Channel #semsol: Logs

This is a public chat log generated from the #semsol IRC channel.

11:08:59 kwijibo: bengee: I parse this: http://paste.lisp.org/display/63203 and get this: http://paste.lisp.org/display/63204 - how come the dc:comment moves onto the (new) document resource?
11:11:16 bengee: because of the <dd< typo perhaps?
11:11:43 bengee: may arc think it's some nested resource?
11:12:14 kwijibo: ah
11:12:19 kwijibo: ta bengee
11:12:21 kwijibo: eagle eyes
11:12:36 kwijibo: :)
11:12:43 bengee: ;)
11:14:51 kwijibo: ace, unit test passes now, ta bengee
11:14:59 bengee: ah, cool
11:18:35 kwijibo: bengee: if you wanted to stop stream parsing after say 1000 triples, how would you do that?
11:18:49 kwijibo: is there a flag you can set or something?
11:18:58 bengee: no
11:19:33 bengee: I think there was an option in ARC1, IIRC
11:19:51 bengee: like "max_triples" or something
11:20:41 kwijibo: could you unset the reader maybe?
11:21:23 kwijibo: $this->reader->closeStream() ?
11:21:25 bengee: I could re-add a config option
11:21:52 kwijibo: well, just wondering really
11:23:40 bengee: another use case for hooks maybe. check for t_count in addT methods, close the stream if needed
11:23:42 kwijibo: like, if I was extending the parser to use my own addT method, and i wanted to stop after a set number of triples, or when a certain triple was found
11:23:43 kwijibo: in fact, the latter is a use case i have
11:24:35 kwijibo: would that work then, closing the stream?
11:25:42 bengee: it may issue an error
11:25:53 bengee: but should stop the pasring process
11:27:04 bengee: to avoid the error, you could perhaps re-initiaize the stream with an empty string
11:28:22 bengee: reader->activate($url, '<!-- irrelevant data -->')
11:30:13 bengee: activate($url, '') would open a new socket, but activate($url, ' ') should create a data stream for ' '
14:06:29 edsu: edsu is seeing a memory error when loading a bunch of data: Fatal error: Allowed memory size of 33554432 bytes exhausted (tried to allocate 359264 bytes) in /var/www/inkdroid.org/sparql/arc/store/ARC2_StoreLoadQueryHandler.php on line 329
14:08:16 kwijibo: edsu: are you loading from Ntriples, or RDF/XML?
14:08:53 kwijibo: kwijibo remembering that only rdf/xml stream parses - though now also SPOG ?
14:12:11 bengee: turtle/ntriples stream, too
14:12:42 bengee: as long as the stream is coming from a file/url
14:12:58 edsu: ahh yeah, i seem to remember having this convo w/ you before bengee :)
14:13:06 edsu: here's the code http://pastebin.com/m7d832783
14:14:15 edsu: edsu is trying to stream ntriples
14:14:38 bengee: bengee tries it locally
14:15:20 edsu: oh maybe it needs to be local?
14:17:43 bengee: bengee tries to remember if we found some conclusion back then
14:21:21 bengee: bengee sparqls the irc log endpoint for "edsu"
14:21:42 bengee: ok, 2008-04-14
14:22:52 edsu: bengee: nice
14:22:53 bengee: it was some erro back then that filled up the mem
14:23:24 edsu: oh right, it's coming back to me now
14:23:29 bengee: it's loading fine, here
14:24:02 edsu: are you loading from the URI
14:24:39 bengee: the store didn't exist: http://arc.semsol.org/community/irc/logs/2008/04/14
14:25:08 bengee: yes
14:26:14 bengee: I'm at 250K triples, mem still constant
14:28:39 edsu: yeah i was doing a $store->drop() instead of a $store-reset()
14:28:47 edsu: $store->reset();
14:29:22 edsu: back then i think
14:29:32 bengee: yeah
14:29:38 edsu: bengee: you using that same code, i pasted roughly?
14:30:58 bengee: yes, just with an additional if (!$store->isSetUp()) $store->setUp();
14:31:08 bengee: to make sure the tables exist
14:32:20 bengee: bengee afk for an hour in a minute :/
14:32:31 bengee: does the insert log give any hints?
14:32:56 bengee: e.g. how far you get before the memory error occurs?
14:33:52 bengee: bengee stops loading at 450k triples
14:36:35 edsu: i get the exception after this in my log: 5000 in 10.10508 = 495 t/s (90000 in 181.01252 = 497 t/s )
14:39:58 edsu: edsu wonders if it's the setup() .... and adds bengee's if statement
14:42:53 edsu: memory is stable
14:43:08 edsu: that was it, bengee++ # again
14:46:16 edsu: edsu tries out rdf-tools on blog
14:52:45 edsu: bengee: how do i populate my wordpress triplestore?
14:53:10 edsu: bengee: i think i may have inadvertenty clicked on the reset the rdf store check box :(
14:53:18 edsu: when configuring rdf-tools
14:53:47 edsu: http://inkdroid.org/journal/rdf-tools/sparql
14:53:58 edsu: edsu is bungling around as you can tell
15:02:44 edsu: oh i get it, it's a store that happens to live inside wordpress space, not tied to the blog content in any way?
15:11:32 danbri: if its the sparqlpress thing, there are weak ties via admin panels, widgets etc
15:14:22 edsu: was trying out rdf-tools from http://arc.semsol.org/download
16:20:06 bengee: edsu, right, it's just some base infrastructure that other plugins can use
16:20:50 bengee: terraces is using it for a couple of interesting things around openid
17:57:24 edsu: bengee: cool
18:18:38 bengee: oh, it was kwijibo who invented the POSH acronym!
18:19:09 bengee: bengee apologises for probably silly dialogue we had a couple of days ago, then