The following instructions are dated Feb 4, 2009. Error handling in the OAIToolkit should have improved since then, however, fast-extract is probably still capable of producing bad MARCXML. This also doesn't tell you how to install the Toolkit, and assumes some specific file paths you'll want to change.

—

To extract records from EG:

PERL5LIB=/home/phasefx/svn/OpenSRF/STABLE_0_9/src/perlmods/:$PERL5LIB time ./fast-extract| \
./import/marc_add_ids -set_001_003 -f deleted -f id -f tcn_source -f tcn_value -f marc > ~/exported-records.marcxml

To start and stop Tomcat:

cd ~/apache-tomcat-6.0.18
bin/startup.sh
bin/shutdown.sh

To load or update records:

cd ~/OAIToolkit
cp some.marc.xml marcxml/
./load.sh

It seems that Tomcat needs to be restarted for an updated Lucene index to become visible.

If you delete ~/OAIToolkit/lucene_index/, it will be recreated on the next invocation of load.sh

The toolkit doesn't like records with bad leaders, or nonvalid subfield codes and indicators.

It can break completely with a "MarcException", and not report in its summary the records it does manage to create and/or update.

To view the web service:

http://hostname:8080/OAIToolkit