UML, Profiles, AGX-Generators and more
UML (Unified Modelling Language) is a graphical language designed to describe software systems visually through diagrams. There are several different types of diagrams available, but at the moment AGX version 3 uses only class diagrams.
ArchGenXML (versions 1 and 2), the predecessor of AGX, was able to use both class diagrams and state machine diagrams. Its sole purpose was to generate Products (= plugins) for the ‘Plone Content Management System <http://plone.org>’_ , where class diagrams would be translated to content types while state machines were used to generate workflow states and transitions of content types for Plone. When it is finished, AGX v.3 will be a multi-purpose tool.
Class diagrams can be used to draw packages (python packages or modules), classes (python-classes) and as well its properties (attributes) and operations (methods) on these. In addition, associations in the diagram show how objects are aggregated within or referenced from one another.
The goal of model-driven development is to create the “blueprints” for your software in a well-defined, easily-communicated format: the UML model and diagram thereof. You can design your model using visual tools until you have a structure which adequately represents your needs. AGX will generate the necessary code.
You probably have to customize that code somewhat, filling in method bodies, creating additional web-templates for your framework of choice etc., but AGX takes care of all the boilerplate for you.
With tagged values and stereotypes you can customize the generated code with a surprising degree of flexibility and control. When you need to hand-code something, AGX keeps your code if you follow some simple rules, i.e. use predefined slots for your modifications.
This manual is not aimed at teaching you UML and object-oriented, model-driven software development. There are several other fine manuals about that on the web. A very good starting point is the OMG UML Resource Page including its web-links to tutorials. For a quick-start reading Practical UML helps.
AGX makes use of UML2 and its profiles.
A profile in the Unified Modeling Language (UML) provides a generic extension mechanism for customizing UML models for particular domains and platforms. Extension mechanisms allow refining standard semantics in strictly additive manner, so that they can’t contradict standard semantics.
Profiles are defined using stereotypes, tag definitions, and constraints that are applied to specific model elements, such as Classes, Attributes, Operations, and Activities. A Profile is a collection of such extensions that collectively customize UML for a particular domain.
from Wikipedia at 2010-07-08 Profile (UML)
AGX provides a specific profile for each generator package. The pyegg profile for example contains support for packages, modules, decorators etc. by providing stereotypes and tagged values.
AGX is extensible: Each generator addresses one specific domain and provides a profile for this domain. Generators and profiles may depend on a more basic domain, e.g. the ZCA-generator (Zope Component Architecture) extends the pyegg-generator.
While creating her UML model the developer applies domain specific features to the various items in the model. Through profiles the relevant stereotypes and tagged values for the domain are made available. After creating a diagram with domain specific information set on the elements AGX is invoked and code is generated.
While AGX is running, domain specific generators are working in sequence. Each generator reads the same source tree and writes to the same target tree, e.g. from a UML model to a tree of files. A domain specific generator can be tailored for any purpose.
AGX already supports some standard structures from the Python world, and has things set up for Python eggs, Zope Component Architecture with its ZCML configuration, GenericSetup profiles, buildout.
Each generator extends or modifies the target tree to fit it’s needs. During the whole transformation process the target tree is held in memory. After the generators have done their work, the target tree dumps the data to the file system. For more information about the generation process see section Under the hood.