Re: GPath, and Builder synergy

On Tue, Oct 05, 2004 at 10:56:56AM +0100, John Wilson wrote:
> I have been playing with new ways of using Builders and GPath in the 
> sandbox (XMLSlurper and StreamingMarkupBuilder). I recently enhanced 
> StreamingMarkupBuilder so that it would understanfdGPath expressions 
> dump subtrees defined by a GPath expression directly into the output of 
> the builder (This was covered by my XMLOpen presentation 
> http://groovy.codehaus.org/Articles).
> 
> It strikes me that if we make GPath understand Builders we could have a 
> very powerful way of specifying complex transforms on documents.
> 
> My idea is to allow closures to be assigned to nodes specified by a 
> GPath expression. The result of this operation is that the node and all 
> subnodes are replaced by the result of running the closure as a Builder 
> closure with the original subnode as its parameter. The builder would 
> create whatever underlying representation of the  document was being 
> used by the GPath implementation.
> 
> Perhaps an example would help:
> 
> given the document
> 
> <root>
>   <node>
>       <someInterestingData attr = '1'>my <b>data</b></someInterestingData>
>       <someBooringData>don't read this</someBooringData >
>   </node>
> </root>
> 
> 
> the GPath statement
> 
> node = {data(myAttr:it. someInterestingData['@attr']) {
>               mkp:yield it. someInterestingData.contents()
>               }
> }
> 
> would produce
> 
> <root>
>   <node>
>       <data myAttr = '1'>my <b>data</b></data>
>   </node>
> </root>
> 
> 
> I'm not quite sure about the notation. Perhaps node.transform{ ....} 
> would be clearer?
> 
> Anybody interested in using Groovy in this way?
> 
> John Wilson
> The Wilson Partnership
> http://www.wilson.co.uk

Cited the whole mail intentionally, as I was suprised that there was no
answer to it. I would be really happy to use this black magic. The syntax
is OK, I studied the GPath statement only for ten minutes, and then
understood what's going on.

incze