discussions and some issue triage, and preparations for the next few BookBrainz projects.

That's it for me ! CatQuest go !

Meow! `😺`

Last week I worked on irombooks, instrument images,

added melophone (a fun instrument that is a bellowed reed that looks like a lute!)

Freaky.

This week I will look at BookBrainz work and edition formats, bookogs people have brought much interesting to the table, and I will research this thoroughly, as agreed with Mr_Monkey

♏

fin, go reosarevok

Hi!

This week I've been working on more React conversions

Including a lot related to cover art, since bitmap teased that he wanted to finally look into the EAA

(that will take a while though)

:O and maybe BAA as well?

EAA = event art archive

CatQuest: is that an archive about sheep?

no books

Oh, books, I guess

I liked the sheep better

but yea i was looking for a sheep emoji

But maybe? :D

Eventually.

whoooo

soooonnnn

🐑

Today I worked a bit on putting info together for the blog post about our messy mess

And also got the sad news the BBC just dropped their artist pages

:(

[musicbrainz-server] reosarevok opened pull request #1805 (master…MBS-11247): MBS-11247: Stop highlighting BBC Music links https://github.com/metabrainz/musicbrainz-serve...

So yeah ^

Asked them whether they still do stuff with MB or what - seems that they have some obscure pages ruaok showed me that do use it, and I got yet another support email today asking for the artist's ID for BBC Introducing so I guess they still ask for them there

So not sure what the BBC / MB status is

But losing the pages is kinda sad. Oh well.

oh sorry.

archive?

there was a part of the email I redacted. the BBC uses our data very much in a lot of places.

we are the ID space that ties quite a few of their systems together.

CatQuest: yeah, some are archived, but I meant just as the user-facing "look ma we're on the BBC" thing

yea , thats

this change, does not mean that the BBC is done using MB.

Oh, neat. I didn't know if they were looking into migrating from those again. Thanks ruaok :)

s/again/as well/

That's about it for me. Freso?

🙋‍♂️

I haven’t been doing so sharp this last week, which means I have also done very little. :( Just a bit of basic things.

fin, yvanzo go!

Hi!

(People still up: bitmap, zas, alastairp, shivam-kapila, _lucifer, jmp_music – anyone else who wants to give review, let me know ASAP!)

Last week, I worked on updating search components according to latest MBS changes.

I rearranged Jira’s system dashboard for MB/GH SSO and correctly reinstalled that add-on.

Plus usual PR reviews, support. Fin.

Go bitmap?

hey

so I of course helped with writing & deploying fixes for the user data leak that ruaok talked about. fortunately, was able to find the issue and fix it on all servers within a couple hours after starting an audit

that all happened yesterday

!m bitmap

You're doing good work, bitmap!

earler in the week I rewrote the CAA-indexer in python (it's currently in perl), extended it to support events for the future event art archive, and greatly simplified the architecture by dropping rabbitmq in favor of a postgres table for storing the indexing events

(the amount of CAA traffic doesn't justify a dedicated message queue, and makes development/deployment significantly harder)

it's still being fleshed out, but the project URL is https://github.com/metabrainz/artwork-indexer

besides supporting events, I tried to keep the codebase extensible in case we ever want to add support for e.g. sheep/bookbrainz in the future. hence "artwork"-indexer instead of "CAA"-indexer

Wee! Thanks !

:D

other than that, I worked on fixing issues discovered on the beta server due to SameSite cookie changes and the recording "first release date" feature

Can Picard be run from CLI?

the recording date feature introduced a performance regression, but I've updated the implementation to resolve that, so hopefully it can still be included in the next release (if we delay that by a week)

agopo: No. Also, we’re in a meeting right now. Please hold follow-up questions for ~half an hour. :)

that's about it I think. zas go!

agopo: we're in meeting, but yes it can, though I guess you look for beets (Picard is mainly an UI app)

hey

zas: Wait, what? Since when? 🤯 /me makes mental note to look into this later

As said earlier, I gave an hand to ruaok for LB scripts

apart that, I investigate google search webmaster tools errors about few MB pages ending in 504s, and reported findings to bitmap/yvanzo

our rankings are not that bad, but we pay the lack of mobile compatibility, I hope the move to react will have visible effects on this side soon enough

I also investigated an issue causing all our services to drop for few seconds, likely related to consul stuff

I set up new telegraf stuff to collect more metrics in this field (and created associated graphs in grafana)

plus a bunch of upgrades (discourse/influxdb/grafana mainly)

and usual user support, along Picard code reviews/PRs

fin. alastairp ?

ZzZzzz hi

hi

I worked on some stuff that's already been mentioned by others, work with Mr_Monkey on brainzplayer iframes and stuff

I spent some time looking into typesense with ruaok, it looks great, and super useful for our needs

I made some changes to troi to improve how we pass commandline arguments to the tools that we are making with it

I started to add support to the dataset hoster to report programming errors to sentry so that we can track down issues more easily

I took up MBS-10388 again, to get more genres in MB, after it became clear that it would be really useful to have them in troi

MBS-10388: Add some genre tags to MBS https://tickets.metabrainz.org/browse/MBS-10388

I started some improvements to how Listenbrainz starts up, and how we use daemontools and consul

with the plan to roll out these changes over all pythonbrainzes

finally, I spent the weekend and some time last week making plans to store playlists in LB and writing a good chunk of code to make this happen, along with Mr_Monkey and ruaok. Hoping that we'll have some good playlists in LB very shortly

shivam-kapila: next up!

Or… _lucifer, maybe?

hi all!

(Only jmp_music (and maybe shivam-kapila?) left on my list – last call for anyone else wanting to give review!)

I worked on fixing a websocket bugs during the weekend hack. I also worked on turning the follow server into something we can use to allow multiple collaborators to update a playlist in realtime.

Hope to get that finished soon and submit a PR.

that's it for me.

jmp_music__: next?

Hey everybody!!

hi!

Last week I worked on outputing the confusion matrices and the results as gaia does. I finished with the CM's dictionary output and now I have already proceeded into the integration part in AB's sklearn tool.

Fin! @shivam-kapila next?

I haven’t seen any word from shivam-kapila, so I don’t think they’re around, so…

That’s all for reviews! Thank you everyone who gave reviews! :heart:

We have one more item on this meeting’s agenda:

zas/bitmap: consul upgrade

ok

being critical to the whole infrastructure it is quite complex to upgrade

more napalm?

it is time to plan an upgrade

and this needs careful planning else it may results to long downtime ;)

I think downtime can be avoided, deploying new versions of consul / git2consul in parallel of current setup

oh no

if the latest versions of consul-template are backwards compatible with the version of consul we use, maybe we can start with that

this ties in with the plans that I wanted to do with pythonbrainzes too, which I just mentioned above, too

in the past we discussed replacing git2consul altogether due to weird issues and not being maintained. but I'm not sure with what

I noticed some docker images/containers expect consul on port 8500 -> we need to ensure it can be easily modified everywhere

yes, git2consul is unmaintained

there are go alternatives, with recent activity

I wonder if this general meeting is the place to have this discussion.

I for one am exhausted. perhaps we should have a dedicated meeting to address this issue?

well, when do you think we can do that?

how urgent is this?

not urgent, but sooner is better

it can wait January (I don't want to mess up with that during Christmas holidays...)

ok, let me pick this thread up tomorrow and organize a meeting.

ok