Topics in a topic map can be about a wide variety of topic types, for instance: people, locations, events, and more. Each of those topic types would likely be represented by a different UIMA
annotator. Thus, in order to read unstructured text and create and maintain topics in a topic map, this proposed topic map harvesting agent is liable to be a
society of agents, each offering resources to the growing topic map.
The
SolrAgentFramework is designed to host a wide variety of such agents in different topologies, such as:
- All agents running under one SolrAgentFramework situated on a single computer
- Agents distributed in many computers, contained by a SolrAgentFramework on each computer, framework of which has a TCP connection over a local area network (eventually, the web) to the SolrAgentCoordinator.