| 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 | use strict;
|
|---|
| 35 |
|
|---|
| 36 | sub print_general_usage {
|
|---|
| 37 | my ($plugin_name) = @_;
|
|---|
| 38 |
|
|---|
| 39 | print STDERR "\n usage: plugin $plugin_name [options]\n\n";
|
|---|
| 40 | print STDERR " -noscaleup Don't scale up small images when making thumbnails\n";
|
|---|
| 41 | print STDERR " -thumbnailtype s Make thumbnails in format 's'\n";
|
|---|
| 42 | print STDERR " -thumbnailsize n Make thumbnails of size nxn\n";
|
|---|
| 43 | print STDERR " -convertto s Convert main inage to (gif|png|jpg)\n";
|
|---|
| 44 | print STDERR " -minimumsize n Ignore images smaller than n bytes\n";
|
|---|
| 45 | print STDERR "\n";
|
|---|
| 46 | }
|
|---|
| 47 |
|
|---|
| 48 | sub new {
|
|---|
| 49 | my ($class) = @_;
|
|---|
| 50 | my $plugin_name = shift (@_);
|
|---|
| 51 | my $self = new BasPlug ("ImagePlug", @_);
|
|---|
| 52 | # general options available to all plugins
|
|---|
| 53 |
|
|---|
| 54 | if (!parsargv::parse(\@_,
|
|---|
| 55 | q^noscaleup^, \$self->{'noscaleup'},
|
|---|
| 56 | q^thumbnailtype/.*/gif^, \$self->{'thumbnailtype'},
|
|---|
| 57 | q^converttotype/.*/^, \$self->{'converttotype'},
|
|---|
| 58 | q^thumbnailsize/[0-9]*/100^, \$self->{'thumbnailsize'},
|
|---|
| 59 | q^minimumsize/[0-9]*/100^, \$self->{'minimumsize'},
|
|---|
| 60 | "allow_extra_options")) {
|
|---|
| 61 |
|
|---|
| 62 | print STDERR "\nThe $plugin_name plugin uses an incorrect general option (general options are those\n";
|
|---|
| 63 | print STDERR "available to all plugins). Check your collect.cfg configuration file.\n";
|
|---|
| 64 | &print_general_usage($plugin_name);
|
|---|
| 65 | die "\n";
|
|---|
| 66 | }
|
|---|
| 67 |
|
|---|
| 68 | return bless $self, $class;
|
|---|
| 69 | }
|
|---|
| 70 |
|
|---|
| 71 | sub get_default_process_exp {
|
|---|
| 72 | my $self = shift (@_);
|
|---|
| 73 |
|
|---|
| 74 | return q^(?i)(\.jpe?g|\.gif|\.png|\.bmp|\.xbm|\.tif?f)$^;
|
|---|
| 75 | }
|
|---|
| 76 |
|
|---|
| 77 | sub run_convert {
|
|---|
| 78 | my $self = shift (@_);
|
|---|
| 79 | my $filename = shift (@_);
|
|---|
| 80 | my $file = shift (@_);
|
|---|
| 81 | my $doc_obj = shift (@_);
|
|---|
| 82 | my $section = $doc_obj->get_top_section();
|
|---|
| 83 |
|
|---|
| 84 | if ($file eq "" || $filename eq "") {
|
|---|
| 85 | return "";
|
|---|
| 86 | }
|
|---|
| 87 | # if ($filename =~ m/thumbnail/) {
|
|---|
| 88 | # return "";
|
|---|
| 89 | # }
|
|---|
| 90 | # if ($filename =~ m/converted/) {
|
|---|
| 91 | # return "";
|
|---|
| 92 | # }
|
|---|
| 93 | if ($filename =~ m/ /) {
|
|---|
| 94 | print STDERR "IamgePlug: \"$filename\" contains a space. choking.\n";
|
|---|
| 95 | return "";
|
|---|
| 96 | }
|
|---|
| 97 |
|
|---|
| 98 | my $minimumsize = $self->{'minimumsize'};
|
|---|
| 99 | my $thumbSize = $self->{'thumbnailsize'};
|
|---|
| 100 | if ($thumbSize eq "") { $thumbSize = 100; };
|
|---|
| 101 | my $converttotype = $self->{'converttotype'};
|
|---|
| 102 | my $thumbnailtype = $self->{'thumbnailtype'};
|
|---|
| 103 | if ($thumbnailtype eq "") { $thumbnailtype = "gif"; };
|
|---|
| 104 | my $originalfilename = ""; # only set if we do a conversion
|
|---|
| 105 | my $thumbnailfilename = "";
|
|---|
| 106 |
|
|---|
| 107 | my $type = "unknown";
|
|---|
| 108 |
|
|---|
| 109 | if (defined $minimumsize && (-s $filename < $minimumsize)) {
|
|---|
| 110 | print STDERR "ImagePlug: \"$filename\" too small, skipping\n"
|
|---|
| 111 | if $self->{'verbosity'} > 1;
|
|---|
| 112 | }
|
|---|
| 113 | #see if we need to convert ...
|
|---|
| 114 | if ($converttotype ne "" && $filename =~ m/$converttotype$/) {
|
|---|
| 115 | $originalfilename = $filename;
|
|---|
| 116 | $filename = &util::get_tmp_filename();
|
|---|
| 117 | $filename = "$filename.$converttotype";
|
|---|
| 118 | $self->{'tmp_filename'} = $filename;
|
|---|
| 119 | if (-e $filename) {
|
|---|
| 120 | print STDERR "File names to convert from and to are the same. choking in Imageplug on \"$filename\"\n";
|
|---|
| 121 | return "";
|
|---|
| 122 | }
|
|---|
| 123 | my $result = "";
|
|---|
| 124 | my $command = "convert -verbose $originalfilename $filename";
|
|---|
| 125 | $result = `$command`;
|
|---|
| 126 | print STDERR "$command\n"
|
|---|
| 127 | if $self->{'verbosity'} > 2;
|
|---|
| 128 | print STDERR "$result\n"
|
|---|
| 129 | if $self->{'verbosity'} > 3;
|
|---|
| 130 | $type = $converttotype;
|
|---|
| 131 | }
|
|---|
| 132 |
|
|---|
| 133 | #check that the thumbnail doesn't already exist ...
|
|---|
| 134 | $thumbnailfilename = &util::get_tmp_filename();
|
|---|
| 135 | $thumbnailfilename = $thumbnailfilename . ".thumbnail.$thumbnailtype";
|
|---|
| 136 | $self->{'tmp_filename2'} = $thumbnailfilename;
|
|---|
| 137 |
|
|---|
| 138 | #make the thumbnail
|
|---|
| 139 | my $result = "";
|
|---|
| 140 | my $command = "convert -verbose -geometry $thumbSize" . "x$thumbSize $filename $thumbnailfilename";
|
|---|
| 141 | $result = `$command` ;
|
|---|
| 142 | print STDERR "$command\n"
|
|---|
| 143 | if $self->{'verbosity'} > 2;
|
|---|
| 144 | print STDERR "$result\n"
|
|---|
| 145 | if $self->{'verbosity'} > 3;
|
|---|
| 146 |
|
|---|
| 147 | if ($result =~ m/([0-9]+)x([0-9]+)=>([0-9]+)x([0-9]+)/) {
|
|---|
| 148 | $doc_obj->add_metadata ($section, "ImageHeight", $1);
|
|---|
| 149 | $doc_obj->add_metadata ($section, "ImageWidth", $2);
|
|---|
| 150 | $doc_obj->add_metadata ($section, "ThumbHeight", $3);
|
|---|
| 151 | $doc_obj->add_metadata ($section, "ThumbWidth", $4);
|
|---|
| 152 | }
|
|---|
| 153 |
|
|---|
| 154 | my $size = "unknown";
|
|---|
| 155 | if ($result =~ m/^[^\n]* ([0-9]+)b/) {
|
|---|
| 156 | $size = $1;
|
|---|
| 157 | }
|
|---|
| 158 | if ($result =~ m/^[^\n]* ([0-9]+)kb/) {
|
|---|
| 159 | $size = 1024 * $1;
|
|---|
| 160 | }
|
|---|
| 161 |
|
|---|
| 162 | if ($result =~ m/^[^\n]*JPE?G/i) {
|
|---|
| 163 | $type = "jpeg";
|
|---|
| 164 | }
|
|---|
| 165 | if ($result =~ m/^[^\n]*GIF/i) {
|
|---|
| 166 | $type = "gif";
|
|---|
| 167 | }
|
|---|
| 168 | if ($result =~ m/^[^\n]*PNG/i) {
|
|---|
| 169 | $type = "png";
|
|---|
| 170 | }
|
|---|
| 171 | if ($result =~ m/^[^\n]*TIF?F/i) {
|
|---|
| 172 | $type = "tiff";
|
|---|
| 173 | }
|
|---|
| 174 | if ($result =~ m/^[^\n]*BMP/i) {
|
|---|
| 175 | $type = "bmp";
|
|---|
| 176 | }
|
|---|
| 177 | if ($result =~ m/^[^\n]*XBM?F/i) {
|
|---|
| 178 | $type = "xbm";
|
|---|
| 179 | }
|
|---|
| 180 |
|
|---|
| 181 | #if there's a leading directory name, eat it...
|
|---|
| 182 | $file =~ s/^[^\/\\]*[\/\\]//;
|
|---|
| 183 |
|
|---|
| 184 | $doc_obj->add_metadata ($section, "ImageType", $type);
|
|---|
| 185 | $doc_obj->add_metadata ($section, "Image", "$file");
|
|---|
| 186 | $doc_obj->add_metadata ($section, "ImageSize", $size);
|
|---|
| 187 |
|
|---|
| 188 | #add the thumbnail as an associated file ...
|
|---|
| 189 | if (-e "$thumbnailfilename") {
|
|---|
| 190 | $doc_obj->associate_file("$thumbnailfilename","thumbnail.$thumbnailtype","image/$thumbnailtype",$section);
|
|---|
| 191 | $doc_obj->add_metadata ($section, "ThumbType", $thumbnailtype);
|
|---|
| 192 | $doc_obj->add_metadata ($section, "Thumb", "thumbnail.$thumbnailtype");
|
|---|
| 193 | } else {
|
|---|
| 194 | print STDERR "ImagePlug: couldn't find \"$thumbnailfilename\"\n";
|
|---|
| 195 | }
|
|---|
| 196 |
|
|---|
| 197 | #add the image as an associated file ...
|
|---|
| 198 | $doc_obj->associate_file($filename,$file,"image/$type",$section);
|
|---|
| 199 |
|
|---|
| 200 | return $type;
|
|---|
| 201 | }
|
|---|
| 202 |
|
|---|
| 203 |
|
|---|
| 204 | # The ImagePlug read() function. This function does all the right things
|
|---|
| 205 | # to make general options work for a given plugin. It calls the process()
|
|---|
| 206 | # function which does all the work specific to a plugin (like the old
|
|---|
| 207 | # read functions used to do). Most plugins should define their own
|
|---|
| 208 | # process() function and let this read() function keep control.
|
|---|
| 209 | #
|
|---|
| 210 | # ImagePlug overrides read() because there is no need to read the actual
|
|---|
| 211 | # text of the file in, because the contents of the file is not text...
|
|---|
| 212 | #
|
|---|
| 213 | # Return number of files processed, undef if can't process
|
|---|
| 214 | # Note that $base_dir might be "" and that $file might
|
|---|
| 215 | # include directories
|
|---|
| 216 |
|
|---|
| 217 | sub read {
|
|---|
| 218 | my $self = shift (@_);
|
|---|
| 219 | my ($pluginfo, $base_dir, $file, $metadata, $processor, $maxdocs) = @_;
|
|---|
| 220 |
|
|---|
| 221 | my $filename = &util::filename_cat($base_dir, $file);
|
|---|
| 222 | return 0 if $self->{'block_exp'} ne "" && $filename =~ /$self->{'block_exp'}/;
|
|---|
| 223 | if ($filename !~ /$self->{'process_exp'}/ || !-f $filename) {
|
|---|
| 224 | return undef;
|
|---|
| 225 | }
|
|---|
| 226 | print STDERR "ImagePlug: called on \"$filename\"\n"
|
|---|
| 227 | if $self->{'verbosity'} > 1;
|
|---|
| 228 |
|
|---|
| 229 | $file =~ s/^[\/\\]+//; # $file often begins with / so we'll tidy it up
|
|---|
| 230 |
|
|---|
| 231 | # create a new document
|
|---|
| 232 | my $doc_obj = new doc ($filename, "indexed_doc");
|
|---|
| 233 |
|
|---|
| 234 | #run convert to get the thumbnail and extract size and type info
|
|---|
| 235 | my $result = run_convert($self, $filename, $file, $doc_obj);
|
|---|
| 236 |
|
|---|
| 237 | if (!defined $result)
|
|---|
| 238 | {
|
|---|
| 239 | print "ImagePlug: couldn't process \"$filename\"\n";
|
|---|
| 240 | return 0;
|
|---|
| 241 | }
|
|---|
| 242 |
|
|---|
| 243 | #create an empty text string so we don't break downstream plugins
|
|---|
| 244 | my $text = "";
|
|---|
| 245 |
|
|---|
| 246 | # include any metadata passed in from previous plugins
|
|---|
| 247 | # note that this metadata is associated with the top level section
|
|---|
| 248 | my $section = $doc_obj->get_top_section();
|
|---|
| 249 | $self->extra_metadata ($doc_obj, $section, $metadata);
|
|---|
| 250 |
|
|---|
| 251 | # do plugin specific processing of doc_obj
|
|---|
| 252 | return undef unless defined ($self->process (\$text, $pluginfo, $base_dir, $file, $metadata, $doc_obj));
|
|---|
| 253 |
|
|---|
| 254 | # do any automatic metadata extraction
|
|---|
| 255 | $self->auto_extract_metadata ($doc_obj);
|
|---|
| 256 |
|
|---|
| 257 | # add an OID
|
|---|
| 258 | $doc_obj->set_OID();
|
|---|
| 259 |
|
|---|
| 260 | # $doc_obj->add_text($section, "<pre>\n$text\n</pre>");
|
|---|
| 261 |
|
|---|
| 262 | # process the document
|
|---|
| 263 | $processor->process($doc_obj);
|
|---|
| 264 |
|
|---|
| 265 | if (defined $self->{'tmp_filename'} &&
|
|---|
| 266 | -e $self->{'tmp_filename'}) {
|
|---|
| 267 | util::rm($self->{'tmp_filename'})
|
|---|
| 268 | }
|
|---|
| 269 | if (defined $self->{'tmp_filename2'} &&
|
|---|
| 270 | -e $self->{'tmp_filename2'}) {
|
|---|
| 271 | util::rm($self->{'tmp_filename2'})
|
|---|
| 272 | }
|
|---|
| 273 |
|
|---|
| 274 | return $result; # processed the file
|
|---|
| 275 | }
|
|---|
| 276 |
|
|---|
| 277 | # do plugin specific processing of doc_obj
|
|---|
| 278 | sub process {
|
|---|
| 279 | my $self = shift (@_);
|
|---|
| 280 | my ($textref, $pluginfo, $base_dir, $file, $metadata, $doc_obj) = @_;
|
|---|
| 281 | my $outhandle = $self->{'outhandle'};
|
|---|
| 282 |
|
|---|
| 283 | return 1;
|
|---|
| 284 | }
|
|---|
| 285 |
|
|---|
| 286 | 1;
|
|---|
| 287 |
|
|---|
| 288 |
|
|---|
| 289 |
|
|---|
| 290 |
|
|---|
| 291 |
|
|---|
| 292 |
|
|---|
| 293 |
|
|---|
| 294 |
|
|---|
| 295 |
|
|---|
| 296 |
|
|---|
| 297 |
|
|---|
