#metabrainz

woo

waa

I have an irrational love of those sites.

geeky and musical all at the same time.

just like my Run GCC sticker on my old laptop.

it tooks us almost an entire 24h to get 1.5m lowlevel files on to our compute cluster

but once they were there, running highlevel for them took only an hour or so

oh silly computing.

next time I do it, I'll just tar the files, not targz them

scp or nc?

it saved 50% size, but took ages to just compress/uncompress

scp is painfully slow. nc is super fast.

I used scp, but I think the transfer was actually the faster bit of it

and you can do compression in line with nc.

granted, those 24h include 8h sleeping too

and fighting with the network at the uni

do nc and bzip in the same go and that will likely yield the fastest setup

ah, no help there

so I think I can do it better the next time

however, I'm still not sure it was faster than running 12 threads on spike

Omnifone says they don’t use our databases...

comments?

ruaok: tell me how we know?

I expected that response.

Phil Sant, one their founders (?), explained in detail on how they use our DB at the last Future Music Forum in Barcelona.

Gentlecat: In the views, at various places, the views deals with the the existing model as a class(for ex: review's class). and then the instance(s) of this class is played around with. for this reason, i feel we should keep the classes even after removing ORM. your views on this?

anyone knows a good markdown to PDF converter which also supports html. Have been looking for a good one since a long time.

i m using this, http://www.markdowntopdf.com/, but it does not format tables, and quoted content as they should be

Gentlecat: and it would be easier to keep them as classes and won't even affect the performance

although, ofcourse, ORM won't be used anywhere

why do we need them?

Leftmost: should we merge master into direct-db? There are quite a few improvements I'm missing in the RE

it would help us handling the instance as a whole or a list of them. moreover at the views, the present implementation is according to class. and it is hence easier to retrieve and pass around any of the instances of the classes by functions.

okay lemme give u a concrete example

in Review model we have a method called get_popular() which returns list of popular reviews

this method is used by views

specifically index view

and itereated to get all the reviews

iterated*

if we keep it as classes, we won't have to fiddle around with these.

and it would work

[listenbrainz-server] tallicamike opened pull request #75: Fix [LB-84] | Listenstore (master...fix-api-does-not-validate-count) https://github.com/metabrainz/listenbrainz-serv...

alternative would be get_popular() returning a strings of text (each of which would be a review)

which would cause the view's implementation to change too.

basically, keeping classes would reduce the changes at view's ends and would somewhat reduce the possible bugs since the implementation changes less

umm did you get it?

Gentlecat: this was just an example, there would be more of such instances

Hey, just wondering is there a way to trigger the rebuilding of search indexes? in the docs it mentions calling bin/reindex but this is for the VM version

hey. other than that script, we have nothing.

and that script is trivial -- just the invocation of the builder and nuking files.

copying that should be simple.

Leftmost: yeah, I'm going to merge master into d-d, we'll have to do the merge anyway at some point and doing most of it now saves time fixing other things

@ruaok where would I find a copy of that file though, can't seem to find it in the musicbrainz-server

that took a while to find: https://github.com/metabrainz/chef-cookbooks/bl...

alas, useless.

here:

dammit, i was close, i was looking through the chef-cookbooks project

CatQuest: if she feels like sharing... I'm ok with it. :)

:D

sure...how shall I share?

and [off]ed link to a paste or google dock or what's convenient to you?

I don’t think it needs to be off...ruaok ?

Sadly no. Probably a he said she said sort of situation... Let me think...

I mean, the uni uses it at least, right? They get it for free because they're a public institution, but if they're doing commercial stuff now... Meh

Let me see if I can find Phil's email. Let's ask him.

note, the person I was corresponding with is the PR person filling in while the other is on maternity leave...and the people she copied appear to be more on the business than the engineering side. She got back to me quickly. I have asked that she consult with the engineering team and get back to me

OK, let me know how that turns out. If that is negative too, we'll talk to Phil.

should be psant@omnifone (based on format of the others) Shall I forward it?

...going to try forwarding my second reply

i have uploaded the proposal but i am unable to view it. Can some mentor verify if its uploaded correctly or not?

On clicking view, the page redirects to a blank page.

Hey, a GSoC question in addition to what I proposed yesterday: new AcousticBrainz client feature or a web interface on AcousticBrainz website that enable community users to import open access datasets used in research with a set of minimum metadata and linked(using the track title or other info) to the original tracks MBID. This last option will improve the datasets quality adding missing metadata and other useful information.

Gentlecat: Whenever you are free, do tell me what do you think about keeping the classes.

LordSputnik, probably.

Freso: final deadline is 25th ? I thought proposals could be changed after that

they can be. but final deadline for new submissions is the 25th

It's just that Fresos post on discourse talks about final PDFs, and I thought that PDFs were unchanhable

Let me get back to the office and I'll try and reread everything

nice, 7 proposals now.

hellska: you should post idea on the forums https://community.metabrainz.org/c/metabrainz/g...

and submit it on GSoC website

alastairp: now you have me confused too. there is a final PDF which we can see until the deadline.

and draft proposals.

Everything was clear to me until I read discourse

it's probably just a formality. nobody stops us from adjusting the plan afterwards, right?

Gentlecat: yes I’m formatting the final document now!

Gentlecat: likely. still odd though.

yeah, now I am fully confused as well. :)

and I've read the faq. may need to read the mailing list archie

+v

I don't see what the problem is

hi guys... can someone confirm if its safe to call release_packaging.id = 7 "vinyl"?

you should look in the release_packaging_type (IIRC) table.

ruaok: i don't have that table... only release_packaging

wait i think i got it... medium_format

what was the release_packaging.name say for id = 7?

"none"

then I would not say that it is safe to call it vinly.

vinyl.

FEH>

NOT SAFE.

ruaok: yeah ill just use the medium_format table instead of release_packaging. i think it will work for my needs

lol thanks for the helrp as usual

alastairp: a small clarification if you have time.

about your comment on my proposal regarding authentication API