| 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 | sub BEGIN {
|
|---|
| 31 | @ISA = ('BasPlug');
|
|---|
| 32 | }
|
|---|
| 33 |
|
|---|
| 34 |
|
|---|
| 35 | sub print_usage {
|
|---|
| 36 | my ($plugin_name) = @_;
|
|---|
| 37 |
|
|---|
| 38 | print STDERR "
|
|---|
| 39 | usage: plugin ImagePlug [options]
|
|---|
| 40 |
|
|---|
| 41 | -noscaleup Don't scale up small images when making thumbnails
|
|---|
| 42 |
|
|---|
| 43 | -thumbnailsize n Make thumbnails of size nxn
|
|---|
| 44 |
|
|---|
| 45 | -thumbnailtype s Make thumbnails in format 's'
|
|---|
| 46 |
|
|---|
| 47 | -screenviewsize n If set, makes an image of size n for screen display
|
|---|
| 48 | and sets Screen, ScreenSize, ScrrenWidth and Screeneight
|
|---|
| 49 | metadata. By default it is not set.
|
|---|
| 50 |
|
|---|
| 51 | -screenviewtype s If -screenviewsize is set, this sets the screen display
|
|---|
| 52 | image type. Defaults to jpg.
|
|---|
| 53 |
|
|---|
| 54 | -convertto s Convert main inage to (gif|png|jpg)
|
|---|
| 55 |
|
|---|
| 56 | -minimumsize n Ignore images smaller than n bytes
|
|---|
| 57 |
|
|---|
| 58 | "
|
|---|
| 59 | }
|
|---|
| 60 |
|
|---|
| 61 |
|
|---|
| 62 | sub new {
|
|---|
| 63 | my ($class) = @_;
|
|---|
| 64 | my $plugin_name = shift (@_);
|
|---|
| 65 | my $self = new BasPlug ("ImagePlug", @_);
|
|---|
| 66 |
|
|---|
| 67 | if (!parsargv::parse(\@_,
|
|---|
| 68 | q^noscaleup^, \$self->{'noscaleup'},
|
|---|
| 69 | q^converttotype/.*/^, \$self->{'converttotype'},
|
|---|
| 70 | q^minimumsize/[0-9]*/100^, \$self->{'minimumsize'},
|
|---|
| 71 |
|
|---|
| 72 | q^thumbnailsize/[0-9]*/100^, \$self->{'thumbnailsize'},
|
|---|
| 73 | q^thumbnailtype/.*/gif^, \$self->{'thumbnailtype'},
|
|---|
| 74 | q^screenviewsize/[0-9]*/0^, \$self->{'screenviewsize'},
|
|---|
| 75 | q^screenviewtype/.*/jpg^, \$self->{'screenviewtype'},
|
|---|
| 76 | "allow_extra_options")) {
|
|---|
| 77 |
|
|---|
| 78 | print STDERR "\nImagePlug uses an incorrect option.\n";
|
|---|
| 79 | print STDERR "Check your collect.cfg configuration file.\n";
|
|---|
| 80 | &print_usage($plugin_name);
|
|---|
| 81 | die "\n";
|
|---|
| 82 | }
|
|---|
| 83 |
|
|---|
| 84 | return bless $self, $class;
|
|---|
| 85 | }
|
|---|
| 86 |
|
|---|
| 87 |
|
|---|
| 88 | sub get_default_process_exp {
|
|---|
| 89 | my $self = shift (@_);
|
|---|
| 90 |
|
|---|
| 91 | return q^(?i)(\.jpe?g|\.gif|\.png|\.bmp|\.xbm|\.tif?f)$^;
|
|---|
| 92 | }
|
|---|
| 93 |
|
|---|
| 94 |
|
|---|
| 95 | # Create the thumbnail and screenview images, and discover the Image's
|
|---|
| 96 | # size, width, and height using the convert utility.
|
|---|
| 97 |
|
|---|
| 98 | sub run_convert {
|
|---|
| 99 | my $self = shift (@_);
|
|---|
| 100 | my $filename = shift (@_); # filename with full path
|
|---|
| 101 | my $file = shift (@_); # filename without path
|
|---|
| 102 | my $doc_obj = shift (@_);
|
|---|
| 103 | my $section = $doc_obj->get_top_section();
|
|---|
| 104 |
|
|---|
| 105 | my $verbosity = $self->{'verbosity'};
|
|---|
| 106 | my $outhandle = $self->{'outhandle'};
|
|---|
| 107 |
|
|---|
| 108 | # check the filename is okay
|
|---|
| 109 | return 0 if ($file eq "" || $filename eq "");
|
|---|
| 110 |
|
|---|
| 111 | # Code now extended to quote filenames in 'convert' commnads
|
|---|
| 112 | # Allows spaces in filenames, but note needs spaces to be escaped in URL as well
|
|---|
| 113 | # if ($filename =~ m/ /) {
|
|---|
| 114 | # print $outhandle "ImagePlug: \"$filename\" contains a space. choking.\n";
|
|---|
| 115 | # return undef;
|
|---|
| 116 | # }
|
|---|
| 117 |
|
|---|
| 118 | my $minimumsize = $self->{'minimumsize'};
|
|---|
| 119 | if (defined $minimumsize && (-s $filename < $minimumsize)) {
|
|---|
| 120 | print $outhandle "ImagePlug: \"$filename\" too small, skipping\n"
|
|---|
| 121 | if ($verbosity > 1);
|
|---|
| 122 | }
|
|---|
| 123 |
|
|---|
| 124 | # Convert the image to a new type (if required).
|
|---|
| 125 | my $converttotype = $self->{'converttotype'};
|
|---|
| 126 | my $originalfilename = ""; # only set if we do a conversion
|
|---|
| 127 | my $type = "unknown";
|
|---|
| 128 |
|
|---|
| 129 | if ($converttotype ne "" && $filename =~ m/$converttotype$/) {
|
|---|
| 130 |
|
|---|
| 131 | $originalfilename = $filename;
|
|---|
| 132 | $filename = &util::get_tmp_filename() . ".$converttotype";
|
|---|
| 133 | $self->{'tmp_filename'} = $filename;
|
|---|
| 134 |
|
|---|
| 135 | my $command = "convert -interlace plane -verbose \"$originalfilename\" \"$filename\"";
|
|---|
| 136 | print $outhandle "$command\n" if ($verbosity > 2);
|
|---|
| 137 | my $result = '';
|
|---|
| 138 | $result = `$command`;
|
|---|
| 139 | print $outhandle "$result\n" if ($verbosity > 3);
|
|---|
| 140 |
|
|---|
| 141 | $type = $converttotype;
|
|---|
| 142 | }
|
|---|
| 143 |
|
|---|
| 144 | # Add the image metadata
|
|---|
| 145 | my $url = $file;
|
|---|
| 146 | $url =~ s/ /%20/g;
|
|---|
| 147 |
|
|---|
| 148 | $doc_obj->add_metadata ($section, "Image", $url);
|
|---|
| 149 | my ($image_type, $image_width, $image_height, $image_size)
|
|---|
| 150 | = &identify($filename, $outhandle, $verbosity);
|
|---|
| 151 |
|
|---|
| 152 | $doc_obj->add_metadata ($section, "ImageType", $image_type);
|
|---|
| 153 | $doc_obj->add_metadata ($section, "ImageWidth", $image_width);
|
|---|
| 154 | $doc_obj->add_metadata ($section, "ImageHeight", $image_height);
|
|---|
| 155 | $doc_obj->add_metadata ($section, "ImageSize", $image_size);
|
|---|
| 156 |
|
|---|
| 157 | # Add the image as an associated file
|
|---|
| 158 | $doc_obj->associate_file($filename,$file,"image/$type",$section);
|
|---|
| 159 |
|
|---|
| 160 | # Make the thumbnail image
|
|---|
| 161 | my $thumbnailsize = $self->{'thumbnailsize'} || 100;
|
|---|
| 162 | my $thumbnailtype = $self->{'thumbnailtype'} || 'gif';
|
|---|
| 163 | my $thumbnailfile = &util::get_tmp_filename() . ".$thumbnailtype";
|
|---|
| 164 | $self->{'tmp_filename2'} = $thumbnailfile;
|
|---|
| 165 |
|
|---|
| 166 | # Generate the thumbnail with convert
|
|---|
| 167 | my $command = "convert -interlace plane -verbose -geometry $thumbnailsize"
|
|---|
| 168 | . "x$thumbnailsize \"$filename\" \"$thumbnailfile\"";
|
|---|
| 169 | print $outhandle "$command\n" if ($verbosity > 2);
|
|---|
| 170 | my $result = '';
|
|---|
| 171 | $result = `$command 2>&1` ;
|
|---|
| 172 | print $outhandle "$result\n" if ($verbosity > 3);
|
|---|
| 173 |
|
|---|
| 174 | # Add the thumbnail as an associated file ...
|
|---|
| 175 | if (-e "$thumbnailfile") {
|
|---|
| 176 | $doc_obj->associate_file("$thumbnailfile", "thumbnail.$thumbnailtype",
|
|---|
| 177 | "image/$thumbnailtype",$section);
|
|---|
| 178 | $doc_obj->add_metadata ($section, "ThumbType", $thumbnailtype);
|
|---|
| 179 | $doc_obj->add_metadata ($section, "Thumb", "thumbnail.$thumbnailtype");
|
|---|
| 180 | }
|
|---|
| 181 |
|
|---|
| 182 | # Extract Thumnail metadata from convert output
|
|---|
| 183 | if ($result =~ m/[0-9]+x[0-9]+=>([0-9]+)x([0-9]+)/) {
|
|---|
| 184 | $doc_obj->add_metadata ($section, "ThumbWidth", $1);
|
|---|
| 185 | $doc_obj->add_metadata ($section, "ThumbHeight", $2);
|
|---|
| 186 | }
|
|---|
| 187 |
|
|---|
| 188 | # Make a screen-sized version of the picture if requested
|
|---|
| 189 | if ($self->{'screenviewsize'}) {
|
|---|
| 190 |
|
|---|
| 191 | # To do: if the actual image smaller than the screenview size,
|
|---|
| 192 | # we should use the original !
|
|---|
| 193 |
|
|---|
| 194 | my $screenviewsize = $self->{'screenviewsize'};
|
|---|
| 195 | my $screenviewtype = $self->{'screenviewtype'} || 'jpeg';
|
|---|
| 196 | my $screenviewfilename = &util::get_tmp_filename() . ".$screenviewtype";
|
|---|
| 197 | $self->{'tmp_filename3'} = $screenviewfilename;
|
|---|
| 198 |
|
|---|
| 199 | # make the screenview image
|
|---|
| 200 | my $command = "convert -interlace plane -verbose -geometry $screenviewsize"
|
|---|
| 201 | . "x$screenviewsize \"$filename\" \"$screenviewfilename\"";
|
|---|
| 202 | print $outhandle "$command\n" if ($verbosity > 2);
|
|---|
| 203 | my $result = "";
|
|---|
| 204 | $result = `$command 2>&1` ;
|
|---|
| 205 | print $outhandle "$result\n" if ($verbosity > 3);
|
|---|
| 206 |
|
|---|
| 207 | # get screenview dimensions, size and type
|
|---|
| 208 | if ($result =~ m/[0-9]+x[0-9]+=>([0-9]+)x([0-9]+)/) {
|
|---|
| 209 | $doc_obj->add_metadata ($section, "ScreenWidth", $1);
|
|---|
| 210 | $doc_obj->add_metadata ($section, "ScreenHeight", $2);
|
|---|
| 211 | }
|
|---|
| 212 |
|
|---|
| 213 | #add the screenview as an associated file ...
|
|---|
| 214 | if (-e "$screenviewfilename") {
|
|---|
| 215 | $doc_obj->associate_file("$screenviewfilename", "screenview.$screenviewtype",
|
|---|
| 216 | "image/$screenviewtype",$section);
|
|---|
| 217 | $doc_obj->add_metadata ($section, "ScreenType", $screenviewtype);
|
|---|
| 218 | $doc_obj->add_metadata ($section, "Screen", "screenview.$screenviewtype");
|
|---|
| 219 | } else {
|
|---|
| 220 | print $outhandle "ImagePlug: couldn't find \"$screenviewfilename\"\n";
|
|---|
| 221 | }
|
|---|
| 222 | }
|
|---|
| 223 |
|
|---|
| 224 | return $type;
|
|---|
| 225 | }
|
|---|
| 226 |
|
|---|
| 227 |
|
|---|
| 228 | # Discover the characteristics of an image file with the ImageMagick
|
|---|
| 229 | # "identify" command.
|
|---|
| 230 |
|
|---|
| 231 | sub identify {
|
|---|
| 232 | my ($image, $outhandle, $verbosity) = @_;
|
|---|
| 233 |
|
|---|
| 234 | # Use the ImageMagick "identify" command to get the file specs
|
|---|
| 235 | my $command = "identify $image 2>&1";
|
|---|
| 236 | print $outhandle "$command\n" if ($verbosity > 2);
|
|---|
| 237 | my $result = '';
|
|---|
| 238 | $result = `$command`;
|
|---|
| 239 | print $outhandle "$result\n" if ($verbosity > 3);
|
|---|
| 240 |
|
|---|
| 241 | # Read the type, width, and height
|
|---|
| 242 | my $type = 'unknown';
|
|---|
| 243 | my $width = 'unknown';
|
|---|
| 244 | my $height = 'unknown';
|
|---|
| 245 |
|
|---|
| 246 | if ($result =~ /^$image (\w+) (\d+)x(\d+)/) {
|
|---|
| 247 | $type = $1;
|
|---|
| 248 | $width = $2;
|
|---|
| 249 | $height = $3;
|
|---|
| 250 | }
|
|---|
| 251 |
|
|---|
| 252 | # Read the size
|
|---|
| 253 | my $size = "unknown";
|
|---|
| 254 | if ($result =~ m/^.* ([0-9]+)b/) {
|
|---|
| 255 | $size = $1;
|
|---|
| 256 | } elsif ($result =~ m/^.* ([0-9]+)kb/) {
|
|---|
| 257 | $size = 1024 * $1;
|
|---|
| 258 | }
|
|---|
| 259 |
|
|---|
| 260 | print $outhandle "file: $image:\t $type, $width, $height, $size\n"
|
|---|
| 261 | if ($verbosity > 2);
|
|---|
| 262 |
|
|---|
| 263 | # Return the specs
|
|---|
| 264 | return ($type, $width, $height, $size);
|
|---|
| 265 | }
|
|---|
| 266 |
|
|---|
| 267 |
|
|---|
| 268 | # The ImagePlug read() function. This function does all the right things
|
|---|
| 269 | # to make general options work for a given plugin. It calls the process()
|
|---|
| 270 | # function which does all the work specific to a plugin (like the old
|
|---|
| 271 | # read functions used to do). Most plugins should define their own
|
|---|
| 272 | # process() function and let this read() function keep control.
|
|---|
| 273 | #
|
|---|
| 274 | # ImagePlug overrides read() because there is no need to read the actual
|
|---|
| 275 | # text of the file in, because the contents of the file is not text...
|
|---|
| 276 | #
|
|---|
| 277 | # Return number of files processed, undef if can't process
|
|---|
| 278 | # Note that $base_dir might be "" and that $file might
|
|---|
| 279 | # include directories
|
|---|
| 280 |
|
|---|
| 281 | sub read {
|
|---|
| 282 | my $self = shift (@_);
|
|---|
| 283 | my ($pluginfo, $base_dir, $file, $metadata, $processor, $maxdocs) = @_;
|
|---|
| 284 |
|
|---|
| 285 | my $outhandle = $self->{'outhandle'};
|
|---|
| 286 |
|
|---|
| 287 | my $filename = &util::filename_cat($base_dir, $file);
|
|---|
| 288 | return 0 if $self->{'block_exp'} ne "" && $filename =~ /$self->{'block_exp'}/;
|
|---|
| 289 | if ($filename !~ /$self->{'process_exp'}/ || !-f $filename) {
|
|---|
| 290 | return undef;
|
|---|
| 291 | }
|
|---|
| 292 | print $outhandle "ImagePlug processing \"$filename\"\n"
|
|---|
| 293 | if $self->{'verbosity'} > 1;
|
|---|
| 294 |
|
|---|
| 295 | #if there's a leading directory name, eat it...
|
|---|
| 296 | $file =~ s/^.*[\/\\]//;
|
|---|
| 297 |
|
|---|
| 298 | # create a new document
|
|---|
| 299 | my $doc_obj = new doc ($filename, "indexed_doc");
|
|---|
| 300 | $doc_obj->set_OIDtype ($processor->{'OIDtype'});
|
|---|
| 301 |
|
|---|
| 302 | #run convert to get the thumbnail and extract size and type info
|
|---|
| 303 | my $result = run_convert($self, $filename, $file, $doc_obj);
|
|---|
| 304 |
|
|---|
| 305 | if (!defined $result)
|
|---|
| 306 | {
|
|---|
| 307 | print "ImagePlug: couldn't process \"$filename\"\n";
|
|---|
| 308 | return 0;
|
|---|
| 309 | }
|
|---|
| 310 |
|
|---|
| 311 | #create an empty text string so we don't break downstream plugins
|
|---|
| 312 | my $text = "Dummy text to sidestep display bug.";
|
|---|
| 313 |
|
|---|
| 314 | # include any metadata passed in from previous plugins
|
|---|
| 315 | # note that this metadata is associated with the top level section
|
|---|
| 316 | my $section = $doc_obj->get_top_section();
|
|---|
| 317 | $self->extra_metadata ($doc_obj, $section, $metadata);
|
|---|
| 318 |
|
|---|
| 319 | # do plugin specific processing of doc_obj
|
|---|
| 320 | return undef unless defined ($self->process (\$text, $pluginfo, $base_dir,
|
|---|
| 321 | $file, $metadata, $doc_obj));
|
|---|
| 322 |
|
|---|
| 323 | # do any automatic metadata extraction
|
|---|
| 324 | $self->auto_extract_metadata ($doc_obj);
|
|---|
| 325 |
|
|---|
| 326 | # add an OID
|
|---|
| 327 | $doc_obj->set_OID();
|
|---|
| 328 | $doc_obj->add_text($section, $text);
|
|---|
| 329 |
|
|---|
| 330 | # process the document
|
|---|
| 331 | $processor->process($doc_obj);
|
|---|
| 332 |
|
|---|
| 333 | # clean up temporary files - we do this here instead of in
|
|---|
| 334 | # run_convert becuase associated files aren't actually copied
|
|---|
| 335 | # until after process has been run.
|
|---|
| 336 | if (defined $self->{'tmp_filename'} &&
|
|---|
| 337 | -e $self->{'tmp_filename'}) {
|
|---|
| 338 | &util::rm($self->{'tmp_filename'})
|
|---|
| 339 | }
|
|---|
| 340 | if (defined $self->{'tmp_filename2'} &&
|
|---|
| 341 | -e $self->{'tmp_filename2'}) {
|
|---|
| 342 | &util::rm($self->{'tmp_filename2'})
|
|---|
| 343 | }
|
|---|
| 344 | if (defined $self->{'tmp_filename3'} &&
|
|---|
| 345 | -e $self->{'tmp_filename3'}) {
|
|---|
| 346 | &util::rm($self->{'tmp_filename3'})
|
|---|
| 347 | }
|
|---|
| 348 |
|
|---|
| 349 | $self->{'num_processed'}++;
|
|---|
| 350 |
|
|---|
| 351 | return 1;
|
|---|
| 352 | }
|
|---|
| 353 |
|
|---|
| 354 | # do plugin specific processing of doc_obj
|
|---|
| 355 | sub process {
|
|---|
| 356 | my $self = shift (@_);
|
|---|
| 357 | my ($textref, $pluginfo, $base_dir, $file, $metadata, $doc_obj) = @_;
|
|---|
| 358 | my $outhandle = $self->{'outhandle'};
|
|---|
| 359 |
|
|---|
| 360 | return 1;
|
|---|
| 361 | }
|
|---|
| 362 |
|
|---|
| 363 | 1;
|
|---|
