Re: ANN: DBpedia Version 2015-10 released

Hi Michael,

On Tue, Apr 5, 2016 at 3:08 PM, Michael Brunnbauer <brunni@netestate.de>
wrote:

>
> Hello Dimitris,
>
> I got DBpedia 2015-04 from
>
>  http://downloads.dbpedia.org/2015-04/core/
>
> It seems that
>
>  http://downloads.dbpedia.org/2015-10/core/
>
> also contains a reasonable and current subset of DBpedia. Is this correct?
>

Yes, there is always a big overlap with the previous releases.
The changes between the previous releases are the wikipedia changes we
capture and some post-processing / quality checks we introduce


> What is the difference to
>
>  http://downloads.dbpedia.org/2015-10/core-i18n/en/


core is the data we host in the public sparql / LDF endpoints .
we additionally include textual data (labels, comments) in other languages
and external links
we exclude some datasets from core-18n/en that (we think) are not useful to
all users (e.g. page-links)


> As a result of your switch to ttl, there is now a mixture of .nt and .ttl
> files in
>
>  http://downloads.dbpedia.org/2015-10/core/


you are right, the external links are still in nt format, we will make all
formats uniform in the next release


> but the .ttl files do not seem to use full Turtle syntax at first glance.
>

yes we do not use abbreviations and have one full triple per line (nt-like)


> Can they be parsed as N-Triple? I ask because I usually had to fix some
> syntax errors before Jena would parse your N-Triples files.
>


this version should have these syntax errors fixed, let us know if there is
still a problem.

Best,
Dimitris


-- 
Kontokostas Dimitris

Received on Tuesday, 5 April 2016 12:24:11 UTC