#metabrainz

Maybe we should add anti-spam prompt notice to 'Create account' page of wiki too, it blocked me when I tried to register several days ago :)

Also confirming email and visual editor seem broken

yyoung, yvanzo: I found one frustrating issue with keyboard navigation on the new URL stuff:

If I just want to add a new URL, I used to be able to just paste, it autoselects, I press Enter, it submits

Now I press Enter, and it locks - and it selects not the next text field, but the [x] button, so I press Enter again trying to submit, and it removes it again :D

I'd expect it to go to the *next* input field automatically - so, the type if unset, maybe? and the next URL text field if autoselected, probably

(and to allow me to at least submit the form if I press Enter on an empty field)

Sounds feasible, I'll have a try

I'm afraid we can't change the behavior of Enter since it's used to submit links too

But we can adjust the focus :)

Why can't it submit links if there's a link, but submit the form if there's nothing? :)

Yes that seems possible too, just need a test for conflicts :)

There should be an easy way to submit the form from the URL section, so that's the easiest one I can think of

[musicbrainz-server] 14y-young opened pull request #2215 (03master…MBS-11680-fix): MBS-11680: Allow submitting form from empty URL input https://github.com/metabrainz/musicbrainz-serve...

mooin!

zas: what do you think of my idea of moving stats & nagios to the same dedicated server?

one of the AX servers ought to be more power than needed and cheaper than both those machines.

I'm OK with it

reosarevok: Opened a patch PR.

we should move more services away from google. I think once single 60eur server from hetzner could replace the $359 setup from google.

the two busiest VMs (nagios, stats, I guess) are $80 before sustained use discount, but before IP address markup.

IP addresses are $55 alone. just the IP address can cover a nice new VM.

it was all fun and games when google was handing out free credits.

We can order a machine, I'll migrate stuff over next week

awesome, thanks. :)

ruaok: should I proceed?

do you have a machine in mind?

AX41 at Helsinki is 34€/month

with 2x512Gb SSD (which is enough for the planned usage)

even with stats?

~80GB on stats.

yerp, seems so. 34/month. we pay more for IPs at google for that.

go for it!

ok, on it

ordered

cool. I like saving $300/month. :)

To start with, I'll migrate both nagios and grafana/influxdb to it, we'll need to update nagios IP on all monitored machines

but I think about moving to prometheus/grafana, getting rid of influxdb and nagios, it will be done in near future

sweeeit. :)

yyoung: changed this to point to beta, because we will want to release it straight there probably :) But that means there's one extra commit (to master) we don't want, can you get rid? :)

Either drop the commit on your branch or cherry-pick on beta I guess

"I wonder if that design overhaul is still coming after everything has been shifted to react?" - I assume this is still the goal? https://community.metabrainz.org/t/adding-cover... if anyone wants to respond.

reosarevok: Done

Freso: I'll tease a response. :)

[listenbrainz-server] 14jdaok opened pull request #1581 (03master…make-pins-listenable-on-brainzplayer): Make pins listenable on BrainzPlayer https://github.com/metabrainz/listenbrainz-serv...

the fins, whom I suspected of drinking more heavily than the bavarians, seem to be more on point, zas. the server is ready.

hey, wasn't it ordered only an hour or so ago?

Is there a way to retrieve a recording mbid from a track mbid besides using the entity search on the musicbrainz website?

lucifer: 90 minutes, yeah. :)

that was awesome! :D

!m hetzner finland

You're doing good work, hetzner finland!

jasondk: do you really mean a track MBID?

which is different from a recording MBID...

ruaok: yes :)

why are you using track MBIDs? how did you arrive there?

Well my listens are imported from last.fm so most of them contain lastfm_track_mbid's

So i'm trying to see if there's a way to link these back to recording_mbids so they are reviewable

are those really track MBIDs or recording MBIDS? I kinda doubt that last.fm knows the difference.

reosarevok: is there an API endpoint for this?

They seem to be distinct

for example https://musicbrainz.org/recording/03ec6e14-40e9... doesn't exist (track-mbid)

but if i search the mbid on the website i get redirected to https://musicbrainz.org/release/01925076-929d-3...

which links to the recording

indeed, that is a track MBID.

I think we should ask reosarevok about this. I'm not sure what to do here.

Sure i just wanted to double check if there was an endpoint

if not, if we decide to add one, adding the implemention to the reviewer is easy :o

reosarevok: Found another bug when adding another relationship to unsubmitted links, updated patch commit in the same PR

We have an /mbid endpoint on the site

I don't remember if the ws too

yyoung: my review was for your first commit, didn’t test the second one.

yvanzo: Saw it, investigating now

Oh yeah, what yvanzo suggested seems smart :)

Updated :)

I have a question about the Spotify API imports in ListenBrainz.

Now, Spotify API allows to query a 'recently listened tracks' with the limit of 50, pagination possible.

When I first linked it, it seemed to take listening history before the account registration-linking.

Now, it only shows listening history beginning from the time I linked the account.

Interesting thing is, that pressing next page on the actual last page loads for a while (didn't look at netw tools, mby timeout), then grays out the 'next page button'. Pressing the last page button returns just gateway timeout, the 'one of our servers is sleepy'.

I looked at the code, didn't find much of interest. Didn't spend hours looking at it either.

Is it supposed to be getting the history before the account linkage? I think it should clarify in the Spotify linking page, if it does.

user in question https://listenbrainz.org/user/d1c14465

..if not, I already wrote a thing to parse ~45k track history from the spotify data takeout, and shove it in to listenbrainz API, so use this instead then? https://github.com/jtagcat/dotfiles/blob/main/s...

param suggested I'd ask here, and suspected that the tracks I saw from the time before linking Spotify were the 'last 50', and ListenBrainz just gets the last 50 (or 20 by default).

hi piwu.

that is something that I've never considered and I don't think any one of us did either -- I suspect that we just go and import the last 50 tracks from when we're given access.

alastairp: lucifer ^^

i used to think this was intentional. that we only import since the time user gave us access.

i also think we import the last 50 tracks, unless something has changed in the past few months

my understanding was that spotify's "recently listened tracks" only gives you 50

at least when we first set it up, it wasn't possible to get more than that

i rewrote some parts of the spotify reader so there may be a regression.

data takeout importer would be incredible, though!

piwu: what do you mean it only shows listening history from the time you linked the account? this is on your user page on LB?

alastairp, i think we can now import from a long time ago as well. we just need an appopriate starting timestamp and then can paginate the api response 50 at a time.

yes, this says that the max limit is 50, but after/before parameters are new to me: https://developer.spotify.com/documentation/web...

fwiw, LB uses them already so when a user connects we could set the latest listen ts to a long time ago (spotify founding or LB founding)

this is also related to a user request on forums to run the Last.FM importer in a specific range of time as in letting the user choose the time from which they want the import to start.

huh, it was supported in 2018 too... no idea why we didn't take advantage of that :( http://web.archive.org/web/20181006015030/https...

AcousticBrainz stat stuck again?

ruaok: Thanks for the forum response. :)

MrClon: yes - the bug is fixed, but we didn't make a release yet. I was waiting until today to make a release with some other bug fixes too

Thanks

It would be great add to "datasets" feature some description. I did not fully understand what it is

we have 3 new datasets features that we hope to release this year. We can definitely add more documentation about what this is for

I already have a planning ticket for it, which I created last month! https://usercontent.irccloud-cdn.com/file/fTHGm...

ruaok: Before I log off for the weekend, what do you think about calling off meetings next week and/or week after next? There was some talk after the meeting this Monday as a number of people are more-or-less holiday'ing, but I don’t think there was any clear conclusion. I also personally have a couple of medical appointments (incl. my 2nd covid shot! 🙌 ) as well as exams these two coming weeks, so having a (tiny) bit less

on my plate would be welcome, but meetings are not a huge workload either way.

I'm ok with cancelling the meetings for the next two weeks. three even -- my holidays don't really start until the 20th,

Alright. Let’s make next meeting Sept. 6th and enjoy the last few summer weeks. :)

:o

Happy holydays MetaBrainziacs!

!m MetaBrainziacs

You're doing good work, MetaBrainziacs!

:D

just hope that nothing breaks catastrophically in the next 3 weeks

https://twitter.com/bryanmjohnson/status/142615... this is very good

Hi Freso, , since the final weeks of GSoC are approaching, I think we can have a meeting during the end of August in case we want updates from everyone. I'm fine with the holidays though, just putting forward a thought. 😄

akshaaatt[m]: see pm please!

Sure thing!

Hi all, I invite everyone to drop their opinions regarding the MusicBrainz website at https://community.metabrainz.org/t/musicbrainz-...

Thank you for your time!

akshaaatt[m]: let me consider this

(I have a few user css. one thing it does it give the whoel webist a grey frame)

Oh okayy