JDBM for gs3
|Reported by:||kjdon||Owned by:||kjdon|
Add in Steve's code for using JDBM instead of GDBM.
Files can be found at kanuka:/research/kjdon/home/steve-jdbm
Here are instructions given by Steve:
- jdbm-unspecified.jar needs to end up in web/WEB-INF/lib for runtime
- JDBMWrapper.java is the filled out version of the skeleton Kathy created in src/java/org/greenstone/gsdl3/util
txt2jdbm takes the path to a collection directory as an argument (absolute or relative) eg
It will look for index/text/db.txt under the given directory and produce index/text/<colname>.db (eg index/text/hcibib.db )
jdbm2txt takes a collection directory path as above
It will look for index/text/<colname>.db and produce index/text/db.txt
I have no elegant error trapping/recovery.
So the workflow for a Java only runtime for a collection would be
- set buildtype to Lucene in collect.cfg
- remove current index dir
- mv building to index
- add <databaseType name="jdbm"/> to every serviceRack element in the buildConfig.xml file for the collection
Add these instructions onto wiki once the code is added into Greenstone.