Hi Tarandeep,
Thanks, it's quite more clear now :-). Have you check if the information
you need (for example dbp-ont:capital) is actually in the index?. You
can check it for example looking for "India" entity directly in your
custom DBpedia site in the EntityHub.
Regards
El 18/07/13 12:30, Sawhney, Tarandeep Singh escribió:
No problems Rafa, may be i didnt explain with details/clarity.
We are using custom ontology to extract custom entities from text and then
we want to them to link with DBpedia entities (in local dbpedia reference
site).
We found dbpedia reference doesnt have enough data that we need, so decided
to download additional data for selected entities (related to fashion
brands, fashion designers, company names) directly from dbpedia.org
We then indexed these individual RDF files and created indexes with new
reference site
We then did not use DBpedia reference site, instead used our new reference
site which has dbpedia data that we need with our new Entityhub linking
engine
But after we followed steps i mentioned in my earlier email, during
enhancement, custom entities are getting de-referenced from my new
reference site but i dont see additional data that i needed which exists in
local cache.
Hope this explains what we are trying to do, please let me know if some
more information is required.
Best regards
tarandeep
On Thu, Jul 18, 2013 at 3:21 PM, Rafa Haro <rh...@zaizi.com> wrote:
Hi Tarandeep,
El 18/07/13 11:18, Sawhney, Tarandeep Singh escribió:
Hi Rafa
Thanks for your response
Yes, we have tried the whole URI of the property (
http://dbpedia.org/ontology/****capital<http://dbpedia.org/ontology/**capital>
)<http://dbpedia.org/**ontology/capital<http://dbpedia.org/ontology/capital>
)>
also
but it didn't help
Yes we are using EntityHub cache to locally store with all the additional
information we pulled from Dbpedia.org
In the documentation provided at
http://stanbol.apache.org/**docs/trunk/customvocabulary.**html<http://stanbol.apache.org/docs/trunk/customvocabulary.html>
it is mentioned --->
*Optionally, if your data do use namespaces that are not present in
prefix.cc (or the server used for indexing does not have internet
connectivity) you can manually define required prefixes by creating/using
the a indexing/config/**namespaceprefix.mappings file
*
*
*
Can we get some inputs on if some changes to this file are required while
using DBpedia data
This file can be used at 'indexing time' when you use the indexing tool
for creating the index for the DBpedia site. I have just seen that dbp-ont
is already included as prefix. What I don't have clear right now is if you
are generating your own dbpedia index including all the dbpedia ontology
properties (that should be a enormous index) or if you are generating an
index each time you need a new entity or even you are trying to retrieve
the entities from dbpedia in a 'live' way :-). Sorry I'm confused about
your workflow.
Also, looks like we are missing on some configurations in the overall
process, so if dev community can please provide help, it will be much
appreciated
best regards
tarandeep
On Thu, Jul 18, 2013 at 1:38 PM, Rafa Haro <rh...@zaizi.com> wrote:
Hi Tarandeep,
Have you tried using the whole URI of the property (
http://dbpedia.org/ontology/****capital<http://dbpedia.org/ontology/**capital>
)<http://dbpedia.org/**ontology/capital<http://dbpedia.org/ontology/capital>
)>
??
Anyway, maybe it is a better idea to change your workflow, because I
suppose that your example about "India" entity is something that could
happen to you with more entities because the default DBpedia site in
Stanbol doesn't contain information about dbp-ont properties. I would
suggest to use EntityHub cache to locally store entities with all the
information you need directly from DBpedia. So, maybe you can try to
directly retrieve the entities from any DBpedia endpoint, store them in
the
EntityHub cache to ensure that you can use it later as your convenience.
Maybe the workflow could be the following:
1. Enhance a document using Stanbol DBpedia site for linking.
2. For each extracted entity:
2.1. If the entity is already store in the EntityHub, get it
using
LDPath for dereferencing.
2.2. If not, retrieve the entity from DBpedia endpoint as RDF
data
and store it in the EntityHub. Then retrieve it
I would day that this is currently possible in Stanbol, but maybe someone
else in the list can give you more light with the issue.
Regards
El 18/07/13 09:48, Sawhney, Tarandeep Singh escribió:
Hi All,
In the stanbol local cache we have limited triples in dbpedia reference
site.
We have a need to get more triples for entities which are present in
dbpedia
reference site. For example entity "India" has limited triples, so when
we
enhance text which has india, it gets us only information which is there
in
dbpedia reference site.
We have followed below mentioned steps to add more RDF data for entity
"India" by creating our own reference site.
1 - Downloaded rdf-data for 'India' from [1].
2 - Generated indexes for this rdf-data as suggested in article [2] with
*Demo
*as a reference site name.
3- Initialized indexes within stanbol instance as per [2].
4- Configured new EntityLinking engine, '*demoLinkingEngine*' with
*Demo
*as
referenced site as per [3].
I have added *dbp-ont:capital *in *'"Fields used for
derefrencing*
"option.
5- Configured new weighted chain (*demoChain*).
6 - Now i am trying to enhance *"India is a country."* I am getting
India
as de-reference entity but unable to get any new information related
to *dbp-ont:capital
*which exists in my new reference site *Demo, *which in this case should
give us URI value of "New Delhi"
[1] http://dbpedia.org/page/India
[2]
http://stanbol.apache.org/****docs/trunk/customvocabulary.****html<http://stanbol.apache.org/**docs/trunk/customvocabulary.**html>
<http://stanbol.apache.**org/docs/trunk/**customvocabulary.html<http://stanbol.apache.org/docs/trunk/customvocabulary.html>
[3]
http://stanbol.apache.org/****docs/trunk/components/****
enhancer/engines/**<http://stanbol.apache.org/**docs/trunk/components/**enhancer/engines/**>
entityhublinking<http://**stanbol.apache.org/docs/trunk/**
components/enhancer/engines/**entityhublinking<http://stanbol.apache.org/docs/trunk/components/enhancer/engines/entityhublinking>
Can you please let me know if i am doing something wrong here or missing
some configurations.
Please let me know in case you need some more information on how we are
trying to do it
best regards
tarandeep
--
------------------------------
This message should be regarded as confidential. If you have received
this
email in error please notify the sender and destroy it immediately.
Statements of intent shall only become binding when confirmed in hard
copy
by an authorised signatory.
Zaizi Ltd is registered in England and Wales with the registration number
6440931. The Registered Office is Brook House, 229 Shepherds Bush Road,
London W6 7AN.
--
------------------------------
This message should be regarded as confidential. If you have received this
email in error please notify the sender and destroy it immediately.
Statements of intent shall only become binding when confirmed in hard copy
by an authorised signatory.
Zaizi Ltd is registered in England and Wales with the registration number
6440931. The Registered Office is Brook House, 229 Shepherds Bush Road,
London W6 7AN.
--
------------------------------
This message should be regarded as confidential. If you have received this
email in error please notify the sender and destroy it immediately.
Statements of intent shall only become binding when confirmed in hard copy
by an authorised signatory.
Zaizi Ltd is registered in England and Wales with the registration number
6440931. The Registered Office is Brook House, 229 Shepherds Bush Road,
London W6 7AN.
