#musicbrainz-devel

Hello!

I have a decoded json variable in perl. The dump is printed here: http://plaintext.mbsandbox.org/logstatistic

But when I try to get the first element of the list ($var[0]), I get an error while running the server

"Error while loading app.psgi: Global symbol "@decoded_data" requires explicit package name at lib/MusicBrainz/Server/Data/LogStatistic.pm line 36."

Could anybody help?

maybe you're using the wrong sigil on @decoded_data? that usually the problem when I get that error

*that's

I tried @var[0] but I got the same error

what's on line 36? it's complaining about @decoded_data

this is what I tried:

$log_stat->data(Dumper($decoded_data[0]));

I can put the code on gist

if that helps more

is decoded_data an array reference? then you need to do $decoded_data->[0]

oh that might be the case

thanks bitmap, I'll try

yay, it works! I really need to learn perl properly :)

I have one more question if it's ok :)

$log_stat->data(Dumper($decoded_data->[0])); this produces this hash on the output:

{ 'count' => '9', 'percent' => '0.900000', '_tc' => '1000', 'mbid' => '60f7d020-9580-47c5-8839-a4f25659687d' }

but this: $log_stat->data(Dumper(keys %{ $decoded_data->[0] } )); produces the single string 'count' for some reason

however when I try $decoded_data->[0]{'percent'} it works without a problem

why cant the keys function return all the keys in a list?

keys returns an array, not an array ref. is Dumper expecting an array ref perhaps?

I thought dumper would just output any variable

it does expect a reference iirc

oh, probably Dumper generates an array if given an array, and $log_stat->data() only uses the first argument.

That may be so

That's right, thanks warp

I was expecting Dumper to output the whole array like when it dumped $decoded_data

which is an array of hashes

plaintext: well, it does. it's just that $log_stat->data() isn't expecting that.

oh I get it

I think

plaintext: install Devel::REPL if you don't have it yet, it makes it easier to try these things.

plaintext: https://frob.nl/tmp/arrays.png

is this sort-of like a shell for perl?

yes

cool, thanks

see the screenshot :)

yeah

that would be helpful

happy birthday adhawkins-away !

can anyone tell me if we can connect to the local musicbrainz database (which I have downloaded form the website ) using the APIs

vinay_: If you set up your own musicbrainz-server you'll have local access to the web service

@navap : Thanks for your reply .. I am using my school network to install it on one of the servers. I read the documentation for setting up the musicbrainz server . Is there any way I can use the existing apache server to host this

There probably is, but I've only used the built in catalyst server so I won't be much help in setting up something else

you'll need to set up apache as a reverse proxy to running the server with FCGI -- apache will be an extra layer "around" the main server

you can't run it with e.g. mod_perl or such, as far as I know, but it's run using Catalyst/PSGI/Plack, so if you want to look at those docs they might inform you

Thanks Guys. I just found out this. Can I use this instead http://wiki.musicbrainz.org/Development/Search_...

that's a separate component entirely

that's used only for search -- the perl server talks to that one (usually just via search.musicbrainz.org, even for external deployments) as a backend for the search functionality

if you want the API, you'll need the perl server setup

Basically I want to use the APIs to extract information , but the APIs need a webservice .

Ok I will try the perl server setup

Thank You

you could also just set up the database, if you just want any old access to the data

but yeah, look into the perl server setup :)

And when I use the APIs musicbrains blocks us after some requests

there's a rate-limit, yes, one per second

I wrote programs to extarct data from the database but wanted to see if by using the APIs will it be any faster

the apis will be slower than directly accessing the data from the database, of course. All they do is add an extra layer between you in the database.

kepstin: probably -- the only exceptions might be if he's getting heavy usage of the cache or of the search indexes

the current etags implementation on the api doesn't actually stop it from doing the db request on 'cached' requests...

unless there's some other caching in the server itself?

yeah, I don't mean etags, I mean object caching

I didn't think there was on the webservice, since invalidating such a cache on db updates is a problem.

I don't know exactly how it plays out or how well it works, but I know that there is some sort of cache

it's most effective for things like the list of languages though

I'm not familiar with that, though - and presumbly an application accessing the db directly could cache relevant stuff internally

yeah, it could

kepstin: https://github.com/metabrainz/musicbrainz-serve... includes the caching stuff for core entities (CoreEntityCache and EntityCacheBase, which the former inheritS)

kepstin: it just does _delete_from_cache after any update, delete, or merge (from perl)

but yeah, in general the database should be faster :)

I was just trying to see the performance but now I dont think its worth it :)

I will stick to the database then

Thanks Guys . I really apreciate it .

Hello

Is it possible to do something like this in an entity?

isa => 'ArrayRef[ArrayRef[Str]]',

Because for some statistics I need to store a whole table of data

I think it should be possible, but I guess look at the Moose documentation to be sure

thanks

yeah, looks like you can

great! I'll try now

actually, I think I'll need ArrayRef[Map[...]],

but if I need a weird structure like this

there must be a better way to do it

hm

probably

I'm not sure what though

where can I find ocharles/warp docs? :P

I think I'll ask for some advice on monday

yeah, that's likely when they're back

in the meantime I'll look into moose docs

but now I'm off to bed

cool

sleep well :)

thank you

bye!