CATS Harvester Web Service v1.0



A Harvester service accepts instructions describing a number of queries to be processed by a number of targets (i.e. nominated item banks). The aim is to harvest up to a maximum number of items - by executing a specific collation algorithm.

In the simplest case, a single query is run against a single target – the collation method being "stop when you have enough". All queries are expressed as CQL (Common Query Language) – a standard query language which allows for simple construction of queries but also allowing for more complex query definitions.

A specific Harvester service implementation is responsible for determining how it will process instructions. Following the CATS model, a Harvester service should pass requests to subordinate Connector services. These in turn process queries for specific Targets, and then collate the results from the Connector(s) using a Collation Algorithm identified in the original CATS profile.

A Harvester service returns a list of items found by the query processing, and a collection of files containing the data for these items. Note that the data from an individual item may require more than a single file.

The current CATS implementation of a Harvester service provides only a simple Cascading Collation Algorithm - where a number of Targets are processed (in the order supplied) until a sufficient number of items is found. (The implementation is further limited by the single implementation of the Connector interface.)

Download site: http://catsproject.cvs.sourceforge.net/catsproject/CATS/HarvesterService/

Related items
This piece of software part of the following systems:
 * part of::CATS System

This piece of software belongs to the following organisations:
 * belongs to::CATS

This piece of software uses the following standards:
 * uses::WSDL v1.1
 * uses::QTI v2.0

This piece of software provides the following service descriptions:
 * provides::Harvesting Items from Item Banks