= Peregrine =
== Contact persons ==
 * H. Mei 
 * Dmitry Katsubo
 * Martijn Schuemie - the original developer of the project
 * Bharat Singh - integrates Peregrine into ConceptWiki [1]

== Code repository ==
 * gForce SVN:  https://gforge.nbic.nl/svn/data-mining/tags/0.3/
 * Publish Peregrine as a WebService,  look at "peregrine/ws" project: https://wiki.nbic.nl/index.php/Peregrine_Releases. The same for RMI.

== Information/Description ==
 * Peregrine is an indexing engine. The main goal of this project is to parse the input text and search for predefined terms. The term may refer one or several concepts and Peregrine tries to disambiguate them. It is based on the source code originally developed in EMC and extended/improved on NBIC initialtive. *
 * more on https://wiki.nbic.nl/index.php/Peregrine
 * Features per cycle and deliver packages in short runs  (agile-like approach).
   * http://projectmanagement.nbic.nl/Ticket/Display.html?id=114
   * http://nbic.cmbi.ru.nl/mediawiki/index.php/Cycle_3

(morris)

== T.M. System Objectives ==
 * A service to recognize terms in unstructured text (terms from pheno ontologies)
 * a visualization that shows the terms hightlighted in html (in MOLGENIS)
 * (future work: not only simple terms but also E/Q type triplets, or gene-pheno relations, etc).

== Use cases ==
 * A medical doctor pastes a patient description from his own notes and then chooses terms to systematically describe the phenotype
 * A researcher has a body of pubmed articles containing patient reports and wants to classify it to phenotype terms

== Question ==
 * Is Peregrine a good candidate for these use cases?
 * how can we best integrate to Peregrine (embeded via a Jar or via a web service or something)?
 * How can we configure Peregrine to use our own body of terms/ontologies (from OntoCAT).