Overview

There are a number of different ways in which code or business logic can be "injected" into the generic Infinit.e framework (see architecture diagrams):

From the above diagram: 

• The are various ways regexes and javascript can be specified on sources to import. That type of customization is discussed here.
• Custom entity extractors can be added. This is described here.
  
  • (the support for this is not currently very good - the medium term aim is to replace the current custom infrastructure with a UIMA-compatible framework, which fits much better into our open philosophy)
• Hadoop-based JARs can be plugged-in to the framework and run either as a one-off or scheduled regularly. 
  
  • We will add more documentation support for this in the future, for now check out the API calls and this example (and command line equivalent).
  • There is also a full example of an Infinit.e-compatible Hadoop JAR hosted on github here.
  • Hadoop JARs can be created following this guide and can be uploaded via the file uploader, described here. 
• Follow-on processes can use the REST interface to collect data. Check out the API reference and more usefully the section with commonly needed examples.
  
  • (Obviously this isn't really injection of logic into Infinit.e, it's just integration of Infinit.e with other infrastructure).
• Visualization widgets can be developed and plugged into the GUI framework. This is described here.

Contributing code back to the Infinit.e community

There are two ways in which this can happen:

• Simply publish your plug-in on one of github, google code, sourceforge, etc and then let us know. In most cases, we will also link to such community projects from our web-site.
• Contribute back to the Infinit.e github trunk - see the "developing core infrastructure" section for details.