Hi Habib,
What you are doing is ok but with some exceptions. First of all, if your
domain ontology is not something dynamic (you are not planning to change it
very often) you might prefer to use a Referenced Site instead of Managed
Site which is more suitable for dynamic uses cases. Basically, you would
need to use the EntityHub genericrdf indexer tool as explained here:
https://stanbol.apache.org/docs/trunk/customvocabulary.html (Using a
Referenced Site section).
By configuring the indexing part, the most important part is to properly
model the fields mappings, specially those fields that must be considered
as labels as you main use case is to Enhance content with that domain
ontology. Regarding this, I can't see any good candidate for labels in the
piece of ontology that you posted. Take into account that the Entity Linker
has been designed for linking mostly concepts names (labels) and not whole
sentences like "Digital Product Design and Factory". That doesn't sound as
a label for me, but you can try anyway.
Hope that helps,
Rafa
On Fri, Mar 17, 2017 at 12:23 PM Habibrahman A <habibrahma...@tcs.com>
wrote:
> Dear Team,
>
> We're using Stanbol for semantic search in one of our projects about
> Knowledge Management.
>
> We need to use the domain specific ontology provided to us.
>
> Below are the steps we did.
>
> 1. Created a new SolrYard
> 2. Create a new Site (Managed Site) and linked the yard
> 3. Using 'curl' command, loaded the ontology to the site
> curl -i -X POST -H "Content-Type:text/turtle" -T {ATRA.ttl} "
> http://localhost:8080/entityhub/site/customSite/entity";
> 4. Created a EntityHub Linking and referred the site created in step 2
> 5. Created a custom Enhancement Chain and added the EntityHub Linking
> created in step 4.
>
> When we run search (with sentences from the ontology), we couldn't see the
> entities defined being identified, but the results are different from when
> default chain is run. So we think something is happening, but not sure
> what and how.
>
> For eg.,
> The below is part of the ontology
> ##################################
> <http://www.sample-xo.com/KM/ATRb#1PDP>
> rdf:type ATRb:ATR ;
> ATRb:acronym "PDP" ;
> ATRb:comment "Digital Product Development Process and fully integrated
> Digitized Production and Supply Chain (> 2x Acceleration)" ;
> ATRb:level "1" ;
> ATRb:refinedBy <http://www.sample-xo.com/KM/ATRb#2FAL> ;
> ATRb:refinedBy <http://www.sample-xo.com/KM/ATRb#2LTA> ;
> ATRb:refinedBy <http://www.sample-xo.com/KM/ATRb#2STA> ;
> ATRb:title "Digital Product Design and Factory" ;
> #############################################
> when we search for 'Digital Product design and factory', we do not see
> anything in the list of entities identified. The enhancement chain just
> identifies the language.
>
> Our enhancemen chain contains the below engines
>
> tika;optional
> langdetect
> opennlp-sentence
> opennlp-token
> opennlp-pos
> opennlp-chunker
> sampleEntityLinking
> dbpedia-disamb-linking
> disambiguation-mlt
> dbpedia-dereference
>
>
> We would like to know, if what we are doing is correct ? Also, why we are
> not seeing the entities from the ontology that is loaded.
>
> Also, we were told that we need to create a pipeline to actually have the
> semantic search implemented.
>
> Pipeline steps being,
> 1. Load input documents to an external Solr
> 2. Configure the Managed site, enhancement engine and the chain
> 3. Hit Enhancement chain with the search text
> 4. Parse the output and extract the entities identified
> 5. Hit Solr, which has the input documents indexed, with the identified
> entities.
>
> Is this correct understanding.
>
> We're very new to Stanbol and search technologies as such.
>
> Your help is much appreciated.
>
> Thanks & Regards,
> Habib
> Thanks & Regards
> Habib Rahman
> Manufacturing TEG - Java CoE
> Tata Consultancy Services
> Ph:- +91446616 9247
> Cell:- 9094765645 <(909)%20476-5645>
> Mailto: habibrahma...@tcs.com
> Website: http://www.tcs.com
> ____________________________________________
> Experience certainty. IT Services
> Business Solutions
> Consulting
> ____________________________________________
> =====-----=====-----=====
> Notice: The information contained in this e-mail
> message and/or attachments to it may contain
> confidential or privileged information. If you are
> not the intended recipient, any dissemination, use,
> review, distribution, printing or copying of the
> information contained in this e-mail message
> and/or attachments to it are strictly prohibited. If
> you have received this communication in error,
> please notify us by reply e-mail or telephone and
> immediately and permanently delete the message
> and any attachments. Thank you
>
>
>
