Re: Introducing Modular Marmotta Backends

[Sergio Fernández <se...@salzburgresearch.at>]

Just to wrap-up all this new stuff, for those who are currently 
upgrading to 3.2.0-SNAPSHOT they would face with broken dependencies. 
You need to change your dependencies to something like the attached patch.

FMPOV the current bundling is a kind of mess, not only for some modules 
of the platform, but specially regarding the launchers. So we must find 
a better solution with Maven...


On 03/10/13 17:43, Sebastian Schaffert wrote:
> Dear all,
>
> now with the 3.1 release out I have finally spent some time on working on
> MARMOTTA-85, the long awaited exchangeable Sesame backends. I have just
> committed a first development prototype of this functionality to the
> develop branch. Please feel free to try it out. For now, I have implemented
> the following example backends:
> - marmotta-backend-kiwi - this is the old database based backend)
> - marmotta-backend-native - this backend is based on a Sesame NativeStore
> located in the MARMOTTA_HOME, 80% finished and untested
> - marmotta-backend-bigdata - this backend is based on the BigData triple
> store; since this triple store is licensed under GPL, these modules are not
> built by default, you need to run Maven with the profile "-Pnon-free"
>
> Exchanging backends is not completely trivial, because not all backends
> support all kinds of functionalities in the same way. For now, I have
> decided to use Maven to resolve the dependencies in a proper way. This
> means you cannot change backends at runtime. Instead, you need to create
> custom launchers with the configuration you want. You can find an example in
>
> launchers/marmotta-webapp-bigdata
>
> You can switch to this directory and simply run "mvn clean tomcat7:run
> -Pcleanall" to experiment with this bundle (after installing all of
> Marmotta with "mvn clean install -Pnon-free").
>
> A note on the BigData backend: this is a bit of a mess, because:
> - BigData only supports the Sesame 2.6 API, so I had to wrap it with the
> new Sesame 2.7 methods (HACK!)
> - BigData essentially ignores the Sail stack concept of Sesame, but we make
> heavy use of it in Marmotta, so I designed the wrapper classes in a way
> that the stack will work in most situations - the exception will probably
> be SPARQL where BigData uses its own direct access
>
> Feel free to try this out, especially if you have the requirement to work
> with bigger amounts of data. But no guarantees from me :-)
>
> Greetings,
>
> Sebastian
>

-- 
Sergio Fernández
Senior Researcher
Knowledge and Media Technologies
Salzburg Research Forschungsgesellschaft mbH
Jakob-Haringer-Straße 5/3 | 5020 Salzburg, Austria
T: +43 662 2288 318 | M: +43 660 2747 925
sergio.fernandez@salzburgresearch.at
http://www.salzburgresearch.at