Semantic Web Interest Group IRC Chat Logs for 2005-09-06

00:00:22 <sbp> wow, someone read the topic!

00:58:34 <Elementalistmage> hey

02:23:00 <myren> is there an emoticon namespace?

02:23:30 <myren> <inquire xmlns="urn:alienintels.com:myren:emoticons"/>

03:01:48 <chimezie_> chimezie_ is now known as chimezie

03:22:37 <crschmidt> myren: not that I'm aware of

03:43:10 <myren> where does one go about announcing such things?

03:46:09 <kasei> schemaweb?

04:57:28 <_aniasis> _aniasis is now known as aniasis

12:49:16 <chrisc_> chrisc_ is now known as chrisc

14:55:26 <danbri> if anyone cares to play with jabber/rdf, this is as far as I've got using the Smack libraries: http://rdfweb.org/viewcvs/viewcvs.cgi/foafproject/htdocs/2005/code/foaftown/jabber/

14:56:19 <danbri> needs Ant, can be run in client or server mode, but haven't quite got the server to teach the Smack code that it can handle sparql-based IQ requests, nor the client to send the correct form of IQ

14:56:41 <danbri> on the plus side, Smack has a lovely graphical debugger for messages sent and received

14:57:25 <danbri> to play, grab the .java, build.xml, Smack's JARs, and get yourself a couple of test accounts on a public jabber server (jabber-hispano.org worked for me...)

15:01:24 <edsu_> edsu_ is now known as edsu

15:11:22 * danbri puts a holding page at http://www.w3.org/2005/09/xmpp-sparql-binding since a uri for the binding is needed by XMPP/Jabber

15:12:13 <erim> libby, you here?

15:13:05 <crschmidt> danbri: any docs around for what you're doing specifically with jabber?

15:13:08 <crschmidt> I've found jabberpy

15:13:20 <danbri> docs, only the blog entries linked from that last page

15:13:20 <erim> .me wonders what the current thoughts for vocabularies for managing conference events, people, etc. are

15:13:45 <danbri> basically, use the IQ mechanism to transmit RDF queries, and their XML resultset'd response

15:14:10 <danbri> haven't got to details of the binding yet, bogged down in xmpp/jabber details re making touch between endpoints

15:14:46 <danbri> i've got a Listener that hears textual messages; but when I send an IQ packet instead (the preferred way to send a query), something in the Smack library sends back a 'can't handle that' msg

15:14:51 <danbri> even though i thought i'd registered a handler

15:15:26 <danbri> when that basic step is behind us, can figure out how the details of the SPARQL protocol would make sense in terms of XMPP iq

15:15:36 <danbri> <iq id="8NGrm-5" to="foaf@jabber-hispano.org/sparqlclient" from="foaf2@jabber-hispano.org/sparqlserver" type="error">

15:15:38 <danbri> <error code="501">feature-not-implemented</error>

15:15:40 <danbri> </iq>

15:16:08 <danbri> there's also a piece of work todo on endpoint feature discovery

15:16:35 <danbri> currently am hardcoding resource IDs /sparqlclient and /sparqlserver, should instead get those from inspecting the roster to see if an rdf-capable endpoint is attached

15:17:21 * libby waves

15:17:25 <crschmidt> currently have my client connected as <crschmidt@jabber-hispano.org/sparql>

15:17:30 <crschmidt> no work done on anything yet

15:17:41 <crschmidt> just the standard jabberpy "test_client.py"

15:17:43 <danbri> I had difficulty earlier with getting a list of presences from the roster

15:18:10 <danbri> If you can bear Java, the Smack graphical debug window really helps demystify

15:18:15 * crschmidt can't :)

15:18:19 <danbri> :)

15:18:28 <crschmidt> the jabberpy stuff seems pretty simple

15:18:35 <crschmidt> was really just wondering what happens if you send a query my way

15:19:08 <danbri> I can try

15:20:20 * crschmidt attempst to subscribe to foaf@'s presence

15:20:21 <crschmidt> [[[

15:20:21 <crschmidt> DEBUG: sent <presence to='foaf@jabber-hispano.org' type='subscribe' />

15:20:21 <crschmidt> DEBUG: got data <iq xmlns='jabber:client' to='crschmidt@jabber-hispano.org/sparql' type='set'><query xmlns='jabber:iq:roster'><item ask='subscribe' subscription='none' jid='foaf@jabber-hispano.org'/></query></iq>

15:20:21 <dc_swig> Label DEBUG not found.

15:20:22 <dc_swig> Label DEBUG not found.

15:20:26 <crschmidt> Tue Sep 6 11:17:40 2005 - RECV: - <iq xmlns='jabber:client' to='crschmidt@jabber-hispano.org/sparql' type='set'><query xmlns='jabber:iq:roster'><item ask='subscribe' subscription='none' jid='foaf@jabber-hispano.org'/></query></iq>

15:20:30 <crschmidt> ]]]

15:20:46 <danbri> I accepted the adition (via gaim)

15:20:49 <danbri> can you do foaf2 while at it?

15:24:45 <danbri> [java] Initiating connection, role is: [client].

15:24:47 <danbri> [java] Connection roster of client foaf@jabber-hispano.org is: org.jivesoftware.smack.Roster@1a99347

15:24:48 <danbri> [java] Roster item: crschmidt@jabber-hispano.org [foafs list of bots]

15:24:49 <danbri> [java] Roster item: foaf2: foaf2@jabber-hispano.org [foafs list of bots]

15:24:51 <danbri> [java] Server crschmidt@jabber-hispano.org has no presences visible to foaf@jabber-hispano.org

15:25:25 <crschmidt> did foaf2 get crschmidt's subscribe request?

15:25:55 <danbri> wasn't logged in as foaf2 sorry, checking

15:26:25 <danbri> done

15:27:36 <danbri> foaf@ sent this, <iq id="Dh051-5" to="crschmidt@jabber-hispano.org/sparql" type="get">SELECT BLAH BLAH BLAH FROM BLAH</iq>

15:27:41 <crschmidt> hm

15:27:46 * crschmidt scrolls back through

15:27:53 <crschmidt> (example_client output with debug is verbose)

15:28:06 <danbri> i think it should have used a <query xmlns="http://bindinguri/..."/> subelement, with sparql payload

15:28:36 <crschmidt> Tue Sep 6 11:22:14 2005 - RECV: - <presence xmlns='jabber:client' to='crschmidt@jabber-hispano.org/sparql' from='foaf@jabber-hispano.org/sparqlclient' id='bvcO1-3'/><message xmlns='jabber:client' from='foaf@jabber-hispano.org/sparqlclient' type='chat' to='crschmidt@jabber-hispano.org' id='bvcO1-4'><body>Howdy from foaf atjabber-hispano.org</body><thread>Akem90</thread></message><iq xmlns='jabber:client' from='foaf@jabber-hispano.org/sparqlclient'

15:29:27 <danbri> so chat messages get through, and we _could_ just stuff SPARQL in those, but it'd be monsteriously tasteless to do so

15:29:37 <danbri> IQ seems harder to get right, but worth it

15:29:58 * crschmidt nods

15:30:05 <crschmidt> The iq did come through there

15:30:19 <crschmidt> it just came attached to the <presence> and <message>, it looks like

15:30:27 <myren> What do you guys use to test IQ's? none of the clients i've seen expose their clients to IQ's

15:30:29 <danbri> the client didn't see a response

15:30:49 <myren> I should make an XMPP client for developers...

15:30:53 <danbri> we're using software libraries not GUI client apps (although Smack has basic GUI)

15:30:57 <crschmidt> DEBUG: DEPTH -> 2 , tag -> jabber:client iq, attrs -> {u'type': u'get', u'from': u'foaf@jabber-hispano.org/sparqlclient', u'id': u'bvcO1-5', u'to': u'crschmidt@jabber-hispano.org/sparql'}

15:30:57 <dc_swig> Label DEBUG not found.

15:31:01 <crschmidt> DEBUG: data-> SELECT BLAH BLAH BLAH FROM BLAH

15:31:01 <dc_swig> Label DEBUG not found.

15:31:23 <danbri> ok, that's promising

15:31:25 <crschmidt> danbri: not sure what exactly I'm supposed to be looking at or expecting, to be honest, all just guesswork

15:31:30 <danbri> same here ;)

15:31:33 <crschmidt> DEBUG: dispatch called for: name->iq ns->

15:31:33 <dc_swig> Label DEBUG not found.

15:31:42 <crschmidt> it looks like it did call a dispatch for that iq

15:31:46 <crschmidt> but no namespace attached?

15:32:06 <danbri> yeah, doesn't seem to be making the full thing

15:32:12 <crschmidt> anyway, i'd assume since it called a dispatch, I could attach something to listen to that

15:32:32 <crschmidt> but not sure how I'd issue a response yet :)

15:33:12 <danbri> class SPARQLIQ extends IQ {

15:33:14 <danbri> public String getChildElementXML() {

15:33:15 <danbri> return "SELECT BLAH BLAH BLAH FROM BLAH"; // todo :)

15:33:17 <danbri> }

15:33:18 <danbri> }

15:33:41 <danbri> I guess that function has to make the <query > element by hand... the namespace regsitration thing i made was for the receiving handler only i suppose

15:33:56 <danbri> I'll update that to look more like stpeter's xml and see how it smells

15:34:22 <danbri> issue a response... would be great if you figure that out!

15:34:38 <danbri> i've a hope that if i can send the right shape IQ, my listener code might handle it

15:35:44 <crschmidt> can you send that IQ again?

15:37:39 <danbri> 2 mins

15:40:08 <danbri> <iq id="73y71-5" to="crschmidt@jabber-hispano.org/sparql" type="get">

15:40:10 <danbri> <query xmlns="http://www.w3.org/2005/09/xmpp-sparql-binding"> PREFIX foaf: <http://xmlns.com/foaf/0.1/> PREFIX rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> SELECT ?jabberID, ?name WHERE { ?x rdf:type foaf:Person . ?x foaf:jabberID ?jabberID . ?x foaf:name ?name }

15:40:11 <danbri> </query>

15:40:13 <danbri> </iq>

15:40:16 <danbri> that's more like it

15:40:31 <crschmidt> woot

15:40:47 <crschmidt> The handler now is taking that data in

15:40:54 <danbri> whee

15:40:58 <crschmidt> so I have a variable with:

15:40:59 <crschmidt> <iq from='foaf@jabber-hispano.org/sparqlclient' type='get' id='73y71-5' to='crschmidt@jabber-hispano.org/sparql'>

15:41:02 <crschmidt> <query xmlns = 'http://www.w3.org/2005/09/xmpp-sparql-binding' > PREFIX foaf: <http://xmlns.com/foaf/0.1/> PREFIX rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> SELECT ?jabberID, ?name WHERE { ?x rdf:type foaf:Person . ?x foaf:jabberID ?jabberID . ?x foaf:name ?name }

15:41:07 <crschmidt> </query></iq>

15:41:13 <crschmidt> not the cleanest thing yet, but i'm sure i can move on from there using some of the other functinos

15:41:21 <danbri> hurrah

15:41:29 <crschmidt> hm, would be nice if we could get an xmlns on the iq though, so i could listen for it specifically

15:41:34 <crschmidt> not sure if i 'm mixing up terms or something though

15:41:45 <danbri> no, the subelement i think is the right way

15:41:48 <danbri> i asked mr jabber

15:41:58 <danbri> well, unless i misunderstood!

15:42:06 <crschmidt> hm

15:42:26 <crschmidt> well, the registerHandler function has: def registerHandler(self,name,handler,type='',ns='',chained=False, makefirst=False, system=False):

15:42:34 <danbri> He suggested: <query xmlns='http://www.w3.org/2005/09/xmpp-sparql-binding'>

15:42:47 <danbri> I wasn't sure if 'query' was an invention of our little protocol, or a convention within Jabber

15:42:59 <danbri> I think more now the former, and hence we should base names on the DAWG protocol spec

15:43:16 <crschmidt> ah, nevermind

15:43:17 <crschmidt> right you are

15:43:19 <crschmidt> DEBUG: dispatch called for: name->iq ns->http://www.w3.org/2005/09/xmpp-sparql-binding

15:43:19 <dc_swig> Label DEBUG not found.

15:43:20 <danbri> which usese that name anyway

15:43:29 <danbri> see http://www.w3.org/TR/rdf-sparql-protocol/#ap

15:43:33 <crschmidt> so, i can register a protocol just for that NS

15:43:54 <danbri> It contains one operation, query, which is used to convey a SPARQL query -- including a SPARQL query string and an RDF dataset -- and a query result between clients (requesters) and services (responders).

15:44:10 <danbri> hmm how to pass in the dataset too? we might need to make subelements? if <iq> allows those

15:44:19 <danbri> guess we'll need subelements for results anyhow

15:44:35 * crschmidt changes client to only listen to iq in that ns

15:45:28 <crschmidt> danbri: can you send one more, just as a check?

15:45:29 <danbri> For now, I think use a single argument

15:45:55 <crschmidt> hm

15:46:05 <crschmidt> seems that the general iq Handler grabbed it too

15:46:19 <crschmidt> so I'll have to make sure not to put anything important in that handler, it handles any namespace, not just the null ns

15:46:45 <danbri> that was from my winxp laptop, but i think its ok. i see an outgoing query with the correct ns and subelement

15:50:59 * crschmidt moves entire setup to athena so he can test against julie's redland db

15:53:52 <crschmidt> okay, danbri, if you send another query, it should actually execute it and spit the results out on my local screen

15:53:56 <crschmidt> so if you want to try that, that'd be cool

15:54:16 <crschmidt> SELECT ?a ?b WHERE { <http://crschmidt.net/foaf.rdf#crschmidt> ?a ?b. } is a good sample query

15:54:19 <crschmidt> if you can alter them

15:55:12 <danbri> can you be "/sparqlserver' instead of /sparql, when being a queryable thing?

15:55:23 <crschmidt> i can be whatever i need to be

15:55:26 <danbri> thx

15:55:29 <crschmidt> it's just provided on command line

15:55:44 * danbri trying to make my code more customisable, but at moment am flipping between targetting yours and mine

15:56:20 <crschmidt> heh

15:59:20 <danbri> 1 min

16:01:43 <crschmidt> danbri: no query with that last message

16:02:03 <danbri> <iq id="BL4n8-5" to="crschmidt@jabber-hispano.org/sparqlserver" type="get"><query xmlns='http://www.w3.org/2005/09/xmpp-sparql-binding'>SELECT ?a ?b WHERE { <http://crschmidt.net/foaf.rdf#crschmidt> ?a ?b. }

16:02:04 <danbri> </query>

16:02:09 <danbri> maybe & < horkage

16:02:11 <crschmidt> hm

16:02:18 <crschmidt> send it once more?

16:02:41 <danbri> that's better

16:02:59 <danbri> <iq id="Q0kSr-5" to="crschmidt@jabber-hispano.org/sparqlserver" type="get">

16:03:01 <danbri> <query xmlns="http://www.w3.org/2005/09/xmpp-sparql-binding">SELECT ?a ?b WHERE { <http://crschmidt.net/foaf.rdf#crschmidt> ?a ?b. }

16:03:02 <danbri> </query>

16:03:04 <danbri> </iq>

16:04:27 <crschmidt> hm, yep, horked my client somehow

16:05:16 * crschmidt maybe fixes

16:05:45 * crschmidt goes to make lunch for four year old, feel free to resend

16:06:46 <danbri> resent

16:07:36 <ndw> Resend the lunch to your four year old?

16:07:38 * ndw grins

16:07:45 <danbri> :)

16:08:01 <crschmidt> ah, okay

16:08:05 <crschmidt> missed a module i need

16:08:48 <crschmidt> danbri: alright, last shot?

16:09:26 <danbri> my server says [java] parseIQ called - Parsing out some sparql...

16:09:32 <danbri> this is good :)

16:09:38 <danbri> ok rerunning it at yours

16:10:19 <crschmidt> hm, redland error

16:10:20 <crschmidt> Error: argument number 4: a 'char *' is expected, 'unicode(SELECT ?a ?b WHERE { <http://crschmidt.net/foaf.rdf#crschmidt> ?a ?b. } )' is received, Query: SELECT ?a ?b WHERE { <http://crschmidt.net/foaf.rdf#crschmidt> ?a ?b. }

16:10:39 <danbri2> is there a space before the final .

16:10:43 * danbri2 waves from laptop

16:10:49 <crschmidt> either way

16:11:10 <danbri2> we've reduced it to a sane problem anyway :)

16:11:56 * crschmidt attempts fix, restarts client

16:14:09 <danbri> public IQ parseIQ(org.xmlpull.v1.XmlPullParser parser) {}

16:14:20 <danbri> ...seems to be getting called when the query hits my server

16:16:53 <crschmidt> okay, if you send a query again now, looks like it *might* try to send back an XML version of the response in a message

16:17:32 <danbri2> no sign of a response

16:17:47 <crschmidt> meh

16:18:04 <crschmidt> not your fault

16:19:00 <danbri2> let me know when to retry. am getting my head around xml pull parser api

16:19:08 <danbri2> then i try plug in jena at this end

16:19:22 <crschmidt> can retry

16:19:22 <danbri2> and expose my itunes mp3 stash (metadata only :)

16:19:52 <crschmidt> meh

16:19:55 <crschmidt> still getting a weird error

16:20:14 <crschmidt> Error: argument number 4: a 'char *' is expected, 'unicode(SELECT ?a ?b WHERE { <http://crschmidt.net/foaf.rdf#crschmidt> ?a ?b. }

16:20:17 <crschmidt> )' is received, Query: SELECT ?a ?b WHERE { <http://crschmidt.net/foaf.rdf#crschmidt> ?a ?b.

16:20:24 <danbri2> does it work locally? jabber aside?

16:23:42 <crschmidt> yeah

16:24:03 <danbri> [java] parseIQ called - Parsing out some sparql...

16:24:04 <danbri> [java] Start tag query

16:24:06 <danbri> [java] Text SELECT ?a ?b WHERE { <http://crschmidt.net/foaf.rdf#crschmidt> ?a ?b. }

16:24:07 <danbri> [java] End tag query

16:24:09 <danbri> [java] Text

16:24:10 <danbri> [java] End tag iq

16:24:25 <danbri> OK my server can listen for queries, more or less...

16:25:14 <danbri> reckon you can send them too, once you've fixed this current problem?

16:25:26 <crschmidt> don't know how to send iq

16:25:37 <crschmidt> (and already spent too many cycles today :/ got paying work need to get done)

16:25:46 <crschmidt> last shot at server is in place if you want to give it a shot

16:26:23 <danbri> retrying

16:26:39 <danbri> hmm my server IQ handler code is getting all the presence events too, that wasn't intended

16:27:04 <crschmidt> yeah, must be an underlying redland something or other

16:27:45 <danbri> thanks for giving it a go, we're getting there

16:27:47 <crschmidt> would guess it's dying for some reason on self._query = Redland.librdf_new_query(_world._world, query_language, ruri, querystring, rbase_uri)

16:27:59 <crschmidt> so for some reason, querystring is not a char *

16:28:51 * crschmidt tries to use a str() aroudn the querystring, just in case that helps

16:32:03 <danbri2> retry?

16:32:20 <crschmidt> sure

16:32:23 <crschmidt> hey!

16:32:24 <crschmidt> that worked!

16:33:10 <danbri2> whee :)

16:33:19 <crschmidt> fixed XML bug

16:33:20 <crschmidt> try again?

16:34:02 <danbri2> how was that?

16:34:14 <crschmidt> hm

16:34:33 <crschmidt> okay, fixed again

16:35:32 <crschmidt> got a response?

16:35:40 <danbri2> my server instance saw a presence event when the query client came online to send its message your way

16:35:59 <danbri2> ah i see the reply

16:36:02 <danbri2> sent as a chat message

16:36:06 <crschmidt> yep

16:36:09 <crschmidt> haven't done anything else yet

16:36:14 <crschmidt> sending messages easy

16:36:20 <crschmidt> sending iqs - well, maybe not any harder, but different :)

16:36:23 <danbri2> entity escaped markup, with no results

16:36:29 <danbri2> IQs, exactly

16:36:29 <crschmidt> yeah

16:36:36 <danbri2> cool

16:36:37 <crschmidt> not sure why there was no results

16:36:59 <danbri2> so, with a bit more fiddling, we can all expose little rdf databases to everyone on our buddylist

16:37:07 <danbri2> there must be some use for such a thing!

16:37:33 <crschmidt> oh, i see it

16:37:51 <danbri2> the other thing apart from sending as an IQ, is that you should send it as a response to that message, so that the IDs tie them together

16:37:55 <danbri2> I think

16:38:23 * crschmidt nods

16:38:34 <crschmidt> Alright, you should at least get some result now

16:38:38 <crschmidt> if you try again

16:39:08 <crschmidt> not ideal, not sure why it's only returning that result

16:39:44 <danbri2> &lt;result> etc

16:39:54 <danbri2> &lt;result>

16:39:56 <danbri2> &lt;a uri="http://purl.org/net/inkel/rdf/schemas/lang/1.1#masters"/>

16:39:57 <danbri2> &lt;b>en&lt;/b>

16:39:59 <danbri2> &lt;/result>

16:40:22 <crschmidt> right

16:40:31 <crschmidt> oh, wait, double encoded

16:40:32 <crschmidt> that's goofy

16:40:38 * crschmidt guesses he doesn't need to do the XML on his end

16:40:49 <danbri2> quick, before Norm sees it

16:41:02 * crschmidt updates that

16:41:55 <crschmidt> ^q SELECT ?a ?b WHERE { <http://crschmidt.net/foaf.rdf#crschmidt> ?a ?b. }

16:42:06 <julie> foaf:icqChatID 47729022, foaf:interest http://www.livejournal.com/interests.bml?int=cell+phones, foaf:mbox_sha1sum bbff51a6d70630daafe242c186a6e27fda3e99c7, foaf:knows http://crschmidt.net/foaf.rdf#rho, foaf:gender neuter, foaf:knows (r1114530965r1209), foaf:knows http://crschmidt.net/~julie/me.rdf#julie, owl:sameAs http://copia.ogbuji.net/files/FOAF/chimezie.rdf#Christopher, foaf:interest http://www.livejournal.com/interests.bml?int=svg, foaf:kn

16:42:06 <julie> ows http://crschmidt.net/foaf.rdf#jpr, tipjar:hasTipjar (r1121976434r4), foaf:mbox mailto:crschmidt@crschmidt.net, foaf:knows http://crschmidt.net/foaf.rdf#loki, foaf:dorkthreshhold 8, foaf:interest http://www.livejournal.com/interests.bml?int=semantic+web, foaf:knows (r1114530965r22814), foaf:knows http://crschmidt.net/foaf.rdf#stace, kiss:kissed http://crschmidt.net/foaf.rdf#jpa, xfn:sweetheart http://crschmidt.net/foaf.rdf#jpa, foaf:name crsch

16:43:15 <crschmidt> ah!

16:43:18 <danbri2> oh crap, i'm in an infinitely long xml document, no wonder i'm not seeing my function return

16:43:19 <crschmidt> figured it out

16:43:32 <crschmidt> okay, you'll get better results now

16:43:38 <crschmidt> wasn't using julie's db

16:44:58 <danbri2> no result yet... is it wedged?

16:46:13 <danbri> meanwhile, when you're debugging i point my laptop client to send to my desktop server instead... which now handles the xml... [java] Extracted query: SELECT ?a ?b WHERE { <http://crschmidt.net/foaf.rdf#crschmidt> ?a ?b. }

16:46:30 <danbri> ...so i'm in same situation as you, ie could run the query, just ain't clear how to respond

16:47:35 <crschmidt> ah, nope, another broken thing, one sec

16:48:41 <crschmidt> fixed

16:48:52 * crschmidt is not so good at this :)

16:49:42 <crschmidt> there you go, much longer response :)

16:50:15 <danbri2> indeed so

16:50:25 <danbri2> can you do it non-escaped?

16:50:42 * crschmidt shakes his head

16:50:44 <crschmidt> i'm not escaping it

16:50:47 <crschmidt> library is

16:50:50 <danbri2> oh, cos a msg

16:50:52 <danbri2> right

16:50:56 <crschmidt> right

16:51:05 <crschmidt> so when it's not a message anymore, can probably do better

16:53:20 <danbri2> ah, when my server is offline, the query bounces back to client as a 404

16:53:28 <danbri2> but the client lib parses it autoamtically

16:53:49 <danbri2> i'll should make sure i only exec queries when in server mode, potential security hole there!

17:01:10 <crschmidt> query is definitely a Jabber specific thing, not a SPARQL/RDF thing

17:01:29 <danbri2> the choice of element name, within iq, is ours...

17:02:01 <danbri2> to do w3c sparql protocol properly, we'll need a convention for passing in choice of zero or more datasets too (by URI i think)

17:02:11 * danbri2 gets ARQ from CVS

17:02:30 <danbri2> i just committed all of Jena .jars into my cvs for 'convenience', tho could prolly survive on a subset

17:02:57 * danbri2 builds ARQ jar

17:27:12 <danbri2> ok I have it working locally, in that a sparql query comes in, and is executed by my code in server mode

17:27:26 <danbri2> haven't fig'd out how to return the results yet, nor do dataset selection

17:55:48 <valmont_> valmont_ is now known as valmont

18:23:09 <mkflow> hi to all

18:23:24 <danbri> hi

18:24:56 <danbri> logger, pointer?

18:24:56 <danbri> See http://www.ilrt.bris.ac.uk/discovery/chatlogs/swig/2005-09-06#T18-24-56

18:25:48 <dajobe> it points right here

18:28:13 <dmiles_afk> oh hey i been not reading up on advancements in RDF(S) lately but ternary predicates supported?

18:28:34 <dmiles_afk> and what happened with negation and negation by failure?

18:29:47 <mkflow> mkflow is now known as Mkflows

18:30:31 <danbri2> check out the www-rdf-rules list

18:30:44 <danbri2> and the rules workshop list that its been crossposted with lately

18:30:55 <danbri2> there's a charter being designed for a new SW/rdf/etc rules WG

18:31:01 * danbri2 isn't following closes

18:31:03 <crschmidt> Sandro's at the head of that, right?

18:31:05 <danbri2> closely

18:31:05 <dmiles_afk> ah yeah.. on the list .. starting to serhc my emails now

18:31:16 <danbri2> Sandro, yup

18:31:41 * sandro waves, but can't stay right now.

18:32:14 <Wikier> Wikier is now known as WikierOFF

18:32:21 <danbri> stringwriters... outputstreams... java ... :(

18:32:31 * crschmidt chuckles... mwhahaha

18:32:32 <danbri> all i want to do is get XML formatted results from ARQ...

18:32:35 <dmiles_afk> you liking the java language though?

18:32:42 <danbri> its been a while

18:33:36 <danbri> ok, if i can do rfo.outputAsXML(System.err) ; to get my XML on the console... and output writers are byte oriented... yet stringwriters are char oriented... how do i get the markup into a string?

18:33:51 <dmiles_afk> java is adictive it allows you to design stuff very well

18:33:53 <danbri> shouldn't be rocket science but i'm chasing around a web of javadocs and google...

18:34:12 <danbri> i don't want to design stuff well, i want to make a cheesy hack that kinda works, and leave it to the completer-finishers to engineer properly :)

18:34:43 <dmiles_afk> hehe.. yeah it lets you do both

18:35:03 <danbri> I'll go try bribe my housemates w/ cup of tea...

18:35:28 <dmiles_afk> rfo.outputAsXML(System.err)

18:35:38 <dmiles_afk> you want two modes correct?

18:36:03 <dmiles_afk> you can make a new PrintWriter(new StringWriter())

18:36:26 <dmiles_afk> erm oops rdfStr = new StringWriter();

18:36:31 <dmiles_afk> then wrapp it with a print writer

18:36:43 <dmiles_afk> ah but you probly knew that

18:38:15 <dmiles_afk> or PrintStream etc

18:40:07 <dmiles_afk> oh i see .. btye/char .. yeah in 1.5 they trying to make people think about it more.. but still not having to

18:40:09 <crschmidt> danbri: couple questions on jabber-stuff when you get a chance

18:44:18 <crschmidt> jabberpy is great stuff

18:44:20 <crschmidt> [[[

18:44:20 <crschmidt> iq = jabber.Iq(fromName,"get", "http://www.w3.org/2005/09/xmpp-sparql-binding")

18:44:24 <crschmidt> iq.setQueryPayload(result)

18:44:26 <crschmidt> print "IQ: <%s> %s" % (JID, iq.getQueryNode())

18:44:29 <crschmidt> con.send(iq)

18:44:31 <crschmidt> ]]]

18:44:50 <crschmidt> danbri: client should now send an Iq instead of msg as response

18:46:42 <crschmidt> not sure if it's how you're expecting it: sending a type='get', with NS as above, sending full SPARQL XML Query results

18:50:27 <danbri> i'll retry

18:51:36 <danbri2> ok i sent the wrong query

18:51:44 <danbri2> did you see it? no results response here yet

18:52:37 <crschmidt> no, my client is offline

18:52:38 <danbri2> resent correct query

18:52:40 <danbri2> ah

18:52:41 <crschmidt> i'm working on getting sending working

18:52:45 <danbri2> lemme know when

18:52:47 <danbri2> ok

18:52:59 <crschmidt> not sure how to include non-XML content in QueryPayload

18:54:12 <danbri2> such as?

18:54:22 <crschmidt> oh wait, my client did get your query

18:54:28 <crschmidt> hah, that'll take about 15 minutes

18:54:38 <crschmidt> okay, maybe not that long, but too long :)

18:54:53 <danbri2> the 1st query you mean? all jabberIDs and names

18:55:10 <crschmidt> yeah, but you selceted "?x rdf:type foaf:Person" first

18:55:26 <crschmidt> you apparently didn't read my blog post on "Optimizing SPARQL queries" :)

18:55:38 <crschmidt> SELECT ?jabberID ?name WHERE { ?x rdf:type foaf:Person . ?x foaf:jabberID ?jabberID . ?x foaf:name ?name }

18:55:43 <danbri2> i didnt' think order was supposed to matter

18:55:58 <crschmidt> ^q SELECT ?jabberID ?name WHERE { ?x foaf:jabberID ?jabberID . ?x foaf:name ?name. }

18:56:04 <crschmidt> that should be much quicker, but will still take a while

18:56:14 <julie> ressu@ressukka.net Sami Haahtinen, piotr.karwasz@jabber.org Piotr Przemysław Karwasz, dirkx@wirelessleiden.nl Dirk Willem van Gulik, masaka@jabber.org 神崎正英 (x2), sad-hu@Jabber.org Matti Lassila, phil@jabbernet.org Phil Wilson, sebastian.dietzold@imise.uni-leipzig.de Sebastian Dietzold, MattCroydon@jabber.com Matt Croydon, ajmitch@jabber.org Andrew "ajmitch" Mitchell, dirkx@jabber.net Asemantics, thecrypto@jabber.org Julian Krause, seitz@

18:56:14 <julie> jabber.com Bill Seitz, dirkx@jabber.net Dirk-Willem van Gulik, dhrasmus@charente.de Dan Rasmussen, schika@jabber.org Stephan Bock, chris@mkp.ca Chris Forsyth, kingmob@jabber.org Philipp Droessler, danbri@gnu.minu.nu Dan Brickley (x3), zcat@jabber.meta.net.nz Bruce "zcat" Kingsbury, danbri@jabber.org Dan Brickly, emacsen@jabber.org Serge Wroclawski, dirkx@wirelessleiden.nl Asemantics, dirkx@jabber.com Asemantics, freso@jabber.org Frederik 'Freso'

18:56:19 <crschmidt> yeah, that wasn't so bad

18:56:25 <crschmidt> Redland follows triples as they happen

18:57:03 <crschmidt> so first it'll find all the ?x which match rdf:type foaf:Person

18:57:09 <crschmidt> which is a big chunk of the dataset

18:57:10 <danbri2> ^q SELECT ?jabberID ?name WHERE { ?x foaf:jabberID ?jabberID . ?x foaf:name ?name . ?x foaf:mbox <mailto:danbri@w3.org> . }

18:57:20 <julie> danbri@jabber.org Dan Brickly, danbri@gnu.minu.nu Dan Brickly, danbri@gnu.minu.nu Dan Brickley (x2), danbri@jabber.org Dan Brickley (x2)

18:57:29 <danbri2> hmm should've put that last bit first? nah, that's the engine's job

18:58:04 <crschmidt> danbri: You can say that, and wait for the results, or you can just do the engine's job for it, which is the most appropriate given the current optimization strategies for engines

18:58:15 <crschmidt> going to kill the client searching for jabberIDs

18:58:24 <crschmidt> (hm, or try, stuck looping)

18:58:38 <crschmidt> 54% CPU, fun stuff

18:58:47 <danbri2> so new query... who spells my name 'Brickly'? :)

19:00:52 <WikierOFF> WikierOFF is now known as Wikier

19:01:00 <crschmidt> youch

19:01:17 <crschmidt> ^q select ?p where (?p foaf:name "Dan Brickly")

19:01:17 <julie> (r1124565507r7)

19:01:31 <crschmidt> ^q select ?n where (?p foaf:name "Dan Brickly") (?p2 foaf:knows ?p) (?p2 foaf:name ?n)

19:01:32 <julie> Morten Frederiksen@en, Jim, Jim Ley (x2), マカシネビル　チャールズ, شارل مككاثينبل, Charles McCathieNeville, crschmidt^^<http://www.w3.org/2001/XMLSchema#string>, شارل مككاثينبل@ar, マカシネビル　チャールズ@ja, Morten Frederiksen^^<http://www.w3.org/2001/XMLSchema#string>, Morten Frederiksen, Jim Ley^^<http://www.w3.org/2001/XMLSchema#string>, Morten, Charles McCathieNevile (Chaals), Christopher S

19:01:33 <julie> chmidt, Charles McCathieNevile

19:01:38 <crschmidt> hm

19:01:42 <crschmidt> no way of knowing. No contexts

19:01:48 <danbri2> ah

19:01:56 <crschmidt> alright, danbri, got an endpoint i can send queries at?

19:02:14 <crschmidt> got a client in place and working

19:02:23 <danbri2> try foaf2@

19:02:30 <crschmidt> example query?

19:02:36 <crschmidt> or does it not matter?

19:02:36 <danbri2> it can only query a fixed dataset, http://danbri.org/foaf.rdf

19:02:44 <danbri2> so the jabber query i sent you would work

19:03:19 <crschmidt> aw, fuck

19:03:24 <crschmidt> well, i got back:

19:03:25 <crschmidt> Tue Sep 6 15:03:12 2005 - RECV: - <iq xmlns='jabber:client' from='foaf2@jabber-hispano.org' type='error' to='crschmidt@jabber-hispano.org/sparqlclient'><error type='cancel' code='501'><feature-not-implemented xmlns='urn:ietf:params:xml:ns:xmpp-stanzas'/></error><query xmlns='http://www.w3.org/2005/09/xmpp-sparql-binding'>SELECT ?a ?b WHERE { ?a ?b ?c. }</query></iq>

19:03:36 <danbri2> i had that woe before

19:03:45 <crschmidt> which my client then caught

19:03:48 <danbri2> maybe i'm offline; that response came from server not my desktop

19:03:50 <crschmidt> and said "ooh, you're asking me a question!"

19:04:01 <danbri2> aha, i had that problem

19:04:04 <crschmidt> so now it's spinning and trying to spit out all of julie's dataset

19:04:05 <danbri2> i thought i was being a doofus w/ the apis

19:04:11 <crschmidt> yeah, remembered you mentioning it

19:04:19 <crschmidt> not sure if it is or not

19:04:34 <danbri2> the query handler is getting called whenever a query-related stanza floats past, and isn't being passed enough app context to know wtf is happening

19:04:40 <crschmidt> looks like errors are still sending the <query> back, so i guess needs to check that iq.firstChild is <query>

19:05:53 <danbri> can you retry?

19:06:10 <crschmidt> one sec

19:06:15 <crschmidt> lemme make it so my client doesn't execute

19:08:29 <crschmidt> DEBUG: sent <iq to='foaf2@jabber-hispano.org/sparqlserver' type='get'><query xmlns = 'http://www.w3.org/2005/09/xmpp-sparql-binding' >SELECT ?a WHERE { ?a ?b ?c. }</query></iq>

19:08:29 <dc_swig> Label DEBUG not found.

19:08:57 <crschmidt> danbri: I think we need to set up a different type for responses (other than "get"): thoughts?

19:10:05 <danbri> see http://www.jivesoftware.org/builds/smack/docs/latest/javadoc/org/jivesoftware/smack/packet/IQ.html

19:10:15 <danbri> Each IQ packet has a specific type that indicates what type of action is being taken: "get", "set", "result", or "error".

19:10:27 <danbri> pretty RESTy

19:10:56 <crschmidt> ah

19:11:08 <crschmidt> ah

19:11:31 <crschmidt> danbri: then should the Iq type for returns be "result"?

19:11:33 <danbri> the way I have to do it is: class SPARQLProvider implements IQProvider

19:11:41 <danbri> "result", yup

19:11:51 * danbri is going to go out on a limb here ;)

19:12:12 <crschmidt> i thinhk i knocked foaf2 sparqlserver offline

19:12:19 <danbri> where 0 results is an error, i guess we check the sparql protocol spec

19:12:38 <crschmidt> DEBUG: sent <iq to='foaf2@jabber-hispano.org/sparqlserver' type='get'><query xmlns = 'http://www.w3.org/2005/09/xmpp-sparql-binding' >SELECT ?jabberID ?name WHERE { ?x foaf:jabberID ?jabberID . ?x foaf:name ?name. }</query></iq>

19:12:38 <dc_swig> Label DEBUG not found.

19:12:41 <crschmidt> DEBUG: got data <presence xmlns='jabber:client' to='crschmidt@jabber-hispano.org/sparqlclient' from='foaf2@jabber-hispano.org/sparqlserver' type='unavailable' id='2SJ96-4'/>

19:12:42 <dc_swig> Label DEBUG not found.

19:12:48 <crschmidt> oh, heh

19:12:50 <crschmidt> no namespaces

19:13:06 <crschmidt> I bet that kicked your sparql query engine in the butt, uncaught exception, boomf, no more foaf2

19:13:27 <danbri> no, it didn't even register here

19:13:59 <crschmidt> hm

19:14:04 <crschmidt> well, something happened there :p

19:14:14 <danbri> [java] parseIQ called - Parsing out some sparql...

19:14:16 <danbri> [java] com.hp.hpl.jena.query.parser.QNameException: Line 1, column 35: Unresolved qname: foaf:jabberID

19:14:18 <danbri> [java] at com.hp.hpl.jena.query.parser.ParserBase.fixupQName(ParserBase.java:209)

19:14:19 <danbri> so it did

19:14:24 <crschmidt> hah, i totally win

19:14:36 <danbri> retry?

19:14:54 <crschmidt> sent another query your way

19:15:25 <crschmidt> no response from you though

19:15:39 <danbri> [java] <results ordered="false" distinct="false">

19:15:41 <danbri> [java] <result>

19:15:42 <danbri> [java] <binding name="jabberID">

19:15:44 <danbri> [java] <literal xml:lang="en">danbri@jabber.org</literal>

19:15:45 <danbri> [java] </binding>

19:15:47 <danbri> [java] <binding name="name">

19:15:48 <danbri> [java] <literal xml:lang="en">Dan Brickley</literal>

19:15:50 <danbri> [java] </binding>

19:15:51 <danbri> [java] </result>

19:15:53 <danbri> [java] </results>

19:16:16 <danbri> went to System.err here, thanks (shellac investigated...) to AndyS having removed the Writer facility from ARQ's output formatter

19:16:25 <crschmidt> ah

19:16:27 <danbri> will make a version with it in... and ask andy

19:16:40 <crschmidt> bah: Error: 'ascii' codec can't encode character u'\xf6' in position 464: ordinal not in range(128), Query: PREFIX foaf: <http://xmlns.com/foaf/0.1/> SELECT ?jabberID ?name WHERE { ?x foaf:jabberID ?jabberID . ?x foaf:name ?name. }

19:16:44 <danbri> however yr query didn't, for some reason, register in the Smack gui debug

19:16:54 <crschmidt> danbri: no <message> component

19:17:01 <crschmidt> Maybe?

19:17:10 <danbri> normally i see a log of all packets, by type

19:17:14 <danbri> as well as raw received packets

19:17:15 <crschmidt> Your Iq packets always include a <message> and <presence> packet too

19:17:15 <crschmidt> ah

19:17:18 <danbri> it shows in the latter

19:18:36 <danbri> ok, i tried my client

19:18:41 <danbri> lets find out the difference...

19:19:45 <danbri> mine: <iq xmlns='jabber:client' from='foaf@jabber-hispano.org/sparqlclient' type='get' to='foaf2@jabber-hispano.org/sparqlserver' id='puPEE-5'><query xmlns='http://www.w3.org/2005/09/xmpp-sparql-binding'>SELECT ?a ?b WHERE { <http://crschmidt.net/foaf.rdf#crschmidt> ?a ?b. } </query>

19:19:58 <danbri> yours: <query xmlns='http://www.w3.org/2005/09/xmpp-sparql-binding'>PREFIX foaf: <http://xmlns.com/foaf/0.1/> SELECT ?jabberID ?name WHERE { ?x foaf:jabberID ?jabberID . ?x foaf:name ?name. }</query></iq>

19:20:07 <danbri> can you set an id= automagically?

19:20:19 <danbri> it'll all mine (in theory) to send a responce that you could filter against

19:20:40 <danbri> ie. you should have some facility roughly like this:

19:20:42 <danbri> PacketCollector collector = con.createPacketCollector(

19:20:44 <danbri> new PacketIDFilter(sparqlRequest.getPacketID()));

19:20:45 <danbri> con.sendPacket(sparqlRequest);

19:21:15 <danbri> ...where sparqlRequest was my outgoing message, and I'm interested in filtering responses to it from the broader stream of xml stanzas coming my way

19:21:46 <crschmidt> i think my <iq> doesn't have an xmlns on it

19:22:35 <danbri> <iq xmlns='jabber:client' from='crschmidt@jabber-hispano.org/sparqlclient' type='get' to='foaf2@jabber-hispano.org/sparqlserver'>

19:22:42 <danbri> it does, sorry didn't copy that bit

19:22:52 <danbri> its the id=

19:22:59 <crschmidt> okay

19:23:03 <crschmidt> so I need to set an id...

19:23:19 <danbri> thats the ony difference i can see, explaining why it doesn't show up in the main pane of Smack debug

19:23:27 <danbri> and i know that that means the app won't be able to respond to it

19:31:54 <crschmidt> Tue Sep 6 15:31:33 2005 - SENT: - <iq to='foaf2@jabber-hispano.org/sparqlserver' type='get' id='5'><query xmlns = 'http://www.w3.org/2005/09/xmpp-sparql-binding' >PREFIX foaf: <http://xmlns.com/foaf/0.1/> SELECT ?jabberID ?name WHERE { ?x foaf:jabberID ?jabberID . ?x foaf:name ?name. }</query></iq>

19:32:35 <danbri> did you choose the value of id ?

19:33:00 * danbri wondering whether it has to match some production like letters-then-numbers

19:33:09 <crschmidt> nope

19:33:14 <crschmidt> con.getAnID

19:33:21 <danbri> <iq id="4" to="foaf2@jabber-hispano.org/sparqlserver" from="crschmidt@jabber-hispano.org/sparqlclient" type="get">

19:33:23 <danbri> <query xmlns="http://www.w3.org/2005/09/xmpp-sparql-binding"> PREFIX foaf: <http://xmlns.com/foaf/0.1/> PREFIX rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> SELECT ?jabberID, ?name WHERE { ?x rdf:type foaf:Person . ?x foaf:jabberID ?jabberID . ?x foaf:name ?name }

19:33:24 <danbri>

19:33:26 <danbri> </query>

19:33:27 <danbri> that arrived

19:33:38 <Wikier> Wikier is now known as WikierOFF

19:33:38 <danbri> hey, that explains why two queries appeared in my console when i sent only 1, doh :)

19:33:53 <crschmidt> heh, heh

19:34:03 <crschmidt> still not getting a response: should I be?

19:34:45 <danbri> no, i've not figured out how to

19:36:46 <crschmidt> okay

19:36:56 <danbri> getting there

19:39:45 <crschmidt> alright, so i have something that acts as both client and server

19:39:51 <crschmidt> and you have something that acts as client, right?

19:40:12 <danbri> yup, and a listens-but-won't-answer server

19:40:20 <crschmidt> heh, heh

19:40:29 <crschmidt> i had one of those hours ago! yet again, python wins over java

19:40:55 * danbri ponders, decides he's happy to blame the tools on this occasion

19:41:09 <danbri> though Smack seems pretty solid

19:41:18 <danbri> i'm just rusty

19:41:59 * crschmidt nods

19:42:14 * danbri re-reads http://www.jivesoftware.org/builds/smack/docs/latest/documentation/providers.html

19:42:19 <crschmidt> I think the fact that in 2 hours of sit-down time, I've hacked it together is still pretty impressive

19:43:02 <danbri> OK i guess i'm trying to do too much inside the IQ provider code

19:44:48 <crschmidt> i'm still not sure how to set a namespace on the <iq>

19:44:51 <crschmidt> not sure if it's required or not

19:45:13 <danbri2> you don't need to

19:45:20 <danbri2> see above, i think yours and mine are same now

19:45:42 <crschmidt> assuming you don't need that, yes

19:45:51 <danbri2> i've figured out what to do... in SPARQLProvider *all* I have to do is generate a SPARQLIQ java object that encapsulates the sparql query

19:45:58 <danbri2> then I re-capture it again with a listener

19:46:05 <danbri2> that'll give me the context i need to respond

19:59:15 * danbri wonders Java's preferred way of escaping text for XML emissions

20:38:24 <WikierOFF> WikierOFF is now known as Wikier

21:29:44 <raxor> raxor is now known as raxor|away

21:48:27 <danbri> [java] Server got a packet of interest: <iq id="BGvDg-5" to="foaf2@crschmidt.net/sparqlserver" from="foaf@crschmidt.net/sparqlclient" type="get">

21:48:29 <danbri> [java] <query xmlns='http://www.w3.org/2005/09/xmpp-sparql-binding'>

21:48:31 <danbri> [java] PREFIX foaf: <http://xmlns.com/foaf/0.1/> PREFIX rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> SELECT ?jabberID, ?name WHERE { ?x rdf:type foaf:Person . ?x foaf:jabberID ?jabberID . ?x foaf:name ?name } [java]

21:48:33 <danbri> [java] </query>

21:48:33 <danbri> [java] </iq>

21:48:45 <danbri> OK fixed all the weirdness, got a custom packet listener wired up, it's all looking good

21:49:14 <danbri> -> http://rdfweb.org/viewcvs/viewcvs.cgi/foafproject/htdocs/2005/code/foaftown/jabber/

21:49:35 * danbri notices the clock's wrong on the webserver, great...

21:51:26 * crschmidt attempts to query against foaf2, fails?

21:51:37 <danbri> its mostly offline

21:51:45 <danbri> just trying to rewire jena in

22:45:41 <raxor|away> raxor|away is now known as raxor

The IRC chat here was automatically logged without editing and contains content written by the chat participants identified by their IRC nick. No other identity is recorded.