#musicbrainz

All of my php is just pushing stuff onto the end of associative arrays of associative arrays of.. etc. :)

Generic solution is generic.

yeah, I guess the problem is I’m trying to do a multi-dimensional associative array, and it’s confusing me :P

er, a nested assoc array ?

you can't really make a mutli-dim array -- wrong terminology. :)

how would you turn these two halves of data chunks into something useful? http://musicbrainz.pastebin.com/d124189ba

I’m still at the “scraping the data into foo stage”, so as long as I end up with some ordered way to store that, right now, I’m not locked into any particular method

an array of assoc arrays?

Ohhh, I see the problem now. Smell of pizza really helps.

vd dinner is still over 1:30 out.

brianfreud: what is the end goal?

I don’t have it in that snippet, but not only do tracks have the 1, 2, 3, thing, but they also store roles that way

I'd have an array of trakcs (or, a map of albums -> an array of tracks, by number), each track being an object with a name, and a map of which artist did what.

ruoak: to be able to present the user with a list of people, to map each to an existing MB artist, and a list of roles, to be able to have the user map each role to an AR

Go through the data you have, add any new information you have to that beastly structure.

and to do it in some way that doesn’t take 10 min to process :P

(you can normalize out artist names if you want, but it's probably pointless at this stage)

so [track][artist, role]?

track containing a set of [artist => role] pairs, yeah.

hmm, ok, I can see how I’d put it in there...

on the other end, then, how would I pull out each unique artist and each unique role? Sounds cpu expensive...

essentially --> for tracks.length( for artists.length( for roles.split(“,”).length( for(interpreted Track list).length ( store arrayFoo[Track][artist][role] )))) ?

hmmm, ok, I think I see how maybe to do it. :)

arrayFoo[Track][artist] = role; yeah. Unique artist/roles can be kept track of while you're building the tree, or you could just scan it (only have to look at everything once, shouldn't be too slow) at the end.

arrayBar[Artist] = true; ;)

was thinking just, at the same time, store two other simple arrays, one of artists, one of roles, and use those values to store in the big array...

http://musicbrainz.pastebin.com/d7db99c82 if anyone has ideas on how to optimize it :)

brianfreud_ is now known as BrianFreud

srotta, you have any experience with JSON?

json

hey

I just emailed you back the queries I’m dealing with... just have no idea how to do anything with them

afaiui you can just eval json in your javascript and it will make some kind of object for u

it’ll automatically send the request and wait for the response?

have you read anything at all about it?

yeah, I was trying to dig out at least some idea of how to handle it, but everything just goes on about how it’s almost javascript already. Didn’t see anything about how to actually request it, or how actually parse out the data

there’s a pretty nice interface to wikipedia data, though, and I’d love to snag just a few crumbs using it... but it’s only JSON

you request it the same way you request xml

the thingy i was working on makes a javascript call to (a php script) which returns a json object

but instead of parsing it as it, you parse it as json

er, parsing it as xml

GM_xmlhttpRequest({ method: 'GET', url: 'http://www.freebase.com/api/service/mqlread?queries={"albums":{"query":{"type":"/people/person","date_of_birth":null,"name":"tori amos"}}}', ....

hurrah

osx 10.5.2 updated without too much pain :)

and picard still works hehe

BrianFreud: r u able to see what your json producer is returning by sending it some kind of query?

BrianFreud: do you have an example how would you do it for xml?

doh, another restart required :(

soccos: yes - I just put the request in as url in Firefox, it renders the response back as text

brianfreud: you normally get json as the response from the server, it've never seen a case where the server expects json as query parameter

BrianFreud: the producer I wrote returns the json in a var assignment i.e. var foo = {$json_string}

luks: http://musicbrainz.pastebin.com/m459819ba

oDiv.innerHTML = responseDetails.responseText;

outsidecontex1: yeah, freebase uses JSON both on the request and the response side (and luckily has a query builder :D)

start with: var obj = eval('(' + responseDetails.responseText + ')');

BrianFreud: i've only ever used it as a response

you want to use an actual json parser

but this is good enough for testing

then obj will have the response as an javascript object, and you can do obj.foo.bar[2].etc

and that then turns each element from the JSON response into oDiv.elements?

no, it's not a dom tree

it's plain old javascript object

ok, so given this as part of the JSON response, "date_of_birth": "1963-08-22", I then can access that as obj.date_of_birth[0] ?

obj.albums.result.date_of_birth

ah, ok, so I can still think of it as turning into “nodes”, cool :)

no, you think of it as objects

you can't use getElementsById or anything DOM-related, because there are no DOM elements

not a problem

Can I still test for the existance of obj.albums.result.date_of_birth ?

the usual JS way

ie, I make the request for date_of_death, obviously, not everyone actually has a date_of_death response

etc.

ok, sounds easy enough, thanks :)

can mb return json now then?

but you definitely want to replace eval with something else later

soccos: no

luks: why’s that?

you can use object.hasOwnProperty("blah") to test for existence (sorry if I'm joining late and not being helpful)

because you don't want to execute arbitrary code from some random server in your script

oh, and eval’s actually executing it, gotcha

not that freebase likely will attempt it, but I definitely see your point

oh yeah, and amazon, but that’s mostly written already

that's the thing i never understodd with json: it is actually JS code, which would make it easy to inlcude it with eval. but that's exactly what you don't want so you need a proper parser. why should you then use json in the first place and not something else?

outsidecontex1: because it's simple enough and widely used

and less verbose than xml

it's human-parseable as well

well, just looking at it from not knowing much of JSON, it would seem to make more sense to use JSON when you control the source (intranets) and XML when you don’t (internet)... I can see outsidecontex1’s point

It's also pretty reliable for places where you control the source... er

Brian X_X

but there are prettier data representations than a programming language

outsidecontex1: As usual, tmtowtdi and uww :)

yeah, and maybe I’m wierd, but XML seems almost more readable to me than JSON

there is no reason to waste bandwidth with <date_of_birth>1963-08-22</date_of_birth> if "date_of_birth":"1963-08-22" will do just fine