The MAGE Software Tool Kit MAGEstk

1
The MAGE Software Tool Kit(MAGEstk)
Angel Pizarro Center for Bioinformatics,
CBIL University of Pennsylvania
2
Overview

• MAGEstk APIs are directly produced from the
  MAGE-OM
• MAGEstk contains four platform dependent
  implementations
• MAGE-Java
• MAGE-Perl
• MAGE-Python
• MAGE-C
• You can (and should) use the MAGE-OM UML diagrams
  as a source of documentation

3
Packages

• A package in the model corresponds to a namespace
  in the API
• BioMaterial org.biomage.BioMaterial
• The base set of classes belong to the
  org.biomage.Common package
• In Perl these are in the root directory
• Also in Perl the namespaces are Bio.MAGE.
• A _package class to access the Identifiable
  classes within the package
• A set of inner classes hold references to the
  contained classes (java.util.Vector)

4
Classes

• Abstract classes in the model are abstract in
  MAGE-java
• All attributes and associations have get and
  set methods
• Associations with ..N cardinality have and
  addTo method.
• ..1 cardinality is an attribute of referenced
  class type
• ..N cardinality is an inner class _list.
• Implemented association constraints as
  interfaces.
• org.biomage.Interface
• All associations are ordered.

5
Description Package
MAGE-Java API
6
MAGE-ML

• Each class is responsible for input and output of
  MAGE-ML.
• Constructor takes as an argument a Xerces SAX
  Attributes node
• Each class has a method to write MAGE-ML
• writeMAGEML(java.io.Writer out).

7
MAGE-Java

• Michael Miller (Rosetta)
• Robert Hubley (ISB)
• Kjell Petersen (UIB)
• The rest of the MAGE-Java group