#metabrainz

Morning 🗿🎉

Morning Pratha-Fish (hope they went well)

kellnerd: yess, they went surprisingly well lol

time to write a GSoC proposal now

Exactly, that's what I'm currently doing as well

kellnerd: best of luck!

Hey reosarevok, iirc you were going on a holiday from 26th right?

Just wanted to verify that so I don't end up disturbing you in between :)

Pratha-Fish: Thanks, you too!

Hi! I am exploring some more projects for sending my GSoC proposal, and stumbled upon "Make Use of auto-tag in Picard". However there is no project description there. Can anyone inform me some details related to it?

mayhem, outsidecontext: ^

moooin!

lucifer: I think the answer to that lies mostly with you. because working on this feature requires the right metadata to be returned from the new artist-release-recording mbid mapping lookup.

Also another thing that came into my mind for this project "Integrate Internet Archive in MusicBrainz" was that instead of integrating Wayback with Musicbrainz, an integration may directly be implemented for Brainzplayer (for the playback support between Wayback and Brainzplayer) yvanzo bitmap

s/this/the/

if that is something that can be completed soonish we could take this on as a project, if outsidecontext is up for it.

mayhem: its done to some extent already, i could complete that in time for GSoC: https://github.com/metabrainz/listenbrainz-serv...

lucifer: when getting images from CAA, is caa_release_mbid any different from normal release_mbid?

O’Moin

Moin!

komarmalade: If you mean “Integrate music streaming from Internet Archive” (not the Wayback) in BrainzPlayer/ListenBrainz, it would certainly be useful as well. However lucifer is more relevant to discuss the details than bitmap and me.

The Wayback Machine serves web pages, rarely with streaming, and even then it doesn’t have any structured metadata about the streaming if any.

The Internet Archive (that supports the Wayback Machine among other projects) supports collections of recordings with structured metadata, streaming capability, and API.

monkey: https://docs.google.com/file/d/0B-klwLEjaXWcZHR...

komarmalade: yes, adding support to BP for playing music from Internet archive could be interesting.

mayhem: hi, i was checking out navidrome's apis and i don't think it supports retrieving user's listen history.

mayhem: good one 💀👌

lucifer: ah, that's a bummer.

but is that needed? navidrome now supports sending listens to LB, so the most important function is present, no?

mayhem: yeah its weird. iiuc, they have an api to call and that then submits listens to lfm.

ohh, i wasn't aware they support sending listens to LB. their docs only mentioned lfm

didnt they just do a release with LB support?

that is how I found out about the project.

oh yeah https://github.com/navidrome/navidrome/pull/1424

looks like they do support it.

then there's only playback support left to be added

a past listens API endpoint would clearly be useful, but let others request that of them in time.

we don't need anything in LB to support listen submission from navidrome then. should work just fine on its own.

agreed. I am thinking about being about to cold start Navidrome users better.

not important for now.

since there has been some interest in the services projects, i am thinking to add a project to build a content resolver for apple music and add playback support to BP

go for it.

oh one more thing

and the detuning engine should be ready to drop in by the time coding starts, so plan for that as well.

deezer removed support for JS playback recently

I haven't gotten very far on it, but just two regexps are solving a lot of issues.

but the history endpoint still works fine so we can get listen submission working

sigh. should i remove that project from the ideas page then?

no its fine, i am thinking to change that to. listen submission and content resolver for deezer

k

instead of listen submission and playback

is there a playlist creation API so we can delivery recs?

i think there is but will have to confirm

k

also, thinking about on the autotag project idea. I don't think that is suited for a gsoc project.

there are tooo many random variables in play.

i see, yeah its more of an experiment

and the API endpoint that underpins auto tag is rather quite heavy in the data it loads.

makes sense

I think we need to have the correct metadata returned from LB, then make a round of improvements and then we can see what should happen next.

sounds good to me

too hazy to try and wrap into a project.

k

the IA sounds cool to me actually

a lot of that music isn't anywhere on spotify iirc

actually, yes!

spotify or other music services

and hopefully soundcloud soon too! :D

yup that too!

maybe we should be prepping for that one more, just in case?

prepping as in?

I think we can convince them to get us an API key.

putting up as a GSoC project.

ah yes, that would be great

the timing could work well for that.

i think ~3 students have expressed interest in those kind of projects so far and also started working on tickets.

so yeah could work out well

we just need to make it clear that we won't know until the 11th of april (or possibly) later that this project is a go.

sweeet!

have we gotten any uptake on the artist graph thing?

reosarevok: are you traveling yet?

yeah i guess we could ask the student to switch to another service later, if soundcloud doesn't work out.

nope no update on artist graph thingy

yeah, but lets make it clear that a switch may happen.

yup sounds cool

odd. I can never predict what students will go for.

and I'll write the daily jams april 1st gag today.

:D

this year's gsoc response has been lukewarm as compared to previous years

btw LB-1242

LB-1242: Dataset Hoster Improvements https://tickets.metabrainz.org/browse/LB-1242

these are the tickets i could think of, feel free to add more

everyone: We're all set with docker. our account has been moved to a free non-profit project that should give use all the resources we need for the immediate future.

lucifer: regarding LB-1242, it seems that we only need to add my idea of using linking MBIDs in results to start new dataset queries, yes?

LB-1242: Dataset Hoster Improvements https://tickets.metabrainz.org/browse/LB-1242

mayhem: oh that sounds cool, please add.

zas, atj: Opened a PR to try using --blkio-weight flag for MB prod cron container running on trille.

Deployed, will be tested for high latency alert on Wed.

mayhem: yes, in Málaga now, online at random intervals

lucifer: Deezer does have a playlist creation API

oh nice

regarding IA integration in MB, what I'd love is for the URL relationship editor to recognise when I paste a link from the IA, parse the original URL, mark it as ended and then have some sort of "(snapshot from YYYY-MM-DD)" link

+1

the way it works at the moment isn't great, e.g https://musicbrainz.org/release/8da7215c-bde6-4...

Do you mean a link from the Wayback Machine.

It matches the first item in IA/MB ideas.

Hey can someone kindly confirm if reo's on a holiday or not?

Since the deadline's coming close, I would've loved to get my PR reviewed if he's around today.. If not, can someone kindly check it out? (https://github.com/metabrainz/musicbrainz-bot/p...) bitmap maybe?

Hey Pratha-Fish: I confirm; Check the backlog: https://chatlogs.metabrainz.org/libera/metabrai...

yvanzo: Thanks a ton :)

Pratha-Fish: I am

Hopefully bitmap can test that? :)

If not, I can try to look tomorrow while at a rest

yvanzo: BTW I also wanted to confirm if you'd be a co-mentor for the "Automate areas management in MusicBrainz" GSoC project since I'd be mentioning all the mentors in the proposal too :)