1 | ###########################################################################
|
---|
2 | #
|
---|
3 | # BNContentePlug.pm -- plugin for import the BN-Portugal Collection
|
---|
4 | # A component of the Greenstone digital library software
|
---|
5 | # from the New Zealand Digital Library Project at the
|
---|
6 | # University of Waikato, New Zealand.
|
---|
7 | #
|
---|
8 | # Copyright (C) 1999 New Zealand Digital Library Project
|
---|
9 | #
|
---|
10 | # This program is free software; you can redistribute it and/or modify
|
---|
11 | # it under the terms of the GNU General Public License as published by
|
---|
12 | # the Free Software Foundation; either version 2 of the License, or
|
---|
13 | # (at your option) any later version.
|
---|
14 | #
|
---|
15 | # This program is distributed in the hope that it will be useful,
|
---|
16 | # but WITHOUT ANY WARRANTY; without even the implied warranty of
|
---|
17 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
---|
18 | # GNU General Public License for more details.
|
---|
19 | #
|
---|
20 | # You should have received a copy of the GNU General Public License
|
---|
21 | # along with this program; if not, write to the Free Software
|
---|
22 | # Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
|
---|
23 | #
|
---|
24 | ###########################################################################
|
---|
25 |
|
---|
26 | # BNContentePlug - 11/2004
|
---|
27 | #
|
---|
28 | #
|
---|
29 | # This plugin takes "mets.xml" and "record/NCB_***.xml: the file contain MARC details
|
---|
30 | # about BN-Portugal ccllection. The intension is to import such a collection into GS2.
|
---|
31 |
|
---|
32 | package BNContentePlug;
|
---|
33 |
|
---|
34 | use BasPlug;
|
---|
35 | use plugin;
|
---|
36 | #use ghtml;
|
---|
37 | use XMLParser;
|
---|
38 | use XML::Parser;
|
---|
39 |
|
---|
40 | use strict;
|
---|
41 | no strict 'refs'; # allow filehandles to be variables and viceversa
|
---|
42 |
|
---|
43 | sub BEGIN {
|
---|
44 | @BNContentePlug::ISA = ('BasPlug');
|
---|
45 | unshift (@INC, "$ENV{'GSDLHOME'}/perllib/cpan");
|
---|
46 | }
|
---|
47 |
|
---|
48 | my $arguments =
|
---|
49 | [ { 'name' => "process_exp",
|
---|
50 | 'desc' => "{BasPlug.process_exp}",
|
---|
51 | 'type' => "string",
|
---|
52 | 'deft' => &get_default_process_exp(),
|
---|
53 | 'reqd' => "no" },
|
---|
54 | { 'name' => "only_first_doc",
|
---|
55 | 'desc' => "{BNContentePlug.only_first_doc}",
|
---|
56 | 'type' => "flag",
|
---|
57 | 'reqd' => "no" },
|
---|
58 | { 'name' => "first_inoder_ext",
|
---|
59 | 'desc' => "{BNContentePlug.first_inorder_ext}",
|
---|
60 | 'type' => "flag",
|
---|
61 | 'reqd' => "no" },
|
---|
62 | { 'name' => "first_inorder_mime",
|
---|
63 | 'desc' => "{BNContentePlug.first_inorder_mime}",
|
---|
64 | 'type' => "flag",
|
---|
65 | 'reqd' => "no" },
|
---|
66 | { 'name' => "block_exp",
|
---|
67 | 'desc' => "{BasPlug.block_exp}",
|
---|
68 | 'type' => "string",
|
---|
69 | 'deft' => &get_default_block_exp(),
|
---|
70 | 'reqd' => "no" }];
|
---|
71 |
|
---|
72 | my $options = { 'name' => "BNContentePlug",
|
---|
73 | 'desc' => "{BNContentePlug.desc}",
|
---|
74 | 'inherits' => "yes",
|
---|
75 | 'args' => $arguments };
|
---|
76 |
|
---|
77 | # Important variation to regular plugin structure. Need to desclare
|
---|
78 | # $self as global variable to file so XMLParser callback routines
|
---|
79 | # can access the content of the object.
|
---|
80 | my ($self);
|
---|
81 |
|
---|
82 | sub get_default_process_exp {
|
---|
83 | my $self = shift (@_);
|
---|
84 |
|
---|
85 | return q^(?i)(metsHTML\.xml)$^;
|
---|
86 | }
|
---|
87 |
|
---|
88 | # block files
|
---|
89 | sub get_default_block_exp {
|
---|
90 | my $self = shift (@_);
|
---|
91 |
|
---|
92 | # Block all files besides contents
|
---|
93 | #return q^(?i)(metsHTML\.xml|)$^;
|
---|
94 | return q^(?i)((.*?)\.(.*?))$^;
|
---|
95 | }
|
---|
96 |
|
---|
97 | sub new {
|
---|
98 | my ($class) = shift (@_);
|
---|
99 | my ($pluginlist,$inputargs,$hashArgOptLists) = @_;
|
---|
100 | push(@$pluginlist, $class);
|
---|
101 |
|
---|
102 | if(defined $arguments){ push(@{$hashArgOptLists->{"ArgList"}},@{$arguments});}
|
---|
103 | if(defined $options) { push(@{$hashArgOptLists->{"OptList"}},$options)};
|
---|
104 |
|
---|
105 | $self = (defined $hashArgOptLists)? new BasPlug($pluginlist,$inputargs,$hashArgOptLists): new BasPlug($pluginlist,$inputargs);
|
---|
106 |
|
---|
107 | #create XML::Parser object for parsing metsHTML.xml, NCB_???.xml files
|
---|
108 | my $mets_parser = new XML::Parser('Style' => 'Stream',
|
---|
109 | 'Handlers' => {'Doctype' => \&METS_Doctype,
|
---|
110 | 'Start' => \&METS_StartTag,
|
---|
111 | 'End' => \&METS_EndTag
|
---|
112 | });
|
---|
113 |
|
---|
114 | my $marc_parser = new XML::Parser('Style' => 'Stream',
|
---|
115 | 'Handlers' => {'Char' => \&Char,
|
---|
116 | 'Doctype' => \&MARC_Doctype,
|
---|
117 | 'Start' => \&MARC_StartTag,
|
---|
118 | 'End' => \&MARC_EndTag
|
---|
119 | });
|
---|
120 | $self->{'mets_parser'} = $mets_parser;
|
---|
121 | $self->{'marc_parser'} = $marc_parser;
|
---|
122 |
|
---|
123 | $self->{'index_file'} = "";
|
---|
124 |
|
---|
125 | return bless $self, $class;
|
---|
126 | }
|
---|
127 |
|
---|
128 | sub read_marc_content {
|
---|
129 | my $self = shift (@_);
|
---|
130 | my ($marc_file) = @_;
|
---|
131 |
|
---|
132 | # parse the Marc_file: NCB_???.xml
|
---|
133 | eval{
|
---|
134 | $self->{'marc_parser'}->parsefile($marc_file);
|
---|
135 | };
|
---|
136 |
|
---|
137 | if ($@) {
|
---|
138 | die "BNContentePlug: ERROR $marc_file is not a well formed XML file ($@)\n";
|
---|
139 | }
|
---|
140 | }
|
---|
141 |
|
---|
142 | # Read metsHTML.xml from BN-Portugal collection
|
---|
143 | sub metadata_read {
|
---|
144 | my $self = shift (@_);
|
---|
145 | my ($pluginfo, $base_dir, $file, $metadata, $extrametakeys, $extrametadata, $processor, $maxdocs, $gli) = @_;
|
---|
146 |
|
---|
147 | my $outhandle = $self->{'outhandle'};
|
---|
148 |
|
---|
149 | my $filename = &util::filename_cat($base_dir, $file);
|
---|
150 |
|
---|
151 | if ($filename !~ /metsHTML\.xml$/ || !-f $filename) {
|
---|
152 | if ($filename =~ /\.xml$/i || $filename =~ /log\.txt$/i || $filename =~ /isbd\.html$/i) {
|
---|
153 | $self->{'file_blocks'}->{$filename}=1;
|
---|
154 | }
|
---|
155 | return undef;
|
---|
156 | } else {
|
---|
157 | $self->{'file_blocks'}->{$filename}=1;
|
---|
158 | }
|
---|
159 |
|
---|
160 | print $outhandle "BNContentePlug: extracting metadata from $filename\n"
|
---|
161 | if $self->{'verbosity'} > 1;
|
---|
162 |
|
---|
163 | my ($dir) = $filename =~ /^(.*?)[^\/\\]*$/;
|
---|
164 | $self->{'dir'} = $dir;
|
---|
165 |
|
---|
166 | eval {
|
---|
167 | $self->{'mets_parser'}->parsefile($filename);
|
---|
168 | };
|
---|
169 |
|
---|
170 | if ($@) {
|
---|
171 | die "BNContentePlug: ERROR $filename is not a well formed XML file ($@)\n";
|
---|
172 | }
|
---|
173 |
|
---|
174 | # read NCB_???.xml to parse MARC records and save as metadata
|
---|
175 | my $marc_file = &util::filename_cat($dir,$self->{'marc_file'});
|
---|
176 | $self->read_marc_content ($marc_file);
|
---|
177 |
|
---|
178 | if (defined $self->{'index_file'} && $self->{'index_file'} ne "") {
|
---|
179 | my $index_file = $self->{'index_file'};
|
---|
180 | push(@$extrametakeys,$index_file);
|
---|
181 | $extrametadata->{$index_file} = $self->{'saved_metadata'};
|
---|
182 | } else {
|
---|
183 | print STDERR "####Warning can't find main index file\n";
|
---|
184 | }
|
---|
185 | return 1;
|
---|
186 | }
|
---|
187 |
|
---|
188 | # The BNContentePlug read() function. This function does all the right things
|
---|
189 | # to make general options work for a given plugin. It calls the process()
|
---|
190 | # function which does all the work specific to a plugin (like the old
|
---|
191 | # read functions used to do). Most plugins should define their own
|
---|
192 | # process() function and let this read() function keep control.
|
---|
193 | #
|
---|
194 | # Return number of files processed, undef if can't process
|
---|
195 | # Note that $base_dir might be "" and that $file might
|
---|
196 | # include directories
|
---|
197 |
|
---|
198 | sub readxxx {
|
---|
199 | my $self = shift (@_);
|
---|
200 | my ($pluginfo, $base_dir, $file, $metadata, $processor, $maxdocs, $total_count, $gli) = @_;
|
---|
201 | my $outhandle = $self->{'outhandle'};
|
---|
202 |
|
---|
203 | my $filename = &util::filename_cat($base_dir, $file);
|
---|
204 | #return 0 if $self->{'block_exp'} ne "" && $filename =~ /$self->{'block_exp'}/;
|
---|
205 |
|
---|
206 | #return 0 if ($filename =~ /metsHTML\.xml$/);
|
---|
207 | return 0 if ($filename =~ /\.xml$/);
|
---|
208 | return 0 if (defined $self->{'file_blocks'}->{'filename'});
|
---|
209 |
|
---|
210 | return undef;
|
---|
211 | }
|
---|
212 |
|
---|
213 | # do plugin specific processing of doc_obj
|
---|
214 | sub process {
|
---|
215 | my $self = shift (@_);
|
---|
216 | my ($textref, $pluginfo, $base_dir, $file, $metadata, $doc_obj) = @_;
|
---|
217 | my $outhandle = $self->{'outhandle'};
|
---|
218 |
|
---|
219 | return 1;
|
---|
220 | }
|
---|
221 |
|
---|
222 | sub METS_Doctype {
|
---|
223 | my ($expat, $name, $sysid, $pubid, $internal) = @_;
|
---|
224 | die if ($name !~ /^metsHTML\.xml$/);
|
---|
225 | }
|
---|
226 |
|
---|
227 | sub MARC_Doctype {
|
---|
228 | my ($expat, $name, $sysid, $pubid, $internal) = @_;
|
---|
229 | print STDERR "###MARC Name =$name\n";
|
---|
230 | #die if ($name !~ /^metsHTML\.xml$/);
|
---|
231 | #die if (!$name);
|
---|
232 | }
|
---|
233 |
|
---|
234 | sub METS_StartTag {
|
---|
235 | my ($expat, $element, %attr) = @_;
|
---|
236 | my @file_blocks;
|
---|
237 | if ($element eq "dmdSec") {
|
---|
238 | $self->{'marc_file'} = "";
|
---|
239 | } elsif ($element eq "mdRef") {
|
---|
240 | my $marc_href = $attr{'xlink:href'};
|
---|
241 | $self->{'marc_file'} = $marc_href;
|
---|
242 | } elsif ($element eq "FLocat"){
|
---|
243 | my $assocfiles = $attr{'xlink:href'};
|
---|
244 | if ($assocfiles =~ /index\.html$/) {
|
---|
245 | my $index_file = &util::filename_cat($self->{'dir'}, $assocfiles);
|
---|
246 | $self->{'index_file'} = $index_file;
|
---|
247 | } else {
|
---|
248 | my $link = &util::filename_cat($self->{'dir'}, $assocfiles);
|
---|
249 | #$self->{'file_blocks'}->{$link} = 1;
|
---|
250 | }
|
---|
251 | }
|
---|
252 | }
|
---|
253 |
|
---|
254 | sub METS_EndTag {
|
---|
255 | my ($expat, $element, %attr) = @_;
|
---|
256 | }
|
---|
257 |
|
---|
258 | sub MARC_StartTag {
|
---|
259 | my ($expat, $element, %attr) = @_;
|
---|
260 |
|
---|
261 | if ($element eq "record") {
|
---|
262 | $self->{'saved_metadata'} = {};
|
---|
263 | }
|
---|
264 | elsif ($element eq "datafield") {
|
---|
265 | $self->{'metaname'} = $element;
|
---|
266 | $self->{'datafield'} = $attr{'tag'};
|
---|
267 | }
|
---|
268 | elsif ($element eq "subfield") {
|
---|
269 | $self->{'subfield'} = $attr{'code'};
|
---|
270 | $self->{'text'} = "";
|
---|
271 | }
|
---|
272 | }
|
---|
273 |
|
---|
274 | sub MARC_EndTag {
|
---|
275 | my ($expat, $element) = @_;
|
---|
276 |
|
---|
277 | if ($element eq "datafield") {
|
---|
278 | $self->{'metaname'} = "";
|
---|
279 | }
|
---|
280 | elsif ($element eq "subfield") {
|
---|
281 | my $mvalue = $self->{'text'};
|
---|
282 | my $mname = $self->{'datafield'}."^".$self->{'subfield'};
|
---|
283 | #print STDERR "**** $mname = $mvalue\n";
|
---|
284 |
|
---|
285 | $mvalue =~ s/\[/&\#91;/g;
|
---|
286 | $mvalue =~ s/\[/&\#93;/g;
|
---|
287 |
|
---|
288 |
|
---|
289 | if (defined $self->{'saved_metadata'}->{$mname}) {
|
---|
290 | # accumulate - add value to existing value(s)
|
---|
291 | if (ref ($self->{'saved_metadata'}->{$mname}) eq "ARRAY") {
|
---|
292 | push (@{$self->{'saved_metadata'}->{$mname}}, $mvalue);
|
---|
293 | } else {
|
---|
294 | $self->{'saved_metadata'}->{$mname} =
|
---|
295 | [$self->{'saved_metadata'}->{$mname}, $mvalue];
|
---|
296 | }
|
---|
297 | } else {
|
---|
298 | # accumulate - add value into (currently empty) array
|
---|
299 | $self->{'saved_metadata'}->{$mname} = [$mvalue];
|
---|
300 | }
|
---|
301 | # store something here
|
---|
302 | $self->{'subfield'} = "";
|
---|
303 | $self->{'text'} = "";
|
---|
304 | }
|
---|
305 |
|
---|
306 | }
|
---|
307 |
|
---|
308 |
|
---|
309 | # This Char function overrides the one in XML::Parser::Stream to overcome a
|
---|
310 | # problem where $expat->{Text} is treated as the return value, slowing
|
---|
311 | # things down significantly in some cases.
|
---|
312 | sub Char {
|
---|
313 | use bytes; # Necessary to prevent encoding issues with XML::Parser 2.31+
|
---|
314 | $_[0]->{'Text'} .= $_[1];
|
---|
315 | if ((defined $self->{'subfield'} && ($self->{'subfield'} ne ""))) {
|
---|
316 | $self->{'text'} .= $_[1];
|
---|
317 | }
|
---|
318 | return undef;
|
---|
319 | }
|
---|
320 |
|
---|
321 | 1;
|
---|