| 1 | ###########################################################################
|
|---|
| 2 | #
|
|---|
| 3 | # OAIMetadataXMLPlugin.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) 2010 DL Consulting Ltd
|
|---|
| 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 | # OAIMetadataXMLPlugin is a child of MetadataXMLPlugin
|
|---|
| 27 | # It processes the metadata.xml file just like MetadataXMLPlugin.
|
|---|
| 28 | # Additionally, it uses the "dc.Identifier" field and extracts OAI metadata from the specified OAI server (-oai_server_http_path)
|
|---|
| 29 |
|
|---|
| 30 | package OAIMetadataXMLPlugin;
|
|---|
| 31 |
|
|---|
| 32 | use strict;
|
|---|
| 33 | no strict 'refs';
|
|---|
| 34 |
|
|---|
| 35 | use MetadataXMLPlugin;
|
|---|
| 36 |
|
|---|
| 37 | sub BEGIN {
|
|---|
| 38 | @OAIMetadataXMLPlugin::ISA = ('MetadataXMLPlugin');
|
|---|
| 39 | unshift (@INC, "$ENV{'GSDLHOME'}/perllib/cpan");
|
|---|
| 40 | }
|
|---|
| 41 |
|
|---|
| 42 | my $arguments = [
|
|---|
| 43 | { 'name' => "oai_server_http_path",
|
|---|
| 44 | 'desc' => "{OAIMetadataXMLPlugin.oai_server_http_path}",
|
|---|
| 45 | 'type' => "string",
|
|---|
| 46 | 'deft' => "" },
|
|---|
| 47 |
|
|---|
| 48 | { 'name' => "metadata_prefix",
|
|---|
| 49 | 'desc' => "{OAIMetadataXMLPlugin.metadata_prefix}",
|
|---|
| 50 | 'type' => "string",
|
|---|
| 51 | 'deft' => "oai_dc" },
|
|---|
| 52 |
|
|---|
| 53 | # If koha_mode flag is specified, the plugin will try to generate the oaiextracted.koharecordlink metadata
|
|---|
| 54 | # This metadata contains the link back to Koha document
|
|---|
| 55 | { 'name' => "koha_mode",
|
|---|
| 56 | 'desc' => "{OAIMetadataXMLPlugin.koha_mode}",
|
|---|
| 57 | 'type' => "flag",
|
|---|
| 58 | 'reqd' => "no" },
|
|---|
| 59 | ];
|
|---|
| 60 |
|
|---|
| 61 | my $options = { 'name' => "OAIMetadataXMLPlugin",
|
|---|
| 62 | 'desc' => "{OAIMetadataXMLPlugin.desc}",
|
|---|
| 63 | 'abstract' => "no",
|
|---|
| 64 | 'inherits' => "yes",
|
|---|
| 65 | 'args' => $arguments };
|
|---|
| 66 |
|
|---|
| 67 |
|
|---|
| 68 | sub new
|
|---|
| 69 | {
|
|---|
| 70 | my ($class) = shift (@_);
|
|---|
| 71 | my ($pluginlist,$inputargs,$hashArgOptLists) = @_;
|
|---|
| 72 | push(@$pluginlist, $class);
|
|---|
| 73 |
|
|---|
| 74 | push(@{$hashArgOptLists->{"ArgList"}},@{$arguments});
|
|---|
| 75 | push(@{$hashArgOptLists->{"OptList"}},$options);
|
|---|
| 76 |
|
|---|
| 77 | my $self = new MetadataXMLPlugin($pluginlist, $inputargs, $hashArgOptLists);
|
|---|
| 78 |
|
|---|
| 79 | return bless $self, $class;
|
|---|
| 80 | }
|
|---|
| 81 |
|
|---|
| 82 |
|
|---|
| 83 | sub metadata_read
|
|---|
| 84 | {
|
|---|
| 85 | my $self = shift (@_);
|
|---|
| 86 | my ($pluginfo, $base_dir, $file, $block_hash, $extrametakeys, $extrametadata,$extrametafile, $processor, $maxdocs, $gli) = @_;
|
|---|
| 87 |
|
|---|
| 88 | # Read in the normal metadata.xml file
|
|---|
| 89 | $self->SUPER::metadata_read(@_);
|
|---|
| 90 |
|
|---|
| 91 | my $outhandle = $self->{'outhandle'};
|
|---|
| 92 |
|
|---|
| 93 | #======================================================================#
|
|---|
| 94 | # Checks to make sure the OAI-PMH server is connectable [START]
|
|---|
| 95 | #======================================================================#
|
|---|
| 96 | print $outhandle "OAIMetadataXMLPlugin: Checking OAI server (" . $self->{"oai_server_http_path"} . ") connection\n" if ($self->{'verbosity'})> 1;
|
|---|
| 97 |
|
|---|
| 98 | # Checks to make sure LWP (5.64) is available, it should always be available if you have Perl installed
|
|---|
| 99 | # However if you are using the Greenstone's cut-down version of Perl, this LWP module will not be included
|
|---|
| 100 | eval { require LWP };
|
|---|
| 101 | if ($@)
|
|---|
| 102 | {
|
|---|
| 103 | print STDERR "Error: Failed to load Perl module LWP: $@\n";
|
|---|
| 104 | return;
|
|---|
| 105 | }
|
|---|
| 106 |
|
|---|
| 107 | # Create the LWP module
|
|---|
| 108 | my $browser = LWP::UserAgent->new;
|
|---|
| 109 | my $response = $browser->get($self->{"oai_server_http_path"});
|
|---|
| 110 |
|
|---|
| 111 | # Do not go further if the OAI server is not accessible
|
|---|
| 112 | if (!$response->is_success)
|
|---|
| 113 | {
|
|---|
| 114 | print $outhandle "OAIMetadataXMLPlugin: Error! OAI server (" . $self->{"oai_server_http_path"} . ") unavailable\n";
|
|---|
| 115 | return;
|
|---|
| 116 | }
|
|---|
| 117 | #======================================================================#
|
|---|
| 118 | # Checks to make sure the OAI-PMH server is connectable [END]
|
|---|
| 119 | #======================================================================#
|
|---|
| 120 |
|
|---|
| 121 | #======================================================================#
|
|---|
| 122 | # Process each fileset [START]
|
|---|
| 123 | #======================================================================#
|
|---|
| 124 | foreach my $one_file (@{$extrametakeys})
|
|---|
| 125 | {
|
|---|
| 126 | # Don't harvest file sets that don't have dc.Identifier set, "dc.Identifier" is usde as the key between Greenstone and OAI Server!
|
|---|
| 127 | next if (!defined($extrametadata->{$one_file}->{"dc.Identifier"}->[0]) || $extrametadata->{$one_file}->{"dc.Identifier"}->[0] eq "");
|
|---|
| 128 |
|
|---|
| 129 | #======================================================================#
|
|---|
| 130 | # Only try to harvest file set with dc.Identifier specified. [START]
|
|---|
| 131 | #======================================================================#
|
|---|
| 132 | # The dc.Identifier has to be the same as the OAI record identifier
|
|---|
| 133 | my $oai_identifier = $extrametadata->{$one_file}->{"dc.Identifier"}->[0];
|
|---|
| 134 |
|
|---|
| 135 | # Now, let's get the OAI metadata
|
|---|
| 136 | my $request = $self->{"oai_server_http_path"} . "?verb=GetRecord&identifier=" . $oai_identifier. "&metadataPrefix=" . $self->{"metadata_prefix"};
|
|---|
| 137 | print $outhandle "OAIMetadataXMLPlugin: OAI Harvesting Request (" . $request . ")\n";
|
|---|
| 138 | $response = undef;
|
|---|
| 139 | $response = $browser->get($request);
|
|---|
| 140 | die "OAIMetadataXMLPlugin: This should never be happening - \"get\" should always be successful unless the OAI server was temporary down (some kind of race condition)\n" unless ($response->is_success);
|
|---|
| 141 | my $reponse_content = $response->content();
|
|---|
| 142 |
|
|---|
| 143 | # Check to make sure there is no error in the OAI response
|
|---|
| 144 | if ($reponse_content =~ /\<error\scode\=[\"\']([^\"\']+)[\"\']>([^\<]*)\<\/error\>/)
|
|---|
| 145 | {
|
|---|
| 146 | print $outhandle "OAIMetadataXMLPlugin: Failed to retrive OAI record (" . $oai_identifier . "). ErrorCode: [$1] ErrorMessage: [$2], skip.\n";
|
|---|
| 147 | next;
|
|---|
| 148 | }
|
|---|
| 149 | print $outhandle "OAIMetadataXMLPlugin: OAI record (" . $oai_identifier . ") found.\n";
|
|---|
| 150 |
|
|---|
| 151 | # Get the oai metadata (We will need to extend this code to support future metadataPrefix)
|
|---|
| 152 | my $oai_content = undef;
|
|---|
| 153 |
|
|---|
| 154 | # Special Note for KOHA OAI Server: there is an error in the KOHA's OAI-PMH server (it is still under development at the time when I am writting this)
|
|---|
| 155 | # The metadata set should be oai_dc:dc tag, but they incorrectly output the tag as oaidc:dc (which doesn't match with the metadataPrefix)
|
|---|
| 156 | if ($self->{"metadata_prefix"} eq "oai_dc" && $reponse_content =~ /\<oai\_?dc:dc[^\>]+\>(.*?)\<\/oai\_?dc\:dc\>/s)
|
|---|
| 157 | {
|
|---|
| 158 | $oai_content = $1;
|
|---|
| 159 | }
|
|---|
| 160 | else
|
|---|
| 161 | {
|
|---|
| 162 | my $reg_match = "\<" . $self->{"metadata_prefix"} . "\:" . $self->{"metadata_prefix"} . "[^\>]+\>(.*?)\<\/" . $self->{"metadata_prefix"} . "\:" . $self->{"metadata_prefix"} . "\>";
|
|---|
| 163 | if ($reponse_content =~ /$reg_match/s)
|
|---|
| 164 | {
|
|---|
| 165 | $oai_content = $1;
|
|---|
| 166 | }
|
|---|
| 167 | else
|
|---|
| 168 | {
|
|---|
| 169 | print $outhandle "OAIMetadataXMLPlugin: Failed to match " . $self->{"metadata_prefix"} . ":" . $self->{"metadata_prefix"} . " metadata set, skip\n " . $reponse_content . "\n";
|
|---|
| 170 | next;
|
|---|
| 171 | }
|
|---|
| 172 | }
|
|---|
| 173 |
|
|---|
| 174 | # Get each metadata field and value
|
|---|
| 175 | while ($oai_content =~ /\<([^\>]+)\>([^\<]+)\<\/[^\>]+\>/g)
|
|---|
| 176 | {
|
|---|
| 177 | my $field_name = "oaiextracted." . lc($1);
|
|---|
| 178 | my $value = $2;
|
|---|
| 179 |
|
|---|
| 180 | # Special hack for Koha data from Nitesh
|
|---|
| 181 | # Some of their data contain " \" as the value... that is pretty wrong.
|
|---|
| 182 | # If the value is empty, ignore it.
|
|---|
| 183 | if ($value =~ /^[^\w]*$/)
|
|---|
| 184 | {
|
|---|
| 185 | print STDERR "Ignore value:[" . $value . "]\n";
|
|---|
| 186 | next;
|
|---|
| 187 | }
|
|---|
| 188 |
|
|---|
| 189 | # Special case for identifier
|
|---|
| 190 | if ($self->{"koha_mode"} == 1 && $1 eq "identifier" && $2 =~ /https?\:\/\//)
|
|---|
| 191 | {
|
|---|
| 192 | $field_name = "oaiextracted.koharecordlink";
|
|---|
| 193 |
|
|---|
| 194 | # Koha OAI server is not up-to-date... so it was still pointing to the old interface
|
|---|
| 195 | # This might need change over once they update the Koha OAI server
|
|---|
| 196 | $value =~ s/\/opac\/opac\-detail\.pl\?bib\=/\/catalogue\/detail\.pl\?biblionumber\=/;
|
|---|
| 197 | }
|
|---|
| 198 | $extrametadata->{$one_file}->{$field_name} = [] if (!defined($extrametadata->{$one_file}->{$field_name}));
|
|---|
| 199 | push(@{$extrametadata->{$one_file}->{$field_name}}, $value);
|
|---|
| 200 | }
|
|---|
| 201 | #======================================================================#
|
|---|
| 202 | # Only try to harvest file set with dc.Identifier specified. [END]
|
|---|
| 203 | #======================================================================#
|
|---|
| 204 | }
|
|---|
| 205 | #======================================================================#
|
|---|
| 206 | # Process each fileset [END]
|
|---|
| 207 | #======================================================================#
|
|---|
| 208 | }
|
|---|
| 209 |
|
|---|
| 210 | 1;
|
|---|
