Hi

I'm working through what feels like a reasonably sensible Use Case, but
find the documentation hard to follow or missing for the things I need. I
can do the basics of getting data in an out in a experimental context, but
making a application-strength approach work raises a few questions:

The Use Case is I want to achieve :
1) generate some statements in another application (using Python) and push
these to a triplestore using a standard protocol (i.e. LDP) that would
allow that application to use any compliant server.
2) Generate some rules to create inferences, as per the kiwi-reasoner module
3) test the application and these rules with some sample data, and clear
the repository so i can re-run tests
4) bundle the rules with the application so on deployment this can be
configured.

I find the documentation rather Java client-centric rather than
Web-centric, so its hard to know how to recreate the promised functionality
in Python when I need to.

i have some specific questions - and if these are not well represented
somewhere in the documentation I'm happy to help add it once I have tested
the answers :-)

1) Firstly, I have found various different loading methods documented :
https://wiki.apache.org/marmotta/ImportData  does not reference the LDP
methods;
http://marmotta.apache.org/platform/ldp-module.html.  What is the
relationship - are they equivalent? Are there specific advantages the
non-standard approach retains over the custom one? Is there any guidance
document?

2) I was aiming at using  LDP, but I find the LDP documentation
inconsistent with the quickstart installation instructions, which naturally
puts the webapp into /marmotta (yes I can do Apache connectors to hide this
- but it would be nicer if the default outcome worked with the
documentation). is this a problem with the install docs or the LDP ones?

3) I cant get the LDP examples to work using :8080/marmotta or
:8080/marmotta/lpd as a root container - what should I be using?

4) Is there a way to flush the repository so i can iteratively test that
insertion and reasoning are behaving - or some other way to partition
things for unit testing? Currently the only way I can see to unit test is
to undeploy marmotta and redeploy, but even then I suspect its going to
require additional work to clear a postgres database.?

5) This link goes nowhere:
http://wiki.apache.org/marmotta/UsageResourceInteraction (in
http://wiki.apache.org/marmotta/DeveloperIntroduction) - is the source page
out of date - and is there an updated equivalent?

6) http://marmotta.apache.org/download.html gives no options or information
about included modules - specifically how does one get the kiwi reasoner?

7) How can you tell if the kiwi reasoner is installed?

8) It is possible to interact with the kiwi reasoner via web services? I
want  to load a set of rules during application deployment.

9) Can reasoning rules be applied to LDP containers - with the default
container having global rules? Is there any other way of limiting or
applying different rule sets via services?

10) Is there python binding for the reasoner - and a documentation place
where python utilities like https://rdflib.github.io/sparqlwrapper/ are
listed?

Cheers
Rob Atkinson

Reply via email to