You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@marmotta.apache.org by Gustavo Mora <gm...@gmail.com> on 2017/08/11 22:21:12 UTC
Re: Road to RDF4J
Hi,
The final date is close. I'm working in the wiki, but I'd like to know
which solution to take to some of the issues that I have pointed out.
Besides that, I tested Marmotta and it's working.
Cheers,
Gustavo.
2017-07-22 12:26 GMT-05:00 FRANCISCO XAVIER SUMBA TORAL <
xavier.sumba93@ucuenca.ec>:
> Hi,
>
> In RDF4J, there’re two new tests [1] to evaluate large transactions of
> size 200000. Those tests are working but takes too long to execute. The
> only concern is compiling time. So, should we include those tests.
>
>
> > On Jul 19, 2017, at 12:56, Sergio Fernández <wi...@apache.org> wrote:
> >
> > Hi again,
> >
> > On Tue, Jul 18, 2017 at 9:25 PM, Gustavo Mora <gm...@gmail.com>
> wrote:
> >>
> >> Other thing about module/classes name conventions. Do we should maintain
> >> Sesame word in Marmotta modules or classes? or should we change module
> >> names like SesameValueBackend to RDF4JValueBackend or Apache Marmotta:
> >> Sesame HTTP Backend to Apache Marmotta: RDF4J HTTP Backend.
> >>
> >
> > My vote would go for s/Sesame/RDF4J. What do other think?
>
> I’d say to change everything for RDF4J since Marmotta is not longer using
> Sesame.
>
> >
> > Also, the dependency (TinkerPop 2.x) used in Titan and Accumulo backends
> >> are deprecated, and GraphSail haven't been migrated to TinkerPop 3.x
> [1].
> >> TinkerPop 2 uses Sesame 2.7.
> >>
> >
> > At some point we shuld get rid of those experimental backends that nobody
> > uses…
>
> So, should we ignore those modules?
>
> >
> > Besides those problems most of the tests are now working, and I'm trying
> to
> >> fix some tests incorporated in RDF4J.
> >>
> >
> > Great!!!
> >
> >
> >
> >
> > [1] https://groups.google.com/forum/#!topic/gremlin-users/CtSLC64gKZA
> >>
> >> 2017-07-16 13:38 GMT-05:00 Gustavo Mora <gm...@gmail.com>:
> >>
> >>> Hi,
> >>>
> >>> Marmotta is working with Sesame 4, I started migrating to RDF4J. Before
> >>> that I encountered two problems.
> >>>
> >>> 1. The test ComplexKiWiSparqlQueryTest -> ComplexSPARQLQueryTest#
> >> testSES1979MinMaxInf
> >>> is failing 'cause KiWi changes literal's datatype at time of insertion.
> >>>
> >>> For example, the following insertion is successful:
> >>>
> >>> PREFIX : <http://example.org/>
> >>> INSERT DATA
> >>> {
> >>> GRAPH <http://test/test> {
> >>> :a14 :float "-3.4428452385e-13"^^xsd:float .
> >>> :a100 :float "-INF"^^xsd:float .
> >>> :a101 :float "INF"^^xsd:float .
> >>> }
> >>> }
> >>>
> >>>
> >>> When querying KiWi store (SELECT * { GRAPH < http://test/test > { ?a
> ?b
> >>> ?c } }), we get the following results:
> >>>
> >>> *A*
> >>>
> >>> *B*
> >>>
> >>> *C*
> >>>
> >>> http://example.org/a14
> >>>
> >>> http://example.org/float
> >>>
> >>> “-3.4428452385E-13” ^^xsd:float
> >>>
> >>> http://example.org/a100
> >>>
> >>> http://example.org/float
> >>>
> >>> “-INF”^^xsd:string
> >>>
> >>> http://example.org/a101
> >>>
> >>> http://example.org/float
> >>>
> >>> “INF”^^xsd:string
> >>>
> >>> The problem is with special values of float (i.e. NaN, INF, -INF [1])
> >>> stored in Marmotta because the XSD float datatype definition is ignored
> >> and
> >>> replaced with the default (xsd:string). This problem is probably
> present
> >>> in both inserting and querying processes within KiWi.
> >>>
> >>>
> >>> 2. Some problems to execute LdpSuiteTest#testRunSuite because the suite
> >>> used (LdpTestSuite[2]) is deprecated; it uses Sesame 2. Major problems
> >>> between URI and IRI.
> >>>
> >>>
> >>> As a solution, I propose creating an issue for each problem. Or any
> other
> >>> ideas to approach this issues?
> >>>
> >>>
> >>> Cheers,
> >>>
> >>> Gustavo.
> >>>
> >>>
> >>> [1] https://www.w3.org/TR/xmlschema-2/#float
> >>>
> >>> [2] https://w3c.github.io/ldp-testsuite/
> >>>
> >>>
> >>>
> >>
>
>
> [1] https://github.com/eclipse/rdf4j/blob/48897c4205fb2ffb6eda7e244b381a
> 1f81fd9ffd/testsuites/store/src/main/java/org/eclipse/rdf4j/sail/
> SailConcurrencyTest.java#L159-L239 <https://github.com/eclipse/rdf4j/blob/
> 48897c4205fb2ffb6eda7e244b381a1f81fd9ffd/testsuites/store/
> src/main/java/org/eclipse/rdf4j/sail/SailConcurrencyTest.java#L159-L239>
Re: Road to RDF4J
Posted by Gustavo Mora <gm...@gmail.com>.
Hi.
I'll keep working on the open issues and I hope you guys can help me too
with them. Specifically, I will focus on finding an alternative for the LDP
test suite that works in RDF4J. Also I could replace the Sesame word for
RDF4J in all the modules' names, if you guys agree.
The other issues that the migration process uncovered on Marmotta are
probably beyond my current knowledge, and I will definitely need your help.
In particular those related with the SPARQL to SQL translations on optional
statements [1], the optimization of Kiwi to handle large transactions more
efficiently [2] and the support for special values of the float datatype
[3]. I will be creating/updating entries in jira for these issues so you
guys can help me out with some insight about how to approach them.
Finally, regarding to the absence of an implementation for GraphSail
(ThinkerPop 3.x) in the experimental backbends, it would take a while to
solve until the community releases a compatible implementation for RDF4J
,but it's comming [4]. I'll be on the lookout of any news in their mailing
lists.
Regards,
Gustavo.
[1] https://issues.apache.org/jira/browse/MARMOTTA-603
[2] https://lists.apache.org/thread.html/2f6108d5759955693e3173ede8fb27
43da887dd92acc4e540d5c3409@%3Cdev.marmotta.apache.org%3E
[3] https://lists.apache.org/thread.html/964d71b7a0b2ab7c21bf6a9483d96b
4eb416f4aebd8acca61c8602dc@%3Cdev.marmotta.apache.org%3E
[4] https://groups.google.com/forum/#!topic/gremlin-users/CtSLC64gKZA
2017-08-28 17:05 GMT-05:00 Sergio Fernández <wi...@apache.org>:
> Hi Gustavo,
>
> thanks, the documentation is something very important for us as outcome of
> the program.
>
> GSoC officially ends tomorrow. We knew the project was difficult. As it
> looks there are so many unknowns in the tasks, it'd be great if you can
> keep working on addressing some of the open issues in the next few weeks.
>
> I hope once we can focus on getting 3.4.0 release out I can move to address
> with you some of them. So please, keep updating PR #31 [1] and MARMOTTA-659
> [2] and discussing here all the development questions. Probably it' d be
> good for everybody if you can refresh on the individual issues we have
> still open.
>
> Thanks! We really appreciate the work so have done so far. And I really
> hope you stay with us for long ;-)
>
> Cheers,
>
> [1] https://github.com/apache/marmotta/pull/31
> [2] https://issues.apache.org/jira/browse/MARMOTTA-659
>
>
> On Wed, Aug 23, 2017 at 3:28 PM, Gustavo Mora <gm...@gmail.com> wrote:
>
> > Hi,
> >
> > I've generated some documentation in the wiki [1]. Also, Marmotta is
> > working with rdfj4. Xavier took a look to the documentation, but I also
> > would like to know what do you think Sergio about the documentation and
> > code changes. Is there something I should change? Also, I'd like to
> > continue working on it after GSOC.
> >
> > Cheers,
> > Gustavo.
> >
> > [1] https://wiki.apache.org/marmotta/GSoC/2017/MARMOTTA-659
> >
> > 2017-08-11 17:21 GMT-05:00 Gustavo Mora <gm...@gmail.com>:
> >
> > > Hi,
> > >
> > > The final date is close. I'm working in the wiki, but I'd like to know
> > > which solution to take to some of the issues that I have pointed out.
> > > Besides that, I tested Marmotta and it's working.
> > >
> > > Cheers,
> > > Gustavo.
> > >
> > > 2017-07-22 12:26 GMT-05:00 FRANCISCO XAVIER SUMBA TORAL <
> > > xavier.sumba93@ucuenca.ec>:
> > >
> > >> Hi,
> > >>
> > >> In RDF4J, there’re two new tests [1] to evaluate large transactions of
> > >> size 200000. Those tests are working but takes too long to execute.
> The
> > >> only concern is compiling time. So, should we include those tests.
> > >>
> > >>
> > >> > On Jul 19, 2017, at 12:56, Sergio Fernández <wi...@apache.org>
> > wrote:
> > >> >
> > >> > Hi again,
> > >> >
> > >> > On Tue, Jul 18, 2017 at 9:25 PM, Gustavo Mora <gm...@gmail.com>
> > >> wrote:
> > >> >>
> > >> >> Other thing about module/classes name conventions. Do we should
> > >> maintain
> > >> >> Sesame word in Marmotta modules or classes? or should we change
> > module
> > >> >> names like SesameValueBackend to RDF4JValueBackend or Apache
> > Marmotta:
> > >> >> Sesame HTTP Backend to Apache Marmotta: RDF4J HTTP Backend.
> > >> >>
> > >> >
> > >> > My vote would go for s/Sesame/RDF4J. What do other think?
> > >>
> > >> I’d say to change everything for RDF4J since Marmotta is not longer
> > using
> > >> Sesame.
> > >>
> > >> >
> > >> > Also, the dependency (TinkerPop 2.x) used in Titan and Accumulo
> > backends
> > >> >> are deprecated, and GraphSail haven't been migrated to TinkerPop
> 3.x
> > >> [1].
> > >> >> TinkerPop 2 uses Sesame 2.7.
> > >> >>
> > >> >
> > >> > At some point we shuld get rid of those experimental backends that
> > >> nobody
> > >> > uses…
> > >>
> > >> So, should we ignore those modules?
> > >>
> > >> >
> > >> > Besides those problems most of the tests are now working, and I'm
> > >> trying to
> > >> >> fix some tests incorporated in RDF4J.
> > >> >>
> > >> >
> > >> > Great!!!
> > >> >
> > >> >
> > >> >
> > >> >
> > >> > [1] https://groups.google.com/forum/#!topic/gremlin-users/
> CtSLC64gKZA
> > >> >>
> > >> >> 2017-07-16 13:38 GMT-05:00 Gustavo Mora <gm...@gmail.com>:
> > >> >>
> > >> >>> Hi,
> > >> >>>
> > >> >>> Marmotta is working with Sesame 4, I started migrating to RDF4J.
> > >> Before
> > >> >>> that I encountered two problems.
> > >> >>>
> > >> >>> 1. The test ComplexKiWiSparqlQueryTest -> ComplexSPARQLQueryTest#
> > >> >> testSES1979MinMaxInf
> > >> >>> is failing 'cause KiWi changes literal's datatype at time of
> > >> insertion.
> > >> >>>
> > >> >>> For example, the following insertion is successful:
> > >> >>>
> > >> >>> PREFIX : <http://example.org/>
> > >> >>> INSERT DATA
> > >> >>> {
> > >> >>> GRAPH <http://test/test> {
> > >> >>> :a14 :float "-3.4428452385e-13"^^xsd:float .
> > >> >>> :a100 :float "-INF"^^xsd:float .
> > >> >>> :a101 :float "INF"^^xsd:float .
> > >> >>> }
> > >> >>> }
> > >> >>>
> > >> >>>
> > >> >>> When querying KiWi store (SELECT * { GRAPH < http://test/test > {
> > ?a
> > >> ?b
> > >> >>> ?c } }), we get the following results:
> > >> >>>
> > >> >>> *A*
> > >> >>>
> > >> >>> *B*
> > >> >>>
> > >> >>> *C*
> > >> >>>
> > >> >>> http://example.org/a14
> > >> >>>
> > >> >>> http://example.org/float
> > >> >>>
> > >> >>> “-3.4428452385E-13” ^^xsd:float
> > >> >>>
> > >> >>> http://example.org/a100
> > >> >>>
> > >> >>> http://example.org/float
> > >> >>>
> > >> >>> “-INF”^^xsd:string
> > >> >>>
> > >> >>> http://example.org/a101
> > >> >>>
> > >> >>> http://example.org/float
> > >> >>>
> > >> >>> “INF”^^xsd:string
> > >> >>>
> > >> >>> The problem is with special values of float (i.e. NaN, INF, -INF
> > [1])
> > >> >>> stored in Marmotta because the XSD float datatype definition is
> > >> ignored
> > >> >> and
> > >> >>> replaced with the default (xsd:string). This problem is probably
> > >> present
> > >> >>> in both inserting and querying processes within KiWi.
> > >> >>>
> > >> >>>
> > >> >>> 2. Some problems to execute LdpSuiteTest#testRunSuite because the
> > >> suite
> > >> >>> used (LdpTestSuite[2]) is deprecated; it uses Sesame 2. Major
> > problems
> > >> >>> between URI and IRI.
> > >> >>>
> > >> >>>
> > >> >>> As a solution, I propose creating an issue for each problem. Or
> any
> > >> other
> > >> >>> ideas to approach this issues?
> > >> >>>
> > >> >>>
> > >> >>> Cheers,
> > >> >>>
> > >> >>> Gustavo.
> > >> >>>
> > >> >>>
> > >> >>> [1] https://www.w3.org/TR/xmlschema-2/#float
> > >> >>>
> > >> >>> [2] https://w3c.github.io/ldp-testsuite/
> > >> >>>
> > >> >>>
> > >> >>>
> > >> >>
> > >>
> > >>
> > >> [1] https://github.com/eclipse/rdf4j/blob/48897c4205fb2ffb6eda7e
> > >> 244b381a1f81fd9ffd/testsuites/store/src/main/java/org/
> > >> eclipse/rdf4j/sail/SailConcurrencyTest.java#L159-L239 <
> > >> https://github.com/eclipse/rdf4j/blob/48897c4205fb2ffb6eda7
> > >> e244b381a1f81fd9ffd/testsuites/store/src/main/
> > >> java/org/eclipse/rdf4j/sail/SailConcurrencyTest.java#L159-L239>
> > >
> > >
> > >
> >
>
Re: Road to RDF4J
Posted by Sergio Fernández <wi...@apache.org>.
Hi Gustavo,
thanks, the documentation is something very important for us as outcome of
the program.
GSoC officially ends tomorrow. We knew the project was difficult. As it
looks there are so many unknowns in the tasks, it'd be great if you can
keep working on addressing some of the open issues in the next few weeks.
I hope once we can focus on getting 3.4.0 release out I can move to address
with you some of them. So please, keep updating PR #31 [1] and MARMOTTA-659
[2] and discussing here all the development questions. Probably it' d be
good for everybody if you can refresh on the individual issues we have
still open.
Thanks! We really appreciate the work so have done so far. And I really
hope you stay with us for long ;-)
Cheers,
[1] https://github.com/apache/marmotta/pull/31
[2] https://issues.apache.org/jira/browse/MARMOTTA-659
On Wed, Aug 23, 2017 at 3:28 PM, Gustavo Mora <gm...@gmail.com> wrote:
> Hi,
>
> I've generated some documentation in the wiki [1]. Also, Marmotta is
> working with rdfj4. Xavier took a look to the documentation, but I also
> would like to know what do you think Sergio about the documentation and
> code changes. Is there something I should change? Also, I'd like to
> continue working on it after GSOC.
>
> Cheers,
> Gustavo.
>
> [1] https://wiki.apache.org/marmotta/GSoC/2017/MARMOTTA-659
>
> 2017-08-11 17:21 GMT-05:00 Gustavo Mora <gm...@gmail.com>:
>
> > Hi,
> >
> > The final date is close. I'm working in the wiki, but I'd like to know
> > which solution to take to some of the issues that I have pointed out.
> > Besides that, I tested Marmotta and it's working.
> >
> > Cheers,
> > Gustavo.
> >
> > 2017-07-22 12:26 GMT-05:00 FRANCISCO XAVIER SUMBA TORAL <
> > xavier.sumba93@ucuenca.ec>:
> >
> >> Hi,
> >>
> >> In RDF4J, there’re two new tests [1] to evaluate large transactions of
> >> size 200000. Those tests are working but takes too long to execute. The
> >> only concern is compiling time. So, should we include those tests.
> >>
> >>
> >> > On Jul 19, 2017, at 12:56, Sergio Fernández <wi...@apache.org>
> wrote:
> >> >
> >> > Hi again,
> >> >
> >> > On Tue, Jul 18, 2017 at 9:25 PM, Gustavo Mora <gm...@gmail.com>
> >> wrote:
> >> >>
> >> >> Other thing about module/classes name conventions. Do we should
> >> maintain
> >> >> Sesame word in Marmotta modules or classes? or should we change
> module
> >> >> names like SesameValueBackend to RDF4JValueBackend or Apache
> Marmotta:
> >> >> Sesame HTTP Backend to Apache Marmotta: RDF4J HTTP Backend.
> >> >>
> >> >
> >> > My vote would go for s/Sesame/RDF4J. What do other think?
> >>
> >> I’d say to change everything for RDF4J since Marmotta is not longer
> using
> >> Sesame.
> >>
> >> >
> >> > Also, the dependency (TinkerPop 2.x) used in Titan and Accumulo
> backends
> >> >> are deprecated, and GraphSail haven't been migrated to TinkerPop 3.x
> >> [1].
> >> >> TinkerPop 2 uses Sesame 2.7.
> >> >>
> >> >
> >> > At some point we shuld get rid of those experimental backends that
> >> nobody
> >> > uses…
> >>
> >> So, should we ignore those modules?
> >>
> >> >
> >> > Besides those problems most of the tests are now working, and I'm
> >> trying to
> >> >> fix some tests incorporated in RDF4J.
> >> >>
> >> >
> >> > Great!!!
> >> >
> >> >
> >> >
> >> >
> >> > [1] https://groups.google.com/forum/#!topic/gremlin-users/CtSLC64gKZA
> >> >>
> >> >> 2017-07-16 13:38 GMT-05:00 Gustavo Mora <gm...@gmail.com>:
> >> >>
> >> >>> Hi,
> >> >>>
> >> >>> Marmotta is working with Sesame 4, I started migrating to RDF4J.
> >> Before
> >> >>> that I encountered two problems.
> >> >>>
> >> >>> 1. The test ComplexKiWiSparqlQueryTest -> ComplexSPARQLQueryTest#
> >> >> testSES1979MinMaxInf
> >> >>> is failing 'cause KiWi changes literal's datatype at time of
> >> insertion.
> >> >>>
> >> >>> For example, the following insertion is successful:
> >> >>>
> >> >>> PREFIX : <http://example.org/>
> >> >>> INSERT DATA
> >> >>> {
> >> >>> GRAPH <http://test/test> {
> >> >>> :a14 :float "-3.4428452385e-13"^^xsd:float .
> >> >>> :a100 :float "-INF"^^xsd:float .
> >> >>> :a101 :float "INF"^^xsd:float .
> >> >>> }
> >> >>> }
> >> >>>
> >> >>>
> >> >>> When querying KiWi store (SELECT * { GRAPH < http://test/test > {
> ?a
> >> ?b
> >> >>> ?c } }), we get the following results:
> >> >>>
> >> >>> *A*
> >> >>>
> >> >>> *B*
> >> >>>
> >> >>> *C*
> >> >>>
> >> >>> http://example.org/a14
> >> >>>
> >> >>> http://example.org/float
> >> >>>
> >> >>> “-3.4428452385E-13” ^^xsd:float
> >> >>>
> >> >>> http://example.org/a100
> >> >>>
> >> >>> http://example.org/float
> >> >>>
> >> >>> “-INF”^^xsd:string
> >> >>>
> >> >>> http://example.org/a101
> >> >>>
> >> >>> http://example.org/float
> >> >>>
> >> >>> “INF”^^xsd:string
> >> >>>
> >> >>> The problem is with special values of float (i.e. NaN, INF, -INF
> [1])
> >> >>> stored in Marmotta because the XSD float datatype definition is
> >> ignored
> >> >> and
> >> >>> replaced with the default (xsd:string). This problem is probably
> >> present
> >> >>> in both inserting and querying processes within KiWi.
> >> >>>
> >> >>>
> >> >>> 2. Some problems to execute LdpSuiteTest#testRunSuite because the
> >> suite
> >> >>> used (LdpTestSuite[2]) is deprecated; it uses Sesame 2. Major
> problems
> >> >>> between URI and IRI.
> >> >>>
> >> >>>
> >> >>> As a solution, I propose creating an issue for each problem. Or any
> >> other
> >> >>> ideas to approach this issues?
> >> >>>
> >> >>>
> >> >>> Cheers,
> >> >>>
> >> >>> Gustavo.
> >> >>>
> >> >>>
> >> >>> [1] https://www.w3.org/TR/xmlschema-2/#float
> >> >>>
> >> >>> [2] https://w3c.github.io/ldp-testsuite/
> >> >>>
> >> >>>
> >> >>>
> >> >>
> >>
> >>
> >> [1] https://github.com/eclipse/rdf4j/blob/48897c4205fb2ffb6eda7e
> >> 244b381a1f81fd9ffd/testsuites/store/src/main/java/org/
> >> eclipse/rdf4j/sail/SailConcurrencyTest.java#L159-L239 <
> >> https://github.com/eclipse/rdf4j/blob/48897c4205fb2ffb6eda7
> >> e244b381a1f81fd9ffd/testsuites/store/src/main/
> >> java/org/eclipse/rdf4j/sail/SailConcurrencyTest.java#L159-L239>
> >
> >
> >
>
Re: Road to RDF4J
Posted by Gustavo Mora <gm...@gmail.com>.
Hi,
I've generated some documentation in the wiki [1]. Also, Marmotta is
working with rdfj4. Xavier took a look to the documentation, but I also
would like to know what do you think Sergio about the documentation and
code changes. Is there something I should change? Also, I'd like to
continue working on it after GSOC.
Cheers,
Gustavo.
[1] https://wiki.apache.org/marmotta/GSoC/2017/MARMOTTA-659
2017-08-11 17:21 GMT-05:00 Gustavo Mora <gm...@gmail.com>:
> Hi,
>
> The final date is close. I'm working in the wiki, but I'd like to know
> which solution to take to some of the issues that I have pointed out.
> Besides that, I tested Marmotta and it's working.
>
> Cheers,
> Gustavo.
>
> 2017-07-22 12:26 GMT-05:00 FRANCISCO XAVIER SUMBA TORAL <
> xavier.sumba93@ucuenca.ec>:
>
>> Hi,
>>
>> In RDF4J, there’re two new tests [1] to evaluate large transactions of
>> size 200000. Those tests are working but takes too long to execute. The
>> only concern is compiling time. So, should we include those tests.
>>
>>
>> > On Jul 19, 2017, at 12:56, Sergio Fernández <wi...@apache.org> wrote:
>> >
>> > Hi again,
>> >
>> > On Tue, Jul 18, 2017 at 9:25 PM, Gustavo Mora <gm...@gmail.com>
>> wrote:
>> >>
>> >> Other thing about module/classes name conventions. Do we should
>> maintain
>> >> Sesame word in Marmotta modules or classes? or should we change module
>> >> names like SesameValueBackend to RDF4JValueBackend or Apache Marmotta:
>> >> Sesame HTTP Backend to Apache Marmotta: RDF4J HTTP Backend.
>> >>
>> >
>> > My vote would go for s/Sesame/RDF4J. What do other think?
>>
>> I’d say to change everything for RDF4J since Marmotta is not longer using
>> Sesame.
>>
>> >
>> > Also, the dependency (TinkerPop 2.x) used in Titan and Accumulo backends
>> >> are deprecated, and GraphSail haven't been migrated to TinkerPop 3.x
>> [1].
>> >> TinkerPop 2 uses Sesame 2.7.
>> >>
>> >
>> > At some point we shuld get rid of those experimental backends that
>> nobody
>> > uses…
>>
>> So, should we ignore those modules?
>>
>> >
>> > Besides those problems most of the tests are now working, and I'm
>> trying to
>> >> fix some tests incorporated in RDF4J.
>> >>
>> >
>> > Great!!!
>> >
>> >
>> >
>> >
>> > [1] https://groups.google.com/forum/#!topic/gremlin-users/CtSLC64gKZA
>> >>
>> >> 2017-07-16 13:38 GMT-05:00 Gustavo Mora <gm...@gmail.com>:
>> >>
>> >>> Hi,
>> >>>
>> >>> Marmotta is working with Sesame 4, I started migrating to RDF4J.
>> Before
>> >>> that I encountered two problems.
>> >>>
>> >>> 1. The test ComplexKiWiSparqlQueryTest -> ComplexSPARQLQueryTest#
>> >> testSES1979MinMaxInf
>> >>> is failing 'cause KiWi changes literal's datatype at time of
>> insertion.
>> >>>
>> >>> For example, the following insertion is successful:
>> >>>
>> >>> PREFIX : <http://example.org/>
>> >>> INSERT DATA
>> >>> {
>> >>> GRAPH <http://test/test> {
>> >>> :a14 :float "-3.4428452385e-13"^^xsd:float .
>> >>> :a100 :float "-INF"^^xsd:float .
>> >>> :a101 :float "INF"^^xsd:float .
>> >>> }
>> >>> }
>> >>>
>> >>>
>> >>> When querying KiWi store (SELECT * { GRAPH < http://test/test > { ?a
>> ?b
>> >>> ?c } }), we get the following results:
>> >>>
>> >>> *A*
>> >>>
>> >>> *B*
>> >>>
>> >>> *C*
>> >>>
>> >>> http://example.org/a14
>> >>>
>> >>> http://example.org/float
>> >>>
>> >>> “-3.4428452385E-13” ^^xsd:float
>> >>>
>> >>> http://example.org/a100
>> >>>
>> >>> http://example.org/float
>> >>>
>> >>> “-INF”^^xsd:string
>> >>>
>> >>> http://example.org/a101
>> >>>
>> >>> http://example.org/float
>> >>>
>> >>> “INF”^^xsd:string
>> >>>
>> >>> The problem is with special values of float (i.e. NaN, INF, -INF [1])
>> >>> stored in Marmotta because the XSD float datatype definition is
>> ignored
>> >> and
>> >>> replaced with the default (xsd:string). This problem is probably
>> present
>> >>> in both inserting and querying processes within KiWi.
>> >>>
>> >>>
>> >>> 2. Some problems to execute LdpSuiteTest#testRunSuite because the
>> suite
>> >>> used (LdpTestSuite[2]) is deprecated; it uses Sesame 2. Major problems
>> >>> between URI and IRI.
>> >>>
>> >>>
>> >>> As a solution, I propose creating an issue for each problem. Or any
>> other
>> >>> ideas to approach this issues?
>> >>>
>> >>>
>> >>> Cheers,
>> >>>
>> >>> Gustavo.
>> >>>
>> >>>
>> >>> [1] https://www.w3.org/TR/xmlschema-2/#float
>> >>>
>> >>> [2] https://w3c.github.io/ldp-testsuite/
>> >>>
>> >>>
>> >>>
>> >>
>>
>>
>> [1] https://github.com/eclipse/rdf4j/blob/48897c4205fb2ffb6eda7e
>> 244b381a1f81fd9ffd/testsuites/store/src/main/java/org/
>> eclipse/rdf4j/sail/SailConcurrencyTest.java#L159-L239 <
>> https://github.com/eclipse/rdf4j/blob/48897c4205fb2ffb6eda7
>> e244b381a1f81fd9ffd/testsuites/store/src/main/
>> java/org/eclipse/rdf4j/sail/SailConcurrencyTest.java#L159-L239>
>
>
>