#musicbrainz

is there any possibility of doing some kind of mass import of data from Naxos Music Library/classicsonline?

i don't mean by a user, but but those behind the scenes?

Discussions about mass-imports have gone Badly in the past, due to the fact that the system barely copes with the manually entered edits, never mind some batch-generated ones.

barely copes.. you mean in terms of processing power etc>

?

there aren't enough voters

presumably you wouldn't need voting for a mass import

the data on classicsonline is straight from the label

but it doesn't necessarily follow mb guidelines, nor can we guarantee they don't have typos

well it's still better than not having the info at all

say for 'low quality' artists

how much do they have abyway?

anyway, even

NML has 36,196 CDs

Classicsonline might have more

NML's mostly classical music

but a considerable amout of jazz and world as well

I remember there was some talk about importing from cdbaby and in the end they were added as cd stubs

ah

well NML has more info, like composer, performer, conductor, etc.

it would save 1000s of man hours to import those as ARs etc.

something like the old cdbaby importer greasemonkey script that lets people import stuff one by one but doing all the various relationships would work, but it would still mean finding someone to write it

and most people who code for mb are busy busy busy doing ngs :/

ah

i guess i still have mixed feelings about spending time doing edits when I imagine a lot of that work will be irrelevant when they do some kind of mass import 'some day'

generally unless you're the one to write the mass importer, you can assume it won't happen :P

unless you have a big supply of good chocolate and can bribe someone ;)

i can definitely assume it will hapeen 'some day'

i.e. for some x, it will happen in x minutes

:p

I can definitely assume the sun will explode, but I'm not going to sit around waiting for it ;)

navap: exactly :P

ruaok will be pleased to know I'm breaking the search server with faye wong again >D

adamolding is now known as adamgolding2

i was ghoste

*ghosted

anwyay:

after I add a release, what's the fastest way to get it straight into picard?

click the tagger button

it doesnt' show when i've just added a release

if it isn't there, add ?tport=8000 (or &tport=8000 if there's already a ? in the url), or open a browser window from picard (e.g. by using the lookup button at the bottom right)

wack

why doesn't it show all the time?

it times out after a while, because most people don't have picard open all the time

damn ghosted again

If the release has english translations of the titles in brackets, shoudl they be included?

we normally split tracklists into two entries if there's multiple languages on the cover

two releases?

yeah, linked with the translation relationship

that's not much help for tagging, if people want both languages in their tags

I hope one day we can finally have something like track aliases for things like that...

luks: will the search stop remembering things for ages like it currently does?

nikki: yes

\o/

it drives me nuts that I can come back hours later and it's still set to something I forgot I ever searched for

so to get an advanced search, I add adv=1 to the query URI?

what URI?

the api query

then no

adv=1 is not mentioned in the XML WS docs either

if you can explain what are you trying to do, I can help you, but I'll not be trying to get the info out of you :)

ok, i'm jumping ahead

first i'm using the perl api wrapper WebService::MusicBrainz

which composes a URI to query MBz

make sense?

yes

ok

there's a method set_url_params()

so if i'm searching on "Cream" for example, with the basic query, I get some japanese band back first

but if, on the form, i tick "advanced", I get the real Cream back as top hit

I guess basic query means http://musicbrainz.org/ws/1/artist/?type=xml&am...

you can use the 'query' argument to specificy literal Lucene query

which is what the "advanced search" on the website allows you to do

I don't know how to do what in WebService::MusicBrainz

*that

ew

horrid

what?

the perl lib

i may change to parsing the XML directly

and composing the queries manually

is there a page on composing the URI's fr the queries?

The wiki page documents it, yeah.

ccol, ta

</n00b>

adamgolding3: I wouldn't mind a dump as stubs, but I find their data so often needs cleanup to meet our guidelines, the only time a mass import would save would be the initial add time - you'd still have all the time spent cleaning up the data, but now as part of the live db.

CD stubs require CD TOCs

I doubt they are available in the archive

luks: Is that still true given Rob's recent changes re: tocs?

yes

luks, why does release_label have a position column? Does the ordering of labels really matter?

srotta_ is now known as srotta

aCiD2: it probably doesn't matter much, but I wanted to have a stable ordering

I think we could get that by ordering on label name and catalog number though

It just makes the edit types a bit needlessly complicated, as they have to do the same position shuffling as the tracklist