Redland adapter

[Kless]

I've seen in the ActiveRDF wiki [1] an adapter very interesting
recommended for large data sets and performance. It's Redland [2],
built in C++ and with bindings for Python and Ruby.


[1] http://wiki.activerdf.org/GettingStartedGuide
[2] http://librdf.org/

[Philip Cooper]

On May 25, 2008, at 5:54 AM, Kless wrote:

>
> I've seen in the ActiveRDF wiki [1] an adapter very interesting
> recommended for large data sets and performance. It's Redland [2],
> built in C++ and with bindings for Python and Ruby.

rdfAlchemy sits atop rdflib so it should be able to handle any of the
same backends. rdflib does provide (some) support for Redland.

At any point you can change the graph that rdfAchemy uses.

from rdflib import ConjunctiveGraph
db1 = ConjunctiveGraph('IOMemory')
db2 = ConjuncttiveGraph('Sleepycat')
db2.open('./db_myBDBdir')

rdfSubject.db=db1
# some stuff
rdfSubject.db = db2
# some more stuff

If you can open and use an rdflib Redland db you can use it in
rdfAlchemy.
... I think. I'll try do test this on a dev machine, I need
something better than Sleepycat for large datasets...

A quick look at the rdflib Redland code looks like it by default
opens a memory store. To open up a persistent model, you have to
pass it in already opened. I need some testing to be able to add
Redland to the create_engine and engine_from_config methods.

Anyone with more Redland knowledge can certainly correct me.

--
Phil