1 |
############################################################################## |
2 |
# |
3 |
# Perl module: XML::XUpdate::XSLT |
4 |
# |
5 |
# By Andreas Motl, andreas.motl@ilo.de |
6 |
# |
7 |
# $Id: XSLT.pm,v 1.6 2003/05/07 03:11:28 joko Exp $ |
8 |
# |
9 |
# $Log: XSLT.pm,v $ |
10 |
# Revision 1.6 2003/05/07 03:11:28 joko |
11 |
# updated pod: new section "References", added item to "Todo" section |
12 |
# |
13 |
# Revision 1.5 2003/05/06 14:24:06 joko |
14 |
# doesn't use Carp any more |
15 |
# updated pod |
16 |
# attempt to rewrite xml to make it more human readable using helper modules from CPAN (node indentation, etc.) |
17 |
# |
18 |
# Revision 1.4 2003/05/01 23:40:32 joko |
19 |
# minor update: commented debugging part |
20 |
# |
21 |
# Revision 1.3 2003/05/01 20:11:49 joko |
22 |
# * added pod from xupdate.pl |
23 |
# - extracted xml to external files |
24 |
# |
25 |
# Revision 1.1 2003/04/30 02:36:32 joko |
26 |
# initial commit |
27 |
# |
28 |
# |
29 |
############################################################################### |
30 |
|
31 |
=pod |
32 |
|
33 |
|
34 |
=head1 NAME |
35 |
|
36 |
XML::XUpdate::XSLT - A perl module for updating xml documents using XUpdate via XSLT. |
37 |
|
38 |
|
39 |
=head3 Overview |
40 |
|
41 |
This is not the same xupdate currently available from CPAN at |
42 |
http://search.cpan.org/author/PAJAS/XML-XUpdate-LibXML-0.4.0/xupdate . |
43 |
|
44 |
Its intention - however - is identical: |
45 |
xupdate - Process XUpdate commands against an XML document. |
46 |
|
47 |
|
48 |
=head4 Their implementations differ: |
49 |
|
50 |
1. xupdate (by Petr Pajas) uses ... |
51 |
XML::XUpdate::LibXML - Simple implementation of XUpdate format |
52 |
|
53 |
... which is based on XML::LibXML which in turn is: |
54 |
[...] |
55 |
This module is an interface to the gnome libxml2 DOM parser (no SAX parser support yet), |
56 |
and the DOM tree. It also provides an XML::XPath-like findnodes() interface, providing |
57 |
access to the XPath API in libxml2. |
58 |
[...] |
59 |
|
60 |
2. This xupdate attempts to implement the XUpdate specs using XSLT only. |
61 |
|
62 |
|
63 |
=head4 Yet another xupdate - facts in short: |
64 |
|
65 |
S: It would be nice to have a pure perl thingy which does (almost) the same stuff.... |
66 |
|
67 |
Q: Can we achieve compliance with its (XML::XUpdate::LibXML) API? (or just a subset ....) |
68 |
|
69 |
Q: Can we achieve the processing using CPAN's XML::XSLT? |
70 |
S: Proposal: XML::XUpdate::XSLT!? |
71 |
|
72 |
Q: Can we mimic/use the interface of the - already established - 'xupdate' program??? |
73 |
|
74 |
Q: Should we follow the CRUD path first? |
75 |
(CRUD is the acronym for the datastore action primitives: Create, Retrieve, Update, Delete) |
76 |
S?: Proposal: XML::XUpdate::XSLT::API uses XML::XUpdate::XSLT::CRUD |
77 |
|
78 |
|
79 |
=head3 References |
80 |
|
81 |
- XUpdate: |
82 |
Requirements: http://www.xmldb.org/xupdate/xupdate-req.html |
83 |
Working Draft: http://www.xmldb.org/xupdate/xupdate-wd.html |
84 |
- XML API: |
85 |
- XML::XUpdate::LibXML: http://search.cpan.org/author/PAJAS/XML-XUpdate-LibXML-0.4.0/lib/XML/XUpdate/LibXML.pm |
86 |
- XSL / XSLT: |
87 |
http://www.w3.org/TR/xslt |
88 |
http://www.xsl-rp.de/ |
89 |
http://xml.klute-thiemann.de/w3c-de/REC-xslt-20020318/ |
90 |
http://xmlxslt.sourceforge.net/ |
91 |
- misc pointers: |
92 |
"Re: Modify XML documents": http://aspn.activestate.com/ASPN/Mail/Message/perl-xml/1265431 |
93 |
XSL Extensions: http://xmlsoft.org/XSLT/extensions.html |
94 |
EXSLT: http://www.exslt.org/set/functions/difference/index.html |
95 |
"how to insert element at required position in document tree?": http://p2p.wrox.com/archive/xslt/2001-06/98.asp |
96 |
XML APIs for Databases: http://www.javaworld.com/javaworld/jw-01-2000/jw-01-dbxml.html |
97 |
|
98 |
|
99 |
=head3 Todo |
100 |
|
101 |
o What about proper encoding? (ISO-8859-1 or UTF-8) |
102 |
o Is it possible to create the required "xsl_template.xml" at runtime via XSL itself? |
103 |
o Cache contents of external files (*.xml). Performance! |
104 |
o Can exception / error-code handling be improved somehow? |
105 |
Esp.: It would be interesting, if the XUpdate payload actually could be applied, or not... |
106 |
o Conditional processing: <xupdate:if>, etc. |
107 |
|
108 |
|
109 |
=cut |
110 |
|
111 |
|
112 |
###################################################################### |
113 |
package XML::XUpdate::XSLT; |
114 |
###################################################################### |
115 |
|
116 |
use strict; |
117 |
use warnings; |
118 |
|
119 |
use Data::Dumper; |
120 |
use File::Basename; |
121 |
use XML::XSLT 0.41; |
122 |
use XML::XUpdate::Rewrite; |
123 |
|
124 |
# Namespace constants |
125 |
|
126 |
use constant NS_XUPDATE => 'http://www.xmldb.org/xupdate'; |
127 |
|
128 |
use vars qw ( $VERSION ); |
129 |
|
130 |
$VERSION = '0.01'; |
131 |
|
132 |
|
133 |
###################################################################### |
134 |
# PUBLIC DEFINITIONS |
135 |
|
136 |
sub new { |
137 |
my $class = shift; |
138 |
my $self = bless {}, $class; |
139 |
my $args = $self->__parse_args(@_); |
140 |
|
141 |
$self->{DEBUG} = $args->{debug}; |
142 |
$self->{WARNINGS} = $args->{warnings}; |
143 |
|
144 |
$self->__init_default_stylesheets(); |
145 |
|
146 |
return $self; |
147 |
|
148 |
} |
149 |
|
150 |
sub get_stylesheet { |
151 |
my $self = shift; |
152 |
my $name = shift; |
153 |
return $self->{XML}->{xsl}->{$name}; |
154 |
} |
155 |
|
156 |
sub set_stylesheet { |
157 |
my $self = shift; |
158 |
my $name = shift; |
159 |
my $xml = shift; |
160 |
my $options = shift; |
161 |
if ($options->{encap}) { |
162 |
my $template = $self->__slurp_file("xsl_template.xml"); |
163 |
# FIXME! What about the quirky? Is there a better one? |
164 |
$template =~ s!<xsl:quirky_placeholder />!$xml!; |
165 |
$xml = $template; |
166 |
} |
167 |
$self->{XML}->{xsl}->{$name} = $xml; |
168 |
} |
169 |
|
170 |
sub open_document { |
171 |
my $self = shift; |
172 |
my $xml = shift; |
173 |
# FIXME: check for filename, filehandle and URL (etc.) |
174 |
$self->{XML}->{document} = $xml; |
175 |
} |
176 |
|
177 |
sub open_xupdate { |
178 |
my $self = shift; |
179 |
my $xml = shift; |
180 |
# FIXME: check for filename, filehandle and URL (etc.) |
181 |
$self->{XML}->{xupdate} = $xml; |
182 |
} |
183 |
|
184 |
sub process { |
185 |
my $self = shift; |
186 |
$self->_calculate(); |
187 |
$self->_apply(); |
188 |
} |
189 |
|
190 |
# First, translate the xupdate payload to xsl. |
191 |
# FIXME: do DOM only! (don't use "->toString") |
192 |
sub _calculate { |
193 |
my $self = shift; |
194 |
$self->{XSLT_ENGINE_PREP} = XML::XSLT->new( |
195 |
Source => $self->get_stylesheet("xupdate2xsl"), |
196 |
debug => $self->{DEBUG}, |
197 |
warnings => $self->{WARNINGS} |
198 |
); |
199 |
$self->{XSLT_ENGINE_PREP}->open_xml( $self->{XML}->{xupdate} ); |
200 |
$self->{XSLT_ENGINE_PREP}->process(); |
201 |
$self->set_stylesheet( "_worker", $self->{XSLT_ENGINE_PREP}->toString(), { encap => 1 } ); |
202 |
} |
203 |
|
204 |
# After that, use this worker xsl to actually apply the changes to the original document. |
205 |
# FIXME: do DOM only! |
206 |
sub _apply { |
207 |
my $self = shift; |
208 |
|
209 |
# debug - print the calculated xsl on STDERR |
210 |
print STDERR $self->get_stylesheet("_worker"), "\n"; |
211 |
|
212 |
#return; |
213 |
$self->{XSLT_ENGINE_LIVE} = XML::XSLT->new( |
214 |
Source => $self->get_stylesheet("_worker"), |
215 |
debug => $self->{DEBUG}, |
216 |
warnings => $self->{WARNINGS} |
217 |
); |
218 |
$self->{XSLT_ENGINE_LIVE}->open_xml( $self->{XML}->{document} ); |
219 |
$self->{XSLT_ENGINE_LIVE}->process(); |
220 |
$self->{XML}->{result} = $self->{XSLT_ENGINE_LIVE}->toString(); |
221 |
} |
222 |
|
223 |
sub toString { |
224 |
my $self = shift; |
225 |
|
226 |
# use rest of argument list as hash of option values |
227 |
my $options = {@_}; |
228 |
|
229 |
# short circuit - just return what we have - don't modify anything |
230 |
return $self->{XML}->{result} unless $options and $options->{rewrite}; |
231 |
|
232 |
# Rewrite the xml document with certain engine to desired style. |
233 |
# 'engine' defaults to "XMLParser" if not specified. |
234 |
my $rewrite = XML::XUpdate::Rewrite->new( style => $options->{mode}, engine => $options->{using} ); |
235 |
$rewrite->set_document( $self->{XML}->{result} ); |
236 |
# TODO: Implement a configurable fallback here to return the un-rewritten payload if desired. |
237 |
$rewrite->process() or die(__PACKAGE__ . ": Error while rewriting XML document."); |
238 |
return $rewrite->get_document(); |
239 |
} |
240 |
|
241 |
|
242 |
###################################################################### |
243 |
# AUXILIARY METHODS |
244 |
|
245 |
# Argument list parsing. |
246 |
# ... from XML::XUpdate::LibXML (w/o backwards compatibility hook). |
247 |
# Could this make up a Class::AutoFill::__read_arglist which transparently |
248 |
# makes object attributes from constructor arguments in a configurable way? |
249 |
sub __parse_args { |
250 |
my $self = shift; |
251 |
my %args; |
252 |
|
253 |
if ( @_ % 2 ) { |
254 |
$args{dummy} = shift; |
255 |
%args = (%args, @_); |
256 |
} else { |
257 |
%args = @_; |
258 |
} |
259 |
|
260 |
return \%args; |
261 |
} |
262 |
|
263 |
|
264 |
sub __init_default_stylesheets { |
265 |
my $self = shift; |
266 |
if (my $payload = $self->__slurp_file("xupdate2xsl.xml")) { |
267 |
$self->{XML}->{xsl}->{xupdate2xsl} = $payload; |
268 |
} |
269 |
} |
270 |
|
271 |
sub __slurp_file { |
272 |
my $self = shift; |
273 |
my $filename = shift; |
274 |
|
275 |
# does file exist? |
276 |
if (! -e $filename) { |
277 |
$filename = dirname(__FILE__) . "/$filename"; |
278 |
if (! -e $filename) { |
279 |
die("File $filename does not exist."); |
280 |
} |
281 |
} |
282 |
|
283 |
open(FH, "<" . $filename); |
284 |
my @lines = <FH>; |
285 |
my $content = join("", @lines); |
286 |
close(FH); |
287 |
return $content; |
288 |
} |
289 |
|
290 |
1; |
291 |
__END__ |