This is a public chat log generated from the #semsol IRC channel.
10:44:11
bengee, I'm wondering how the label consolidation work on ARC2
10:44:34
I loaded my foaf profile, then ran $store->consolidate();
10:44:54
but no results when I'm SPARQLing for my rdfs:label (which is originally a foaf:name)
10:45:02
I don't think the label stuff is working atm
10:45:52
it was just part of DESCRIBE, where an arc:label was generated on the fly from available label props
10:48:27
ah ok
10:50:15
an is there a way at the moment to make inferencing re. subproperties / subclasses ?
10:50:33
or even something simple like 'property consolidation' ?
10:50:35
no
10:51:12
you can write your own INSERT+CONSTRUCTs, though
10:52:15
humm, indeed. using a trigger on LOAD might be an option
10:52:21
I'll check this
10:52:45
'k
11:21:57
bengee: http://arc.semsol.org/docs/v2/misc/writing_triggers - small detail to fix in the doc, this example works for an INSERT operation but not for a LOAD one
11:23:39
hmm, you mean I should mention that LOADs have to be registered separately?
11:25:36
maybe, or at least, change insert to LOAD in the first snippet of code, it might be confusing
11:26:06
bengee adds LOAD to the example
11:26:11
thx
12:20:40
terraces just send a trigger patch to bengee
12:21:31
oh, was there a bug?
12:25:23
oops sorry, that's a 'feature-patch' actually
12:26:08
cool, thx, will check my mails later today :)
14:38:42
does ARC allow me to serialize rdf using the old-school rdf:Seq and rdf:li for containers? Mozilla's being non-compliant and I need to appease it.
14:39:46
don't think so, no
14:40:56
it'll use the internal rdf:_n props generated during parsing
14:41:21
it'll use the internal rdf:_n props generated during parsing
14:43:29
hmm. I haven't had a lot of success with the rdf:_n syntax. something buggy in their rdf caching (probably the way they serialize/unserialize what I send them).
14:43:51
I'll see what I can cook up that's appetizing.
14:44:16
bengee wonders how tricky it'd be to create a restricted serializer for arc
14:46:47
prolly not too hard, the serializer is based on a resource index, i.e., I'd just have to check if the current node is a seq/alt/bag, and then sort the predicates and use li
14:48:20
do you need <rdf:Seq> instead of <rdf:Description><rdf:type rdf:resource="..." />, too?
14:48:42
potentially. I'm trying out a few variations.
14:52:48
unfortunately moz swallows this fine:
14:52:51
<rdf:Seq rdf:about="http://foo.com/bar">
14:52:53
<rdf:li rdf:resource="http://foo.com/bat"/>
14:52:54
</rdf:Seq>
14:53:00
but not this:
14:53:08
<rdf:Description rdf:about="http://foo.com/bar">
14:53:09
<rdf:type rdf:resource="http://www.w3.org/1999/02/22-rdf-syntax-ns#Seq"/>
14:53:11
<rdf:_1 rdf:resource="http://foo.com/bat"/>
14:53:12
</rdf:Description>
14:53:30
bengee is trying to patch the rdfxml serializer
14:54:03
I should know in 5 mins if it's easy to do or not..
14:54:27
thanks. let me know if it's a pita.
15:09:40
ok, I have seq/alt/bag, now looking at li
15:19:31
I can't guarantee proper sorting, but I have a patch
15:20:49
great. i'd be interested to see how you've constructed your triples too (if you went that route).
15:30:34
bengee prepares a new revision
15:41:57
ok, new rev is online at http://arc.semsol.org/download
15:42:31
you have to set a configuration option to trigger pretty containers
15:43:30
The RDF/XML Serializer accepts a "serializer_prettyprint_containers" config option which will lead to pretty-printed RDF Containers (rdf:Seq/Alt/Bag) if set to true. (http://arc.semsol.org/docs/v2/serializing)
15:44:36
bit in a rush, hope I didn't introduce bugs..
15:44:56
I'll try it on and let you know how it goes. Much obliged!
15:45:09
cool, good luck
15:53:22
bengee off for late lunch
16:07:14
hi
16:07:48
is here anybody..
16:07:57
i have a short question about ARC
16:08:04
for php
16:10:30
sure612, I have to run, but will be back in about 30mins, just post your q
16:10:40
ok thx
16:11:19
with arc there's the possibillity to setup an endpoint
16:11:33
this endpoint is using the own mysql db
16:11:44
but i want to use SPARQL for remote files
16:11:50
is that possible
16:12:21
(i hope you understand me) - i'm german maybe somebody's too...
16:43:05
sure621, remote files or remote SPARQL endpoints?
16:43:59
arc can't run in-memory queries on files directly yet, but it can query remote sparql endpoints usng the RemoteStore component
16:45:00
there is now UI component for the RemotStore (like for the local one), though
16:49:10
remote files
16:49:42
i can load files from other domains etc. and return the tripels
16:49:53
but i want to use SPARQL on the data then
16:50:20
you know what i mean?
16:51:02
bengee is german, too, in case you feel more comfortable with DE, but I think I know what you mean
16:51:21
you don't want to store the files locally at all, right?
16:51:23
hehe
16:51:28
exactly
16:51:34
i want to use it with FOAF
16:51:42
but the foaf data is not in any db
16:51:49
especially not in my own
16:52:26
but to get the facts i want, like the name, i want to use SPARQL commands
16:53:03
then you'd have to first do a LOAD and then a SELECT in ARC
16:53:46
yes
16:53:54
thats possible right?
16:54:04
i.e. you have to buffer the files in a local (or remote) rdf store
16:54:17
because i couldnt find examples for that on the arc site
16:54:23
here's an example
16:54:24
http://demo.openlinksw.com/sparql/
16:54:31
like i wanted to use it
16:54:51
there u can specify the uri of the file and the sparql stuff
16:55:31
buffer?
16:55:32
really
16:55:34
mmm
16:56:52
building something like the demo above is trivial in arc, but you need a local store
16:57:25
ok
16:57:36
then first store the remote file in the db
16:57:43
right
16:57:44
and then i can use sparql
16:57:47
ok
16:57:53
i try to do that :-)
16:58:03
maybe I'll add in-memory sparqling some time, but I personally never really needed it
16:58:17
is it to complicated to do that without storing?
16:58:39
the query execution is different
16:58:53
ok
16:59:07
thats a not so good - but ok
16:59:10
arc translates sparql to sql, which wouldn't be of much use w/o a DB
16:59:12
thats how its got to be
16:59:21
ok
17:00:02
thre's also PHP-based RAP which has an in-memory store
17:00:09
s/thre/there/
17:00:21
rap?
17:00:28
it's a bit more complicated to use that arc, though
17:00:33
aha
17:00:39
do you have a link?
17:00:43
http://www4.wiwiss.fu-berlin.de/bizer/rdfapi/
17:00:52
and is it good working like arc?
17:01:04
or more some kind of beta something
17:02:31
i'll check it out
17:02:32
it's mature, but there is only support and project activity when one of their students needs it for a project
17:02:50
ok
17:02:58
that sounds not that good
17:03:01
it has a lot of nice features, though. definitely worth a try
17:04:32
when i was starting to "work" with rdf i thought i build my own parser :-D
17:04:40
or for foaf
17:04:47
but i think that was not a good idea
17:05:25
i want to build a foaf explorer (or something like that that) in adobe flex
17:05:38
so i wanted to use the action script 3 xml support
17:05:43
but thats too hard
17:05:59
or do you know something about rdf implementations in action script?
17:06:15
achso der eine meinte du verstehst ja auch deusch
17:06:18
t
17:06:26
;)
17:06:47
aldo is working on a flex thingy
17:07:13
http://code.google.com/p/semanticflash/
17:07:35
nice
17:07:40
i'll check it
17:07:55
he's looking for collaborators
17:08:28
hehe
17:08:43
but i have to do this for foaf
17:09:46
or at least thats what i was planning to do
17:09:50
for my master thesis
17:10:26
ah, cool
17:11:07
but i think implementing a foaf parser is too hard for me
17:11:18
do you have a foaf file?
17:11:24
no
17:11:25
:-)
17:11:30
not yet
17:11:51
mine is gone, too ;)
17:12:27
because I'll launch that cool foaf editor just the other week (tm) ;)
17:13:29
(tm) since 2003 or so
17:14:07
bengee, I'm struggling a bit with using getSerializedIndex (had been using getSerializedTriples, so I'm a little unfamiliar)... when you whipped the patch, did you happen to have a testcase that produced something like:
17:14:08
<rdf:Seq rdf:about="http://foo.com/bar">
17:14:09
<rdf:li rdf:resource="http://foo.com/bat"/>
17:14:11
</rdf:Seq>
17:19:12
$parser->toRFXML($triples)
17:21:19
the component you are using getSerializedIndex on has to have that prettyprint config option
17:22:51
right. previously I had been serializing from $triples (internal representation)... now since I see you patched the getSerializedIndex method, I'm moving over to $index (internal representation)... unless you're suggesting that I can keep using triples.
17:24:14
the serializer always builds an index first, you can keep your code as is
17:28:38
great. I think I'm on the right track. thanks again.
17:28:56
peterdm, http://pastebin.com/m6dc83562 has my test file
17:29:08
cool, I'll take a look.
17:39:28
So far it works like a charm. Danke schön! I'll let you know if I encounter any problems down the line.
17:41:39
ah, great