Menu
▾
▴
Re: [SWObjects-devel] Question about SWObjects
|
From: Eric Prud'h. <er...@w3...> - 2011-01-18 15:36:05
|
* Andrea Splendiani <and...@bb...> [2011-01-18 01:15+0000]
> Hi,
>
> I have some problem with SWObjects.
>
> first: with debug on it prints a lot of information, but not all so
> informative. It prints a lot of html (possibly the one that is served) but I
> don't easily find the query translation (which is pretty useful).
> Also, I think it prints the binary png all the time the browser request
> it... which is a bit annoying.
It looks like it's time to put in a real logger with debug levels and that sort of thing. There's a package proposed for the boost libraries; I've been hoping it would be incorporated into the mainstream before I needed it, but I can get around this by incorporating the source directly into the package. Due to lots of demands, this will take me a couple days.
You can reduce the noise but invoking the query you want to investigate from the command line, à la: /Applications/Sparql.app/Contents/MacOS/Sparql --debug 1 -m YourMapFile -npe 'SomeQuery'
> I'm trying to put SWObjects in the middle of RelFinder on one side and
> Joseki on the other. SWObjects starts with a mapping rule which has a
> Service clause which match to the Joseki endpoint.
>
> So far I had mixed experiences.
>
> Re: RelFinder.
> I've used reverse proxy to get around the crosdmain.xml issue.
> I didn't manage to have RelFinder working on the SWObjects endpoint yet
> (still have to play a bit with configurations). However SWObjects easily
> gets to consume all CPU for hours without any debug message. Maybe it gets
> in some loop ?
Yeah, sounds like an issue with the HTTP server (which needs a lot of hardening). I'm guessing we'll see better reliability if I just do the Apache module hack (as writing a real HTTP 1.1 server is a sizable project). There's also the possibility of using JNI (I've never compiled it on a mac, but is hopefully not too hard) or calling the sparql executable as a CGI script.
> Re: Joseki.
>
> I had mixed results.
>
> Sometimes it works.
>
> For a query like:
>
> select ?x ?y
> where {?x <http://www.test.org/causativeOf> ?y }
>
> (the predicate is constructed in a mapping file, and that's the only rule).
>
>
>
> With this query:
>
> select ?x ?xl ?y
> where {?x <http://www.test.org/causativeOf> ?y .
> ?x ?p1 ?xl }
>
> JOSEKI gives this:
>
> 01:10:43 INFO SPARQL :: OK/select: SELECT DISTINCT
> ?_0x10100e800_0_specGene ?x ?_0x10100e800_0_pert ?_0x10100e800_0_interaction
> ?_0x10100e800_0_phenotype ?y WHERE { ?_0x10100e800_0_specGene
> <http://localhost:8180/phibase/hasReferenceGene> ?x . ?_0x10100e800_0_pert
> <http://localhost:8180/phibase/relativeToExperimentSpecificEntity>
> ?_0x10100e800_0_specGene . ?_0x10100e800_0_interaction
> <http://localhost:8180/phibase/Perturbation2Interaction>
> ?_0x10100e800_0_pert . ?_0x10100e800_0_phenotype
> <http://localhost:8180/phibase/GenericPhenotype_interactionID>
> ?_0x10100e800_0_interaction . ?_0x10100e800_0_phenotype
> <http://localhost:8180/phibase/GenericPhenotype_CV-ProcessPhenotype> ?y . }
> 01:10:47 WARN log :: handle failed
> java.io.IOException: FULL head
> at org.mortbay.jetty.HttpParser.parseNext(HttpParser.java:276)
> at org.mortbay.jetty.HttpParser.parseAvailable(HttpParser.java:205)
> at org.mortbay.jetty.HttpConnection.handle(HttpConnection.java:380)
> at
> org.mortbay.jetty.nio.BlockingChannelConnector$Connection.run(BlockingChannelConnector.java:163)
> at
> org.mortbay.thread.QueuedThreadPool$PoolThread.run(QueuedThreadPool.java:488)
Let's see if I understand this failure; JOSEKI is crashing when reading the query, presumably 'cause Sparql is sending a broken HTTP packet? Can you run a packet sniffer like http://www.wireshark.org/download.html ? That should help us see exactly what's going wrong.
> Note that the translated query works directly on JOSEKI. It is possibly
> something that has to do with the http messanging.
>
>
> Sometimes (not easily reproducible) SWOBjects yield nothing but the query
> doesn't even go to JOSEKI (but this is a bit vague)
>
>
> Any hint ?
>
> Regarding deployment. In the long run, it would be nice to have one (or
> more) SPARQL module(s) in apache conf. But that's not something so urgent.
... unless it's the shortest path to satisfaction on the above HTTP issues...
> ciao,
> Andrea
>
>
> Il giorno 16/gen/2011, alle ore 17.59, Eric Prud'hommeaux ha scritto:
>
> > * Andrea Splendiani <and...@bb...> [2011-01-16
> 14:45+0000]
> >> Hi,
> >>
> >> now it works, though at times I had some strange error where in the logs
> you see some binary blurp (like an image). But that's not so common and
> reproducible, may be something else around the system.
> >
> > Is that only when you have --debug enabled?
> > If so, it's probably your web browser asking for the favicon and the
> > server responding with a silly "mr yuck" face.
> >
> >
> >> a question: is there a way to serve some file through the same http
> server which serves sparql ?
> >> I would need a crossdomain.xml file for flash!
> >
> > It would be pretty trivial to hack the source to do that (I could help
> there).
> > I'd like to add this functionality in general, but I haven't worked out
> the packaging issues. Specifically, I'd like to take the favicon and the
> generated HTML at the root out and put them into the filesystem somewhere.
> Issues:
> > where's that somewhere and how does the executable find it? (it should
> have a convenient zero-config behavior)
> > what's the interaction between named graphs and the filesystem? GET and
> PUT of some named graph has nice behaviors now; should they be shadowed in
> the filesystem?
> >
> > TimBL's leaning on me for the filesystem shadow. I have to write some
> serializers to make it work.
> >
> >
> >> ciao,
> >> Andrea
> >>
> >> Il giorno 16/gen/2011, alle ore 04.42, Eric Prud'hommeaux ha scritto:
> >>
> >>> * Andrea Splendiani <and...@bb...> [2011-01-15
> 17:30+0000]
> >>>> Hi,
> >>>>
> >>>> I was trying to use the SWObjects that you (tele)presented at the
> SWAT4LS tutorial in December.
> >>>>
> >>>> I have a few problems, perhaps I'm getting something wrong:
> >>>>
> >>>> So, I'm starting a Sparql server with the following command:
> >>>>
> >>>> ./Sparql --debug 2 -npm
> /Users/andreasplendiani/Dropbox/WorkingProjects/PhiBase2/phibaseRDFMap.mp
> --serve http://localhost:8200/SPARQL -i
> /Users/andreasplendiani/Dropbox/WorkingProjects/PhiBase2/d2r-server-0.7/PhiBaseRDFv003.n3
>
> >>>>
> >>>> The maps is something as:
> >>>>
> >>>> CONSTRUCT
> >>>> {
> >>>> ?gene <http://www.test.org/causativeOf> ?specGene
> >>>> }
> >>>> WHERE {
> >>>> ?specGene <http://localhost:8180/phibase/hasReferenceGene> ?gene
> >>>> }
> >>>>
> >>>> When I query the endpoint, I get:
> >>>>
> >>>> (endpoint result):
> >>>>
> >>>> Q&D SPARQL Server Error
> >>>>
> >>>> SELECT ?x ?y ?z
> >>>> where {?x ?y ?z}
> >>>>
> >>>> yeilded
> >>>>
> >>>> failed to match triples prefixed by "!" in
> >>>> SELECT ?x ?y ?z
> >>>> WHERE
> >>>> {
> >>>> ! ?x ?y ?z .
> >>>> }
> >>>
> >>> that version didn't support wildcard predicates over mapped rules. could
> you try the latest?
> >>>
> >>>> (log):
> >>>>
> >>>> 127.0.0.1 - - [2011-00-15T17:23:00]"GET /favicon.ico 1.1" 200 620
> >>>> GET /favicon.ico HTTP/1.1
> >>>> Host: 127.0.0.1:8200
> >>>> User-Agent: Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10.6; en-US;
> rv:1.9.2.13) Gecko/20101203 Firefox/3.6.13
> >>>> Accept:
> text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8
> >>>> Accept-Language: en-us,en;q=0.5
> >>>> Accept-Encoding: gzip,deflate
> >>>> Accept-Charset: ISO-8859-1,utf-8;q=0.7,*;q=0.7
> >>>> Keep-Alive: 115
> >>>> Connection: keep-alive
> >>>> Cookie:
> __utma=96992031.761815913.1280272954.1288624527.1291566200.4;
> __utmz=96992031.1280272954.1.1.utmcsr=(direct)|utmccn=(direct)|utmcmd=(none)
> >>>>
> >>>>
> >>>> HTTP/1.0 200 OK
> >>>> Content-Type: image/x-icon
> >>>>
> >>>> ?PNG
> >>>>
> >>>> (then binary blurp).
> >>>>
> >>>> Not sure what's going on, any idea ?
> >>>> Same things if I try via SERVICE (instead of actually pointing to a
> local dataset, which I suspect is not read)
> >>>>
> >>>> Is there a mailing list for SWObjects ?
> >>>>
> >>>> ciao,
> >>>> Andrea
> >>>>
> >>>>
> >>>>
> >>>> Andrea Splendiani
> >>>> Senior Bioinformatics Scientist
> >>>> Centre for Mathematical and Computational Biology
> >>>> +44(0)1582 763133 ext 2004
> >>>> and...@bb...
> >>>>
> >>>>
> >>>>
> >>>
> >>> --
> >>> -ericP
> >>
> >> Andrea Splendiani
> >> Senior Bioinformatics Scientist
> >> Centre for Mathematical and Computational Biology
> >> +44(0)1582 763133 ext 2004
> >> and...@bb...
> >>
> >>
> >>
> >
> > --
> > -ericP
>
> Andrea Splendiani
> Senior Bioinformatics Scientist
> Centre for Mathematical and Computational Biology
> +44(0)1582 763133 ext 2004
> and...@bb...
>
>
>
>
--
-ericP
|
