#metabrainz

Yippee, build fixed!

Project musicbrainz-server_beta build #853: FIXED in 19 min: https://ci.metabrainz.org/job/musicbrainz-serve...

Michael Wiencek: MBS-8969: Beta: can't close a track AC bubble except by clicking Done

Project musicbrainz-server_beta build #854: SUCCESS in 19 min: https://ci.metabrainz.org/job/musicbrainz-serve...

* Michael Wiencek: Update translations from transifex.

* Michael Wiencek: Update pot files using current code and production database.

haha

I like the completely fancy-pants use of an Estonian domain :p

I wonder if there's a browser extension for web developers to delay JS for X seconds

kartikgupta0909: Are you using Chrome?

no its firefox

kartikgupta0909: https://community.metabrainz.org/t/how-to-log-i...

It was happening when I was registering on the wiki page

kartikgupta0909: If it's Firefox, it should tell you right there in the dialog what to use as username and password.

(Hint: it's not *your* username nor password.)

oh okay, I have to enter NOTSPAM as the username as well as the password?

Yes. :)

oh okay. I was trying my own username and NOTSPAM as the password.

now it worked

gentlecat: the legacy.py is the curent API, while the api/v1/core.py is the new one you are implementing, is it right?

Yes

so I’s ok if I work on core.py to change the relevant methods to allow lowlevel submission without MBID

I wonder if we should have a separate endpoint for this

because in our general case I think we still want people to submit data with mbids

yesterday we where discussing the possibility to remove the mbid from the URL and get it from the lowlevel data

Hmm, i didnt see that. Let me read it

With the legacy API it made sense in terms of being the same for get and post

let me see if I can extract from the chatlog, not very confident with the tool yet :P

But for the new API, perhaps it's a good idea

I think it should work in this way: if the lowlevel data contains an MBID it is used to store the submission in the database, if it does not contains the MBID we check the submission usind the md5!

but for now I'm not sure we should have a generic endpoint. I think it makes sense to have our normal one, and one for the datasets

Then later, if we want to accept non mbid submissions to general ab, we can merge them

The main reason here is that in general submissions we should not discard duplicate md5 submissions

Our policy is to always keep duplicates

So for now I just follow the endpoints as defined in the API doc using an url_prefix /api/v1/datasets

However, for the dataset system, we have slightly different requirements. Because we don't have mbids to group by, we need to use the md5s as our deduplication system

Yeah, that's how I would do it

for duplicates actually the md5 check if a submission is already present and discard the data, but according to your comment we should keep the data even if a duplicae exists?!

Hmm, really?

Are you confusing this with the data sha256?

no! in fact this not yet implemented :P but this was my idea…

Ah, right. Sorry

I think we should discard duplicates based on md5 for the dataset submission, but not for the general submission

Ok, cool, I start implementing the dataset item submission!

in the API I mean …

Cool

I open a ticket in JIRA for that

Great

alastairp: reminder(PR's). :)

Hey guys, quick question. I see from logs a couple of days ago someone else was having problems getting search server running and it looks like we were having the same error

I finally got all the indexes built on our server

and now it's telling me it couldn't load the "artist index: null" (does the same for all the others too, recording, release etc...)

and then shows errors like "Caused by: java.lang.RuntimeException: javax.xml.bind.JAXBException: Provider com.sun.xml.bind.v2.ContextFactory could not be instantiated: javax.xml.bind.JAXBException: "com.jthink.brainz.mmd" doesnt contain ObjectFactory.class or jaxb.index"

this is within tomcat7 error log when it's trying to startup the server

seems to result in 500 errors: http://ec2-52-208-78-116.eu-west-1.compute.amaz...

armalcolite: 👍

Mike_Aiir: I just merged the fix for that. go ahead pull search-server master and rebuild/redeploy.

Project search_server build #72: STILL FAILING in 20 sec: https://ci.metabrainz.org/job/search_server/72/

themineo: Change the scope of 'mmd' back to compile

ruaok: awesome, i shall give it a try now. cheers

ruaok: is it mvn install i need or mvn package?

ruaok: oh wow, amazing! i feel like a kid at christmas. I'm getting actual xml results now from the search server

thanks for all the assistance!

:)

Thank mineo for that fix. :-)

well thank you to everyone who got us this far!

bitmap: I think this is back? http://tickets.musicbrainz.org/browse/MBS-8898

I'll reopen

ruaok: hey. what are you up to today? Got some time to talk about some jamendo stuff?

Possibly in a bit. Headed to BCN at 15:30.

Ok. Ping me if you have time

Ping.

For a few, anyway...

I'm at lunch :(

Perhaps I'll forward you the email.

We had some discussion around making sure their core catalog is in acoustid and MB. I think Martin is a little confused about how the parts all fit together still

k. I'll be online and off here and there. 18h25 - 20h VST for sure (VST = Vueling Standard Time)

and yeah, there are some odd things happening with jamendo we need to compare notes on.

OK, let's punt to later in the week if you want. Perhaps over a beer?

thats fine. zas arrives tomorrow. maybe the 4 of us can go for one.

Freso: we're downloading their data for audiocommons

alastairp: In FLAC?

Ofc

🙌

hey, what's the status of the bitbucket mbserver mirror?

our mb install is so old that bb is still our origin, I was confused that the schema release wasn't there

should it be automatically updated along with gh?

ruaok: i arrive on thursday, not tomorrow

I love ruaok's understanding of when zas arrives

3 days late, 2 days early, ...