[2810] | 1 | ###########################################################################
|
---|
| 2 | #
|
---|
[15872] | 3 | # GAPlugin.pm
|
---|
[2810] | 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) 2001 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 |
|
---|
[3540] | 26 | # Processes GreenstoneArchive XML documents. Note that this plugin does no
|
---|
| 27 | # syntax checking (though the XML::Parser module tests for
|
---|
| 28 | # well-formedness). It's assumed that the GreenstoneArchive files conform
|
---|
| 29 | # to their DTD.
|
---|
[2810] | 30 |
|
---|
[15872] | 31 | package GAPlugin;
|
---|
[2810] | 32 |
|
---|
[15872] | 33 | use ReadXMLFile;
|
---|
[2810] | 34 |
|
---|
[10254] | 35 | use strict;
|
---|
| 36 | no strict 'refs'; # allow filehandles to be variables and viceversa
|
---|
| 37 |
|
---|
[2810] | 38 | sub BEGIN {
|
---|
[15872] | 39 | @GAPlugin::ISA = ('ReadXMLFile');
|
---|
[2810] | 40 | }
|
---|
| 41 |
|
---|
[7900] | 42 |
|
---|
| 43 | sub get_default_process_exp {
|
---|
| 44 | my $self = shift (@_);
|
---|
| 45 |
|
---|
| 46 | return q^(?i)doc\.xml$^;
|
---|
| 47 | }
|
---|
| 48 |
|
---|
[14004] | 49 | my $arguments =
|
---|
| 50 | [ { 'name' => "process_exp",
|
---|
[15872] | 51 | 'desc' => "{BasePlugin.process_exp}",
|
---|
[14004] | 52 | 'type' => "regexp",
|
---|
| 53 | 'deft' => &get_default_process_exp(),
|
---|
| 54 | 'reqd' => "no" } ];
|
---|
[10254] | 55 |
|
---|
[15872] | 56 | my $options = { 'name' => "GAPlugin",
|
---|
| 57 | 'desc' => "{GAPlugin.desc}",
|
---|
[6408] | 58 | 'abstract' => "no",
|
---|
[14004] | 59 | 'inherits' => "yes",
|
---|
| 60 | 'args' => $arguments };
|
---|
[3540] | 61 |
|
---|
[2810] | 62 | sub new {
|
---|
[10218] | 63 | my ($class) = shift (@_);
|
---|
| 64 | my ($pluginlist,$inputargs,$hashArgOptLists) = @_;
|
---|
| 65 | push(@$pluginlist, $class);
|
---|
[2810] | 66 |
|
---|
[15872] | 67 | push(@{$hashArgOptLists->{"ArgList"}},@{$arguments});
|
---|
| 68 | push(@{$hashArgOptLists->{"OptList"}},$options);
|
---|
[3540] | 69 |
|
---|
[15872] | 70 | my $self = new ReadXMLFile($pluginlist, $inputargs, $hashArgOptLists);
|
---|
[10218] | 71 |
|
---|
[2810] | 72 | $self->{'section'} = "";
|
---|
| 73 | $self->{'section_level'} = 0;
|
---|
| 74 | $self->{'metadata_name'} = "";
|
---|
| 75 | $self->{'metadata_value'} = "";
|
---|
| 76 | $self->{'content'} = "";
|
---|
[12844] | 77 |
|
---|
| 78 | # # Currently used to store information for previous values controls. In
|
---|
| 79 | # # the next contract I'll move to using information directly from Lucene.
|
---|
| 80 | # $self->{'sqlfh'} = 0;
|
---|
[7900] | 81 |
|
---|
[2810] | 82 | return bless $self, $class;
|
---|
| 83 | }
|
---|
| 84 |
|
---|
| 85 | sub xml_start_document {
|
---|
| 86 | }
|
---|
| 87 |
|
---|
| 88 | sub xml_end_document {
|
---|
| 89 | }
|
---|
| 90 |
|
---|
[13147] | 91 | sub get_doctype {
|
---|
| 92 | my $self = shift(@_);
|
---|
| 93 |
|
---|
| 94 | return "(Greenstone)?Archive";
|
---|
| 95 | }
|
---|
| 96 |
|
---|
| 97 |
|
---|
[2810] | 98 | sub xml_doctype {
|
---|
| 99 | my $self = shift(@_);
|
---|
[7900] | 100 |
|
---|
[2810] | 101 | my ($expat, $name, $sysid, $pubid, $internal) = @_;
|
---|
| 102 |
|
---|
[2925] | 103 | # allow the short-lived and badly named "GreenstoneArchive" files to be processed
|
---|
| 104 | # as well as the "Archive" files which should now be created by import.pl
|
---|
| 105 | die "" if ($name !~ /^(Greenstone)?Archive$/);
|
---|
| 106 |
|
---|
[2810] | 107 | my $outhandle = $self->{'outhandle'};
|
---|
[15872] | 108 | print $outhandle "GAPlugin: processing $self->{'file'}\n" if $self->{'verbosity'} > 1;
|
---|
| 109 | print STDERR "<Processing n='$self->{'file'}' p='GAPlugin'>\n" if $self->{'gli'};
|
---|
[9468] | 110 |
|
---|
[2810] | 111 | }
|
---|
| 112 |
|
---|
| 113 |
|
---|
| 114 | sub xml_start_tag {
|
---|
| 115 | my $self = shift(@_);
|
---|
| 116 | my ($expat, $element) = @_;
|
---|
| 117 |
|
---|
| 118 | $self->{'element'} = $element;
|
---|
| 119 | if ($element eq "Section") {
|
---|
| 120 | if ($self->{'section_level'} == 0) {
|
---|
| 121 | $self->open_document();
|
---|
[7900] | 122 | } else {
|
---|
[2810] | 123 | my $doc_obj = $self->{'doc_obj'};
|
---|
| 124 | $self->{'section'} =
|
---|
| 125 | $doc_obj->insert_section($doc_obj->get_end_child($self->{'section'}));
|
---|
| 126 | }
|
---|
[7900] | 127 |
|
---|
[2810] | 128 | $self->{'section_level'} ++;
|
---|
| 129 | }
|
---|
| 130 | elsif ($element eq "Metadata") {
|
---|
| 131 | $self->{'metadata_name'} = $_{'name'};
|
---|
| 132 | }
|
---|
| 133 | }
|
---|
| 134 |
|
---|
| 135 | sub xml_end_tag {
|
---|
| 136 | my $self = shift(@_);
|
---|
| 137 | my ($expat, $element) = @_;
|
---|
| 138 |
|
---|
| 139 | if ($element eq "Section") {
|
---|
| 140 | $self->{'section_level'} --;
|
---|
| 141 | $self->{'section'} = $self->{'doc_obj'}->get_parent_section ($self->{'section'});
|
---|
| 142 | $self->close_document() if $self->{'section_level'} == 0;
|
---|
| 143 | }
|
---|
| 144 | elsif ($element eq "Metadata") {
|
---|
[7900] | 145 | $self->{'doc_obj'}->add_utf8_metadata($self->{'section'}, $self->{'metadata_name'},$self->{'metadata_value'});
|
---|
[12844] | 146 | # Ensure this value is added to the allvalues database in gseditor.
|
---|
| 147 | # Note that the database constraints prevent multiple occurances of the
|
---|
| 148 | # same key-value pair.
|
---|
| 149 | # We write these out to a file, so they can all be commited in one
|
---|
| 150 | # transaction
|
---|
| 151 | #if (!$self->{'sqlfh'})
|
---|
| 152 | # {
|
---|
| 153 | # my $sql_file = $ENV{'GSDLHOME'} . "/collect/lld/tmp/gseditor.sql";
|
---|
| 154 | # # If the file doesn't already exist, open it and begin a transaction
|
---|
| 155 | # my $sql_fh;
|
---|
| 156 | # if (!-e $sql_file)
|
---|
| 157 | # {
|
---|
| 158 | # open($sql_fh, ">" . $sql_file);
|
---|
| 159 | # print $sql_fh "BEGIN TRANSACTION;\n";
|
---|
| 160 | # }
|
---|
| 161 | # else
|
---|
| 162 | # {
|
---|
| 163 | # open($sql_fh, ">>" . $sql_file);
|
---|
| 164 | # }
|
---|
| 165 | # print STDERR "Opened SQL log\n";
|
---|
| 166 | # $self->{'sqlfh'} = $sql_fh;
|
---|
| 167 | # }
|
---|
| 168 |
|
---|
| 169 | #my $mvalue = $self->{'metadata_value'};
|
---|
| 170 | #$mvalue =~ s/\'/\'\'/g;
|
---|
| 171 | #$mvalue =~ s/_claimantsep_/ \& /g;
|
---|
| 172 |
|
---|
| 173 | #my $fh = $self->{'sqlfh'};
|
---|
| 174 | #if ($fh)
|
---|
| 175 | # {
|
---|
| 176 | # print $fh "INSERT INTO allvalues (mkey, mvalue) VALUES ('" . $self->{'metadata_name'} . "', '" . $mvalue . "');\n";
|
---|
| 177 | # }
|
---|
| 178 |
|
---|
| 179 | # Clean Up
|
---|
[2810] | 180 | $self->{'metadata_name'} = "";
|
---|
| 181 | $self->{'metadata_value'} = "";
|
---|
| 182 | }
|
---|
| 183 | elsif ($element eq "Content" && $self->{'content'} ne "") {
|
---|
| 184 | $self->{'doc_obj'}->add_utf8_text($self->{'section'}, $self->{'content'});
|
---|
| 185 | $self->{'content'} = "";
|
---|
| 186 | }
|
---|
| 187 | $self->{'element'} = "";
|
---|
| 188 | }
|
---|
| 189 |
|
---|
| 190 | sub xml_text {
|
---|
| 191 | my $self = shift(@_);
|
---|
| 192 | my ($expat) = @_;
|
---|
| 193 |
|
---|
| 194 | if ($self->{'element'} eq "Metadata") {
|
---|
| 195 | $self->{'metadata_value'} .= $_;
|
---|
| 196 | }
|
---|
| 197 | elsif ($self->{'element'} eq "Content") {
|
---|
| 198 | $self->{'content'} .= $_;
|
---|
| 199 | }
|
---|
| 200 | }
|
---|
| 201 |
|
---|
| 202 | sub open_document {
|
---|
| 203 | my $self = shift(@_);
|
---|
| 204 |
|
---|
| 205 | # create a new document
|
---|
| 206 | $self->{'doc_obj'} = new doc ();
|
---|
| 207 | $self->{'section'} = "";
|
---|
| 208 | }
|
---|
| 209 |
|
---|
| 210 | sub close_document {
|
---|
| 211 | my $self = shift(@_);
|
---|
[12844] | 212 |
|
---|
[2810] | 213 | # add the associated files
|
---|
| 214 | my $assoc_files =
|
---|
| 215 | $self->{'doc_obj'}->get_metadata($self->{'doc_obj'}->get_top_section(), "gsdlassocfile");
|
---|
| 216 |
|
---|
[3629] | 217 | # for when "assocfilepath" isn't the same directory that doc.xml is in...
|
---|
| 218 | my $assoc_filepath_list= $self->{'doc_obj'}->get_metadata($self->{'doc_obj'}->get_top_section(), "assocfilepath");
|
---|
[2810] | 219 |
|
---|
[3629] | 220 | my $assoc_filepath=shift (@$assoc_filepath_list);
|
---|
[12844] | 221 |
|
---|
| 222 | #rint STDERR "Filename is: " . $self->{'filename'} . "\n";
|
---|
| 223 | #rint STDERR "Initially my assoc_filepath is: $assoc_filepath\n";
|
---|
| 224 | #rint STDERR "Custom archive dir is: " . $self->{'base_dir'} . "\n";
|
---|
| 225 | # Correct the assoc filepath if one is defined
|
---|
| 226 | if (defined ($assoc_filepath))
|
---|
| 227 | {
|
---|
| 228 | # Check whether the assoc_filepath already includes the base dir
|
---|
| 229 | if (index($assoc_filepath, $self->{'base_dir'}) == -1)
|
---|
| 230 | {
|
---|
| 231 | # And if not, append it so as to make this absolute
|
---|
| 232 | $assoc_filepath = &util::filename_cat($self->{'base_dir'}, $assoc_filepath);
|
---|
| 233 | }
|
---|
| 234 | }
|
---|
| 235 | else
|
---|
| 236 | {
|
---|
[3629] | 237 | $assoc_filepath = $self->{'filename'};
|
---|
| 238 | $assoc_filepath =~ s/[^\\\/]*$//;
|
---|
[12844] | 239 | }
|
---|
| 240 | #rint STDERR "Goned and made it absolute: $assoc_filepath\n";
|
---|
[3629] | 241 |
|
---|
[2810] | 242 | foreach my $assoc_file_info (@$assoc_files) {
|
---|
| 243 | my ($assoc_file, $mime_type, $dir) = split (":", $assoc_file_info);
|
---|
[12844] | 244 | #rint STDERR "assoc_file: $assoc_file\n";
|
---|
| 245 | #rint STDERR "mime_type: $mime_type\n";
|
---|
| 246 | #rint STDERR "dir: $dir\n";
|
---|
[3629] | 247 | my $real_dir = &util::filename_cat($assoc_filepath, $assoc_file),
|
---|
[2810] | 248 | my $assoc_dir = (defined $dir && $dir ne "")
|
---|
| 249 | ? &util::filename_cat($dir, $assoc_file) : $assoc_file;
|
---|
| 250 | $self->{'doc_obj'}->associate_file($real_dir, $assoc_dir, $mime_type);
|
---|
[12844] | 251 | #rint STDERR "According to me the real assoc_filepath is: $real_dir\n";
|
---|
[2810] | 252 | }
|
---|
| 253 | $self->{'doc_obj'}->delete_metadata($self->{'doc_obj'}->get_top_section(), "gsdlassocfile");
|
---|
| 254 |
|
---|
| 255 | # process the document
|
---|
| 256 | $self->{'processor'}->process($self->{'doc_obj'}, $self->{'file'});
|
---|
| 257 | }
|
---|
| 258 |
|
---|
| 259 |
|
---|
| 260 | 1;
|
---|
[3540] | 261 |
|
---|
| 262 |
|
---|