XML/XUpdate/XSLT.pm

##############################################################################
#
# Perl module: XML::XUpdate::XSLT
#
# By Andreas Motl, andreas.motl@ilo.de
#
#    $Id: XSLT.pm,v 1.5 2003/05/06 14:24:06 joko Exp $
#
#    $Log: XSLT.pm,v $
#    Revision 1.5  2003/05/06 14:24:06  joko
#    doesn't use Carp any more
#    updated pod
#    attempt to rewrite xml to make it more human readable using helper modules from CPAN (node indentation, etc.)
#
#    Revision 1.4  2003/05/01 23:40:32  joko
#    minor update: commented debugging part
#
#    Revision 1.3  2003/05/01 20:11:49  joko
#    * added pod from xupdate.pl
#    - extracted xml to external files
#
#    Revision 1.1  2003/04/30 02:36:32  joko
#    initial commit
#
#
###############################################################################

=pod


=head1 NAME

  XML::XUpdate::XSLT - A perl module for updating xml documents using XUpdate via XSLT.


=head3 Overview

  This is not the same xupdate currently available from CPAN at
  http://search.cpan.org/author/PAJAS/XML-XUpdate-LibXML-0.4.0/xupdate .
  
  Its intention - however - is identical:
  xupdate - Process XUpdate commands against an XML document.


=head4 Their implementations differ:

    1. xupdate (by Petr Pajas) uses ...
    XML::XUpdate::LibXML - Simple implementation of XUpdate format 
    
    ... which is based on XML::LibXML which in turn is:
    [...]
    This module is an interface to the gnome libxml2 DOM parser (no SAX parser support yet), 
    and the DOM tree. It also provides an XML::XPath-like findnodes() interface, providing 
    access to the XPath API in libxml2. 
    [...]

    2. This xupdate attempts to implement the XUpdate specs using XSLT only.
    

=head4 Yet another xupdate - facts in short:

  S: It would be nice to have a pure perl thingy which does (almost) the same stuff....
  
  Q: Can we achieve compliance with its (XML::XUpdate::LibXML) API? (or just a subset ....)
  
  Q: Can we achieve the processing using CPAN's XML::XSLT?
  S: Proposal: XML::XUpdate::XSLT!?
  
  Q: Can we mimic/use the interface of the - already established - 'xupdate' program???
  
  Q: Should we follow the CRUD path first?
      (CRUD is the acronym for the datastore action primitives: Create, Retrieve, Update, Delete)
  S?: Proposal: XML::XUpdate::XSLT::API uses XML::XUpdate::XSLT::CRUD


=head3 References

    - XUpdate:
      Requirements: http://www.xmldb.org/xupdate/xupdate-req.html
      Working Draft: http://www.xmldb.org/xupdate/xupdate-wd.html
    - XML API:
    - XML::XUpdate::LibXML: http://search.cpan.org/author/PAJAS/XML-XUpdate-LibXML-0.4.0/lib/XML/XUpdate/LibXML.pm
    - XSL / XSLT:
      http://www.w3.org/TR/xslt
      http://www.xsl-rp.de/
      http://xml.klute-thiemann.de/w3c-de/REC-xslt-20020318/
      http://xmlxslt.sourceforge.net/
    - misc pointers:
      "Re: Modify XML documents": http://aspn.activestate.com/ASPN/Mail/Message/perl-xml/1265431
      XSL Extensions: http://xmlsoft.org/XSLT/extensions.html
      EXSLT: http://www.exslt.org/set/functions/difference/index.html
      "how to insert element at required position in document tree?": http://p2p.wrox.com/archive/xslt/2001-06/98.asp
      XML APIs for Databases: http://www.javaworld.com/javaworld/jw-01-2000/jw-01-dbxml.html


=head3 Todo

  o What about proper encoding? (ISO-8859-1 or UTF-8)
  o Is it possible to create the required "xsl_template.xml" at runtime via XSL itself?
  o Cache contents of external files (*.xml). Performance!
  o Can exception / error-code handling be improved somehow?
      Esp.: It would be interesting, if the XUpdate payload actually could be applied, or not...
  o <xupdate:if>, etc.
  

=cut


######################################################################
package XML::XUpdate::XSLT;
######################################################################

use strict;
use warnings;

use Data::Dumper;
use File::Basename;
use XML::XSLT 0.41;
use XML::XUpdate::Rewrite;

# Namespace constants

use constant NS_XUPDATE         => 'http://www.xmldb.org/xupdate';

use vars qw ( $VERSION );

$VERSION = '0.01';


######################################################################
# PUBLIC DEFINITIONS

sub new {
  my $class = shift;
  my $self = bless {}, $class;
  my $args = $self->__parse_args(@_);
  
  $self->{DEBUG} = $args->{debug};
  $self->{WARNINGS} = $args->{warnings};

  $self->__init_default_stylesheets();

  return $self;

}

sub get_stylesheet {
  my $self = shift;
  my $name = shift;
  return $self->{XML}->{xsl}->{$name};
}

sub set_stylesheet {
  my $self = shift;
  my $name = shift;
  my $xml = shift;
  my $options = shift;
  if ($options->{encap}) {
    my $template = $self->__slurp_file("xsl_template.xml");
    # FIXME! What about the quirky? Is there a better one?
    $template =~ s!<xsl:quirky_placeholder />!$xml!;
    $xml = $template;
  }
  $self->{XML}->{xsl}->{$name} = $xml;
}

sub open_document {
  my $self = shift;
  my $xml = shift;
  # FIXME: check for filename, filehandle and URL (etc.)
  $self->{XML}->{document} = $xml;
}

sub open_xupdate {
  my $self = shift;
  my $xml = shift;
  # FIXME: check for filename, filehandle and URL (etc.)
  $self->{XML}->{xupdate} = $xml;
}

sub process {
  my $self = shift;
  $self->_calculate();
  $self->_apply();
}

# First, translate the xupdate payload to xsl.
# FIXME: do DOM only! (don't use "->toString")
sub _calculate {
  my $self = shift;
  $self->{XSLT_ENGINE_PREP} = XML::XSLT->new(
    Source => $self->get_stylesheet("xupdate2xsl"), 
    debug => $self->{DEBUG},
    warnings => $self->{WARNINGS}
  );
  $self->{XSLT_ENGINE_PREP}->open_xml( $self->{XML}->{xupdate} );
  $self->{XSLT_ENGINE_PREP}->process();
  $self->set_stylesheet( "_worker", $self->{XSLT_ENGINE_PREP}->toString(), { encap => 1 } );
}

# After that, use this worker xsl to actually apply the changes to the original document.
# FIXME: do DOM only!
sub _apply {
  my $self = shift;
  
  # debug - print the calculated xsl on STDERR
  print STDERR $self->get_stylesheet("_worker"), "\n";
  
  #return;
  $self->{XSLT_ENGINE_LIVE} = XML::XSLT->new(
    Source => $self->get_stylesheet("_worker"), 
    debug => $self->{DEBUG},
    warnings => $self->{WARNINGS}
  );
  $self->{XSLT_ENGINE_LIVE}->open_xml( $self->{XML}->{document} );
  $self->{XSLT_ENGINE_LIVE}->process();
  $self->{XML}->{result} = $self->{XSLT_ENGINE_LIVE}->toString();
}  
  
sub toString {
  my $self = shift;
  
  # use rest of argument list as hash of option values
  my $options = {@_};
  
  # short circuit - just return what we have - don't modify anything
  return $self->{XML}->{result} unless $options and $options->{rewrite};
  
  # Rewrite the xml document with certain engine to desired style.
  # 'engine' defaults to "XMLParser" if not specified.
  my $rewrite = XML::XUpdate::Rewrite->new( style => $options->{mode}, engine => $options->{using} );
  $rewrite->set_document( $self->{XML}->{result} );
  # TODO: Implement a configurable fallback here to return the un-rewritten payload if desired.
  $rewrite->process() or die(__PACKAGE__ . ": Error while rewriting XML document.");
  return $rewrite->get_document();
}


######################################################################
# AUXILIARY METHODS

# Argument list parsing.
# ... from XML::XUpdate::LibXML (w/o backwards compatibility hook).
# Could this make up a Class::AutoFill::__read_arglist which transparently 
# makes object attributes from constructor arguments in a configurable way?
sub __parse_args {
  my $self = shift;
  my %args;

  if ( @_ % 2 ) {
    $args{dummy} = shift;
    %args = (%args, @_);
  } else {
    %args = @_;
  }

  return \%args;
}


sub __init_default_stylesheets {
  my $self = shift;
  if (my $payload = $self->__slurp_file("xupdate2xsl.xml")) {
    $self->{XML}->{xsl}->{xupdate2xsl} = $payload;
  }
}

sub __slurp_file {
  my $self = shift;
  my $filename = shift;
  
  # does file exist?
  if (! -e $filename) {
    $filename = dirname(__FILE__) . "/$filename";
    if (! -e $filename) {
      die("File $filename does not exist.");
    }
  }
  
  open(FH, "<" . $filename);
  my @lines = <FH>;
  my $content = join("", @lines);
  close(FH);
  return $content;
}

1;
__END__
1	joko	1.1	##############################################################################
2			#
3			# Perl module: XML::XUpdate::XSLT
4			#
5			# By Andreas Motl, andreas.motl@ilo.de
6			#
7	joko	1.6	# $Id: XSLT.pm,v 1.5 2003/05/06 14:24:06 joko Exp $
8	joko	1.1	#
9			# $Log: XSLT.pm,v $
10	joko	1.6	# Revision 1.5 2003/05/06 14:24:06 joko
11			# doesn't use Carp any more
12			# updated pod
13			# attempt to rewrite xml to make it more human readable using helper modules from CPAN (node indentation, etc.)
14			#
15	joko	1.5	# Revision 1.4 2003/05/01 23:40:32 joko
16			# minor update: commented debugging part
17			#
18	joko	1.4	# Revision 1.3 2003/05/01 20:11:49 joko
19			# * added pod from xupdate.pl
20			# - extracted xml to external files
21			#
22	joko	1.2	# Revision 1.1 2003/04/30 02:36:32 joko
23			# initial commit
24			#
25	joko	1.1	#
26			###############################################################################
27
28	joko	1.3	=pod
29
30
31	joko	1.1	=head1 NAME
32
33	joko	1.5	XML::XUpdate::XSLT - A perl module for updating xml documents using XUpdate via XSLT.
34	joko	1.1
35	joko	1.3
36			=head3 Overview
37
38			This is not the same xupdate currently available from CPAN at
39			http://search.cpan.org/author/PAJAS/XML-XUpdate-LibXML-0.4.0/xupdate .
40
41			Its intention - however - is identical:
42			xupdate - Process XUpdate commands against an XML document.
43
44
45	joko	1.6	=head4 Their implementations differ:
46	joko	1.3
47			1. xupdate (by Petr Pajas) uses ...
48			XML::XUpdate::LibXML - Simple implementation of XUpdate format
49
50			... which is based on XML::LibXML which in turn is:
51			[...]
52			This module is an interface to the gnome libxml2 DOM parser (no SAX parser support yet),
53			and the DOM tree. It also provides an XML::XPath-like findnodes() interface, providing
54			access to the XPath API in libxml2.
55			[...]
56
57			2. This xupdate attempts to implement the XUpdate specs using XSLT only.
58
59
60	joko	1.6	=head4 Yet another xupdate - facts in short:
61	joko	1.3
62			S: It would be nice to have a pure perl thingy which does (almost) the same stuff....
63
64			Q: Can we achieve compliance with its (XML::XUpdate::LibXML) API? (or just a subset ....)
65
66			Q: Can we achieve the processing using CPAN's XML::XSLT?
67			S: Proposal: XML::XUpdate::XSLT!?
68
69			Q: Can we mimic/use the interface of the - already established - 'xupdate' program???
70
71	joko	1.5	Q: Should we follow the CRUD path first?
72			(CRUD is the acronym for the datastore action primitives: Create, Retrieve, Update, Delete)
73	joko	1.3	S?: Proposal: XML::XUpdate::XSLT::API uses XML::XUpdate::XSLT::CRUD
74
75
76	joko	1.6	=head3 References
77
78			- XUpdate:
79			Requirements: http://www.xmldb.org/xupdate/xupdate-req.html
80			Working Draft: http://www.xmldb.org/xupdate/xupdate-wd.html
81			- XML API:
82			- XML::XUpdate::LibXML: http://search.cpan.org/author/PAJAS/XML-XUpdate-LibXML-0.4.0/lib/XML/XUpdate/LibXML.pm
83			- XSL / XSLT:
84			http://www.w3.org/TR/xslt
85			http://www.xsl-rp.de/
86			http://xml.klute-thiemann.de/w3c-de/REC-xslt-20020318/
87			http://xmlxslt.sourceforge.net/
88			- misc pointers:
89			"Re: Modify XML documents": http://aspn.activestate.com/ASPN/Mail/Message/perl-xml/1265431
90			XSL Extensions: http://xmlsoft.org/XSLT/extensions.html
91			EXSLT: http://www.exslt.org/set/functions/difference/index.html
92			"how to insert element at required position in document tree?": http://p2p.wrox.com/archive/xslt/2001-06/98.asp
93			XML APIs for Databases: http://www.javaworld.com/javaworld/jw-01-2000/jw-01-dbxml.html
94
95
96			=head3 Todo
97	joko	1.3
98			o What about proper encoding? (ISO-8859-1 or UTF-8)
99			o Is it possible to create the required "xsl_template.xml" at runtime via XSL itself?
100			o Cache contents of external files (*.xml). Performance!
101	joko	1.5	o Can exception / error-code handling be improved somehow?
102			Esp.: It would be interesting, if the XUpdate payload actually could be applied, or not...
103	joko	1.6	o <xupdate:if>, etc.
104	joko	1.3
105
106	joko	1.1	=cut
107
108
109			######################################################################
110			package XML::XUpdate::XSLT;
111			######################################################################
112
113			use strict;
114			use warnings;
115
116	joko	1.5	use Data::Dumper;
117			use File::Basename;
118	joko	1.1	use XML::XSLT 0.41;
119	joko	1.5	use XML::XUpdate::Rewrite;
120	joko	1.1
121			# Namespace constants
122
123			use constant NS_XUPDATE => 'http://www.xmldb.org/xupdate';
124
125			use vars qw ( $VERSION );
126
127	joko	1.3	$VERSION = '0.01';
128	joko	1.1
129
130			######################################################################
131			# PUBLIC DEFINITIONS
132
133			sub new {
134			my $class = shift;
135			my $self = bless {}, $class;
136			my $args = $self->__parse_args(@_);
137
138	joko	1.2	$self->{DEBUG} = $args->{debug};
139			$self->{WARNINGS} = $args->{warnings};
140
141	joko	1.3	$self->__init_default_stylesheets();
142	joko	1.1
143	joko	1.2	return $self;
144	joko	1.1
145	joko	1.2	}
146	joko	1.1
147	joko	1.2	sub get_stylesheet {
148			my $self = shift;
149			my $name = shift;
150			return $self->{XML}->{xsl}->{$name};
151	joko	1.1	}
152
153	joko	1.2	sub set_stylesheet {
154	joko	1.1	my $self = shift;
155			my $name = shift;
156	joko	1.2	my $xml = shift;
157			my $options = shift;
158			if ($options->{encap}) {
159	joko	1.3	my $template = $self->__slurp_file("xsl_template.xml");
160			# FIXME! What about the quirky? Is there a better one?
161			$template =~ s!<xsl:quirky_placeholder />!$xml!;
162			$xml = $template;
163	joko	1.2	}
164			$self->{XML}->{xsl}->{$name} = $xml;
165	joko	1.1	}
166
167			sub open_document {
168			my $self = shift;
169			my $xml = shift;
170			# FIXME: check for filename, filehandle and URL (etc.)
171			$self->{XML}->{document} = $xml;
172			}
173
174			sub open_xupdate {
175			my $self = shift;
176			my $xml = shift;
177	joko	1.6	# FIXME: check for filename, filehandle and URL (etc.)
178	joko	1.1	$self->{XML}->{xupdate} = $xml;
179			}
180
181			sub process {
182			my $self = shift;
183	joko	1.2	$self->_calculate();
184			$self->_apply();
185	joko	1.1	}
186
187	joko	1.2	# First, translate the xupdate payload to xsl.
188	joko	1.5	# FIXME: do DOM only! (don't use "->toString")
189	joko	1.2	sub _calculate {
190			my $self = shift;
191			$self->{XSLT_ENGINE_PREP} = XML::XSLT->new(
192			Source => $self->get_stylesheet("xupdate2xsl"),
193			debug => $self->{DEBUG},
194			warnings => $self->{WARNINGS}
195			);
196			$self->{XSLT_ENGINE_PREP}->open_xml( $self->{XML}->{xupdate} );
197			$self->{XSLT_ENGINE_PREP}->process();
198			$self->set_stylesheet( "_worker", $self->{XSLT_ENGINE_PREP}->toString(), { encap => 1 } );
199			}
200
201			# After that, use this worker xsl to actually apply the changes to the original document.
202			# FIXME: do DOM only!
203			sub _apply {
204			my $self = shift;
205	joko	1.4
206			# debug - print the calculated xsl on STDERR
207	joko	1.6	print STDERR $self->get_stylesheet("_worker"), "\n";
208	joko	1.4
209	joko	1.2	#return;
210			$self->{XSLT_ENGINE_LIVE} = XML::XSLT->new(
211			Source => $self->get_stylesheet("_worker"),
212			debug => $self->{DEBUG},
213			warnings => $self->{WARNINGS}
214			);
215			$self->{XSLT_ENGINE_LIVE}->open_xml( $self->{XML}->{document} );
216			$self->{XSLT_ENGINE_LIVE}->process();
217			$self->{XML}->{result} = $self->{XSLT_ENGINE_LIVE}->toString();
218			}
219
220	joko	1.1	sub toString {
221			my $self = shift;
222	joko	1.5
223			# use rest of argument list as hash of option values
224			my $options = {@_};
225
226			# short circuit - just return what we have - don't modify anything
227			return $self->{XML}->{result} unless $options and $options->{rewrite};
228
229			# Rewrite the xml document with certain engine to desired style.
230			# 'engine' defaults to "XMLParser" if not specified.
231			my $rewrite = XML::XUpdate::Rewrite->new( style => $options->{mode}, engine => $options->{using} );
232			$rewrite->set_document( $self->{XML}->{result} );
233			# TODO: Implement a configurable fallback here to return the un-rewritten payload if desired.
234			$rewrite->process() or die(__PACKAGE__ . ": Error while rewriting XML document.");
235			return $rewrite->get_document();
236	joko	1.1	}
237	joko	1.5
238	joko	1.1
239			######################################################################
240			# AUXILIARY METHODS
241
242	joko	1.5	# Argument list parsing.
243			# ... from XML::XUpdate::LibXML (w/o backwards compatibility hook).
244			# Could this make up a Class::AutoFill::__read_arglist which transparently
245			# makes object attributes from constructor arguments in a configurable way?
246	joko	1.1	sub __parse_args {
247			my $self = shift;
248			my %args;
249
250	joko	1.5	if ( @_ % 2 ) {
251	joko	1.1	$args{dummy} = shift;
252			%args = (%args, @_);
253			} else {
254			%args = @_;
255			}
256
257			return \%args;
258			}
259
260
261	joko	1.3	sub __init_default_stylesheets {
262			my $self = shift;
263			if (my $payload = $self->__slurp_file("xupdate2xsl.xml")) {
264			$self->{XML}->{xsl}->{xupdate2xsl} = $payload;
265			}
266			}
267
268			sub __slurp_file {
269	joko	1.1	my $self = shift;
270	joko	1.3	my $filename = shift;
271	joko	1.1
272	joko	1.3	# does file exist?
273			if (! -e $filename) {
274			$filename = dirname(__FILE__) . "/$filename";
275			if (! -e $filename) {
276	joko	1.5	die("File $filename does not exist.");
277	joko	1.3	}
278			}
279	joko	1.1
280	joko	1.3	open(FH, "<" . $filename);
281			my @lines = <FH>;
282			my $content = join("", @lines);
283			close(FH);
284			return $content;
285	joko	1.1	}
286
287			1;
288			__END__
MailToCvsAdmin">MailToCvsAdmin	ViewVC Help
Powered by ViewVC 1.1.26