Re: [Fwd: [greenstone-users] OAI records and documents problem]

From George Buchanan
DateTue, 21 Mar 2006 06:06:36 +0000
Subject Re: [Fwd: [greenstone-users] OAI records and documents problem]
In-Reply-To (441F56A7-1010009-cs-waikato-ac-nz)
> Dear colleagues:
>
> I'm working in an OAI collection, with OAI records from differents
> sources. I haven't any problem with the records, but, what about the
> documents? I used Eprints servers as sources, and can't get the
> documents the records identified... because there is a intermediate
> html page with the document data, and the identifier field in the
> records points to this page, so my collection identifies as document
> this page.
The OAI Protocol for metadata harvesting (OAI-PMH) which the OAI Plugin
uses does not define how a "document" should be obtained. All
Greenstone can do is to follow any URLs given by the OAI server - which
is the current behaviour you are seeing.

The current OAI Plugin is generic, and makes no assumptions about which
DL software is running on the OAI server (indeed, the OAI repository
may not be a DL at all!).

If you want to extend the Plugin to 'tweak' the behaviour for specific
types of OAI server (e.g., EPrints in this example), then you can
extend it using Perl. The OAI Plugin is inside your
gsdl/perllib/plugins folder.

I am afraid that there is no OAI protocol trick that you can play, and
we haven't extended the OAI plugin to try and detect which underlying
DL server (if any) is running, and thus tailor the document detection
to EPrints, DSpace or, indeed, Greenstone servers. If there is simply
a 'click through' on the intermediate page, then some simple Perl
parsing should give the result you want. If you wish to explore how to
do that, then the greenstone developer's list may be a good place to
post any subsequent queries. There is also online developer's
documentation on the www.greenstone.org web page under the
'documentation' tab.

Greenstone thrives on being open-source, so if you do achieve the
tailored behaviour you're interested in, then it would be great if you
could send your improved OAI Plugin to us at Waikato so we can
integrate it into future releases of Greenstone!

Sorry not to have an instant solution - good luck with improving your
collection!
>
Dr. George Buchanan
Future Interaction Technology Laboratory
University of Wales, Swansea