FAQ

• What is RAT?

  RAT is a framework for performing graph-based analysis.

• How do I obtain source code for RAT?

  Sourceforge has the most recent version of the system. The cvs repository is located at graph-rat.cvs.sf.net/projects/graph-rat

• Where do I start?

  There are two entry points.

  • The scheduler accepts an xml file as an argument, creates the specified graph using a graph factory, creates the specified scheduler using a scheduler factory, loads a Data Aquisition module using a data aquisition factory, then loads and executes algorithms against the graph, created through the algorithm factory.
  • The other method is to use graphs directly with graphs, actors, and links and then execute algorithms by hand
• What visualization capabilities does RAT have?

  Rat has a couple of visualization algorithms under algorithms/visual that use prefuse visualization tools

• What are graphs, actors, and links?
  
  
  • Graphs - are collections of actors and links with a globally unique id. Graphs can be nested in a tree structure where every child graph is a subset of the parent graph. Graphs can have properties.
  • Actors - are entities in a graph. Each actor has a type (referred to as a mode in Social Networking Analysis) that determines what kind of actor it is. Examples of actors include paper actors for papers published in a proceedings and author actors who are authors of papers. Each actor also has a id that is unique within a type. Actors can also have properties.
  • Links - are connections between actors. Every link has a type (referred to as a relation in Social Network Analysis) that determines what kind of link it is. Links are directional and have a unique source and target actor for every type. Links have a strength (which is a double value) and can also have properties.
  
  
  
• What are Data Aquisition modules?

  Data aquisition modules load data into graphs and are the first executed. This can be acessing a network of web pages, reading from a directory, parsing an xml file, or any other form of data entry. They routinely use crawlers to either read files or crawl the internet.