-
chillware
welcome to another 2 weeks of straight analyzing
-
ruaok
FauxFaux: have you setup DBDefs.pm ?
-
FauxFaux
Yeah, I tried, slave.sh and initdb etc. work, which kind of suggests it's okay? /me examines.
-
Lhademmor has quit
-
Aha, it needs READONLY set to something, cheers ruaok.
-
aCiD2_ joined the channel
-
ruaok
np
-
ruaok has quit
-
ruaok joined the channel
-
grrr.
-
stupid desktop pagers on os x
-
ruaok can't wait for leopard
-
aCiD2_
lol
-
Cord
hmmm... this Label-code thing ... is now the second time that i enter a name and a label code, and the label code belongs to another label...
-
BrianFreud
aCiD2: long time no see!
-
aCiD2_
indeed!
-
Exams finally over :)
-
BrianFreud
Cord: more and more ppl are seeing that - sublabels using the same LC as the parents
-
assumption had been that that didn't happen... there's a thread on it in the forums
-
BrianFreud reminds aCiD2 about panda's scan and a little sentence waiting for the front page :D :D :D
-
Cord
hmmm.
-
FauxFaux
Jeez, the reports take long enough to run.
-
BrianFreud
running the standard ones?
-
FauxFaux
Yeah.
-
BrianFreud
I'm actually kind of curious what the match string for the ASIN one looks like... still misses a few, and catches all the audiobooks
-
FauxFaux
I'm guessing NOT B000%, like mine. To the code!
-
BrianFreud
there's a bug ticket mo created and I've added to - a few examples in there of ones that don't get caught currently
-
FauxFaux
AND url !~ 'http://(www\.)?amazon\.[^/]*/(.*)/[B0][0-9A-Z]{9}(\/){0,1}$'
-
luks joined the channel
-
BrianFreud
so it's assuming all ASINs start with B0?
-
that'd be why the 7 pages of audiobooks get snagged
-
FauxFaux
[B0] implies starting with B /or/ 0, /me checks the report.
-
Yeah, that sounds right. How do you pick out an audio book? There seems to be no consistnet prefix or suffix.
-
I suppose you could just select releases that aren't [ Audiobook, * ]
-
Aankhen``
Judging from the Wikipedia entry, audiobooks use their ISBNs, so anything starting with a 0 ought to be an audiobook. Nyet?
-
BrianFreud
-
audiobooks use book ASINs
-
Aankhen``
BrianFreud: Which are == their ISBNs.
-
FauxFaux
-
Aankhen``
(ISBN-10, not ISBN-13.)
-
BrianFreud
actually, also *any*
www.joyo.com listing uses strictly numeric ASINs too
-
Aankhen``
FauxFaux: Hmm. So much for that theory. Stupid useless Wikipedia. *shakes a fist*
-
BrianFreud
though we have so few (if any) joyo ASINs, it's not showing up
-
hey luks: ruaok was hoping he'd run into you
-
luks
hm?
-
Aankhen``
Sounds painful.
-
BrianFreud
[16:13] * ruaok wonders if luks will be around
-
FauxFaux
Run in, eh.
-
ruaok pokes his head up
-
ruaok
hi luks.
-
luks
hi
-
ruaok
I can't get the "tag this <entity> link" to work on my system.
-
is there anything else I need to setup to make it work?
-
-
BrianFreud
FauxFaux: I assume rowIDs are identical on the MB server and your setup?
-
luks
-
apt-get install libjson-perl, I think :)
-
ruaok
heh
-
luks
or is it libperl-json
-
FauxFaux
BrianFreud: Yus (unless someone's added and deleted stuff since my dump, I guess).
-
BrianFreud
Could you look at urlID # 139680 and see if there's anything wierd about it?
-
FauxFaux guesses it's too long before looking. :p
-
ruaok
luks: thanks.
-
next question.
-
BrianFreud
it's one of the ones not showing up on the report, but it ought to be, acc to the string you listed
-
ruaok
the tag parsing, even though it uses \w still removes valid word characters.
-
frød should be a valid tag, but it gets parsed into two words
-
luks
I know, it needs utf-8 decoding/encoding
-
ruaok
k
-
ruaok hates that in perl
-
FauxFaux
BrianFreud: "http://www.amazon.de/Souls-Black-Testament/dp/B000002IPZ"? Doesn't look like that regex'd catch it to me.
-
BrianFreud re-looks at the regex
-
Cord
ah. great. the albums i have here list 'Polymedia Marketing Group' with LC3230 (which is Motor Music) and has according to discogs an own LC5617
-
FauxFaux
amazon.NOSLASHES/ANYTHING/ASIN OPTIONALSLASH?
-
luks
I'm off for about hour, bbl
-
ruaok waves luks off and returns to hacking
-
BrianFreud
right, but shouldn't ANYTHING be stopping at the / in Souls-Black-Testament/dp/ ?
-
FauxFaux
No, anything is literally anything (in the regex). The fact that guesscase strips out the name stuff from the urls suggests it's wrong, though, yes.
-
BrianFreud
yes, for the correct ASIN standard, the "Souls-Black-Testament" ought to be stripped
-
FauxFaux
Onnit. ;p
-
BrianFreud
it's the one subset - where everything but that is right - that the report currently misses
-
Aankhen``
<luks> apt-get install libjson-perl, I think :) # if that's JSON.pm, stay away from it.
-
Use JSON::XS instead.
-
FauxFaux
BrianFreud: Mm, there are loads of dodgy urls inside amazon, /me wonders how it filters them.
-
BrianFreud
-
sorry, actually
-
-
FauxFaux tries to spot the difference.
-
gp, not dp
-
FauxFaux
Slow query is slowwwwwwwwwwwwwwwwwwwwwwwwwwwwww.
-
BrianFreud
for example, one I worked on earlier...
-
FauxFaux
Reaaaaally slow.
-
BrianFreud
-
FauxFaux
My report throws up "http://www.amazon.com/exec/obidos/ASIN/B00005JHY5/" and things like it, are they allowed?
-
BrianFreud
nope
-
-
we fixed all the ones with crap *after* the asin
-
FauxFaux -> PERL (aaa haa!)
-
but the ones with the crap beween the AZN and the ASIN got missed by the report
-
chillware has quit
-
so hmmm, not a regex expert, but oughtn't it to be something like: AND url !~ 'http://(www\.)?amazon\.[^/]*/gp/product/[B0136][0-9A-Z]{9}(\/){0,1}$'
-
ruaok
if you guys have a minute or two... can you help me tag some artists on my server:
-
-
BrianFreud
sure
-
ruaok
I just want to have a few artists tagged so I can work on the WS to kick out tags
-
BrianFreud
umm, how do we tag?
-
do we need to log in?
-
got it
-
ruaok
yes.
-
separate tags with commas.
-
ruaok watches the db change
-
"confusingname 60s mellow" -> one tag
-
BrianFreud
lol, I figured it out :)
-
ruaok
you want "confusingname, 60s, mellow"
-
:-)
-
much better
-
Aankhen``
Or "confusing name, 60s, mellow"?
-
ruaok
depends on his state of intoxication, really.
-
ruaok wants this feature live sooon
-
FauxFaux
-
BrianFreud
yeah!
-
holy christ
-
Faux just fixed bug 2961
-
FauxFaux
16123 hits.
-
BrianFreud
christ that's gonna be a lot to clean
-
not sure if there's one in there for the audiobooks thing too
-
FauxFaux
Heh, it's worth fixing the reports anyway.
-
BrianFreud
what regex?
-
FauxFaux
That's AND url !~ 'http://(www\.)?amazon\.[^/]*/(.*)product/[B01396][0-9A-Z]{9}(\/){0,1}$'
-
BrianFreud
so dp/product stills slips through?
-
FauxFaux
Oh god. :P
-
BrianFreud
I think you just gave SenRepus a project for the next year with that fix :D
-
FauxFaux
Insisting on gp/product/ASIN gets you 6 more hits, bah.
-
BrianFreud
hehe
-
you know eventually someone would find one of those 6 and put in a ticket on it :P
-
FauxFaux goes to grugingly fix them all.
-
hehe
-
how many do you want ruaok?
-
ruaok
that's a good start, thanks!
-
BrianFreud
:)
-
any way to see a tag cloud yet?
-
curious how it looks :)
-
FauxFaux
... . .. .. BRIANFREUD ... .. .. musicaboutponies .. ......
-
There you go!
-
BrianFreud
lol
-
.... not on last.fm... ... ficticious ... ... EMO .... ... ponies ...
-
FauxFaux
The best tag on last.fm has to be the guiltypleasures tag.
-
BrianFreud looks
-
BrianFreud
hmm, wait, ruoak, does this mean the database now also supports freeform info?
-
FauxFaux listens to 'Brother Brown' - 'Star Catching Girl (feat. Frank'ee) (Brother Brown's Rock Bottom)' from 'Ministry of Sound: Clubber's Guide to... Ibiza Summer 2001 (disc 1)' [2001-05-28] at 400kbps vorbis.
