Need help transforming some deep structures

Hi all!

I'm working on a new module RDF::RDFa::Template, which is an idea I have 
described here:
http://www.kjetil.kjernsmo.net/software/rat/
and discussed at some length in the RDF community. So, I figured, rather 
than talk about it, I write the code and discuss from there. After all, 
rough consensus and running code beats total consensus and no code any day. 
:-) 

I have most of the code for a first iteration of the idea now, but what I 
need is to transform this:
http://svn.kjernsmo.net/RDF-RDFa-Template/trunk/t/data/dbpedia-
comment.input.xhtml
into that:
http://svn.kjernsmo.net/RDF-RDFa-Template/trunk/t/data/dbpedia-
comment.expected.xhtml

Here, the rat:graph elements are removed, and rat:variable elements has its 
values substituted based on the results of a SPARQL Query, which is done 
elsewhere in the code (and now works). A doctype is also inserted and some 
namespace declarations are removed. 

The input document is in the form of an XML::LibXML::Document object when 
this process starts, so I would like to stick to stuff based on 
XML::LibXML.

My idea is write a SAX Filter, and I have started doing that, based on 
Kip's articles from way back, and I have understood that I should somehow 
accumulate events as I go, and finally add them in the final end_element, 
but since the structure is somewhat deep, I haven't understood the details.

I had a look at XML::Filter::DOMFilter::LibXML too, but I quite see if this 
would be the right thing to use; although I copy most of the children of 
rat:graph, which would be nice to use DOM, I also do the substitution in 
there.

Once I release this, I will solicit comments of the RDF(a) community and if 
people like the idea, to make this useful, it has to be several rat:graph 
elements in a document, and the filter should process all these elements. 
Also, importantly, if the SPARQL query returns multiple results, the child 
elements of rat:graph should be copied for each solution. But this is the 
next iteration, so it is not a priority now.

Can anyone please advice on how I should do this?

Cheers,

Kjetil
-- 
Kjetil Kjernsmo
[email protected]
http://www.kjetil.kjernsmo.net/
_______________________________________________
Perl-XML mailing list
[email protected]
To unsubscribe: http://listserv.ActiveState.com/mailman/mysubs