Hi Duane,


On Fri, May 30, 2014 at 4:18 PM, Duane Bayly <duane.ba...@gmail.com> wrote:
> Hi,
>
> We are evaluating Stanbol for a project at work which would need to apply a
> custom ontology to training materials.  The training materials can be
> accessed from Jackrabbit JCR.  Fully utilizing the ontology to search and
> categorize the data is our main objective.
>

Do you also use Apache Sling or just Jackrabbit? In case you are using
Sling you should know that Apache Stanbol does use the Sling launcher
[1] and the Sling Installer API [2] so you might want to consider to
use the JCR installer provider [3] to manage your Stanbol
configuration.

There is also an own section on the Stanbol Webpage that provides
information on how to customize your Stanbol launcher [4]

> We've been able to get all of the high level concepts from the
> documentation fairly easy.
> We are having a difficult time customizing the demo with enhancement
> engines/chains.

For the configuration of (Named) Entity Linking - detection of Entity
mentions in texts - I would suggest to read

* Custom Vocabulary Usage scenario [5]. Today I updated the
"Configuring the Stanbol Enhancer for your custom Vocabularies" so
that it also refers the possibility to use the FST linking engine.
* The Apache Con Europe 2012 presentation [6] about the same topic
* For detailed information about the configuration of the different
engines see their documentation. For that the List of all Enhancement
Engines is the best starting point [7]


For Topic Classification - content classification - there is not yet a
full documentation on the Stanbol Webpage, but this presentation by
ogrisel [8] provides detailed instructions how to train and use the
topic engine.

> Very little experience with OSGI.

For the configuration of Enhancement Engines you do not need to have
much knowledge of OSGI.

* For manually creating configurations you can use the configuration
tab of the Apache Felix Web Console
(http://localhost:8080/system/console/configMgr)
* Have a look at the log file (stanbol/log/error.log). If a provided
configuration is not valid you will find errors and/or warnings in the
log.
* You can use the components tab of the Felix Web Console
(http://localhost:8080/system/console/components) to
deactivate/activate unsatisfied components. This might be necessary
after fixing an invalid configuration.
* if you create/change configurations via the Apache Felix Web Console
you will also find the serialized files in the "stanbol/fileinstall"
folder. They will use "{component}.{component}-{uuid}.config" as file
name. You can take those files rename them to
"{component}-{name}.config". If can provide such files to a Sling OSGI
installer [2] [3] of an other Apache Stanbol instance. In production
you should work with such config files instead of using the Felix Web
Console.

hope this helps

best
Rupert

[1] http://sling.apache.org/documentation/configuration.html
[2] http://sling.apache.org/documentation/bundles/osgi-installer.html
[3] http://sling.apache.org/documentation/bundles/jcr-installer-provider.html
[4] http://stanbol.apache.org/docs/trunk/production-mode/
[5] http://stanbol.staging.apache.org/docs/trunk/customvocabulary
[6] 
http://stanbol.apache.org/presentations/stanbol_custom-vocabularies_ApacheCon_Europe-2012-11_rw.pdf
[7] http://stanbol.apache.org/docs/trunk/components/enhancer/engines/list.html
[8] http://stanbol.apache.org/presentations/Topic-Classification.pdf


>
> Any suggestions on blogs, papers with detailed technical documentation.
>
> Regards, Duane



-- 
| Rupert Westenthaler             rupert.westentha...@gmail.com
| Bodenlehenstraße 11                              ++43-699-11108907
| A-5500 Bischofshofen
| REDLINK.CO 
..........................................................................
| http://redlink.co/

Reply via email to