| 1 | ###########################################################################
|
|---|
| 2 | #
|
|---|
| 3 | # ImagePlug.pm -- simple text plugin
|
|---|
| 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 | package ImagePlug;
|
|---|
| 27 |
|
|---|
| 28 | use BasPlug;
|
|---|
| 29 |
|
|---|
| 30 | use strict;
|
|---|
| 31 | no strict 'refs'; # allow filehandles to be variables and viceversa
|
|---|
| 32 |
|
|---|
| 33 | sub BEGIN {
|
|---|
| 34 | @ImagePlug::ISA = ('BasPlug');
|
|---|
| 35 | }
|
|---|
| 36 |
|
|---|
| 37 | my $arguments =
|
|---|
| 38 | [ { 'name' => "process_exp",
|
|---|
| 39 | 'desc' => "{BasPlug.process_exp}",
|
|---|
| 40 | 'type' => "regexp",
|
|---|
| 41 | 'deft' => &get_default_process_exp(),
|
|---|
| 42 | 'reqd' => "no" },
|
|---|
| 43 | { 'name' => "noscaleup",
|
|---|
| 44 | 'desc' => "{ImagePlug.noscaleup}",
|
|---|
| 45 | 'type' => "flag",
|
|---|
| 46 | 'reqd' => "no" },
|
|---|
| 47 | { 'name' => "thumbnailsize",
|
|---|
| 48 | 'desc' => "{ImagePlug.thumbnailsize}",
|
|---|
| 49 | 'type' => "int",
|
|---|
| 50 | 'deft' => "100",
|
|---|
| 51 | 'range' => "1,",
|
|---|
| 52 | 'reqd' => "no" },
|
|---|
| 53 | { 'name' => "thumbnailtype",
|
|---|
| 54 | 'desc' => "{ImagePlug.thumbnailtype}",
|
|---|
| 55 | 'type' => "string",
|
|---|
| 56 | 'deft' => "gif",
|
|---|
| 57 | 'reqd' => "no" },
|
|---|
| 58 | { 'name' => "screenviewsize",
|
|---|
| 59 | 'desc' => "{ImagePlug.screenviewsize}",
|
|---|
| 60 | 'type' => "int",
|
|---|
| 61 | 'deft' => "0",
|
|---|
| 62 | 'range' => "1,",
|
|---|
| 63 | 'reqd' => "no" },
|
|---|
| 64 | { 'name' => "screenviewtype",
|
|---|
| 65 | 'desc' => "{ImagePlug.screenviewtype}",
|
|---|
| 66 | 'type' => "string",
|
|---|
| 67 | 'deft' => "jpg",
|
|---|
| 68 | 'reqd' => "no" },
|
|---|
| 69 | { 'name' => "converttotype",
|
|---|
| 70 | 'desc' => "{ImagePlug.converttotype}",
|
|---|
| 71 | 'type' => "string",
|
|---|
| 72 | 'deft' => "",
|
|---|
| 73 | 'reqd' => "no" },
|
|---|
| 74 | { 'name' => "minimumsize",
|
|---|
| 75 | 'desc' => "{ImagePlug.minimumsize}",
|
|---|
| 76 | 'type' => "int",
|
|---|
| 77 | 'deft' => "100",
|
|---|
| 78 | 'range' => "1,",
|
|---|
| 79 | 'reqd' => "no" } ];
|
|---|
| 80 |
|
|---|
| 81 | my $options = { 'name' => "ImagePlug",
|
|---|
| 82 | 'desc' => "{ImagePlug.desc}",
|
|---|
| 83 | 'abstract' => "no",
|
|---|
| 84 | 'inherits' => "yes",
|
|---|
| 85 | 'args' => $arguments };
|
|---|
| 86 |
|
|---|
| 87 |
|
|---|
| 88 |
|
|---|
| 89 | sub new {
|
|---|
| 90 | my ($class) = shift (@_);
|
|---|
| 91 | my ($pluginlist,$inputargs,$hashArgOptLists) = @_;
|
|---|
| 92 | push(@$pluginlist, $class);
|
|---|
| 93 |
|
|---|
| 94 | if(defined $arguments){ push(@{$hashArgOptLists->{"ArgList"}},@{$arguments});}
|
|---|
| 95 | if(defined $options) { push(@{$hashArgOptLists->{"OptList"}},$options)};
|
|---|
| 96 |
|
|---|
| 97 | my $self = new BasPlug($pluginlist, $inputargs, $hashArgOptLists);
|
|---|
| 98 |
|
|---|
| 99 | # Check that ImageMagick is installed and available on the path (except for Windows 95/98)
|
|---|
| 100 | if (!($ENV{'GSDLOS'} eq "windows" && !Win32::IsWinNT())) {
|
|---|
| 101 | my $result = `identify 2>&1`;
|
|---|
| 102 | if ($? == -1 || $? == 256) { # Linux and Windows return different values for "program not found"
|
|---|
| 103 | $self->{'imagemagick_not_installed'} = 1;
|
|---|
| 104 | }
|
|---|
| 105 | }
|
|---|
| 106 |
|
|---|
| 107 |
|
|---|
| 108 | return bless $self, $class;
|
|---|
| 109 | }
|
|---|
| 110 |
|
|---|
| 111 | sub get_default_process_exp {
|
|---|
| 112 | my $self = shift (@_);
|
|---|
| 113 |
|
|---|
| 114 | return q^(?i)(\.jpe?g|\.gif|\.png|\.bmp|\.xbm|\.tif?f)$^;
|
|---|
| 115 | }
|
|---|
| 116 |
|
|---|
| 117 | # this makes no sense for images
|
|---|
| 118 | sub block_cover_image
|
|---|
| 119 | {
|
|---|
| 120 | my $self =shift (@_);
|
|---|
| 121 | my ($filename) = @_;
|
|---|
| 122 |
|
|---|
| 123 | return;
|
|---|
| 124 | }
|
|---|
| 125 | # Create the thumbnail and screenview images, and discover the Image's
|
|---|
| 126 | # size, width, and height using the convert utility.
|
|---|
| 127 |
|
|---|
| 128 | sub run_convert {
|
|---|
| 129 | my $self = shift (@_);
|
|---|
| 130 | my $filename = shift (@_); # filename with full path
|
|---|
| 131 | my $file = shift (@_); # filename without path
|
|---|
| 132 | my $doc_obj = shift (@_);
|
|---|
| 133 | my $section = $doc_obj->get_top_section();
|
|---|
| 134 |
|
|---|
| 135 | my $verbosity = $self->{'verbosity'};
|
|---|
| 136 | my $outhandle = $self->{'outhandle'};
|
|---|
| 137 |
|
|---|
| 138 | # check the filename is okay
|
|---|
| 139 | return 0 if ($file eq "" || $filename eq "");
|
|---|
| 140 |
|
|---|
| 141 | # Code now extended to quote filenames in 'convert' commnads
|
|---|
| 142 | # Allows spaces in filenames, but note needs spaces to be escaped in URL as well
|
|---|
| 143 | # if ($filename =~ m/ /) {
|
|---|
| 144 | # print $outhandle "ImagePlug: \"$filename\" contains a space. choking.\n";
|
|---|
| 145 | # return undef;
|
|---|
| 146 | # }
|
|---|
| 147 |
|
|---|
| 148 | my $minimumsize = $self->{'minimumsize'};
|
|---|
| 149 | if (defined $minimumsize && (-s $filename < $minimumsize)) {
|
|---|
| 150 | print $outhandle "ImagePlug: \"$filename\" too small, skipping\n"
|
|---|
| 151 | if ($verbosity > 1);
|
|---|
| 152 | }
|
|---|
| 153 |
|
|---|
| 154 |
|
|---|
| 155 | # Convert the image to a new type (if required).
|
|---|
| 156 | my $converttotype = $self->{'converttotype'};
|
|---|
| 157 | my $originalfilename = ""; # only set if we do a conversion
|
|---|
| 158 | my $type = "unknown";
|
|---|
| 159 |
|
|---|
| 160 |
|
|---|
| 161 | if ($converttotype ne "" && $filename =~ m/$converttotype$/) {
|
|---|
| 162 |
|
|---|
| 163 | $originalfilename = $filename;
|
|---|
| 164 | $filename = &util::get_tmp_filename() . ".$converttotype";
|
|---|
| 165 | $self->{'tmp_filename'} = $filename;
|
|---|
| 166 |
|
|---|
| 167 | my $command = "convert -interlace plane -verbose \"$originalfilename\" \"$filename\"";
|
|---|
| 168 | print $outhandle "$command\n" if ($verbosity > 2);
|
|---|
| 169 | my $result = '';
|
|---|
| 170 | $result = `$command`;
|
|---|
| 171 | print $outhandle "RESULT = $result\n" if ($verbosity > 2);
|
|---|
| 172 |
|
|---|
| 173 | $type = $converttotype;
|
|---|
| 174 | }
|
|---|
| 175 |
|
|---|
| 176 |
|
|---|
| 177 | # Add the image metadata
|
|---|
| 178 | my $url = $file;
|
|---|
| 179 | $url =~ s/ /%20/g;
|
|---|
| 180 |
|
|---|
| 181 | $doc_obj->add_metadata ($section, "Image", $url);
|
|---|
| 182 |
|
|---|
| 183 | # Also want to set filename as 'Source' metadata to be
|
|---|
| 184 | # consistent with other plugins
|
|---|
| 185 | $doc_obj->add_metadata ($section, "Source", $url);
|
|---|
| 186 |
|
|---|
| 187 | my ($image_type, $image_width, $image_height, $image_size)
|
|---|
| 188 | = &identify($filename, $outhandle, $verbosity);
|
|---|
| 189 |
|
|---|
| 190 | if ($image_type ne " ") {
|
|---|
| 191 | $type = $image_type;
|
|---|
| 192 | }
|
|---|
| 193 |
|
|---|
| 194 | $doc_obj->add_metadata ($section, "FileFormat", $type);
|
|---|
| 195 | $doc_obj->add_metadata ($section, "FileSize", $image_size);
|
|---|
| 196 |
|
|---|
| 197 | $doc_obj->add_metadata ($section, "ImageType", $image_type);
|
|---|
| 198 | $doc_obj->add_metadata ($section, "ImageWidth", $image_width);
|
|---|
| 199 | $doc_obj->add_metadata ($section, "ImageHeight", $image_height);
|
|---|
| 200 | $doc_obj->add_metadata ($section, "ImageSize", $image_size);
|
|---|
| 201 |
|
|---|
| 202 | $doc_obj->add_metadata ($section, "srclink",
|
|---|
| 203 | "<a href=\"_httpprefix_/collect/[collection]/index/assoc/[assocfilepath]/[Image]\">");
|
|---|
| 204 | $doc_obj->add_metadata ($section, "/srclink", "</a>");
|
|---|
| 205 |
|
|---|
| 206 | $doc_obj->add_metadata ($section, "srcicon", "<img src=\"_httpprefix_/collect/[collection]/index/assoc/[assocfilepath]/[Image]\" width=100>");
|
|---|
| 207 |
|
|---|
| 208 |
|
|---|
| 209 | # Add the image as an associated file
|
|---|
| 210 | $doc_obj->associate_file($filename,$file,"image/$type",$section);
|
|---|
| 211 |
|
|---|
| 212 |
|
|---|
| 213 | # Make the thumbnail image
|
|---|
| 214 | my $thumbnailsize = $self->{'thumbnailsize'} || 100;
|
|---|
| 215 | my $thumbnailtype = $self->{'thumbnailtype'} || 'gif';
|
|---|
| 216 |
|
|---|
| 217 | my $thumbnailfile = &util::get_tmp_filename() . ".$thumbnailtype";
|
|---|
| 218 | $self->{'tmp_filename2'} = $thumbnailfile;
|
|---|
| 219 |
|
|---|
| 220 | # Generate the thumbnail with convert
|
|---|
| 221 | my $command = "convert -interlace plane -verbose -geometry $thumbnailsize"
|
|---|
| 222 | . "x$thumbnailsize \"$filename\" \"$thumbnailfile\"";
|
|---|
| 223 | print $outhandle "THUMBNAIL: $command\n" if ($verbosity > 2);
|
|---|
| 224 | my $result = '';
|
|---|
| 225 | $result = `$command 2>&1` ;
|
|---|
| 226 | print $outhandle "THUMB RESULT: $result\n" if ($verbosity > 2);
|
|---|
| 227 |
|
|---|
| 228 | # Add the thumbnail as an associated file ...
|
|---|
| 229 | if (-e "$thumbnailfile") {
|
|---|
| 230 | $doc_obj->associate_file("$thumbnailfile", "thumbnail.$thumbnailtype",
|
|---|
| 231 | "image/$thumbnailtype",$section);
|
|---|
| 232 | $doc_obj->add_metadata ($section, "ThumbType", $thumbnailtype);
|
|---|
| 233 | $doc_obj->add_metadata ($section, "Thumb", "thumbnail.$thumbnailtype");
|
|---|
| 234 |
|
|---|
| 235 | $doc_obj->add_metadata ($section, "thumbicon", "<img src=\"_httpprefix_/collect/[collection]/index/assoc/[assocfilepath]/[Thumb]\" width=[ThumbWidth] height=[ThumbHeight]>");
|
|---|
| 236 | }
|
|---|
| 237 |
|
|---|
| 238 | # Extract Thumnail metadata from convert output
|
|---|
| 239 | if ($result =~ m/[0-9]+x[0-9]+=>([0-9]+)x([0-9]+)/) {
|
|---|
| 240 | $doc_obj->add_metadata ($section, "ThumbWidth", $1);
|
|---|
| 241 | $doc_obj->add_metadata ($section, "ThumbHeight", $2);
|
|---|
| 242 | }
|
|---|
| 243 |
|
|---|
| 244 | # Make a screen-sized version of the picture if requested
|
|---|
| 245 | if ($self->{'screenviewsize'}) {
|
|---|
| 246 |
|
|---|
| 247 | # To do: if the actual image smaller than the screenview size,
|
|---|
| 248 | # we should use the original !
|
|---|
| 249 |
|
|---|
| 250 | my $screenviewsize = $self->{'screenviewsize'};
|
|---|
| 251 | my $screenviewtype = $self->{'screenviewtype'} || 'jpeg';
|
|---|
| 252 | my $screenviewfilename = &util::get_tmp_filename() . ".$screenviewtype";
|
|---|
| 253 | $self->{'tmp_filename3'} = $screenviewfilename;
|
|---|
| 254 |
|
|---|
| 255 | # make the screenview image
|
|---|
| 256 | my $command = "convert -interlace plane -verbose -geometry $screenviewsize"
|
|---|
| 257 | . "x$screenviewsize \"$filename\" \"$screenviewfilename\"";
|
|---|
| 258 | print $outhandle "$command\n" if ($verbosity > 2);
|
|---|
| 259 | my $result = "";
|
|---|
| 260 | $result = `$command 2>&1` ;
|
|---|
| 261 | print $outhandle "$result\n" if ($verbosity > 3);
|
|---|
| 262 |
|
|---|
| 263 | # get screenview dimensions, size and type
|
|---|
| 264 | if ($result =~ m/[0-9]+x[0-9]+=>([0-9]+)x([0-9]+)/) {
|
|---|
| 265 | $doc_obj->add_metadata ($section, "ScreenWidth", $1);
|
|---|
| 266 | $doc_obj->add_metadata ($section, "ScreenHeight", $2);
|
|---|
| 267 | }
|
|---|
| 268 | else {
|
|---|
| 269 | $doc_obj->add_metadata ($section, "ScreenWidth", $image_width);
|
|---|
| 270 | $doc_obj->add_metadata ($section, "ScreenHeight", $image_height);
|
|---|
| 271 | }
|
|---|
| 272 |
|
|---|
| 273 | #add the screenview as an associated file ...
|
|---|
| 274 | if (-e "$screenviewfilename") {
|
|---|
| 275 | $doc_obj->associate_file("$screenviewfilename", "screenview.$screenviewtype",
|
|---|
| 276 | "image/$screenviewtype",$section);
|
|---|
| 277 | $doc_obj->add_metadata ($section, "ScreenType", $screenviewtype);
|
|---|
| 278 | $doc_obj->add_metadata ($section, "Screen", "screenview.$screenviewtype");
|
|---|
| 279 |
|
|---|
| 280 | $doc_obj->add_metadata ($section, "screenicon", "<img src=\"_httpprefix_/collect/[collection]/index/assoc/[assocfilepath]/[Screen]\" width=[ScreenWidth] height=[ScreenHeight]>");
|
|---|
| 281 | } else {
|
|---|
| 282 | print $outhandle "ImagePlug: couldn't find \"$screenviewfilename\"\n";
|
|---|
| 283 | }
|
|---|
| 284 | }
|
|---|
| 285 |
|
|---|
| 286 | return $type;
|
|---|
| 287 |
|
|---|
| 288 |
|
|---|
| 289 | }
|
|---|
| 290 |
|
|---|
| 291 |
|
|---|
| 292 |
|
|---|
| 293 | # Discover the characteristics of an image file with the ImageMagick
|
|---|
| 294 | # "identify" command.
|
|---|
| 295 |
|
|---|
| 296 | sub identify {
|
|---|
| 297 | my ($image, $outhandle, $verbosity) = @_;
|
|---|
| 298 |
|
|---|
| 299 | # Use the ImageMagick "identify" command to get the file specs
|
|---|
| 300 | my $command = "identify \"$image\" 2>&1";
|
|---|
| 301 | print $outhandle "$command\n" if ($verbosity > 2);
|
|---|
| 302 | my $result = '';
|
|---|
| 303 | $result = `$command`;
|
|---|
| 304 | print $outhandle "$result\n" if ($verbosity > 3);
|
|---|
| 305 |
|
|---|
| 306 | # Read the type, width, and height
|
|---|
| 307 | my $type = 'unknown';
|
|---|
| 308 | my $width = 'unknown';
|
|---|
| 309 | my $height = 'unknown';
|
|---|
| 310 |
|
|---|
| 311 | my $image_safe = quotemeta $image;
|
|---|
| 312 | if ($result =~ /^$image_safe (\w+) (\d+)x(\d+)/) {
|
|---|
| 313 | $type = $1;
|
|---|
| 314 | $width = $2;
|
|---|
| 315 | $height = $3;
|
|---|
| 316 | }
|
|---|
| 317 |
|
|---|
| 318 | # Read the size
|
|---|
| 319 | my $size = "unknown";
|
|---|
| 320 | if ($result =~ m/^.* ([0-9]+)b/) {
|
|---|
| 321 | $size = $1;
|
|---|
| 322 | }
|
|---|
| 323 | elsif ($result =~ m/^.* ([0-9]+)(\.([0-9]+))?kb?/) {
|
|---|
| 324 | $size = 1024 * $1;
|
|---|
| 325 | if (defined($2)) {
|
|---|
| 326 | $size = $size + (1024 * $2);
|
|---|
| 327 | # Truncate size (it isn't going to be very accurate anyway)
|
|---|
| 328 | $size = int($size);
|
|---|
| 329 | }
|
|---|
| 330 | }
|
|---|
| 331 |
|
|---|
| 332 | print $outhandle "file: $image:\t $type, $width, $height, $size\n"
|
|---|
| 333 | if ($verbosity > 2);
|
|---|
| 334 |
|
|---|
| 335 | # Return the specs
|
|---|
| 336 | return ($type, $width, $height, $size);
|
|---|
| 337 | }
|
|---|
| 338 |
|
|---|
| 339 |
|
|---|
| 340 | # The ImagePlug read() function.
|
|---|
| 341 | # ImagePlug overrides read() because there is no need to read the actual
|
|---|
| 342 | # text of the file in, because the contents of the file is not text...
|
|---|
| 343 | #
|
|---|
| 344 | # Return number of files processed, undef if can't process
|
|---|
| 345 | # Note that $base_dir might be "" and that $file might
|
|---|
| 346 | # include directories
|
|---|
| 347 |
|
|---|
| 348 | sub read {
|
|---|
| 349 | my $self = shift (@_);
|
|---|
| 350 | my ($pluginfo, $base_dir, $file, $metadata, $processor, $maxdocs, $total_count, $gli) = @_;
|
|---|
| 351 |
|
|---|
| 352 | my $outhandle = $self->{'outhandle'};
|
|---|
| 353 |
|
|---|
| 354 | #check process and block exps, smart block, etc
|
|---|
| 355 | my ($block_status,$filename) = $self->read_block(@_);
|
|---|
| 356 | return $block_status if ((!defined $block_status) || ($block_status==0));
|
|---|
| 357 |
|
|---|
| 358 | print STDERR "<Processing n='$file' p='ImagePlug'>\n" if ($gli);
|
|---|
| 359 | print $outhandle "ImagePlug processing $file\n"
|
|---|
| 360 | if $self->{'verbosity'} > 1;
|
|---|
| 361 |
|
|---|
| 362 | # None of this works very well on Windows 95/98...
|
|---|
| 363 | if ($ENV{'GSDLOS'} eq "windows" && !Win32::IsWinNT()) {
|
|---|
| 364 | if ($gli) {
|
|---|
| 365 | print STDERR "<ProcessingError n='$file' r='Windows 95/98 not supported'>\n";
|
|---|
| 366 | }
|
|---|
| 367 | print $outhandle "ImagePlug: Windows 95/98 not supported\n";
|
|---|
| 368 | return -1;
|
|---|
| 369 | }
|
|---|
| 370 |
|
|---|
| 371 | # None of this is going to work very well without ImageMagick...
|
|---|
| 372 | if ($self->{'imagemagick_not_installed'}) {
|
|---|
| 373 | if ($gli) {
|
|---|
| 374 | print STDERR "<ProcessingError n='$file' r='ImageMagick not installed'>\n";
|
|---|
| 375 | }
|
|---|
| 376 | print $outhandle "ImagePlug: ImageMagick not installed\n";
|
|---|
| 377 | return -1;
|
|---|
| 378 | }
|
|---|
| 379 |
|
|---|
| 380 | #if there's a leading directory name, eat it...
|
|---|
| 381 | $file =~ s/^.*[\/\\]//;
|
|---|
| 382 |
|
|---|
| 383 | # create a new document
|
|---|
| 384 | my $doc_obj = new doc ($filename, "indexed_doc");
|
|---|
| 385 | $doc_obj->set_OIDtype ($processor->{'OIDtype'}, $processor->{'OIDmetadata'});
|
|---|
| 386 | $doc_obj->add_utf8_metadata($doc_obj->get_top_section(), "Plugin", "$self->{'plugin_type'}");
|
|---|
| 387 |
|
|---|
| 388 | #run convert to get the thumbnail and extract size and type info
|
|---|
| 389 | my $result = run_convert($self, $filename, $file, $doc_obj);
|
|---|
| 390 |
|
|---|
| 391 | if (!defined $result)
|
|---|
| 392 | {
|
|---|
| 393 | if ($gli) {
|
|---|
| 394 | print STDERR "<ProcessingError n='$file'>\n";
|
|---|
| 395 | }
|
|---|
| 396 | print $outhandle "ImagePlug: couldn't process \"$filename\"\n";
|
|---|
| 397 | return -1; # error during processing
|
|---|
| 398 | }
|
|---|
| 399 |
|
|---|
| 400 |
|
|---|
| 401 | #create an empty text string so we don't break downstream plugins
|
|---|
| 402 | &gsprintf::output_strings_in_UTF8;
|
|---|
| 403 | my $text = &gsprintf::lookup_string("{BasPlug.dummy_text}");
|
|---|
| 404 |
|
|---|
| 405 | # include any metadata passed in from previous plugins
|
|---|
| 406 | # note that this metadata is associated with the top level section
|
|---|
| 407 | my $section = $doc_obj->get_top_section();
|
|---|
| 408 | $self->extra_metadata ($doc_obj, $section, $metadata);
|
|---|
| 409 |
|
|---|
| 410 | # do plugin specific processing of doc_obj
|
|---|
| 411 | unless (defined ($self->process(\$text, $pluginfo, $base_dir, $file, $metadata, $doc_obj))) {
|
|---|
| 412 | print STDERR "<ProcessingError n='$file'>\n" if ($gli);
|
|---|
| 413 | return -1;
|
|---|
| 414 | }
|
|---|
| 415 |
|
|---|
| 416 | # do any automatic metadata extraction
|
|---|
| 417 | $self->auto_extract_metadata ($doc_obj);
|
|---|
| 418 |
|
|---|
| 419 | # if we haven't found any Title so far, assign one
|
|---|
| 420 | # this was shifted to here from inside read()
|
|---|
| 421 | $self->title_fallback($doc_obj,$section,$file);
|
|---|
| 422 | # add an OID
|
|---|
| 423 | $doc_obj->set_OID();
|
|---|
| 424 | $doc_obj->add_utf8_text($section, $text);
|
|---|
| 425 |
|
|---|
| 426 | # process the document
|
|---|
| 427 | $processor->process($doc_obj);
|
|---|
| 428 |
|
|---|
| 429 | # clean up temporary files - we do this here instead of in
|
|---|
| 430 | # run_convert becuase associated files aren't actually copied
|
|---|
| 431 | # until after process has been run.
|
|---|
| 432 | if (defined $self->{'tmp_filename'} &&
|
|---|
| 433 | -e $self->{'tmp_filename'}) {
|
|---|
| 434 | &util::rm($self->{'tmp_filename'})
|
|---|
| 435 | }
|
|---|
| 436 |
|
|---|
| 437 | if (defined $self->{'tmp_filename2'} &&
|
|---|
| 438 | -e $self->{'tmp_filename2'}) {
|
|---|
| 439 | &util::rm($self->{'tmp_filename2'})
|
|---|
| 440 | }
|
|---|
| 441 | if (defined $self->{'tmp_filename3'} &&
|
|---|
| 442 | -e $self->{'tmp_filename3'}) {
|
|---|
| 443 | &util::rm($self->{'tmp_filename3'})
|
|---|
| 444 | }
|
|---|
| 445 |
|
|---|
| 446 | $self->{'num_processed'}++;
|
|---|
| 447 |
|
|---|
| 448 | return 1;
|
|---|
| 449 | }
|
|---|
| 450 |
|
|---|
| 451 | # do plugin specific processing of doc_obj
|
|---|
| 452 | sub process {
|
|---|
| 453 | my $self = shift (@_);
|
|---|
| 454 | my ($textref, $pluginfo, $base_dir, $file, $metadata, $doc_obj) = @_;
|
|---|
| 455 | my $outhandle = $self->{'outhandle'};
|
|---|
| 456 |
|
|---|
| 457 | return 1;
|
|---|
| 458 | }
|
|---|
| 459 |
|
|---|
| 460 | 1;
|
|---|
| 461 |
|
|---|
| 462 |
|
|---|
| 463 |
|
|---|
| 464 |
|
|---|
| 465 |
|
|---|
| 466 |
|
|---|
| 467 |
|
|---|
| 468 |
|
|---|
| 469 |
|
|---|
| 470 |
|
|---|
| 471 |
|
|---|
