AW: [smila-dev] SSS & Persistence Questions

[<Daniel.Stucky@xxxxxxxxxxx>]

Hi Markus,

concerning your 2nd question I would like to make some comments:

> - What concreteley would I have to do to change the "5 minutes to
> success" example within the wiki in order to persist the content of
> crawled files (either to the filesystem or to a database) instead of
> just indexing them?

I'm not exactly sure what your use case is about. SMILA records
(attributes and attachments) are automatically persisted while being
processed. Attributes are stored in the XMLStorage, Attachments in the
BinaryStorage. So you don't have to take care about persisting the data
yourself.

However, if you would like to persist the data somewhere else (like it
is done with the Lucene index), you are free to do so. You have to
implement your own Pipelet (or ProcessingService) to achieve this. In
your implementation you are free to store the data in the filesystem, a
JDBC database or anything else you can do with Java.
Please take a look at
http://wiki.eclipse.org/SMILA/Howto_integrate_a_component_in_SMILA,
especially chapter 2.2 "Default: Integrating local SMILA pipelets or
processing services". This and the linked pages provide information on
how to write your own Pipelet. I think you should do the tutorial
http://wiki.eclipse.org/SMILA/Development_Guidelines/How_to_integrate_th
e_HelloWorld_webservice_as_a_Pipelet. In the part "Process input data"
would provide you code for persisting the data. Not every step will make
sense for your Pipelet, but you should get a general idea how SMILA
works.


If you have any further questions don't hesitate to ask!

Bye,
Daniel