#musicbrainz-devel

JesseW: I've done works and recordings, but making it all work for everything else is pending

plus a comment on the current PR as far as works/recordings, but :)

the types are added, and once it works for everything that has a type there's a hack that can be removed

I wasn't sure how far you'd gotten with some of this stuff, there's some that looks like it could be factored out into a role, which I also haven't really done any of (e.g. the Data:: classes' find_by_collection are all pretty similar)

ok, thanks, that makes sense

anyone knows where the picard script is located?

Nyanko-sensei: not sure what you mean

https://github.com/musicbrainz/picard is the source code, if that's what you mean. if you mean where does it store something like the taggerscript, I'm not sure but I assume just in the usual settings file

the tagger script from "advanced settings"

it's just stored in a config variable with the rest

on a linux system with standard settings it's in ~/.config/MusicBrainz/Picard.conf, forget where it puts it on non-xdg system

s

thanks! will search for that file ^^;

all these symlinks make my 120GB SSD host 10TB of files. OTL

o_O no such file on windows

yeah, that's a linux pathname, I'm sure it's somewhere else there

not sure

so... is it just me or did replication for musicbrainz-server change recently? I'm missing bin/replicate on a fresh build.

ah... or bin/replicate is in the VM only.... that's what it's starting to look like.

That script isn't documented anywhere? Or checked into a repo? I'd download it, but I'm on a plane.

randybias: https://github.com/metabrainz/chef-cookbooks/bl... is the source file for that script exactly, the rest is elsewhere in chef-cookbooks or in musicbrainz-server itself

perfect, thanks.

most notably there's a crontab added as well

That was the hint I needed.

you can, of course, just write your own scripts too :) but anyway

(most of us just put admin/cron/slave.sh or admin/replication/LoadReplicationChanges into crontab directly, since that replicate thing has handling for a lock file, start/stop, etc. so that you don't have to edit the crontab

)

Yeah, I just had used that before and now I'm building from scratch without the VM handy for reference.

Got it.

Gentlecat: the new site is paying dividends. :)

and I've got a long list of people to still hit. :)

you've said that like 3 times this week :)

great!

can you tell I'm excited? I think pretty soon we can start thinking about hiring a sysadmin/devops person.

I think that will be really nice for all of us. :)

hopefully I won't have to deploy CB and related stuff then, which is good :)

we still need to figure out what exactly we want this person to do.

but deploying some updates may be something you won't escape for a while. :)

ruaok: fyi if you get a notification, I'm going to download json dumps from ab

I dialed those way back. should be all good.

oh crap

forgot. the internet here is terrible

ruaok: how is performance, btw? I remember you wanted to attach some monitoring

3 days ETA

thanks guys

performance of what the site itself?

yeah

and replication packet stuff

sakura remins utterly bored.

great

bump https://github.com/musicbrainz/picard-plugins/p...

it's awesome seeing acousticbrainz data improve whenever i resync my library

hey dwni-randy

well, there's more of it. I don't know if that means it's better :)

alastairp, the high level data, i'm looking at a change to a track in my library right now

http://acousticbrainz.org/eb756146-9ceb-46ba-98... for example

used to be marked as happy for some reason

Project metabrainz.org build #92: SUCCESS in 40 sec: http://ci.musicbrainz.org/job/metabrainz.org/92/

tsukanovroman: Adjust max-height for logos on Supporters page

Gentlecat:you read my mind. :)

ariscop: right, that's not actually a change due to improvement :)

it's a change due to randomness

oh?

yeah. my guess is that there was another submission of the same track, and there were enough differences that the high level results differed

Gentlecat: deployed

alastairp: we should add a way to view different high-level data, maybe compare it

and we probably return the most recently submitted when asking for it

ruaok: looks better

Gentlecat: yeah, we were waiting on some stuff here to go ahead with that subproject, but maybe we should just do it

now that we have a good number of duplicates we should be able to start characterising differences, and seeing if we can find an influencing factor

would be a good addition to this http://acousticbrainz.org/770cc467-8dde-4d22-bc...

the fact that high level can change based on slight differences in encoding is enough to show that this is a bit of a joke

and we need a better way of doing genre, etc

what are you interested in showing?

"we have 3 genres for this track"

maybe just N top predictions

well, someone could just see that there's 4, and get the the 4 high levels

I think this would only be useful for the web view

(!) warning: we have conflicting estimates for the genre

OK, new companies on my radar this week: AXS and maybe Vevo.

Lets see what they do. :)

hello, I'm back again. I am going to release isrcsubmit 2.0.1 tomorrow or at the weekend. If anything needs a change before that, now is a good time to bug me.

Freso, Sebastinas: ^

welcome back, JonnyJD !

ruaok: blocked out

;)

ruaok: I just did a whois, and I'm chuckling quite a bit now

registration date, especially

so I'm having a little noob trouble. I'm trying to port the search server's boosting/analysis configuration to solr

I need to use fully qualified Java class names in the solr schema.xml file

i've tried tinkering with my classpath...

do you have experience doing java stuff?

yes...but apparently not enough

lots in school, but this is a different beast

weeksio: first thing to do is to paste some error messages into pastebin or github gist and post the link here.

can you post some errors? it sounds to me like you don't have the correct directory tree set up for the declared path in the source file

how are you running solr? did you download a jar? war?

are you using a container servlet, or built-in jetty?

(just say if you don't understand any of these words, I'm just trying to work out your level :) )

I'll get some errors up

solr 5.1 doesn't need a container...

solr 5!

news to me

I maintain https://github.com/alastair/solr-mvn-template, which is an empty maven template for a solr server

it has a directory tree which lets you put in your own source files, and then compile the whole thing into a war

perhaps it could be useful - it means you won't need to compile source manually and put it in the right place in your classpath

perhaps...

right - that looks like lucene isn't in your classpath

I'm finding out what I learned in my last career...you really start learning when you get a gig