[greenstone-users] Greenstone image collection documentation

From UdcsWeb3
DateTue, 27 Sep 2005 12:53:51 +0200
Subject [greenstone-users] Greenstone image collection documentation
>We don't have a good system at the moment for user supplied
>documentation. The best thing to do is to post it to this mailing list,
>then it will be included (at some stage) in the mailing list archives
>collection. Also we can add it into our FAQ as well.
>We are considering setting up a wiki to allow users to add this kind of

Thanks for a great product once again

This is the documentation that I've set up during the build of a image collection and might be useful for someone else...

General Information about the Collection:

I've build a collection called ghandi2 on my pc locally using a combination of command line scripts (Greenstone scripts that came with the installation) as well as the Greenstone Librarian Interface (here after refered to as GLI), accessing any collection in the following manner: http://localhost/gsdl/cgi-bin/library
I used the perl scripts to make and build the collection and the GLI to assign metadata to the documents an to setup the search criteria. (Searches are based on the description of a specific image).

Although there is already a file with metadata for the 254 images (data describing the content of the images, the creator, dates and so forth - called: gdsl-data.txt), I was not able to incorporate this file into the GReenstone environment. I used the GLI (Greenstone Librarian Interface) to input data for a particular document. This being the reason that I've built the collection based on only 25 images.
(I haven't found a sollution yet to incorporate a tab delimitted text file into greenstone and to generate the metadata.xml file on the fly)
Greenstone generates a xml file for each and every image containing the metadata as well as a thumbnail (thumbnails in other words do not need to exist since they are generated when the collection is build).

The Greenstone Web Interface (look and feel) is very customizable but one has to have a general idea of how Greenstone works and how the directory structure is setup.
To access the demo ghandi collection please visit: http://greenstone.unisa.ac.za/cgi-bin/library.

One can search the collection by entering keywords in the search field. Searches are based on 'image description' metadata.

Procedures followed:

I installed Greenstone 2.56 on my pc locally. Installed in the /usr/local/gsdl folder which is the default for a greenstone installation. You can download Greenstone from www.greenstone.org as well as the installation documentation.

I created a collection called 'ghandi2' making use of the mkcol.pl script using the following command:
/usr/local/gsdl/bin/script/./mkcol.pl -creator youremailaddress ghandi2.
Before running the script make sure that the following 2 environment variables are set GSDLHOME and GSDLOS.
export GSDLHOME=/usr/local/gsdl
export GSDLOS=linux
After the collection was created I copied all the ghandi images to the import directory (/usr/local/gsdl/collect/ghandi2/import/.)
I've build the collection: /usr/local/gsdl/bin/script/./build
Apache was started up. /usr/local/apache/./apachect1 start (Refer to the greenstone developers guid to configure apache httpd.conf file)

The collection was now visible using Mozilla as my browser on my pc locally: http://localhost/gsdl/cgi-bin/library
When one views the collection you'll see thumbnail images, if you click on these images you'll be taken to the original document with its original size.
Please take note that the thumbnails gets auto generated by the greenstone engine
There does not exist any contextual metadata yet, so to create image specific metadata was the next step and I used GLI.

Creating Metadata
GLI is a Java GUI application and is run using the gli.sh shell script: cd /usr/local/gsdl/gli ./gli.sh (Take note you have to be in the gli directory to be able to issue the script)
Make sure that your greenstone GSDLHOME and GSDLOS varaibles are set before envoking the gli application. (These variables are supposed to be set up on installation but with me it was not the case)

Java 1.4.2 or higher should be installed as well.
Once the GLI has loaded you need to open the collection you want to work with.
Click <File> and then <Open>, choose the collection you would like to manipulate/change/enhance.
If a metadata set has not yet been specified, you can do so now...I chose Dublin Core for this basic image collection Click the <Enrich> tab.
Select an image on from the left pane. Select the element you want to capture the metadata for on the right pane Insert a value for the element and click <Append>. If you would like to replace a value rather, just click the <Replace> button instead.
After you've added the metadata for the files, save the file and then click the <Create> tab.

Build the collection.
Metadata does now exist which can be used to be displayed as well as searched.
The Collection Configuration file Each collection has a collection configuration file which can be found in /usr/local/gsdl/collect/nameofthecollection/etc/collect.cfg
This file can be manipulated thru the GLI or just using a text editor. In brief I've altered the collect.cfg file to include Description Creator Date Rights management
data which will be displayed next to each thumbnail
There will also be a description of the original image when the thumbnail is clicked.

Search Criteria
Searching the collection is based on the 'image description' metadata that was captured during the 'metadata process' described earlier
The search functionality was set up using the GLI
start gli
open the collection
select the <Design> tab
select <Browsing classifiers>
select the AZList classifier
select the <Configure classifier> button
type dc.Description in the metadata field. (Specifying that the search is based on the Description metadata)
click <OK>
build the collection
change the collect.cfg file if you want to display any metadata on your collection pages. Take note the the AZList classifier should be listed as the first classifier in the collect.cfg file, otherwise the data might not be displayed.

Local collection to the server
After completing the above steps, I decided to upload the collection on to the server. (greenstone.unisa.ac.za). I just had to copy the collection folder from my pc to the server (I used ssh to establish a connection to the server and then used scp)

Visit http://greenstone.unisa.ac.za/cgi-bin/library to see the 'demo' collection.

Further documentation
Excellent documentation is available at www.greenstone.org

Alwyn Nolte

This message (and attachments) is subject to restrictions and a disclaimer.
Please refer to http://www.unisa.ac.za/disclaimer for full details.
<<<< gwavasig >>>>