[jdom-interest] Compression and JDOM

[Laurent Bihanic]

David Morris wrote:

> Group,
> 
> We have run into a scalability issue while creating a XSL-FO document. 
> The volume of data required to describe a simple columnar report makes 
> it impossible for us to use transform documents like we want. For example, 
> when the JDOM tree for a 70 page report is serialized it takes about 250M. 
> I am looking for suggestions on how we can reduce the size of our output 
> passed to FOP, the tags, etc. are highly redundant and would seem to be 
> a good candidate for compression.
> 
Hi,


As you do not describe how you pass information to FOP, it's quite hard to 
give an answer. But...
What you can do to reduce the size of the XSL-FO document is to use the 
shortest possible prefix for XSL-FO either by making XSL-FO namespace the 
default or by using a one-character prefix.

Have you considered using a SAX pipeline so that the whole process happens in 
memory: JDOM -> XSL-FO -> FOP? The preferred XML input for FOP is SAX (except 
for FOP 0.20.x which is has a show-stopper bug when not parsing from an XML 
text stream).
Using the TrAX API to transform a JDOM document into an XSL-FO document and 
FOP 0.19 to render it in PDF, your code can be as simple as:

       // Allocate and configure a new driver instance.
       Driver driver = new Driver();

       driver.setRenderer(Driver.RENDER_PDF);
       driver.setOutputStream(new BufferedOutputStream(out));

       // Apply XSL Transformation redirecting the
       // result to the FOP rendering engine.
       getXSLTransformer(styleSheet).transform(
			new JDOMSource(doc),
			new SAXResult(driver.getContentHandler()));

       // Render PDF document
       driver.format();
       driver.render();


Hope this helps,


Laurent