SRB Semantic Search System

Author: Stephen Jeffrey
Email: sjj@maths.uq.edu.au

Version: 0.1 Alpha

The semantic augmentation consists of two components:
1. The semantically augmented MySRB web interface
2. The semantic search system

Ontology information (ie. the user-defined metadata that will
be used for semantic searching) is stored inside SRB's metadata 
system, without any modification to that system. The UDSMD type 
metadata fields store the ontology information in the following format:

Metadata name          value                units
-------------------------------------------------
OntologyClass          class_name           empty
OntologyProperty       property_name        value

These fields can be populated by the following:
   Smeta -u 0 UDSMD0="OntologyClass"    filename
   Smeta -u 0 UDSMD1="class_name"       filename
   Smeta -u 1 UDSMD0="OntologyProperty" filename
   Smeta -u 1 UDSMD1="property_name"    filename
   Smeta -u 1 UDSMD2="property_value"   filename
   ...

   Comments are stored in the objects "comment field", 
   which can be set by: 
       SmodD -c "OntologyComment:comment value" object_name
   and retrieved by:
       SgetD -C object_name
   The comment string must begin with "OntologyComment:"

To use the system:

1. Add metadata to the SRB objects
2. Build the instance data
3. Commence using the search system

Installation
============

Overview
--------
1.The modified MySRB interface must be compiled
  and the "mysrb331.cgi" executable put in the cgi-bin directory.

2.The semantic search system must be compiled
- The semantic search system is implemented using a
  client-server architecture.
- The server must be started. The user can optionally
  specify on the command line the name of an ontology
  to load. When the server has finished loading the
  ontology (if specified) it will output the port
  number that the server is running on.
- When the user interacts with the web interface,
  the search parameters are passed to the search client. 
  The client connects to the server,
  the server runs the required task and then returns
  the results to the client. The web interface is
  updated accordingly.

Steps
-----
1. Compile the modified MySRB source code
   See the source code in /usr/local/src/SRB3_3_1/mySRB/src
   NOTE: The Java classpath to the semantic search client is 
   hardwired in alpha1.c (line 20):
   INFERENCING_CLIENT = "java -classpath your_path OWL_inferencing_client";

2. Copy the CGI executable into the cgi-bin directory

4. Compile both the search-server and search-client
   applications. See the source code in OWL_Search/

5. Start the search server See OWL_Search/RUN_SERVER.
   NOTE: The Java classpath and location of the ontology is 
   hardwired in OWL_Search/RUN_SERVER

6. Start the MySRB interface and logon

7. Specify the machine name and port no. where the
   search server is running (See the left-hand side of
   the MySRB web interface), and then click "Display Ontologies"

Usage
=====

1. Add metadata to the objects in the SRB collection.
- only metadata that is semantically relevant is exported
  to the reasoning engine. The format is:

  OntologyClass    class_name
  OntologyProperty property_name property_value
  OntologyComment:comment
  OntologyAnnotation:annotation

  The MySRB web interface can be used to add metadata

2. Build the instance data
- all SRB objects having semantically relevant metadata
  are exported to the search system

3. If desired, save the instance data. It will be saved
   in XML format and could then be reloaded when the
   search system is used again

4. Commence using the system. The search interface is
   under the "Browse Query" icon.