1 | ###########################################################################
|
---|
2 | #
|
---|
3 | # GMLPlug.pm --
|
---|
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 | # plugin which processes a GML format document
|
---|
27 | # assumes that gml tags are all in lower-case.
|
---|
28 |
|
---|
29 | package GMLPlug;
|
---|
30 |
|
---|
31 | use BasPlug;
|
---|
32 | use util;
|
---|
33 | use doc;
|
---|
34 |
|
---|
35 | sub BEGIN {
|
---|
36 | @ISA = ('BasPlug');
|
---|
37 | }
|
---|
38 |
|
---|
39 | use strict;
|
---|
40 |
|
---|
41 | sub new {
|
---|
42 | my ($class) = @_;
|
---|
43 | my $self = new BasPlug ("GMLPlug", @_);
|
---|
44 |
|
---|
45 | return bless $self, $class;
|
---|
46 | }
|
---|
47 |
|
---|
48 | sub get_default_process_exp {
|
---|
49 | my $self = shift (@_);
|
---|
50 |
|
---|
51 | return q^(?i)\.gml?$^;
|
---|
52 | }
|
---|
53 |
|
---|
54 | # return number of files processed, undef if can't process
|
---|
55 | # Note that $base_dir might be "" and that $file might
|
---|
56 | # include directories
|
---|
57 | sub read {
|
---|
58 | my $self = shift (@_);
|
---|
59 | my ($pluginfo, $base_dir, $file, $metadata, $processor, $maxdocs) = @_;
|
---|
60 |
|
---|
61 | my $filename = &util::filename_cat($base_dir, $file);
|
---|
62 | return 0 if $self->{'block_exp'} ne "" && $filename =~ /$self->{'block_exp'}/;
|
---|
63 | if ($filename !~ /$self->{'process_exp'}/ || !-f $filename) {
|
---|
64 | return undef;
|
---|
65 | }
|
---|
66 | $file =~ s/^[\/\\]+//; # $file often begins with / so we'll tidy it up
|
---|
67 |
|
---|
68 | print STDERR "GMLPlug: processing $file\n";
|
---|
69 |
|
---|
70 | my $parent_dir = $file;
|
---|
71 | $parent_dir =~ s/[^\\\/]*$//;
|
---|
72 | $parent_dir = &util::filename_cat ($base_dir, $parent_dir);
|
---|
73 |
|
---|
74 | if (!open (INFILE, $filename)) {
|
---|
75 | print STDERR "GMLPlug::read - couldn't read $filename\n";
|
---|
76 | return 0;
|
---|
77 | }
|
---|
78 |
|
---|
79 | undef $/;
|
---|
80 | my $gml = <INFILE>;
|
---|
81 | $/ = "\n";
|
---|
82 | close (INFILE);
|
---|
83 |
|
---|
84 | my @gml_sections = split("</gsdlsection>",$gml);
|
---|
85 | $gml = shift(@gml_sections);
|
---|
86 |
|
---|
87 | my $no_docs = 0;
|
---|
88 |
|
---|
89 | while (1) {
|
---|
90 | # create a new document
|
---|
91 | my $doc_obj = new doc ();
|
---|
92 | my $section = $doc_obj->get_top_section();
|
---|
93 |
|
---|
94 | # process the document
|
---|
95 | my $firstsection = 1;
|
---|
96 | while (1) {
|
---|
97 | my ($tags, $text) = ("", "");
|
---|
98 |
|
---|
99 | my @indenting_sections = split("<gsdlsection", $gml);
|
---|
100 | shift(@indenting_sections); # first entry is trivially empty
|
---|
101 |
|
---|
102 | foreach $gml (@indenting_sections) {
|
---|
103 |
|
---|
104 | if ($gml =~ /^\s*([^>]*)>(.*)$/so) {
|
---|
105 | $tags = $1 if defined $1;
|
---|
106 | $text = &GMLPlug::_unescape_text($2);
|
---|
107 |
|
---|
108 | } else {
|
---|
109 | print STDERR "GMLPlug::read - error in file $filename\n";
|
---|
110 | print STDERR "text: \"$gml\"\n";
|
---|
111 | last;
|
---|
112 | }
|
---|
113 |
|
---|
114 | # create the section (unless this is the first section)
|
---|
115 | if ($firstsection) {
|
---|
116 | $firstsection = 0;
|
---|
117 | # $tags =~ /gsdlsourcefilename\s*=\s*(?:\"([^\"]*)\")/o;
|
---|
118 | # $src_filename = $2 || $3;
|
---|
119 |
|
---|
120 | } else {
|
---|
121 |
|
---|
122 | $tags =~ s/gsdlnum\s*=\s*\"?(\d+)\"?//o;
|
---|
123 | if (defined $1) {
|
---|
124 | $section .= ".$1";
|
---|
125 | $doc_obj->create_named_section($section);
|
---|
126 | } else {
|
---|
127 | $section = $doc_obj->insert_section($doc_obj->get_end_child($section));
|
---|
128 | }
|
---|
129 | }
|
---|
130 |
|
---|
131 | # add the tags
|
---|
132 | while ((defined $tags) && ($tags =~ s/^\s*(\S+)=\"([^\"]*)\"//o)) {
|
---|
133 | $doc_obj->add_utf8_metadata($section, $1, &GMLPlug::_unescape_text($2))
|
---|
134 | if (defined $1 and defined $2);
|
---|
135 |
|
---|
136 | }
|
---|
137 |
|
---|
138 | # add the text
|
---|
139 | $doc_obj->add_utf8_text($section, $text)
|
---|
140 | if ((defined $text) && ($text ne ""));
|
---|
141 | }
|
---|
142 |
|
---|
143 | $gml = shift(@gml_sections); # get next bit of data
|
---|
144 | last unless defined $gml;
|
---|
145 | last if $section eq ""; # back to top level again (more than one document in gml file)
|
---|
146 | $section = $doc_obj->get_parent_section ($section);
|
---|
147 | } # while (1) section level
|
---|
148 |
|
---|
149 | # add the associated files
|
---|
150 | my $assoc_files = $doc_obj->get_metadata($doc_obj->get_top_section(), "gsdlassocfile");
|
---|
151 | my ($assoc_file_info, $afile);
|
---|
152 | foreach $assoc_file_info (@$assoc_files) {
|
---|
153 | my ($assoc_file, $mime_type, $dir) = split (":", $assoc_file_info);
|
---|
154 | if (defined $dir && $dir =~ /\w/) {
|
---|
155 | $afile = &util::filename_cat($dir, $assoc_file);
|
---|
156 | } else {
|
---|
157 | $afile = $assoc_file;
|
---|
158 | }
|
---|
159 | $doc_obj->associate_file(&util::filename_cat($parent_dir, $assoc_file),
|
---|
160 | $afile, $mime_type);
|
---|
161 | }
|
---|
162 | $doc_obj->delete_metadata($doc_obj->get_top_section(), "gsdlassocfile");
|
---|
163 |
|
---|
164 | # add metadata passed in from elsewhere
|
---|
165 | $self->extra_metadata ($doc_obj, $doc_obj->get_top_section(), $metadata);
|
---|
166 |
|
---|
167 | # do any automatic metadata extraction
|
---|
168 | $self->auto_extract_metadata ($doc_obj);
|
---|
169 |
|
---|
170 | # assume the document has an OID already
|
---|
171 |
|
---|
172 | # process the document
|
---|
173 | $processor->process($doc_obj, $file);
|
---|
174 |
|
---|
175 | $no_docs++;
|
---|
176 | last if ($maxdocs > -1 && $no_docs >= $maxdocs);
|
---|
177 | last unless defined $gml && $gml =~ /\w/;
|
---|
178 | } # while(1) document level
|
---|
179 |
|
---|
180 | return $no_docs; # no of docs processed
|
---|
181 | }
|
---|
182 |
|
---|
183 | sub _unescape_text {
|
---|
184 | my ($text) = @_;
|
---|
185 |
|
---|
186 | # special characters in the gml encoding
|
---|
187 | $text =~ s/</</g;
|
---|
188 | $text =~ s/>/>/g;
|
---|
189 | $text =~ s/"/\"/g;
|
---|
190 | $text =~ s/&/&/g; # this has to be last...
|
---|
191 |
|
---|
192 | return $text;
|
---|
193 | }
|
---|
194 |
|
---|
195 | 1;
|
---|