Millards Corner » DSpace

Adventures in DSpace 1.6

millarj — Fri, 15 Jan 2010 21:56:32 +0000

I thought I’d use this space to document my experience installing and running DSpace 1.6 RC1. First off, we’ve been anxiously awaiting 1.6 for it seems like forever. We’ve been running DSpace 1.5.x in the Miami instance at the OhioLINK DRC but were really looking forward to some of the features promised in 1.6. The short list?

batch metadata editing
statistics integrated into xmlui
built in support for qualified dublin core in the OAI module (BONUS: ORE support!
browser-based import and export of collections

Building the SandBox

Since I wanted to run this in a sandbox, I decided to start with Virtualbox running on a Apple iMac. Into a new Virtual machine, I installed the latest Ubuntu, 9.10 or Karmic Koala. Since I wanted to be able to ssh into the box, and access my DSpace installation from the outside world, I also setup port forwarding in VirtualBox using instructions similar to these. For the impatient, to enable port forwarding of port 8080 on the host to port 8080 on the guest (called ubuntu), type these commands at the hosts command line:

VBoxManage setextradata ubuntu “VBoxInternal/Devices/pcnet/0/LUN#0/Config/http8080/HostPort” 8080
VBoxManage setextradata ubuntu “VBoxInternal/Devices/pcnet/0/LUN#0/Config/http8080/GuestPort” 8080
VBoxManage setextradata ubuntu “VBoxInternal/Devices/pcnet/0/LUN#0/Config/http8080/Protocol” TCP

Installation of Ubuntu on a Virtual machine is an exercise left to the reader.

Installing DSpace 1.6

The install documentation is pretty good actually. I was able to follow the DSpace 1.6 Documentation to get everything to install. Use apt-get to install any dependencies, like the JAVA JDK, maven, postgres, tomcat6, etc. the only glitch was getting Tomcat to start (ouch, pretty big glitch) It turns out that it’s not too hard to fix See This for DSpace 1.5, but the same principle applies.

Make sure Ubuntu is using the right Sun JDK, change the permissions to Tomcat’s files, and set the user that Tomcat runs as:

sudo update-alternatives –set java /usr/lib/jvm/java-6-sun/jre/bin/java

sudo chown -R dspace /var/cache/tomcat6
sudo chown -R dspace /var/lib/tomcat6
sudo chown -R dspace /var/log/tomcat6
sudo chown -R dspace /etc/tomcat6

Edit /etc/default/tomcat6 to set these parameters:
TOMCAT6_USER=dspace
TOMCAT6_SECURITY=no

Start tomcat6 and if you followed all the other installation steps correctly you should have a DSpace1.6RC1 instance running.

Next time: First impressions.

Harvesting CONTENTdm by way of OAI-PMH

millarj — Thu, 20 Aug 2009 13:43:44 +0000

I needed to find a quick way to harvest the dublin core metadata from a CONTENTdm collection, without being able to export it out of the admin interface. A little PHP, the CONTENTdm OAI-PMH interface, and some experimentation on the Open Archives Repository Explorer (http://re.cs.uct.ac.za/) and I was able to come up with this quick hack in php. Presented here in case it is ever useful to anyone besides myself.

http://staff.lib.muohio.edu/~millarj/cdmtab2dspace/oai-harvest.php

CONTENTdm to DSpace

millarj — Wed, 19 Aug 2009 16:17:51 +0000

There are probably lots of ways to do this, but this is mine. As part of our migration to the OhioLINK DRC I’ve created a quick php script to 1) take the default export format of CONTENTdm, 2) translate a mapping of fields to the appropriate dSpace ones, and 3) create the DSpace bulk submission package for it. You will need php5+ running on a unix-like system with allow_fopen set to true. It may work on Windows, but I haven’t tested it and I probably won’t. You could upgrade your Windows Installation and have better luck though.

You can find the scripts here. CAVEAT: These are ugly hacks at best http://staff.lib.muohio.edu/~millarj/cdmtab2dspace/

Instructions:

Export your data from CONTENTdm in the control panel. Select a Tab Delimited export and make sure to check “Return field names in first record”
Next, open the resulting text file in a spreadsheet application of your choice. If you don’t want to munge dates and other non-text data, make sure each column is imported as the “Text” type. Excel for example, has a hard time with ISO dates.
Add a row to the top of the spreadsheet and place the appropriate DSpace fieladn mes (minus the dc. prefix) above the CONTENTdm field to be mapped. If you want to skip a field, just use “skip” as the DSpace field label.
Save it back out as a tab-delimited file.
Download the script and edit it to customize to your hearts content.
Run the script from a unix/linux/macosx commandline “php cdmtab2dspace.php”

At the top of the script are various customizations/configurations. To see how things will look, set $test_mode = “on” When ready to run it live, change that bit to “off”

In particular, customize/rewrite the “move_bitstreams” function to, you guessed it, move the bitstreams (digital objects) from wherever you have them into the submission package.