#musicbrainz-devel

Any devs familiar with coverart here?

Making a request to 'http://coverartarchive.org/release/189d7863-a23c-49d9-ae7e-031b413b2805'

Then 'http://archive.org/download/mbid-189d7863-a23c-49d9-ae7e-031b413b2805/index.json'

The latter gives a 500 ISE in libcaa, but not if I do it in a browser...

Any thoughts?

It's the test.cc program in the tests directory of libcaa if anyone else could try it?

Hmm...something strange going on. It's working now...

Works for me, so probably a transient problem at the IA.

the test program works here as well and the request it sends looks good in wireshark

ianmcorvidae: I made http://tickets.musicbrainz.org/browse/MBH-387, I hope that's the right place for it!

Ok, thanks for checking guys.

alastairp: I think you did the toma.hk stuff, looks like it's doing something broken with unicode: http://acousticbrainz.org/60a2abef-6260-45ac-8e... (if you have any idea how to fix it)

It's an "&" that should be an "&", AFAICT.

Or maybe not. Maybe it's just that the tomahawk player isn't that CJK unicode friendly. :/

ianmcorvidae: Does it work for CJK characters elsewhere?

I have no idea

I mean, for other MBIDs.

But I guess it's simple to mebed the toma.hk player into a quick HTML page to test the theory.

*embed

I can't take tomahawk/playdar seriously since they seem to do everything via track/album/artist titles

ianmcorvidae: resolving via mbid is certainly an option, except that no service supports it

Freso: utf8 should be fine

huh. it seems like beets at least should

Freso: ah you're directly using the embed. let me check that, we have a bigger upgrade for that coming

ianmcorvidae: if they do, i'm happy to integrate that. the problem exists at both ends tho

whoever provides metadata would then also need to provide the matching mbid

for us to be able to properly resolve it

ianmcorvidae: that's the spirit :)

the 'resolve' function in https://github.com/tomahawk-player/tomahawk-res... seems to only support artist/album/title

ianmcorvidae: yeah as i said, there's currently no such resovle method with mbid because i don't know a single service that supports it

but more than happy to integrate that

ah, I see. the beets one at least presumably supports a bunch of things

it'd be a start yeah :)

I suspect nobody deals with MBID redirects all that well though, heh, even MB projects aside from the main server don't really deal well with those

heh

anyway, just wanted to point out we (tomahawk) would love to do all our resolving using sane mbids instead of strings

it'd make a ton of things a whole lot easier for us

alright. oh. should also do this -- sampsyo ^ you may be interested to hear this :)

(I don't actually use beets or tomahawk myself, hence having trouble finding a lot of this stuff)

oh, and I'm quite sure at least acousticbrainz will happily pass toma.hk MBIDs

obviously :)

heh :)

alastairp: we should scheme which talks we want to attend.

oh, if you want

I've been doing it just-in-time

there's no other way I can choose otherwise

both look interesting.

(at least to me)

haha, of couse you'd like that first one :)

the linking one might be good for learning how to rope new datasets into the fold.

and I'm doing the intros on the ethics track -- now sure if I am also doing the questions at the end of the sessions.

I hope not. I want to attend other tracks.

most of the sessions I went to the speaker has done questions

yeah, someone is supposed to run the mic around.

as long as there's a staff member to walk the mic around

yeah, as long as that that isn't me, I'm quite happy.

but I'm going to get over there and see if I can see the three stooges start the day.

cuz, I'm pathetic like that

heh

yeah, I'm just finishing up breakfast

see you there

thanks Freso and ianmcorvidae for the covers!

alastairp: great

ianmcorvidae, alastairp: the tomahawk unicode stuff is https://bugs.tomahawk-player.org/browse/THK-92

oh, and muesli as well: ^

ianmcorvidae: did you report it?

alastairp: Report what?

alastairp: The toma.hk player thing?

Hi all!

Freso: yes. I know there's one for tomahk now, but it's worth us fixing it

and on monday I'm not going to remember something from 3 days ago

alastairp: I'm not sure we *can* fix it.

sure we can. we can unidecode names before we pass it to tomahawk

just because you don't think we can, it doesn't mean we shouldn't look into it

Hey, I'm not saying ianmcorvidae shouldn't file a ticket.

what do we have to do with tomahawk and what do we send there that isn't a bytestream?

JonnyJD: http://acousticbrainz.org/60a2abef-6260-45ac-8e...

so we can't just "not encode" it. But we need to use urlencoding or something *and* some kind of string encoding, right?

alastairp: I would start using urllib.quote_plus() and see if that helps already with the tomahawk problem

not sure if I can test that myself, since I don't know how fast I can setup an acousticbrainz test server

TBH i have no clue how I get that to work that repository at all :-D

looks like flask, and I don't have a clue about flask

you need to clone the repository and then use admin/setup.sh in the admin folder to setup the db

Mineo: but I don't have to download the complete db, right?

no, you don't need the complete db

cool, trying it then

you can just generate a json file by running the extractor manually and then submit it to your locally running server with "curl -d @<json-file-name> localhost:8080/<track-mbid-here>/low-level"

thx

JonnyJD: I have a PR with a slightly fleshed out README you can follow for install instructions.

In abz-server.

It doesn't mention how to do the db, but that's mostly because there are no db dumps yet. Rob has a smaller db dump available, but you should also be able to hack -client to submit to your local test server.

Freso: thanks, although I got that far by now already :-D

except I was calling server.py inside acousticbrainz folder

still missing the css somehow, at least it looks like missing all formatting

yes, everything from /static gives 404, regardless of me doing "python2 acousticbrainz/server.py" or just "python2 server.py" (not that this is relevant for my tests.. though)

Everything? Or CSS?

For CSS, you need to compile the less.

oh, yes. non-css only gives 304, not 404

JonnyJD: Do you have the db running?

"essentia_build_sha' was not found in submitted data" is that something introduced later or because of not using abzsubmit?

Freso: yes, db and server running

(self-built essentia, same commit as the first linux binary released)

JonnyJD: I tihnk the build_sha is part of the JSON file generated by the submitter and not part of the essentia output.

*think

Mineo: submitting per curl doesn'

t work

submitting per abzsubmit (changing [acousticbrainz] host) works

it did work for me

but if abzsubmit works, then everything's okay :)

hm, were you using the essentia-acousticbrainz from AUR or something different?

no, a locally built one - `git describe HEAD` prints "v2.1_beta2-3-gd5e3a6e"

Mineo: so you have a later build, possibly that includes the build_sha by itself (notes that for later to check)

anyways, thanks guys, I have a running acousticbrainz server now and will check out that encoding problem

JonnyJD: Also, if essentia-acousticbrainz is still using the Git SHA of the original Linux "release", it probably ought to get updated anyway...