Menu
▾
▴
swobjects-devel
|
From: Eric Prud'h. <er...@w3...> - 2011-01-16 17:59:48
|
* 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
|
|
From: Andrea S. <and...@bb...> - 2011-01-18 02:01:13
|
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.
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 ?
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)
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.
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...
|
|
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
|
|
From: Andrea S. <and...@bb...> - 2011-01-19 12:59:32
|
Hi,
Yes, a real logger would be nice.
I have to restart my laptop to use WireShark... I usually restart it every 2-3 months. So... give me a couple of days! (I do long term ultra multi-tasking...)
ciao,
Andrea
Il giorno 18/gen/2011, alle ore 15.35, Eric Prud'hommeaux ha scritto:
> * 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
Andrea Splendiani
Senior Bioinformatics Scientist
Centre for Mathematical and Computational Biology
+44(0)1582 763133 ext 2004
and...@bb...
|
|
From: Eric Prud'h. <er...@w3...> - 2011-01-16 18:29:18
|
* Eric Prud'hommeaux <er...@w3...> [2011-01-16 12:59-0500]
> * 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.
I forgot to add, there's another possible strategy here, which is to make the sparql executable also be an apache module, leaning on apache for the generic file serving and registering a handler which takes a partition of the web space for named graphs with relative names (e.g. "tree/and/leaf" vs. "/root/tree/and/leaf" or "http://server/root/tree/and/leaf"). I have no idea how easy it is to compile a .so (or .dynlib) for one version of apache and run it on another.
> > 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
--
-ericP
|
×
Want the latest updates on software, tech news, and AI?
Get latest updates about software, tech news, and AI from SourceForge directly in your inbox once a month.
Thanks for helping keep SourceForge clean.
X