#musicbrainz-devel

in the mb server schema description (http://musicbrainz.org/doc/MusicBrainz_Database...) the abbreviation AR is used where it discusses the link tables. what does AR mean?

probably Advanced Relationship

not sure how up to date this page is

tom[], AR refers to the way relationships are implemented as shown @ https://wiki.musicbrainz.org/MusicBrainz_Databa...

Why "advanced"? I'm not sure. Maybe to differentiate them from sql relations?

i think it was just advanced as in at one time the only relationships we had were release-artist and track-artist, so all the release credits were really advanced from what we had before (or at least I always assumed it was this)

luks: are you the person to direct questions about the Picard cuesheet plugin to?

(we got a question at support but I have no idea about it :) )

reosarevok: no, I'm not

direct questions about picard should be forwarded here http://forums.musicbrainz.org/viewforum.php?id=2

ruaok: don't know how much time you have for that, but I've set up a sentry instance on sakura

now I just need a subdomain to run it on

something like sentry.meb.org

hey.

I've added it to my todo list.

thanks :)

not sure when I'll get to that though. other more pressing things on my list.

wait? what? I didn't hear anything

alastairp: huh?

hi ruaok

how's the trip?

good so far. yours?

headed to Chaos Camp tomorrow morning. :)

we spent 5 days in the mountains. it was nice

I'm back here now. it's no way near as nice

i'll take another break in the last week of august

smrt.

the weather in BCN has been consistently cooler than in northern europe. again.

happened to us last year too.

have fun at Chaos Camp. I'm a little jealous

did you take anything with you?

not of my own stuff.

but we will have a cocktail bot bicycle, an ice machine, a grill and too much meat.

kinda of a bachelor trip to the camp. :)

Gentlecat: https://github.com/metabrainz/metabrainz-logos/...

?

ruaok: anything else I should be saying when answering to possible-customer emails btw? :)

better? monkeydo improved the lettering that was bugging you.

reosarevok: that was perfect.

Ok :)

The dots on the is look a bit weird, but couldn't say why :)

ruaok: hm, I'm not sure what's changed. still looks weird

yeah

"d" in Picard too

I guess it might be that the lowercase is almost as big as the uppercase, so the upper bits of it look a bit crammed

this font is strange in general

the dots make the letter higher than the upper-case

reosarevok: snap

seems incoherent

gah.

all of our nit picking is getting expensive. :(

at this point we can't change the font.

so, we just need to figure out where things need to be manually fixed.

I'm happy with the new picard logo, tbh.

Gentlecat: have a look at the latest commits from monkeydo. See if you can find concrete things that still need fixing.

I mean, logo looks nice. I just noticed that problem with the font and thought that it might be improved a bit

if it can't be changed, so be it

it can be changed, but we need to come to a stopping point.

so, please have a look and then if there is anything quite serious, make a new ticket.

I don't have anything else

so, case closed then?

because monkeydo has closed the tickets.

so, unless we re-open or make new tickets, we'll call it done.

we can always make adjustments ourselves, right?

sure, I don't see why not.

CallerNo6: drsaunder: tnx

CallerNo6: to the RDBMS and to SQL there's nothing different or advanced about these relations. these relations are advanced in the sense that you can represent advanced concepts with them or that it takes advanced application software to make sense of them

i guess

<sniff> bye bye wifi

http://tickets.musicbrainz.org/browse/SEARCH-397 says Instrument has support for tags but /ws/2/instrument?fmt=json&query=tag:(electronic)&offset=0 does not return any results

Are there any actual instruments with that tag?

I tried several ones like classical pop rock jazz but none of them gave anything

Is it safe to think the query would work when the db has some instruments tagged?

I would expect that if it works for areas and/or series it will work for all 3

areas also does not return any results for the above tags

but for series it works

Then I'd imagine it is fine and the others just don't have the tag

Because they all work the same way so it'd be weird if one got added fine and the other two wrongly :)

oh right then thanks :)

alastairp: can you clarify https://github.com/metabrainz/acousticbrainz-se... please?

getting False there if dictionary is empty is what we need

yes, but I don't like the double-check

alastairps if statement catches both cases ("highlevel" is not a key and s_d["highlevel"] is {}) at once

the "if x and y"

I prefer one statement

ohh, I see

it actually seems funny to me that the get_summary_data fills in the "highlevel" key if there is no highlevel data (if the hl extractor fails)

an alternative fix would be to change that

it took me way too long to understand your "if summary_data.get("highlevel", {})"

yeah, that's better :)

hmm, right. I thought it was pretty standard python

but if you don't use it it can be confusing

yeah, I'm just not used to .get()

on second thoughts, you don't actually need the , {}

because get returns None if the key isn't there

normally you use the second argument if you're assigning it to a value

and you need the types to match

right

Mineo: is there a way to edit the pom for the solrquerywriter so the solr-core and lucene-core jars aren't included in the jar-with-dependencies?

we need the jar with dependencies....just not those

because they are already included in solr and because the duplication causes weird request handler warnings

I don't know if that's possible, have you looked at the maven docs?

yes...but have I understood them fully? ...no :P

It almost fixes the problem to use the jar without dependencies, but I still need mmd2

well, they are not included if you build the jar from my master branch, so some change between my pom.xml and yours must make maven include them

ah....ooops!

I'd guess solr itself includes solr-analysis-extras, so you could try setting its scope to provided

but that's just a guess

alastairp: are you sure history file generation is supposed to be slow?

it took 0.08 seconds on a dataset with 40 items

no, I'm not sure

well

I'm trying to understand if we can just do that during initial evaluation

pretty bad 502s :/

because that would significantly simplify this thing

Gentlecat: you remember that the evaluation process runs 720 evaluations?

with parameter permutations

I guess, that's what it says: "number of results ==== 720"

right

the final history file generation is the equivalent of one more of these