#metabrainz

"Attribute (data) does not pass the type constraint because" is gonna haunt me for years

I'll gladly switch to Moo if it means we get error messages like https://gist.github.com/tobyink/5781007

Project pytools build #12: SUCCESS in 33 sec: https://ci.metabrainz.org/job/pytools/12/

uh, finally done

I'm actually like that more now. properly tested and portable

r/like/liking

reosarevok: regarding the auto-editors election process, there were some problems when you tried in January right? do you know if it was fixed?

nice

Project pytools build #13: SUCCESS in 31 sec: https://ci.metabrainz.org/job/pytools/13/

reosarevok, I have nominated loujine again, if it has not been fixed??? And seconders required if it has :)

ok, I'm sold, Moo + Types::Standard is everything I ever dreamed of

except MooseX::ABC, MooseX::Role::Parameterized are used extensively and probably break

sigh, so tired of OO in perl

:(

Project pytools build #14: SUCCESS in 28 sec: https://ci.metabrainz.org/job/pytools/14/

bitmap, do you see any way forward for getting rid of Perl incrementally, or does it have to be all-in?

hmm, well, it's not something I've thought about too much, since I've been too busy working on features

needs a lot more discussion to come up with a plan, if we ever want to do that

Fair enough. It's been in the back of my mind for a while, but it's non-trivial.

when we finish porting templates to React, that will remove some large chunks of perl

ws3 may be written in a different language, too, but that's still vaporware

perl is the language for 2016 and beyond

perl 6, specifically

it's going to be great

right, lol. though I'll admit perl 6 is actually really cool

Freso: what's the situation re: autoeditor group?

(oh wait, I guess he was at some festival or something)

loujine, MajorLurker: I don't think there's any automatic mailing yet, but *maybe* we have a way to contact all autoeditors now. Freso will know, once he's back...

He'll probably even know before then, but it will be more difficult for him to tell us.

... you're a terrible human being, Leftmost

Also correct :p

ruaok: I have no idea about how any of the donation stuff works so please take that one :)

(but not today I guess, it's DO NOT WORK day :D )

reosarevok: lol.

the bottom line: make a donation, get credit for $donation / $4 per month credit.

$4 = 1 month, $8 = 2 months.

so, tell him he is right, new donation needed. :)

Tricking me to work on Labour day. Dirty capitalist.

Ok, will do :)

ha. that was just info, not a demand to work.

:)

But you know I'll forget it all by tomorrow otherwise :D

Ok, answered

suppressing snarks sounds almost like work

Leftmost: we did come up with a plan for ditching Perl in Berlin :) I'll try to remember what it was

zas, you are going with SOLR for the new search servers right?

reosarevok: Situation re: autoeditor group is that we need someone to write a plugin for Discourse to deal with it.

zas: Please do add your name when you add a topic so I know why I have to poke about talking about it. :)

loujine, MajorLurker: so, there's still no good way to let all autoeditors know there's an election going on :(

bugger, so should I just cancel it again? That´s unfortunate.

Leftmost: Re: gamification, I agree with your sentiment.

LordSputnik Leftmost, re: Google Fonts API, we could possibly copy the fonts over to staticbrainz once that comes around. Or does Google Fonts provide something more than just the font?

Freso: I think google font provides he best option depending on what the browser supports iirc

i.e. OT, EOT, WOF, or SVG

also if you are using a popular google font like open sans or lato then it's likely to be cached on the users machine already anyway so no extra download

Ah.

KodeStar: The second part is kind of what staticbrainz would also alleviate. :) https://tickets.musicbrainz.org/browse/MBH-413

bitmap: Is https://community.metabrainz.org/t/musicbrainz-... something you can answer?

Or Gentlecat perhaps?

Freso: yeah, that would work for regular visitors

hi all, just out of curiosity (I am busy with freemusicarchive seeder - testing it now) but something I am thinking off is ... how do we link things like the FreeMusicArchive album id to the MB release id ? else we just end up with another release of the same album with a different external link... surely catching / storing the freemusicarchive id's againgst MB id's make more sense ? or am i on the wrong track here

like discogs id's

madmouser1: Are FMA id's "URLable"?

If so, it might make sense to have FMA added to the "other databases" whitelist and you could link up MBID→FMA ids.

hi Freso. no but it is in the API

Freso: let me quickly check if tere is a correlation

Freso: I can't find a correlation between FMA URL & FMAid, but teh FMA id is in FMA API for each album & track and that can be referenced in the api.

can I pass (via web seed iow. add realease a "other database" id ? that way I can seed the FMA id for each track and album.

I can get the FMA URL from the FMA id via the API ...

madmouser1: Can you get the FMA id via an FMA URL?

Freso: No you can't, currently I get it via jQuery in the page then i make an FMA API call to get all the info

two examples below just to demonstrate

FMa URL : http://freemusicarchive.org/music/Gorowski/Marm...

FMa API for that album: https://freemusicarchive.org/api/get/albums.jso...

the APi have a album_url property that link to FMA URL

I'm thinking about centralizing logs from all servers we run, currently log rate is about 1Mbytes/s (all servers cumulated), this will multiplied by 3 at least soon, and to be safe i target 10Mbytes/s . If someone has experience with a tool able to centralize syslog logs as well as user-defined logs (our nginx logs are not in standard format, and we have

various apps around with very specific log format), i'm curious about that. For now, we just want to centralize, but in the future parsing/analyzing/etc ...

so what I want to do is push (link that FMA album_id to a MBID

zas: what about logstash into elasticsearch for the logs

zas: logstash have grok patterns you can use for logs

madmouser1: yes, that's a possibility, do you use it atm ? at which rate ?

ofc i saw fluentd, nxlog, etc... but my question is more about actual experience with those tools and performance/resources/hardware

zas: yes i do, not at a 10mb rate, at most i do few 1000 log lines from about 10 sources

per second

but logstash can handle it and Elasticsearch can be in a cluster (for scaling)

madmouser1: I don't think there's a good way to have FMA ids in MB at this moment though. We'd either need to add a new field for it (unlikely to happen) or get attributes ("single point releationships") ported to releases. Either approach would be a schema change and would thus not happen before November at the earliest.

s/though/then/

But we can probably link to the FMA pages anyway, using either "download for free" relationships or something else. reosarevok would be the one to know that.

Yeah, I've linked to FMA before I think with download for free

Freso: noted, yes I do a "download for free" link at the moment, but just thought it is such a waste to loose a opportunity to pass the FMA id to MB

(you can also store the license as a URL)

Is there a way to get from the URL to the ID?

no

currently i get the id from the "embed in page" and tehn to a API call

but as I have both pieces of info in the seeder I thought .. let me ask :-0

)

I mean there's nothing wrong with seeding the ID to the annotation at least

do MB store the notes for each import somewhere ?

or a "comments field" then I can seed the FMA id to that, else I will just leave it for mow and revisit later