#metabrainz

reosarevok, bitmap: Slightly related to API docs, does MBS-12093 seem reasonable to you?

MBS-12093: API endpoints for entity attributes https://tickets.metabrainz.org/browse/MBS-12093

Gah, thanks Apple.

monkey: to be fair, it's not like it was registered in the public ports database as "for LB use only" either ;)

For what it's worth, I'm in favor of the change to 8100 (or whatever port is free and generally available)

Got it, thanks for the help!

But for some reason its not taking the input.

Also, I should mention that I havent setup my local development environment yet 👁

Oh nvm, it worked somehow

reosarevok: The rebased PR has conflicts on PaginatedResults.js

Hmm. Wonder why it didn't complain

Will check

Oh, or maybe it's just new ones

Yeah. Rebased again, yvanzo :)

hello, everyone. i can't set up a development environment for listenbrainz. it raises an exception https://0x0.st/o-nB.png

i did everything according to the readthedocs page

treeshateorcs[m]: uhh, sorry that looks like one of my recent patches gone wrong. i'll try to fix it soon.

outsidecontext: will look into it.

oh, sure! can't wait. thank you

lucifer: thanks. it just rather randomly imported a single listen again. Interestingly I had paused playback for a while, started again. The very first listen got imported, but the rest fails again

Alrighty, another roadblock.

Uploaded file: https://uploads.kiwiirc.com/files/9ad4f62dbcca6...

I dont think this script can be run on windows

I tried running it on gitbash, but its giving errors

Trying powershell now.

you'll need docker installed PrathameshG.

so WSL

lucifer: yes, I've installed docker.

Ah crap. Havent initialized docker yet

*WSL yet

docker for linux right?

I am on windows, so I've installed docker for windows. I havent completed the WSL2 setup yet

right windows docker has 2 modes, 1 for linux other for windows.

I think my 4GB RAM laptop's gonna die by the end of all this 😭

@luci

oh yeah docker/WSL is very memory intensive :(

lucifer: is there a way to get it running without WSL? I need as little overhead as possible

I can try doing it all in linux dual boot as well, but damn that's gonna take some time to get setup.

don't think so its possible without WSL. there used to be a docker app for that but afaik its even more memory consuming.

yea I directly downloaded the docker app for windows from their site. Is that a bad news?

also, by WSL, do you mean enabling WSL and then directly installing docker only in WSL?

install wsl, then install docker, it has docs on site how to make it work with WSL.

On it

but if you can dual boot, definitely do it,

looks like the best way to get around it TBH. It's gonna be more resource efficient than windows anyway.

The only problem is that half of my stuff doesn't run well with linux 😑

PrathameshG: give wine a try

treeshateorcs[m] wine's great, but such a hassle to setup. I also found a snap for foobar2000, but it comes with its own set of problems

I'd love to continue using linux, but goddamn it creates a ton of compatibility problems

deadbeef on linux has kinda the same feel as foobar

The biggest Problem is that I cant use anything otherthan foobar2000 due to my crackhead modified setup (especially the stuff with last.fm)

Uploaded file: https://uploads.kiwiirc.com/files/ef399c806dbc3...

No other player provides such field mappings for last.fm afaik

If you know any, please let me know I'd be happy to move👀

you may try your luck with cmus and cmusfm (both are open source, so you can edit the cmusfm's source code to match your requirements, it's a smol program)

but let's not offtop too much

oh right. Forgot this is not discord lol

asubedy: hi, a SQL query is the only way to get recently-added releases at the moment. to get an MBID you need some kind of metadata to search for; we have a search API at https://wiki.musicbrainz.org/MusicBrainz_API/Se... or if you have direct database access you can of course just do a SQL query for that too

Howdy; Querying the musicbrainz API for recordings with a certain word in `recording` key, and when using offsets to loop thourhg pages of results we get back a whole lot more than the `count` value sent by the result. We seem to get a lot of dupes, too, judging by the recording id value. Any suggestions on how to improve? I'll post the query in a sec

bitmap: hi! looking at https://github.com/metabrainz/musicbrainz-serve... again - when you suggest doing all with SQL, would that still involve doing the cursor stuff, but just running SQL deletes instead of $c->model('Editor')->delete?

reosarevok: that's correct, the cursor stuff is still needed to keep each delete in its own transaction

Ok :)

cursors?! oh no.

So when I get this back in a query: {"created":"2022-01-17T17:26:09.742Z","count":272,"offset":0,"recordings":[...]...} What does the "count" key represent? number of total items (recordings) in this case matching the result, or total number of pages of results at whatever the "limit" is?

IIRC it's total items, but IIRC also someone complained recently that the search server didn't keep the order with offset...

yvanzo, do you remember?

Also hello, and thanks for your help and any work you do related to the MB db & API. =)

The second question -- why do I get a lot of repeated "id" values for recordings in the results? I think even within a single page's worth, there are dupes. Is there a way to force DISTINCT on id or something along those lines?

EyePulp: can you link to the query you're seeing this behavior on?

@bitmap I'm fine posting an example here if that's acceptable, or a gist if preferred

<bitmap> "asubedy: hi, a SQL query is..." <- Oh ok then there would be no way to get the recently added releases on the MB revamp without actually connecting it with the database // akshaaatt

So can we do this, instead of recently added releases we can show cover arts of the most listened musics, for which I guess there is an API or something like it through ListenBrainz. (I saw in the chat earlier). Any suggestions on this?

@bitmap https://musicbrainz.org/ws/2/recording?offset=0...

EyePulp: thanks, just the URL is fine

bitmap: what about stuff like $self->c->model('Editor')->unsubscribe_to($editor_id); ?

Run those as code or as SQL?

(to remove other people's subscription to this editor)

I don't expect many to exist, but :)

fish

Or $self->c->model('EditorLanguage')->delete_editor($editor_id);

I'd personally prefer SQL for everything for clarity

Ok

Again asubedy, the revamp repo is not official but I do like that idea. I would be fine with adding that there, nice learning.

squilfish

EyePulp: if I do offset=271 for that I only get one result

you said you are getting more than `count` somehow? I guess from the duplicates?

@bitmap is the count the total number of pages then?

it's the total number of results

CatQuest: sunfish?

moonfish!

fish!

bitmap: Hmm, I think we're not checking whether their collections have any collaborators - not that that is a likely case

EyePulp: so to get the next page you should increment offset by the limit (50)

@bitmap Ah! offset isn't a page offset but individual result counts.

yup

EyePulp, cc reosarevok: yes, total items. The issue about order with offset is SEARCH-667

SEARCH-667: Missing Release Groups and Inconsistent Responses when Searching Official Release Groups for Artist via API https://tickets.metabrainz.org/browse/SEARCH-667

I think any inconsistencies you'd see otherwise are in theory due to index updates

bitmap: I think if the collections are empty, it should be fine to delete any collaborators as well, or would you rather I checked there are none in the query and left those unremoved?

hmm. if they are empty I think it's fine to delete...especially since it'd have to match all these other criteria for the editor being empty, which should be rare

Hmm, although I think our idea earlier was "the collections are untouched"

So maybe we should discard collaborator ones. I expect the number difference will be 0 or 1 at most...

yea, I think it will hardly matter either way, so whatever you prefer

@bitmap - beautiful! results are matching expectations now! Thanks so much for the clarification.

cool, glad it's working now!

<BANG>

It’s Martin Luther King, Jr. Monday!

I have received no mailed in reviews or notices of people outside the usual crowd who want to give review, so… people up for reviews: lucifer, monkey, reosarevok, bitmap, akshaaatt, zas, alastairp, mayhem, yvanzo, CatQuest, Freso – anyone else who wish to give review, let me know ASAP. :)

lucifer: Go!

woa

oh

hi all!

i worked on seeking and implementing feedback on various already open PRs last week. we deployed the long due listen user id PR. and started work on the next steps.

another long due PR to fix listen counts in LB also progressed along nicely. other than that, worked on a few improvement of the artist origins map. when hovering over the country, it now shows the artists name as well.

finally some misc bug fixes and discussions. fin.

monkey: next?

Hi everyone !

Happy new year !

I just started working again today, so don't have a lot to say about the past three weeks. I was vacationing in the US, got married, caught covid again, managed to get back after more delays.

Before the holidays, I mostly worked on the Year in Music project.

This week, docs for BB!

That's it for me, bitmap go !

last week I was mainly working on speeding up area-related queries for the website, since we've been getting alerts about those pages causing excess load

my plan is to bring back area_containment, which was previously a view, but will be brought back as a materialized table. (those have worked amazingly for the artist pages)

I have most of the triggers done and am writing pgtap tests now

Ooh, neat

other than that that, I opened a couple other PRs to fix spammy warnings in our website container logs, and to remove support for Amazon cover art as decided last year (which lets us delete a ton of code, woohoo)

that's about it I think...fin, go reosarevok

Hi! I was on holiday and traveling for most of the week

(People still up: akshaaatt, zas, alastairp, mayhem, yvanzo, CatQuest, Freso – anyone else who wish to give review, let me know ASAP. :))

Then I worked on a few easy tickets created while I was away

And now looking into some changes to the MBS-9356 script so we can run it