HyKSS is set up to be run as an appserver. You can run HyKSS anywhere from the command line by using the “java -jar [jar file name]” command. Additionally, HyKSS allows for the following command line options:

• -p(ort) [port number] - the port the HyKSS will listen for requests on
• -t(hreads) [num threads] - the number of threads for HyKSS to use.
• -i(ndex) - tells HyKSS to index the document collection. The previous index will be deleted and the indexing will be done in batch.
• -d(ebug) - tells HyKSS to run in debug mode. This simply means that more information will be printed to the command line during query execution.

However, currently HyKSS expects the ontology library and document collection to be in a specific location relative to the jar file. It may be prudent to add command line options to specify the locations of these directories at a later time. HyKSS will look for folders named document_collection and ontology_library in the same directory it is placed. HyKSS must also have the ability to right as it will create two directories called keyword_index and semantic_index respectively.

There are currently two locations on dithers designed to run HyKSS. The first is /data/www/wok/hykss and the second is /data/www/wok/hykss/sandbox. The first is designed to be the main site for running HyKSS (the one we reference in papers) and the second is designed to be a sandbox for testing. Each of these locations includes and index.php file that runs the web UI for HyKSS to communicate with. The files differ only in the constants defined at the top of the file. The main HyKSS site uses the default port for communication with HyKSS and thus you should not have to specify any parameters to run HyKSS. The sandbox site will specify the port to use in the index.php file. Ensure that the index.php file and HyKSS are communicating on the same port or you will not get any functionality.

This section will be filled out after the thesis is completed.

• Ontologies in the ontology library must have unique names. This is more than just the file name. Each ontology has an attribute called called “ontologyName” on the OSM tag.
• Ontology names must not include the file extension (.xml).
• Object sets within a single ontology must all have unique names.