source: gsdl/trunk/perllib/plugins/ImagePlug.pm@ 14664

Last change on this file since 14664 was 14664, checked in by anna, 15 years ago

Undoing commit of r14657.

  • Property svn:keywords set to Author Date Id Revision
File size: 14.7 KB
Line 
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
26package ImagePlug;
27
28use BasPlug;
29
30use strict;
31no strict 'refs'; # allow filehandles to be variables and viceversa
32
33sub BEGIN {
34 @ImagePlug::ISA = ('BasPlug');
35}
36
37my $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
81my $options = { 'name' => "ImagePlug",
82 'desc' => "{ImagePlug.desc}",
83 'abstract' => "no",
84 'inherits' => "yes",
85 'args' => $arguments };
86
87
88
89sub new {
90 my ($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
111sub 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
118sub 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
128sub 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 $type = "unknown";
159
160 if ($converttotype ne "" && $filename !~ m/$converttotype$/) {
161 $originalfilename = $filename;
162 $filename = &util::get_tmp_filename() . ".$converttotype";
163 $self->{'tmp_filename'} = $filename;
164
165 my $command = "convert -interlace plane -verbose \"$originalfilename\" \"$filename\"";
166 print $outhandle "$command\n" if ($verbosity > 2);
167 my $result = '';
168 $result = `$command`;
169 print $outhandle "RESULT = $result\n" if ($verbosity > 2);
170
171 $type = $converttotype;
172 $file =~ s/\..*$/\.$type/;
173 }
174
175
176 # Add the image metadata
177 my $url = $file;
178
179 ##not know why it is required at the first place, it seems all works fine without it, so I comment it out
180 ##$url =~ s/ /%20/g;
181
182 $doc_obj->add_metadata ($section, "Image", $url);
183
184 # Also want to set filename as 'Source' metadata to be
185 # consistent with other plugins
186 $doc_obj->add_metadata ($section, "Source", $url);
187
188 my ($image_type, $image_width, $image_height, $image_size)
189 = &identify($filename, $outhandle, $verbosity);
190
191 if ($image_type ne " ") {
192 $type = $image_type;
193 }
194
195 $doc_obj->add_metadata ($section, "FileFormat", $type);
196 $doc_obj->add_metadata ($section, "FileSize", $image_size);
197
198 $doc_obj->add_metadata ($section, "ImageType", $image_type);
199 $doc_obj->add_metadata ($section, "ImageWidth", $image_width);
200 $doc_obj->add_metadata ($section, "ImageHeight", $image_height);
201 $doc_obj->add_metadata ($section, "ImageSize", $image_size);
202 $doc_obj->add_metadata ($section, "NoText", "1");
203
204 $doc_obj->add_metadata ($section, "srclink",
205 "<a href=\"_httpprefix_/collect/[collection]/index/assoc/[assocfilepath]/[Image]\">");
206 $doc_obj->add_metadata ($section, "/srclink", "</a>");
207
208 $doc_obj->add_metadata ($section, "srcicon", "<img src=\"_httpprefix_/collect/[collection]/index/assoc/[assocfilepath]/[Image]\" width=100>");
209
210
211 # Add the image as an associated file
212 $doc_obj->associate_file($filename,$file,"image/$type",$section);
213
214
215 # Make the thumbnail image
216 my $thumbnailsize = $self->{'thumbnailsize'} || 100;
217 my $thumbnailtype = $self->{'thumbnailtype'} || 'gif';
218
219 my $thumbnailfile = &util::get_tmp_filename() . ".$thumbnailtype";
220 $self->{'tmp_filename2'} = $thumbnailfile;
221
222 # Generate the thumbnail with convert
223 my $command = "convert -interlace plane -verbose -geometry $thumbnailsize"
224 . "x$thumbnailsize \"$filename\" \"$thumbnailfile\"";
225 print $outhandle "THUMBNAIL: $command\n" if ($verbosity > 2);
226 my $result = '';
227 $result = `$command 2>&1` ;
228 print $outhandle "THUMB RESULT: $result\n" if ($verbosity > 2);
229
230 # Add the thumbnail as an associated file ...
231 if (-e "$thumbnailfile") {
232 $doc_obj->associate_file("$thumbnailfile", "thumbnail.$thumbnailtype",
233 "image/$thumbnailtype",$section);
234 $doc_obj->add_metadata ($section, "ThumbType", $thumbnailtype);
235 $doc_obj->add_metadata ($section, "Thumb", "thumbnail.$thumbnailtype");
236
237 $doc_obj->add_metadata ($section, "thumbicon", "<img src=\"_httpprefix_/collect/[collection]/index/assoc/[assocfilepath]/[Thumb]\" width=[ThumbWidth] height=[ThumbHeight]>");
238 }
239
240 # Extract Thumnail metadata from convert output
241 if ($result =~ m/[0-9]+x[0-9]+=>([0-9]+)x([0-9]+)/) {
242 $doc_obj->add_metadata ($section, "ThumbWidth", $1);
243 $doc_obj->add_metadata ($section, "ThumbHeight", $2);
244 }
245
246 # Make a screen-sized version of the picture if requested
247 if ($self->{'screenviewsize'}) {
248
249 # To do: if the actual image smaller than the screenview size,
250 # we should use the original !
251
252 my $screenviewsize = $self->{'screenviewsize'};
253 my $screenviewtype = $self->{'screenviewtype'} || 'jpeg';
254 my $screenviewfilename = &util::get_tmp_filename() . ".$screenviewtype";
255 $self->{'tmp_filename3'} = $screenviewfilename;
256
257 # make the screenview image
258 my $command = "convert -interlace plane -verbose -geometry $screenviewsize"
259 . "x$screenviewsize \"$filename\" \"$screenviewfilename\"";
260 print $outhandle "$command\n" if ($verbosity > 2);
261 my $result = "";
262 $result = `$command 2>&1` ;
263 print $outhandle "$result\n" if ($verbosity > 3);
264
265 # get screenview dimensions, size and type
266 if ($result =~ m/[0-9]+x[0-9]+=>([0-9]+)x([0-9]+)/) {
267 $doc_obj->add_metadata ($section, "ScreenWidth", $1);
268 $doc_obj->add_metadata ($section, "ScreenHeight", $2);
269 }
270 else {
271 $doc_obj->add_metadata ($section, "ScreenWidth", $image_width);
272 $doc_obj->add_metadata ($section, "ScreenHeight", $image_height);
273 }
274
275 #add the screenview as an associated file ...
276 if (-e "$screenviewfilename") {
277 $doc_obj->associate_file("$screenviewfilename", "screenview.$screenviewtype",
278 "image/$screenviewtype",$section);
279 $doc_obj->add_metadata ($section, "ScreenType", $screenviewtype);
280 $doc_obj->add_metadata ($section, "Screen", "screenview.$screenviewtype");
281
282 $doc_obj->add_metadata ($section, "screenicon", "<img src=\"_httpprefix_/collect/[collection]/index/assoc/[assocfilepath]/[Screen]\" width=[ScreenWidth] height=[ScreenHeight]>");
283 } else {
284 print $outhandle "ImagePlug: couldn't find \"$screenviewfilename\"\n";
285 }
286 }
287
288 return $type;
289
290
291}
292
293
294
295# Discover the characteristics of an image file with the ImageMagick
296# "identify" command.
297
298sub identify {
299 my ($image, $outhandle, $verbosity) = @_;
300
301 # Use the ImageMagick "identify" command to get the file specs
302 my $command = "identify \"$image\" 2>&1";
303 print $outhandle "$command\n" if ($verbosity > 2);
304 my $result = '';
305 $result = `$command`;
306 print $outhandle "$result\n" if ($verbosity > 3);
307
308 # Read the type, width, and height
309 my $type = 'unknown';
310 my $width = 'unknown';
311 my $height = 'unknown';
312
313 my $image_safe = quotemeta $image;
314 if ($result =~ /^$image_safe (\w+) (\d+)x(\d+)/) {
315 $type = $1;
316 $width = $2;
317 $height = $3;
318 }
319
320 # Read the size
321 my $size = "unknown";
322 if ($result =~ m/^.* ([0-9]+)b/) {
323 $size = $1;
324 }
325 elsif ($result =~ m/^.* ([0-9]+)(\.([0-9]+))?kb?/) {
326 $size = 1024 * $1;
327 if (defined($2)) {
328 $size = $size + (1024 * $2);
329 # Truncate size (it isn't going to be very accurate anyway)
330 $size = int($size);
331 }
332 }
333
334 print $outhandle "file: $image:\t $type, $width, $height, $size\n"
335 if ($verbosity > 2);
336
337 # Return the specs
338 return ($type, $width, $height, $size);
339}
340
341
342# The ImagePlug read() function.
343# ImagePlug overrides read() because there is no need to read the actual
344# text of the file in, because the contents of the file is not text...
345#
346# Return number of files processed, undef if can't process
347# Note that $base_dir might be "" and that $file might
348# include directories
349
350sub read {
351 my $self = shift (@_);
352 my ($pluginfo, $base_dir, $file, $metadata, $processor, $maxdocs, $total_count, $gli) = @_;
353
354 my $outhandle = $self->{'outhandle'};
355
356 #check process and block exps, smart block, etc
357 my ($block_status,$filename) = $self->read_block(@_);
358 return $block_status if ((!defined $block_status) || ($block_status==0));
359
360 print STDERR "<Processing n='$file' p='ImagePlug'>\n" if ($gli);
361 print $outhandle "ImagePlug processing $file\n"
362 if $self->{'verbosity'} > 1;
363
364 # None of this works very well on Windows 95/98...
365 if ($ENV{'GSDLOS'} eq "windows" && !Win32::IsWinNT()) {
366 if ($gli) {
367 print STDERR "<ProcessingError n='$file' r='Windows 95/98 not supported'>\n";
368 }
369 print $outhandle "ImagePlug: Windows 95/98 not supported\n";
370 return -1;
371 }
372
373 # None of this is going to work very well without ImageMagick...
374 if ($self->{'imagemagick_not_installed'}) {
375 if ($gli) {
376 print STDERR "<ProcessingError n='$file' r='ImageMagick not installed'>\n";
377 }
378 print $outhandle "ImagePlug: ImageMagick not installed\n";
379 return -1;
380 }
381
382 #if there's a leading directory name, eat it...
383 $file =~ s/^.*[\/\\]//;
384
385 # create a new document
386 my $doc_obj = new doc ($filename, "indexed_doc");
387 $doc_obj->set_OIDtype ($processor->{'OIDtype'}, $processor->{'OIDmetadata'});
388 $doc_obj->add_utf8_metadata($doc_obj->get_top_section(), "Plugin", "$self->{'plugin_type'}");
389
390 #run convert to get the thumbnail and extract size and type info
391 my $result = run_convert($self, $filename, $file, $doc_obj);
392
393 if (!defined $result)
394 {
395 if ($gli) {
396 print STDERR "<ProcessingError n='$file'>\n";
397 }
398 print $outhandle "ImagePlug: couldn't process \"$filename\"\n";
399 return -1; # error during processing
400 }
401
402
403 #create an empty text string so we don't break downstream plugins
404 my $text = &gsprintf::lookup_string("{BasPlug.dummy_text}",1);
405
406 # include any metadata passed in from previous plugins
407 # note that this metadata is associated with the top level section
408 my $section = $doc_obj->get_top_section();
409 $self->extra_metadata ($doc_obj, $section, $metadata);
410
411 # do plugin specific processing of doc_obj
412 unless (defined ($self->process(\$text, $pluginfo, $base_dir, $file, $metadata, $doc_obj))) {
413 print STDERR "<ProcessingError n='$file'>\n" if ($gli);
414 return -1;
415 }
416
417 # do any automatic metadata extraction
418 $self->auto_extract_metadata ($doc_obj);
419
420 # if we haven't found any Title so far, assign one
421 # this was shifted to here from inside read()
422 $self->title_fallback($doc_obj,$section,$file);
423 # add an OID
424 $doc_obj->set_OID();
425 $doc_obj->add_utf8_text($section, $text);
426
427 # process the document
428 $processor->process($doc_obj);
429
430 # clean up temporary files - we do this here instead of in
431 # run_convert becuase associated files aren't actually copied
432 # until after process has been run.
433 if (defined $self->{'tmp_filename'} &&
434 -e $self->{'tmp_filename'}) {
435 &util::rm($self->{'tmp_filename'})
436 }
437
438 if (defined $self->{'tmp_filename2'} &&
439 -e $self->{'tmp_filename2'}) {
440 &util::rm($self->{'tmp_filename2'})
441 }
442 if (defined $self->{'tmp_filename3'} &&
443 -e $self->{'tmp_filename3'}) {
444 &util::rm($self->{'tmp_filename3'})
445 }
446
447 $self->{'num_processed'}++;
448
449 return 1;
450}
451
452# do plugin specific processing of doc_obj
453sub process {
454 my $self = shift (@_);
455 my ($textref, $pluginfo, $base_dir, $file, $metadata, $doc_obj) = @_;
456 my $outhandle = $self->{'outhandle'};
457
458 return 1;
459}
460
4611;
462
463
464
465
466
467
468
469
470
471
472
Note: See TracBrowser for help on using the repository browser.