[22439] | 1 | ###########################################################################
|
---|
| 2 | #
|
---|
| 3 | # jSongMinerExtractor - helper plugin that identifies audio through
|
---|
| 4 | # external web services based on either an audio
|
---|
| 5 | # computed fingerprint or ID3 title and album
|
---|
| 6 | #
|
---|
| 7 | # A component of the Greenstone digital library software
|
---|
| 8 | # from the New Zealand Digital Library Project at the
|
---|
| 9 | # University of Waikato, New Zealand.
|
---|
| 10 | #
|
---|
| 11 | # Copyright (C) 2010 New Zealand Digital Library Project
|
---|
| 12 | #
|
---|
| 13 | # This program is free software; you can redistribute it and/or modify
|
---|
| 14 | # it under the terms of the GNU General Public License as published by
|
---|
| 15 | # the Free Software Foundation; either version 2 of the License, or
|
---|
| 16 | # (at your option) any later version.
|
---|
| 17 | #
|
---|
| 18 | # This program is distributed in the hope that it will be useful,
|
---|
| 19 | # but WITHOUT ANY WARRANTY; without even the implied warranty of
|
---|
| 20 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
---|
| 21 | # GNU General Public License for more details.
|
---|
| 22 | #
|
---|
| 23 | # You should have received a copy of the GNU General Public License
|
---|
| 24 | # along with this program; if not, write to the Free Software
|
---|
| 25 | # Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
|
---|
| 26 | #
|
---|
| 27 | ###########################################################################
|
---|
| 28 | package jSongMinerExtractor;
|
---|
| 29 |
|
---|
| 30 | use BaseMediaConverter;
|
---|
| 31 |
|
---|
| 32 | use Cwd;
|
---|
[22468] | 33 | use URI::Escape;
|
---|
[22439] | 34 |
|
---|
| 35 | use strict;
|
---|
| 36 | no strict 'refs'; # allow filehandles to be variables and viceversa
|
---|
| 37 |
|
---|
| 38 |
|
---|
| 39 | BEGIN {
|
---|
| 40 | @jSongMinerExtractor::ISA = ('BaseMediaConverter');
|
---|
| 41 | }
|
---|
| 42 |
|
---|
| 43 |
|
---|
| 44 | my $arguments = [
|
---|
| 45 | { 'name' => "track_identification",
|
---|
| 46 | 'desc' => "{jSongMinerExtractor.track_identification}",
|
---|
| 47 | 'type' => "enum",
|
---|
| 48 | 'list' => [{'name' => "Fingerprint then ID3 tags", 'desc' => "{jSongMinerExtractor.fingerprint_first}"},
|
---|
| 49 | {'name' => "ID3 tags only", 'desc' => "{jSongMinerExtractor.only_ids}"},
|
---|
| 50 | {'name' => "Disabled", 'desc' => "{jSongMinerExtractor.off}"} ],
|
---|
| 51 | 'deft' => 'Fingerprint then ID3 tags',
|
---|
| 52 | 'reqd' => "no" }
|
---|
| 53 | ];
|
---|
| 54 |
|
---|
| 55 | my $options = { 'name' => "jSongMinerExtractor",
|
---|
| 56 | 'desc' => "{jSongMinerExtractor.desc}",
|
---|
| 57 | 'abstract' => "yes",
|
---|
| 58 | 'inherits' => "yes",
|
---|
| 59 | 'args' => $arguments };
|
---|
| 60 |
|
---|
| 61 | sub new {
|
---|
| 62 | my ($class) = shift (@_);
|
---|
| 63 | my ($pluginlist,$inputargs,$hashArgOptLists) = @_;
|
---|
| 64 | push(@$pluginlist, $class);
|
---|
| 65 |
|
---|
| 66 | push(@{$hashArgOptLists->{"ArgList"}},@{$arguments});
|
---|
| 67 | push(@{$hashArgOptLists->{"OptList"}},$options);
|
---|
| 68 |
|
---|
| 69 | my $self = new BaseMediaConverter($pluginlist, $inputargs, $hashArgOptLists, 1);
|
---|
| 70 |
|
---|
| 71 | # Set controlling variables
|
---|
| 72 | my $gsdl_home = $ENV{'GSDLHOME'};
|
---|
| 73 | my $music_ir_home = $ENV{'GEXT_MUSICIR'};
|
---|
| 74 |
|
---|
[22441] | 75 | $self->{'jmir_directory'} = &util::filename_cat($music_ir_home,"lib","java"); # Set the directory holding the jMIR .jar files
|
---|
| 76 |
|
---|
[22439] | 77 | return bless $self, $class;
|
---|
| 78 | }
|
---|
| 79 |
|
---|
[22468] | 80 | sub urlEncode{
|
---|
| 81 | # ARG 1: $to_encode is the string to URL encode
|
---|
| 82 | my ($to_encode) = @_;
|
---|
| 83 | return uri_escape($to_encode);
|
---|
| 84 | }
|
---|
| 85 |
|
---|
| 86 | # URL Decode the given string
|
---|
| 87 | sub urlDecode{
|
---|
| 88 | # ARG 1: $to_decode is the string to URL decode
|
---|
| 89 | my ($to_decode) = @_;
|
---|
[24432] | 90 | my $decoded= uri_unescape($to_decode);
|
---|
| 91 |
|
---|
| 92 | return $decoded;
|
---|
[22468] | 93 | }
|
---|
| 94 |
|
---|
[28549] | 95 | sub map_id3v1_genre_num
|
---|
| 96 | {
|
---|
| 97 | my ($genre_num) = @_;
|
---|
| 98 |
|
---|
| 99 | $genre_num =~ s/(/\(/g;
|
---|
| 100 | $genre_num =~ s/)/\)/g;
|
---|
| 101 | $genre_num =~ s/^[\(\[\{](.*)[\)\]\}]$/$1/;
|
---|
| 102 |
|
---|
| 103 | if (($genre_num =~ m/^\d+$/) && ($genre_num >= 148)) {
|
---|
| 104 | return "Unknown";
|
---|
| 105 | }
|
---|
| 106 |
|
---|
| 107 | my $lookup_id3v1_genre = {
|
---|
| 108 | "0" => "Blues",
|
---|
| 109 | "1" => "Classic Rock",
|
---|
| 110 | "2" => "Country",
|
---|
| 111 | "3" => "Dance",
|
---|
| 112 | "4" => "Disco",
|
---|
| 113 | "5" => "Funk",
|
---|
| 114 | "6" => "Grunge",
|
---|
| 115 | "7" => "Hip-Hop",
|
---|
| 116 | "8" => "Jazz",
|
---|
| 117 | "9" => "Metal",
|
---|
| 118 | "10" => "New Age",
|
---|
| 119 | "11" => "Oldies",
|
---|
| 120 | "12" => "Other",
|
---|
| 121 | "13" => "Pop",
|
---|
| 122 | "14" => "R&B",
|
---|
| 123 | "15" => "Rap",
|
---|
| 124 | "16" => "Reggae",
|
---|
| 125 | "17" => "Rock",
|
---|
| 126 | "18" => "Techno",
|
---|
| 127 | "19" => "Industrial",
|
---|
| 128 | "20" => "Alternative",
|
---|
| 129 | "21" => "Ska",
|
---|
| 130 | "22" => "Death Metal",
|
---|
| 131 | "23" => "Pranks",
|
---|
| 132 | "24" => "Soundtrack",
|
---|
| 133 | "25" => "Euro-Techno",
|
---|
| 134 | "26" => "Ambient",
|
---|
| 135 | "27" => "Trip-Hop",
|
---|
| 136 | "28" => "Vocal",
|
---|
| 137 | "29" => "Jazz+Funk",
|
---|
| 138 | "30" => "Fusion",
|
---|
| 139 | "31" => "Trance",
|
---|
| 140 | "32" => "Classical",
|
---|
| 141 | "33" => "Instrumental",
|
---|
| 142 | "34" => "Acid",
|
---|
| 143 | "35" => "House",
|
---|
| 144 | "36" => "Game",
|
---|
| 145 | "37" => "Sound Clip",
|
---|
| 146 | "38" => "Gospel",
|
---|
| 147 | "39" => "Noise",
|
---|
| 148 | "40" => "Alternative Rock",
|
---|
| 149 | "41" => "Bass",
|
---|
| 150 | "42" => "Soul",
|
---|
| 151 | "43" => "Punk",
|
---|
| 152 | "44" => "Space",
|
---|
| 153 | "45" => "Meditative",
|
---|
| 154 | "46" => "Instrumental Pop",
|
---|
| 155 | "47" => "Instrumental Rock",
|
---|
| 156 | "48" => "Ethnic",
|
---|
| 157 | "49" => "Gothic",
|
---|
| 158 | "50" => "Darkwave",
|
---|
| 159 | "51" => "Techno-Industrial",
|
---|
| 160 | "52" => "Electronic",
|
---|
| 161 | "53" => "Pop-Folk",
|
---|
| 162 | "54" => "Eurodance",
|
---|
| 163 | "55" => "Dream",
|
---|
| 164 | "56" => "Southern Rock",
|
---|
| 165 | "57" => "Comedy",
|
---|
| 166 | "58" => "Cult",
|
---|
| 167 | "59" => "Gangsta",
|
---|
| 168 | "60" => "Top 40",
|
---|
| 169 | "61" => "Christian Rap",
|
---|
| 170 | "62" => "Pop/Funk",
|
---|
| 171 | "63" => "Jungle",
|
---|
| 172 | "64" => "Native US",
|
---|
| 173 | "65" => "Cabaret",
|
---|
| 174 | "66" => "New Wave",
|
---|
| 175 | "67" => "Psychadelic",
|
---|
| 176 | "68" => "Rave",
|
---|
| 177 | "69" => "Showtunes",
|
---|
| 178 | "70" => "Trailer",
|
---|
| 179 | "71" => "Lo-Fi",
|
---|
| 180 | "72" => "Tribal",
|
---|
| 181 | "73" => "Acid Punk",
|
---|
| 182 | "74" => "Acid Jazz",
|
---|
| 183 | "75" => "Polka",
|
---|
| 184 | "76" => "Retro",
|
---|
| 185 | "77" => "Musical",
|
---|
| 186 | "78" => "Rock & Roll",
|
---|
| 187 | "79" => "Hard Rock",
|
---|
| 188 | "80" => "Folk",
|
---|
| 189 | "81" => "Folk-Rock",
|
---|
| 190 | "82" => "National Folk",
|
---|
| 191 | "83" => "Swing",
|
---|
| 192 | "84" => "Fast Fusion",
|
---|
| 193 | "85" => "Bebob",
|
---|
| 194 | "86" => "Latin",
|
---|
| 195 | "87" => "Revival",
|
---|
| 196 | "88" => "Celtic",
|
---|
| 197 | "89" => "Bluegrass",
|
---|
| 198 | "90" => "Avantgarde",
|
---|
| 199 | "91" => "Gothic Rock",
|
---|
| 200 | "92" => "Progressive Rock",
|
---|
| 201 | "93" => "Psychedelic Rock",
|
---|
| 202 | "94" => "Symphonic Rock",
|
---|
| 203 | "95" => "Slow Rock",
|
---|
| 204 | "96" => "Big Band",
|
---|
| 205 | "97" => "Chorus",
|
---|
| 206 | "98" => "Easy Listening",
|
---|
| 207 | "99" => "Acoustic",
|
---|
| 208 | "100" => "Humour",
|
---|
| 209 | "101" => "Speech",
|
---|
| 210 | "102" => "Chanson",
|
---|
| 211 | "103" => "Opera",
|
---|
| 212 | "104" => "Chamber Music",
|
---|
| 213 | "105" => "Sonata",
|
---|
| 214 | "106" => "Symphony",
|
---|
| 215 | "107" => "Booty Bass",
|
---|
| 216 | "108" => "Primus",
|
---|
| 217 | "109" => "Porn Groove",
|
---|
| 218 | "110" => "Satire",
|
---|
| 219 | "111" => "Slow Jam",
|
---|
| 220 | "112" => "Club",
|
---|
| 221 | "113" => "Tango",
|
---|
| 222 | "114" => "Samba",
|
---|
| 223 | "115" => "Folklore",
|
---|
| 224 | "116" => "Ballad",
|
---|
| 225 | "117" => "Power Ballad",
|
---|
| 226 | "118" => "Rhythmic Soul",
|
---|
| 227 | "119" => "Freestyle",
|
---|
| 228 | "120" => "Duet",
|
---|
| 229 | "121" => "Punk Rock",
|
---|
| 230 | "122" => "Drum Solo",
|
---|
| 231 | "123" => "Acapella",
|
---|
| 232 | "124" => "Euro-House",
|
---|
| 233 | "125" => "Dance Hall",
|
---|
| 234 | "126" => "Goa",
|
---|
| 235 | "127" => "Drum & Bass",
|
---|
| 236 | "128" => "Club - House",
|
---|
| 237 | "129" => "Hardcore",
|
---|
| 238 | "130" => "Terror",
|
---|
| 239 | "131" => "Indie",
|
---|
| 240 | "132" => "BritPop",
|
---|
| 241 | "133" => "Negerpunk",
|
---|
| 242 | "134" => "Polsk Punk",
|
---|
| 243 | "135" => "Beat",
|
---|
| 244 | "136" => "Christian Gangsta Rap",
|
---|
| 245 | "137" => "Heavy Metal",
|
---|
| 246 | "138" => "Black Metal",
|
---|
| 247 | "139" => "Crossover",
|
---|
| 248 | "140" => "Contemporary Christian",
|
---|
| 249 | "141" => "Christian Rock",
|
---|
| 250 | "142" => "Merengue",
|
---|
| 251 | "143" => "Salsa",
|
---|
| 252 | "144" => "Thrash Metal",
|
---|
| 253 | "145" => "Anime",
|
---|
| 254 | "146" => "JPop",
|
---|
| 255 | "147" => "Synthpop"
|
---|
| 256 | };
|
---|
| 257 |
|
---|
| 258 | my $mapped_genre;
|
---|
| 259 |
|
---|
| 260 | if (defined $lookup_id3v1_genre->{$genre_num}) {
|
---|
[30433] | 261 | # print STDERR "*** changing $genre_num -> ", $lookup_id3v1_genre->{$genre_num}, "\n";
|
---|
[28549] | 262 | $mapped_genre = $lookup_id3v1_genre->{$genre_num};
|
---|
| 263 | }
|
---|
| 264 | else {
|
---|
| 265 | $mapped_genre = $genre_num;
|
---|
| 266 | }
|
---|
| 267 |
|
---|
| 268 | return $mapped_genre;
|
---|
| 269 | }
|
---|
| 270 |
|
---|
| 271 | sub check_for_existing_id3_genre
|
---|
| 272 | {
|
---|
| 273 | my $self = shift @_;
|
---|
| 274 | my ($doc_obj) = @_;
|
---|
| 275 |
|
---|
| 276 | my $top_section=$doc_obj->get_top_section();
|
---|
| 277 |
|
---|
| 278 | # Look for ex.ID3.Genre as well, as special case
|
---|
| 279 | my $genre_md_list = $doc_obj->get_metadata($top_section,"ex.ID3.Genre");
|
---|
| 280 | my @new_genre_md_list = ();
|
---|
| 281 |
|
---|
| 282 | foreach my $gv (@$genre_md_list) {
|
---|
[30433] | 283 | # print STDERR "*** got match on ex.ID3.Genre -> '$gv'\n";
|
---|
[28549] | 284 | my $new_gv = map_id3v1_genre_num($gv);
|
---|
| 285 |
|
---|
| 286 | push(@new_genre_md_list,$new_gv);
|
---|
| 287 | }
|
---|
| 288 |
|
---|
| 289 | $doc_obj->delete_metadata($top_section,"ex.ID3.Genre");
|
---|
| 290 | foreach my $gv (@new_genre_md_list) {
|
---|
| 291 | $doc_obj->add_utf8_metadata($top_section,"ex.ID3.Genre",$gv);
|
---|
| 292 | }
|
---|
| 293 | }
|
---|
| 294 |
|
---|
| 295 |
|
---|
[22468] | 296 | sub parse_txt_metadata
|
---|
| 297 | {
|
---|
| 298 | my $self = shift @_;
|
---|
| 299 | my ($doc_obj,$target_txt_file_path) = @_;
|
---|
| 300 |
|
---|
| 301 | if (open(MIN,"<$target_txt_file_path")) {
|
---|
| 302 |
|
---|
| 303 | my ($md_name, $md_value);
|
---|
| 304 |
|
---|
| 305 | while (defined($md_name=<MIN>) && defined($md_value=<MIN>)) {
|
---|
| 306 |
|
---|
| 307 | chomp $md_name;
|
---|
| 308 | chomp $md_value;
|
---|
| 309 |
|
---|
| 310 | my $top_section=$doc_obj->get_top_section();
|
---|
| 311 |
|
---|
| 312 | $md_name =~ s/\+//g;
|
---|
[24432] | 313 | $md_value =~ s/\+/ /g;
|
---|
[22468] | 314 |
|
---|
| 315 | $md_name = urlDecode($md_name);
|
---|
| 316 | $md_value = urlDecode($md_value);
|
---|
| 317 |
|
---|
[24432] | 318 | # $md_name =~ s/\s+/ /sg;
|
---|
| 319 | $md_name =~ s/\(.*?\)$//s; # can stretch over multiple lines
|
---|
[22468] | 320 | $md_name =~ s/Last\.FM/LastFM/g;
|
---|
| 321 | $md_name =~ s/:/^/g;
|
---|
| 322 | $md_name =~ s/(API)?\^/./;
|
---|
| 323 |
|
---|
[28549] | 324 | # print STDERR "*** md_name = '$md_name'\n";
|
---|
| 325 |
|
---|
| 326 | if ($md_name =~ m/genre$/i) {
|
---|
[30433] | 327 | # print STDERR "*** got match on $md_name -> $md_value\n";
|
---|
[28549] | 328 | $md_value = map_id3v1_genre_num($md_value);
|
---|
| 329 | }
|
---|
| 330 |
|
---|
[24432] | 331 | $doc_obj->add_utf8_metadata($top_section,$md_name,$md_value);
|
---|
[22468] | 332 | }
|
---|
| 333 |
|
---|
| 334 | close(MIN);
|
---|
[28549] | 335 |
|
---|
| 336 | $self->check_for_existing_id3_genre($doc_obj);
|
---|
| 337 |
|
---|
[22468] | 338 | }
|
---|
| 339 | else {
|
---|
| 340 | print STDERR "Error: Failed to open $target_txt_file_path\n";
|
---|
| 341 | print STDERR " !$\n";
|
---|
| 342 | }
|
---|
| 343 | }
|
---|
| 344 |
|
---|
| 345 |
|
---|
[22439] | 346 | sub retrieve_metadata
|
---|
| 347 | {
|
---|
| 348 | my $self = shift(@_);
|
---|
| 349 | my ($source_file_path,$id3_title,$id3_artist,$convert_options) = @_;
|
---|
| 350 |
|
---|
| 351 | $convert_options = "" if (!defined $convert_options);
|
---|
| 352 |
|
---|
| 353 | my $outhandle = $self->{'outhandle'};
|
---|
| 354 | my $verbosity = $self->{'verbosity'};
|
---|
| 355 |
|
---|
| 356 | my $source_file_no_path = &File::Basename::basename($source_file_path);
|
---|
| 357 |
|
---|
| 358 | $self->init_cache_for_file($source_file_path);
|
---|
| 359 |
|
---|
[22468] | 360 | my $target_txt_file_path;
|
---|
| 361 | my $target_acexml_file_path;
|
---|
| 362 |
|
---|
[22439] | 363 | if ($self->{'enable_cache'}) {
|
---|
| 364 | my $cached_dir = $self->{'cached_dir'};
|
---|
| 365 | my $file_root = $self->{'cached_file_root'};
|
---|
| 366 |
|
---|
[22468] | 367 | my $target_txt_file = "${file_root}_metadata.txt";
|
---|
| 368 | my $target_acexml_file = "${file_root}.xml";
|
---|
[22439] | 369 |
|
---|
[22468] | 370 | $target_txt_file_path = &util::filename_cat($cached_dir,$target_txt_file);
|
---|
| 371 | $target_acexml_file_path = &util::filename_cat($cached_dir,$target_acexml_file);
|
---|
[22439] | 372 | }
|
---|
| 373 | else {
|
---|
[22468] | 374 | $target_txt_file_path = &util::get_tmp_filename("_metadata.txt");
|
---|
| 375 | $target_acexml_file_path = &util::get_tmp_filename(".xml");
|
---|
[22439] | 376 | }
|
---|
| 377 |
|
---|
[22441] | 378 | my $jmir_directory = $self->{'jmir_directory'};
|
---|
[22439] | 379 |
|
---|
[22441] | 380 | my $store_cwd = cwd();
|
---|
| 381 |
|
---|
| 382 | if (!-d $jmir_directory) {
|
---|
| 383 | print STDERR "Error: Unable able to find directory '$jmir_directory'\n";
|
---|
| 384 | print STDERR " Cannot run jAudio\n";
|
---|
| 385 | }
|
---|
| 386 | elsif (chdir($jmir_directory)) {
|
---|
| 387 |
|
---|
[26276] | 388 | my $source_file_path_os = $source_file_path;
|
---|
| 389 | if ($^O eq "cygwin") {
|
---|
| 390 | $source_file_path_os = `cygpath -w "$source_file_path"`;
|
---|
| 391 | $source_file_path_os =~ s/\s+$//;
|
---|
| 392 | }
|
---|
| 393 | my $target_txt_file_path_os = $target_txt_file_path;
|
---|
| 394 | if ($^O eq "cygwin") {
|
---|
| 395 | $target_txt_file_path_os = `cygpath -w "$target_txt_file_path"`;
|
---|
| 396 | $target_txt_file_path_os =~ s/\s+$//;
|
---|
| 397 | }
|
---|
| 398 | my $target_acexml_file_path_os = $target_acexml_file_path;
|
---|
| 399 | if ($^O eq "cygwin") {
|
---|
| 400 | $target_acexml_file_path_os = `cygpath -w "$target_acexml_file_path"`;
|
---|
| 401 | $target_acexml_file_path_os =~ s/\s+$//;
|
---|
| 402 | }
|
---|
[22441] | 403 |
|
---|
[29656] | 404 |
|
---|
| 405 | my $jsongminer_cmd = "java -Xmx1024M -jar \"" . $jmir_directory . "/jSongMiner.jar\" $convert_options";
|
---|
[26276] | 406 | $jsongminer_cmd .= " -title \"$id3_title\"" if defined $id3_title;
|
---|
| 407 | $jsongminer_cmd .= " -artist \"$id3_artist\"" if defined $id3_artist;
|
---|
| 408 | $jsongminer_cmd .= " -audio \"$source_file_path_os\"";
|
---|
| 409 | $jsongminer_cmd .= " -savetxtfile \"$target_txt_file_path_os\"";
|
---|
| 410 | $jsongminer_cmd .= " -saveacexmlfile \"$target_acexml_file_path_os\"";
|
---|
[22439] | 411 |
|
---|
[22441] | 412 | if ($verbosity>2) {
|
---|
| 413 | print $outhandle "jSongMinerExtractor: Running ...\n";
|
---|
| 414 | print $outhandle "jSongMinerExtractor: $jsongminer_cmd\n";
|
---|
| 415 | }
|
---|
[22439] | 416 |
|
---|
[22441] | 417 | my $print_info = { 'message_prefix' => "jSongMiner",
|
---|
[22468] | 418 | 'message' => "jSongMinerExtractor: Retrieving audio metadata for $source_file_no_path" };
|
---|
[29656] | 419 |
|
---|
[22441] | 420 | my ($regenerated,$result,$had_error)
|
---|
[22468] | 421 | = $self->autorun_general_cmd($jsongminer_cmd,$source_file_path,$target_txt_file_path,$print_info);
|
---|
[22441] | 422 |
|
---|
| 423 | if ($verbosity>2) {
|
---|
| 424 | print $outhandle "jSongMinerExtractor: ...done\n";
|
---|
| 425 | }
|
---|
| 426 | }
|
---|
| 427 | else {
|
---|
| 428 | print STDERR "Error: failed to change directory to '$jmir_directory'\n";
|
---|
| 429 | print STDERR " Cannot run jAudio\n";
|
---|
| 430 | }
|
---|
| 431 |
|
---|
| 432 | chdir($store_cwd);
|
---|
| 433 |
|
---|
[22468] | 434 | return ($target_acexml_file_path,$target_txt_file_path);
|
---|
[22439] | 435 | }
|
---|
| 436 |
|
---|
| 437 |
|
---|
| 438 |
|
---|
| 439 |
|
---|
| 440 | 1;
|
---|