Pangu - Grammar driven data generation library

Introduction

This library is capable of generating various types of data from an XML Grammar such as DTD, XSD or RNG. Once you compile your XML grammar into a PanguGen you can then pass it the Decorator you'd like along with the GenInfo object and it will generate your output document. One thing to note about this library is that it will attempt to generate documents of the size you specify but it must firstly conform to the grammar used otherwise the document itself would not be valid. Through some testing we've been able to verify that the margin of error is not more than 10% unless you try to generate an object much smaller than your grammar would ever allow. For example, if your grammar had an attribute which had a restriction of a minimum of 100 bytes but you wanted a 32 byte document obviously there isn't any 32 byte document possible

Currently Implemented

We can currently generate documents from XSDs and we support only a subset of XSD types which are the following:

• string, integer, positiveInteger, decimal and boolean

We also only support length and pattern restrictions for the time being. I will put together a roadmap soon on adding the other features but for now just wanted to put the code up.

How to use

You need to first compile your grammar into a PanguGen object and this can be done for XSD objects like so:

PanguGen gen = XSDCompiler.compile(doc);

The doc is of course a Document object parsed from your XSD file. You can then use the gen object directly and pass it the OutputStream and GenInfo object that describe the type of data to generate and you will have a new document generated.

Here is a small example of generating an XML document:

ByteArrayOutputStream os = new ByteArrayOutputStream(); PanguDecorator formatter = new XMLDecorator(); gen.generate(os, formatter, "list", 1024);

Notice that to generate a JSON document you'd just need to change the formatter object used.

Building

You should be able to build pangu from the command line by just issuing:

ant build

You will then have your pangu.jar in the build directory ready to be included in one of your existing Java projects.

Why is it called Pangu?

Well since this library is responsible for "creating" other documents and being the geek I am I looked up a few of the names of gods of creation and found pangu in chinese mythology that pangu was the creator of the universe and that he balanced the ying & yang of the world by holding up the sky away from melding with the earth. Seemed like a good name for a library that was responsible for generating documents in a clean and balanced way.