Re: Greenstone

From John R. McPherson
DateFri, 07 Feb 2003 10:19:37 +1300
Subject Re: Greenstone
In-Reply-To (Pine-GSO-4-44-0302041717510-15015-100000-qew-cs)
"Natalia N. Modjeska" wrote:
>
> John,
>
> My question is related to your reply to Donald D. Dollberg's
> question a few days ago about fielded searches.
>
> We would like to use Greenstone as a back-end search engine
> in a health care project which aims to bring evidence-based
> medical resources to doctors at point of care. We would like
> to be able to point the engine to specific parts of the
> documents (e.g., sections tagged <treatment> or <benefits>;
> the documents are in XML, but we know how to handle that).
> It seems that the mgpp backend you mentioned would suit our
> purpose. Would you have an idea about how easy/difficult it
> would be to extend it to a different domain? Also, where can
> we find mgpp? Thank you.
>

Hi,
mgpp is already part of greenstone (for version 2.38), but the
upcoming 2.39 has lots of improvements to the mgpp integration
with greenstone. See the documentation and/or the mailing list
archives (http://www.nzdl.org/cgi-bin/library?c=gsarch&p=about)
for hints about setting up the collect.cfg file to use mgpp.

If you have data in your own format, you might have to write
a greenstone import plugin. These aren't *too* bad if you already
know how to code in perl. The plugins basically say which bits
of text are metadata, and format the main text.

Hope this helps

John