Seeking peer review of new fast high level XML pull parser

Hi XML folks,

I'm putting the finishing touches on XML::TreePuller which is a high
level very fast XML pull processor but I'd like to solicit the
feedback of the public before I publish the first version onto CPAN.
It is functional and almost ready for a release so feel free to test
it if you would like. I would appreciate any feedback people have
especially on the
documentation and interface - I'm aiming to be understandable at a
glance and easy to use.

You can see the code for the module at
https://triddle.projecthut.com/svn/triddle/XML-TreePuller/lib/XML/TreePuller.pm
and there is a real-world example of the same interface (with an
insignificant change) in use in MediaWiki::DumpFile::Pages at
http://cpansearch.perl.org/src/TRIDDLE/MediaWiki-DumpFile-0.0.14/lib/MediaWiki/DumpFile/Pages.pm
(XML::TreePuller and MediaWiki::DumpFile::XML have a nearly identical
interface).

How fast is it? In my benchmarks
(https://triddle.projecthut.com/svn/triddle/XML_Speed_Test/test_cases/)
it performs as follows:

---
- filename: datastore/20-simplewiki-20091021-pages-articles.xml
  markup_density: 0.202659609191331
  md5sum: 8fa1e9de18b8da7523ebfe2dac53482a
  size: 227681797
  tests:
    - MiB/sec: 25.6054577962408
      md5sum: 8fa1e9de18b8da7523ebfe2dac53482a
      name: test/XML-LibXML-Reader.t
      percentage: 100
      runtimes:
        system: 0.74
        total: 8.48
        user: 7.74
    - MiB/sec: 12.450360212851
      md5sum: 8fa1e9de18b8da7523ebfe2dac53482a
      name: test/XML-TreePuller_element.t
      percentage: 205
      runtimes:
        system: 0.51
        total: 17.44
        user: 16.93
    - MiB/sec: 9.98318538446536
      md5sum: 8fa1e9de18b8da7523ebfe2dac53482a
      name: test/XML-TreePuller_config.t
      percentage: 256
      runtimes:
        system: 0.54
        total: 21.75
        user: 21.21
    - MiB/sec: 6.67489339416298
      md5sum: 8fa1e9de18b8da7523ebfe2dac53482a
      name: test/XML-SAX-ExpatXS.t
      percentage: 383
      runtimes:
        system: 0.41
        total: 32.53
        user: 32.12
    - MiB/sec: 5.03441414588735
      md5sum: 8fa1e9de18b8da7523ebfe2dac53482a
      name: test/XML-Parser.t
      percentage: 508
      runtimes:
        system: 0.71
        total: 43.13
        user: 42.42
    - MiB/sec: 3.21537512382825
      md5sum: 8fa1e9de18b8da7523ebfe2dac53482a
      name: test/XML-LibXML-SAX.t
      percentage: 796
      runtimes:
        system: 0.82
        total: 67.53
        user: 66.71
    - MiB/sec: 1.06375799584618
      md5sum: 8fa1e9de18b8da7523ebfe2dac53482a
      name: test/XML-Twig.t
      percentage: 2407
      runtimes:
        system: 1.73
        total: 204.12
        user: 202.39

Also here is an example just for good measure:

#!/usr/bin/env perl

use strict;
use warnings;

use XML::TreePuller;

sub gen_xml {
  	return <<EOF
  	
<wiki version="0.3">

<!-- schema says that there is always 1 siteinfo and zero or more page
  elements follow -->
<siteinfo>
  <sitename>ExamplePedia</sitename>
  <url>http://example.pedia/</url>
  <namespaces>
    <namespace key="-1">Special</namespace>
    <namespace key="0" />
    <namespace key="1">Talk</namespace>
  </namespaces>
</siteinfo>

<page>
  <title>A good article</title>
  <text>Some good content</text>
</page>

<page>
  <title>A bad article</title>
  <text>Some bad content</text>
</page>

</wiki>
  	  	
EOF
}

sub element_example {
	my $xml = XML::TreePuller->new(string => gen_xml());
	
	print "Printing namespace names using configuration style:\n";
	
	$xml->config('/wiki/siteinfo/namespaces/namespace' => 'short');
	
	while(defined(my $element = $xml->next)) {
		print $element->attribute('key'), ": ", $element->text, "\n";
	}
	
	print "End of namespace names\n";
}

sub subtree_example {
	my $xml = XML::TreePuller->new(string => gen_xml());
	
	print "Printing titles using a subtree:\n";
	
	$xml->config('/wiki/page' => 'subtree');

	while(defined(my $element = $xml->next)) {
		print "Title: ", $element->get_elements('title')->text, "\n";
	}	
	
	print "End of titles\n";
}

sub path_example {
	my $xml = XML::TreePuller->new(string => gen_xml());
	
	print "Printing path example:\n";
	
	$xml->config('/wiki/siteinfo', 'subtree');
	$xml->config('/wiki/page/title', 'short');
	
	while(my ($matched_path, $element) = $xml->next) {
		print "Path: $matched_path\n";
	}
	
	print "End path example\n";
}


element_example(); print "\n";
subtree_example(); print "\n";
path_example(); print "\n";

__END__

Output:

Printing namespace names using configuration style:
-1: Special
0:
1: Talk
End of namespace names

Printing titles using a subtree:
Title: A good article
Title: A bad article
End of titles

Printing path example:
Path: /wiki/siteinfo
Path: /wiki/page/title
Path: /wiki/page/title
End path example









Thanks everyone and I'm looking forward to seeing any feedback!

Tyler Riddle

-- 
If you wish to make an apple pie from scratch you must first invent
the universe. -- Carl Sagan
_______________________________________________
Perl-XML mailing list
[email protected]
To unsubscribe: http://listserv.ActiveState.com/mailman/mysubs