XOM API for Processing XML with Java

[Wolfgang Hoschek <whoschek AT lbl.gov>]

You can do such kind of things with XQuery. Since XQuery is a superset 
of XPath 2.0, it can also be used with plain XPath expressions as 
queries. Also note that XPath 2.0 supports regular expressions via the 
matches, replace, and tokenize functions.

For example:
// find the links of all images in a XHTML-like document
XQuery xquery = new XQuery("//img/@src", null);

// find the links of all JPG images in a XHTML-like document
// XQuery xquery = new XQuery("//img/@src[fn:matches(string(.), 
'.jpg')]", null);

Nodes nodes = xquery.evaluate(doc);
for (int i=0; i < nodes.size(); i++) {
   System.out.println("node "+i+": "+nodes.get(i).toXML());
}

I'm working on XOM extensions for that. A working prototype is at

http://dsd.lbl.gov/~hoschek/xom-contrib-download

and javadoc is here:

http://dsd.lbl.gov/~hoschek/xom-contrib/

One can query some hundreds of 10K documents per second with that 
approach,
but in terms of performance it certainly cannot compete with a 
NodeFactory filter.
Rather, the primary motivation here is power with simplicity.

Wolfgang.

> Elliotte, be brave and put some ideas down for us to see.
> With your understanding of XML *and* developers, you
> are one of the best positioned to do this.
> It might be what makes XOM rock even more one day.
> Add regexps to XPATH? or callbacks?
> a new way to analyse XML which is more powerful
> than regexps alone, but which doesn't turn entities
> into landmines?  be creative! go nuts!
>
> Luca
>
>
> > What Bray is suggesting here is a sort of XML-aware regular 
> expression
> > language. He very well understands that current regular-expression
> > languages are inadequate for handling XML.
> >
> > One thing I have seen some people do is define a new, very inflexible
> > syntax for XML infosets that resolves things like character entity
> > references, insignificant white space, quotes around attribute 
> values,
> > and so forth; then write a program to convert XML documents into this
> > new syntax; and then query the new syntax with regular expressions. 
> It's
> > a hack, but it works.
> >
> > In essence Bray is here proposing to build the alternate, 
> less-flexible
> > syntax into the regular expression engine. Perhaps a good idea, but 
> it
> > does need a new engine/language. The existing ones can't do this.
>
>
> http://www.tbray.org/ongoing/When/200x/2003/03/16/XML-Prog
>
> in particular notice the place where the guy does
> (which is very close to what I was trying to do with
> XOM):
>
> while (<STDIN>) {
>   next if (X<meta>X);
>   if    (X<h1>|<h2>|<h3>|<h4>X)
>   { $divert = 'head'; }
>   elsif (X<img src="/^(.*\.jpg)$/i>X)
>   { &proc_jpeg($1); }
>   # and so on...
> }
>
> Luca