Hello Vincent,

We identified the problem and added back a lot of missing coordinates. The
problem will be 100% fixed in the next release which is due soon.

We apologize for any inconvenience.

Cheers,
Dimitris

Sent from my mobile, please excuse my brevity.
Hello Vincent,

I ran the GeoExtractor for the English dump files again, as a first test.
This resulted in the same geo-coordinates files as produced by the main
extraction.
The problem seems not to be near the surface.

We will have a closer look at this in the coming weeks.

Regards,

Markus

2015-08-05 16:47 GMT+02:00 Vincent Malic <vma...@indiana.edu>:

> Dear Dimitris,
>
> Thanks for the info, I really appreciate you looking into this. I've been
> working with the MilitaryConflict data so that's the window through which I
> noticed the missing triples, but since you mentioned there weren't any
> changes to the conflict mapping I poked around a little more and I'm lead
> to believe this isn't an issue with dbpedia extracting conflict data and
> may have to do with geographic coordinates in general. As way of
> illustration I took a look at the geographic coordinates dataset from the
> 3.9 downloads, which has 1,987,961 triples, while the corresponding
> geographic dataset on the 4.0 downloads page has only 2284 triples, which
> seems like a significant drop for geographic data across the whole dbpedia
> dataset. Perhaps there were some changes made to the extraction framework
> for the general extraction of the Wikipedia "coord" tag? Hope this
> information helps.
>
> Once again, thanks for looking into this issue.
>
> Best,
> Vincent
>
> On Wed, Aug 5, 2015 at 3:25 AM Dimitris Kontokostas <jimk...@gmail.com>
> wrote:
>
>> Thank you for the report Vincent,
>>
>> The online data are based on the new to-be-announced release 2015-04
>>
>> What you found is indeed quite strange as there were no changes in
>> -
>> http://mappings.dbpedia.org/index.php/Mapping_en:Infobox_military_conflict
>>
>> - https://en.wikipedia.org/wiki/Template:Infobox_military_conflict
>> - and from a sample of pages with missing coords I checked there were no
>> substantial changes
>>
>> On the other hand, running a sample extraction of a page returns correct
>> coordinates which complies with the DBpedia Live data
>>
>> http://mappings.dbpedia.org/server/extraction/en/extract?title=Battle_of_Augusta&revid=&format=trix&extractors=custom
>>
>> We will look further into this
>>
>> Best,
>> Dimitris
>>
>>
>> On Mon, Aug 3, 2015 at 10:08 PM, Vincent Malic <vma...@indiana.edu>
>> wrote:
>>
>>> Dear all,
>>>
>>> I've only recently started working with dbpedia as a research resource,
>>> so I apologize in advance if I'm asking a silly question with an obvious
>>> answer. Dbpedia is a wonderful resource, a huge thanks to all of you on the
>>> list who are putting hard work into it.
>>>
>>> It seems to me that there was a recent change to dbpedia that caused a
>>> good chunk of information I've been working with to go missing. I've been
>>> working with the geographical locations of military conflicts over the past
>>> few months, stitching together a dataset from dbpedia triples that linked
>>> several thousand battles to their locations via the georss:point predicate.
>>>
>>> I started rebuilding this dataset this week and it seems that much of
>>> that information I saw before is missing. To illustrate, the following
>>> query to the SPARQL endpoint at http://dbpedia.org/sparql produces only
>>> 11 hits:
>>>
>>> select COUNT(?battle) as ?count
>>> where {
>>> ?battle rdf:type dbo:MilitaryConflict .
>>> ?battle georss:point ?location .
>>> }
>>>
>>> Just to make sure I wasn't going crazy I decided to do essentially the
>>> same query to the dbpedia live SPARQL endpoint (
>>> http://live.dbpedia.org/sparql):
>>>
>>> select COUNT(?battle) as ?count
>>> where {
>>> ?battle rdf:type dbpedia-owl:MilitaryConflict .
>>> ?battle georss:point ?location .
>>> }
>>>
>>> This returns 4225 hits, which is more reminiscent of the amount of
>>> results I got working with non-live dbpedia a few months ago. So, a few
>>> months ago, vanilla dbpedia gave me several thousand results, and dbpedia
>>> live right now does the same, but as of recently dbpedia vanilla only
>>> returns 11 results. Any reason that this may be the case? Perhaps I'm
>>> missing something quite simple? I appreciate any feed back.
>>>
>>> Thank you and best regards,
>>> Vincent Malic
>>> Ph.D Student in Information Science
>>> School of Informatics and Computing
>>> Indiana University, Bloomington
>>>
>>>
>>> ------------------------------------------------------------------------------
>>>
>>> _______________________________________________
>>> Dbpedia-discussion mailing list
>>> Dbpedia-discussion@lists.sourceforge.net
>>> https://lists.sourceforge.net/lists/listinfo/dbpedia-discussion
>>>
>>>
>>
>>
>> --
>> Kontokostas Dimitris
>>
>
>
> ------------------------------------------------------------------------------
>
> _______________________________________________
> Dbpedia-discussion mailing list
> Dbpedia-discussion@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/dbpedia-discussion
>
>
------------------------------------------------------------------------------
_______________________________________________
Dbpedia-discussion mailing list
Dbpedia-discussion@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/dbpedia-discussion

Reply via email to