This is a public chat log generated from the #semsol IRC channel.
06:35:32
bengee: hi. is there a way to SPARQL query a graph in memory?
06:35:55
hi, no
06:35:57
it seems MySQL is required for any SPARQL query?
06:36:02
hum..
06:36:20
yes, no in-memory sparqling yet, although demand is growing
06:37:00
what do you say to pgsql users willing to use arc2?
06:38:18
any plan to support pgsql? (even if it is not as optimized as on MySQL)
06:38:44
I believe some features should be possible if you abstract the db
06:38:47
they could maybe fund a pg backend, not sure
06:41:44
a bzr repository is going up today, btw, should make code contributions more practical in the future
07:31:09
bengee: fantastic
07:32:08
bengee: the email I just sent to arc-dev got delivered to one of my mailboxes broken. hope you get the attachments ok
07:32:44
that's really awkward
07:33:31
got it, thx
07:34:12
bengee has to write a better mailing list module
07:38:54
bengee: why did you write a module? do you parse them to generate some RDF, or is it just for templating them to appear on your site?
07:39:45
it's for relaying, storing, displaying
07:41:36
for shared hosts where you can't or don't want to install software
07:42:06
I see. is it part of trice? open source?
07:42:16
yes
08:16:53
is there any way to set up an ARC sparql endpoint with a custom default query?
08:17:20
you mean, in the textarea?
08:17:26
yeah
08:17:29
bengee: scor does arc do streaming n triples parsing yet?
08:18:01
I see conditionnal $q in getHTMLDocForm() but can't figure out where to set it
08:18:07
I added some layout options, lemme see if the query can be customized
08:18:21
the one letter functions are still obscure to me :(
08:18:38
that'd be great
08:18:54
kwijibo looks forward to the bzr repository
08:19:23
bengee++
08:19:25
in fact what I want is to include a list of prefixes in the begining of the default query
08:19:34
kwijibo, the turtle(nt parser should stream
08:19:38
hello kwijibo
08:19:43
:)
08:19:53
hi scor, hi bengee excellent :)
08:20:20
i can use the same trick with the RDF/XML parser? extend the parser and over ride the addT method ?
08:20:52
the issue with the turtle parser is that when (the end of) an object can't be properly detected, the buffer grows and grows
08:21:29
that still happens with many DBPedia URIs
08:21:46
s/object/triple term/
08:22:35
bengee: any idea what kind of content that happens with? weird unicode characters? or ?
08:22:44
yes
08:23:13
or just ' or { or } in URIs, I think
08:23:32
is that legal URI syntax?
08:23:40
I made the parser more forgiving already
08:23:52
I don't think so, but DBPedia is full of those
08:23:57
or |
08:25:09
paul houle kept sending me examples, still haven't looked at all of them
08:27:12
bengee I can do class StreamingNtriples extends ARC2_NTriplesParser { function addT($s, $p, $o, $t, $dt, $lang){ /*do stuff */ } } ?
08:27:20
scor, query is currently not configurable, but I'll change the endpoint class. sounds like a reasonable config option
08:28:31
for now, you can prolly do if (!isset($_GET['query])) $_GET['query'] = 'default query here'
08:28:55
bengee ducks
08:31:01
I don't think there is a dedicated NTriples_Parser class, but for ARC2_TurtleParser, the sub-classing should work
08:32:56
bengee: sorry, yeah, that's what i meant, thanks :)
08:41:55
bengee: you mean, I should hack the ARC2 code?
08:42:12
to add if (!isset($_GET['query])) $_GET['query'] =
08:43:49
scor, you can do that on the script level before the endpoint is instantiated
08:44:27
it's a hack, though
08:44:44
and you'd have to check both GET and POST
08:45:05
I've already tweaked the code locally, will upload it later
08:45:25
there is now a config option: 'endpoint_default_query'
08:45:36
Could not properly handle "get" in ARC2_SPARQLPlusParser
08:46:35
I have:
08:46:37
if (!isset($_GET['query'])) $_GET['query'] = 'get';
08:46:39
$ep = ARC2::getStoreEndpoint($config);
08:46:42
$ep->go();
08:46:54
s/get/your default query here/
08:47:35
ah, wait, it'll be auto-executed then
08:47:43
oh, it has to parse! sorry
08:47:54
ok, you have to wait for the patch, I guess
08:48:09
no worries, looking forward to the bzr repo!