[22362] | 1 | ###########################################################################
|
---|
| 2 | #
|
---|
| 3 | # MediainfoOGVPlugin.pm -- Plugin for OGV multimedia files
|
---|
| 4 | #
|
---|
| 5 | # A component of the Greenstone digital library software from the New
|
---|
| 6 | # Zealand Digital Library Project at the University of Waikato, New
|
---|
| 7 | # Zealand.
|
---|
| 8 | #
|
---|
[24546] | 9 | # Copyright (C) 2010 Arnaud Yvan
|
---|
[22362] | 10 | #
|
---|
| 11 | # This program is free software; you can redistribute it and/or modify
|
---|
| 12 | # it under the terms of the GNU General Public License as published by
|
---|
| 13 | # the Free Software Foundation; either version 2 of the License, or
|
---|
| 14 | # (at your option) any later version.
|
---|
| 15 | #
|
---|
| 16 | # This program is distributed in the hope that it will be useful, but
|
---|
| 17 | # WITHOUT ANY WARRANTY; without even the implied warranty of
|
---|
| 18 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
|
---|
| 19 | # General Public License for more details.
|
---|
| 20 | #
|
---|
| 21 | # You should have received a copy of the GNU General Public License
|
---|
| 22 | # along with this program; if not, write to the Free Software
|
---|
| 23 | # Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
|
---|
| 24 | #
|
---|
| 25 | ###########################################################################
|
---|
| 26 |
|
---|
| 27 | # MediainfoOGVPlugin - a plugin for OGV multimedia files
|
---|
[24546] | 28 | # contributed to Greenstone by Arnaud Yvan
|
---|
[22362] | 29 |
|
---|
| 30 | # This is a simple Plugin for importing OGV multimedia files.
|
---|
| 31 | # Mediainfo will retrieve the metadata of the file. A fictional document will
|
---|
| 32 | # be created for every such file, and the file itself will be passed
|
---|
| 33 | # to Greenstone as the "associated file" of the document.
|
---|
| 34 |
|
---|
| 35 | # Here's an example where it is useful: I have a collection of
|
---|
| 36 | # ogg movie files with names like conference_20080402.ogv.
|
---|
| 37 | # I add this line to the collection configuration file:
|
---|
| 38 |
|
---|
| 39 | # plugin MediainfoOGVPlugin -process_exp "*.ogv" -assoc_field "movie"
|
---|
| 40 |
|
---|
| 41 | # A document is created for each movie, with the associated movie
|
---|
| 42 | # file's name in the "movie" metadata field.
|
---|
| 43 |
|
---|
[24546] | 44 | # The plugin also add some metadata :
|
---|
[22362] | 45 | # Duration (in seconds), filesize, title, artist, location, organization,
|
---|
| 46 | # date, contact, copyright.
|
---|
| 47 |
|
---|
| 48 |
|
---|
| 49 |
|
---|
| 50 | package MediainfoOGVPlugin;
|
---|
| 51 |
|
---|
[31492] | 52 | use BaseImporter;
|
---|
[22362] | 53 |
|
---|
| 54 | use strict;
|
---|
| 55 | no strict 'refs'; # allow filehandles to be variables and viceversa
|
---|
| 56 | no strict 'subs';
|
---|
| 57 |
|
---|
| 58 | sub BEGIN {
|
---|
[31492] | 59 | @MediainfoOGVPlugin::ISA = ('BaseImporter');
|
---|
[22362] | 60 | }
|
---|
| 61 |
|
---|
[22401] | 62 | my $arguments =
|
---|
| 63 | [ { 'name' => "process_exp",
|
---|
[31492] | 64 | 'desc' => "{BaseImporter.process_exp}",
|
---|
[22401] | 65 | 'type' => "regexp",
|
---|
| 66 | 'deft' => &get_default_process_exp(),
|
---|
| 67 | 'reqd' => "no" },
|
---|
| 68 | { 'name' => "assoc_field",
|
---|
[22362] | 69 | 'desc' => "{MediainfoOGVPlugin.assoc_field}",
|
---|
| 70 | 'type' => "string",
|
---|
[22401] | 71 | 'deft' => "Movie",
|
---|
[22362] | 72 | 'reqd' => "no" },
|
---|
[22401] | 73 | ];
|
---|
[22362] | 74 |
|
---|
| 75 | my $options = { 'name' => "MediainfoOGVPlugin",
|
---|
| 76 | 'desc' => "{MediainfoOGVPlugin.desc}",
|
---|
| 77 | 'abstract' => "no",
|
---|
| 78 | 'inherits' => "yes",
|
---|
| 79 | 'args' => $arguments };
|
---|
| 80 |
|
---|
| 81 |
|
---|
| 82 | sub new {
|
---|
| 83 | my ($class) = shift (@_);
|
---|
| 84 | my ($pluginlist,$inputargs,$hashArgOptLists) = @_;
|
---|
| 85 | push(@$pluginlist, $class);
|
---|
| 86 |
|
---|
| 87 | push(@{$hashArgOptLists->{"ArgList"}},@{$arguments});
|
---|
| 88 | push(@{$hashArgOptLists->{"OptList"}},$options);
|
---|
| 89 |
|
---|
[31492] | 90 | my $self = new BaseImporter($pluginlist, $inputargs, $hashArgOptLists);
|
---|
[22362] | 91 |
|
---|
[22401] | 92 | return bless $self, $class;
|
---|
| 93 | }
|
---|
[22362] | 94 |
|
---|
[22401] | 95 | sub init {
|
---|
| 96 | my $self = shift(@_);
|
---|
| 97 | my ($verbosity, $outhandle, $failhandle) = @_;
|
---|
| 98 |
|
---|
[31492] | 99 | $self->BaseImporter::init($verbosity, $outhandle, $failhandle);
|
---|
[22401] | 100 |
|
---|
[22362] | 101 | # Check that Mediainfo is installed and available on the path (except for Windows 95/98)
|
---|
| 102 | if (!($ENV{'GSDLOS'} eq "windows" && !Win32::IsWinNT())) {
|
---|
| 103 | my $result = `mediainfo 2>&1`;
|
---|
| 104 | if ($? == -1 || $? == 256) { # Linux and Windows return different values for "program not found"
|
---|
| 105 | $self->{'mediainfo_not_installed'} = 1;
|
---|
| 106 | }
|
---|
| 107 | }
|
---|
| 108 |
|
---|
| 109 | }
|
---|
[22401] | 110 | sub get_default_process_exp {
|
---|
| 111 | return q^(?i)\.ogv$^;
|
---|
| 112 | }
|
---|
[22362] | 113 |
|
---|
| 114 | # MediainfoOGVPlugin processing of doc_obj.
|
---|
| 115 |
|
---|
| 116 | sub process {
|
---|
| 117 | my $self = shift (@_);
|
---|
| 118 | my ($pluginfo, $base_dir, $file, $metadata, $doc_obj, $gli) = @_;
|
---|
| 119 |
|
---|
| 120 | #print STDERR "\n\n**** START of processing MediainfoOGVPlugin\n\n";
|
---|
| 121 |
|
---|
| 122 | my ($filename_full_path, $filename_no_path) = &util::get_full_filenames($base_dir, $file);
|
---|
| 123 | my $outhandle = $self->{'outhandle'};
|
---|
| 124 | my $verbosity = $self->{'verbosity'};
|
---|
| 125 |
|
---|
| 126 | # check the filename is okay - do we need this??
|
---|
| 127 | if ($filename_full_path eq "" || $filename_no_path eq "") {
|
---|
[22401] | 128 | print $outhandle "MediainfoOGVPlugin: couldn't process \"$filename_no_path\"\n";
|
---|
[22362] | 129 | return undef;
|
---|
| 130 | }
|
---|
| 131 |
|
---|
[22401] | 132 | # Add the file as an associated file ...
|
---|
| 133 | my $section = $doc_obj->get_top_section();
|
---|
| 134 | my $file_format = "ogv";
|
---|
| 135 | my $mime_type = "video/ogg";
|
---|
| 136 | my $assoc_field = $self->{'assoc_field'};
|
---|
| 137 |
|
---|
| 138 | # The assocfilename is the url-encoded version of the utf8 filename
|
---|
| 139 | my $assoc_file = $doc_obj->get_assocfile_from_sourcefile();
|
---|
| 140 |
|
---|
| 141 | $doc_obj->associate_file($filename_full_path, $assoc_file, $mime_type, $section);
|
---|
| 142 | $doc_obj->add_metadata ($section, "FileFormat", $file_format);
|
---|
| 143 | $doc_obj->add_metadata ($section, "dc.Format", $file_format);
|
---|
| 144 | $doc_obj->add_metadata ($section, "MimeType", $mime_type);
|
---|
| 145 | $doc_obj->add_utf8_metadata ($section, $assoc_field, $doc_obj->get_source()); # Source metadata is already in utf8
|
---|
| 146 |
|
---|
[24225] | 147 | # srclink_file is now deprecated because of the "_" in the metadataname. Use srclinkFile
|
---|
[22663] | 148 | $doc_obj->add_metadata ($section, "srclink_file", $doc_obj->get_sourcefile());
|
---|
[24225] | 149 | $doc_obj->add_metadata ($section, "srclinkFile", $doc_obj->get_sourcefile());
|
---|
[22401] | 150 | $doc_obj->add_metadata ($section, "srcicon", "_iconogg_");
|
---|
| 151 |
|
---|
| 152 | # we have no text - add dummy text and NoText metadata
|
---|
| 153 | $self->add_dummy_text($doc_obj, $section);
|
---|
| 154 |
|
---|
| 155 | if (defined $self->{'mediainfo_not_installed'}) {
|
---|
| 156 | # can't do any extracted
|
---|
| 157 | print STDERR "Mediainfo not installed, so can't extract metadata\n";
|
---|
| 158 | return 1;
|
---|
| 159 | }
|
---|
| 160 |
|
---|
| 161 |
|
---|
[22362] | 162 | # Retrieve the file metadata
|
---|
[22364] | 163 | my $command = "mediainfo --inform=\"General;%Duration/String2%|%Duration%|%FileSize/String2%|%Performer%|%Title%|%Recorded_Date%|%Recorded/Location%|%Producer%|%Copyright%|%LICENSE%|%Publisher%\" \"$filename_full_path\"";
|
---|
[22362] | 164 | print $outhandle "$command\n" if ($verbosity > 2);
|
---|
[22401] | 165 |
|
---|
[22362] | 166 | # execute the mediainfo command and store the output of execution in $videodata
|
---|
[22363] | 167 | # my $video_metadata = `$command`; # backticks operator way
|
---|
[22364] | 168 | # Another way to execute the command: better experience with this than with backticks operator above:
|
---|
| 169 | # use open() to read the output of executing the command (which is piped through to a handle using the | operator)
|
---|
[22363] | 170 | my $video_metadata;
|
---|
| 171 | if (open(VMIN,"$command|")) {
|
---|
| 172 | my $line;
|
---|
[22401] | 173 |
|
---|
[22364] | 174 | # we read a line of output at a time
|
---|
[22363] | 175 | while (defined ($line=<VMIN>)) {
|
---|
[22401] | 176 | #print STDERR "***** line = $line\n";
|
---|
| 177 |
|
---|
| 178 | $video_metadata .= $line;
|
---|
[22363] | 179 | }
|
---|
[22401] | 180 |
|
---|
[22363] | 181 | close(VMIN);
|
---|
[22364] | 182 | }
|
---|
[22401] | 183 |
|
---|
[22362] | 184 | print $outhandle "$video_metadata\n" if ($verbosity > 2);
|
---|
| 185 |
|
---|
| 186 | # There are 10 fields separated by |, split these into ordered, individual-named variables
|
---|
[22364] | 187 | my ($FormattedDuration,$Duration,$Filesize,$Artist,$Title,$Date,$Location,$Organization,$Copyright,$License,$Contact) = split(/\|/,$video_metadata);
|
---|
| 188 | $Duration = int($Duration/1000);
|
---|
| 189 | $Artist =~ s/\\047/\'/g; # Perl's way of doing: $artist = `echo $artist | sed \"s/\\047/'/g\"`;
|
---|
| 190 | $Title =~ s/\\047/\'/g; # $title = `echo $title | sed \"s/\\047/'/g\"`;
|
---|
[22362] | 191 |
|
---|
[22364] | 192 | print $outhandle "RESULT = $FormattedDuration\n$Duration\n$Filesize\n$Artist\n$Title\n$Date\n$Location\n$Organization\n$Copyright\n$License\n$Contact\n" if ($verbosity > 2);
|
---|
[22362] | 193 |
|
---|
| 194 |
|
---|
[22364] | 195 | $doc_obj->add_metadata ($section, "FormattedDuration", $FormattedDuration);
|
---|
| 196 | $doc_obj->add_metadata ($section, "Duration", $Duration);
|
---|
[22362] | 197 |
|
---|
| 198 |
|
---|
[22364] | 199 | $doc_obj->add_metadata($section,"FileSize",$Filesize);
|
---|
| 200 | $doc_obj->add_utf8_metadata($section,"dc.Creator",$Artist);
|
---|
| 201 | $doc_obj->add_utf8_metadata($section,"Artist",$Artist);
|
---|
| 202 | $doc_obj->add_utf8_metadata($section,"dc.Title",$Title);
|
---|
| 203 | $doc_obj->add_utf8_metadata($section,"Title",$Title);
|
---|
| 204 | $doc_obj->add_utf8_metadata($section,"dc.Date",$Date);
|
---|
| 205 | $doc_obj->add_utf8_metadata($section,"Date",$Date);
|
---|
| 206 | $doc_obj->add_utf8_metadata($section,"dc.Coverage",$Location);
|
---|
| 207 | $doc_obj->add_utf8_metadata($section,"Location",$Location);
|
---|
| 208 | $doc_obj->add_utf8_metadata($section,"dc.Publisher",$Organization);
|
---|
| 209 | $doc_obj->add_utf8_metadata($section,"Organization",$Organization);
|
---|
| 210 | $doc_obj->add_utf8_metadata($section,"dc.Rights",$Copyright);
|
---|
| 211 | $doc_obj->add_utf8_metadata($section,"Copyright",$Copyright);
|
---|
| 212 | $doc_obj->add_utf8_metadata($section,"dc.accessRights",$License);
|
---|
| 213 | $doc_obj->add_utf8_metadata($section,"License",$License);
|
---|
| 214 | $doc_obj->add_utf8_metadata($section,"dc.Contributor",$Contact);
|
---|
| 215 | $doc_obj->add_utf8_metadata($section,"Contact",$Contact);
|
---|
[22362] | 216 |
|
---|
| 217 |
|
---|
[22401] | 218 | #print STDERR "\n\n**** END of MediainfoOGVPlugin\n\n";
|
---|
[22362] | 219 |
|
---|
| 220 | return 1;
|
---|
| 221 | }
|
---|
| 222 |
|
---|
| 223 |
|
---|
| 224 | 1;
|
---|
| 225 |
|
---|
| 226 |
|
---|
| 227 |
|
---|
| 228 |
|
---|
| 229 |
|
---|
| 230 |
|
---|
| 231 |
|
---|
| 232 |
|
---|
| 233 |
|
---|
| 234 |
|
---|
| 235 |
|
---|