source: gsdl/trunk/perllib/plugins/PagedImagePlugin.pm@ 16827

Last change on this file since 16827 was 16827, checked in by davidb, 13 years ago

Edited needed to make pageImagePlugin work again after plugin restructuring

  • Property svn:executable set to *
  • Property svn:keywords set to Author Date Id Revision
File size: 21.5 KB
Line 
1###########################################################################
2#
3# PagedImagePlugin.pm -- plugin for sets of images and OCR text that
4# make up a document
5# A component of the Greenstone digital library software
6# from the New Zealand Digital Library Project at the
7# University of Waikato, New Zealand.
8#
9# Copyright (C) 1999 New Zealand Digital Library Project
10#
11# This program is free software; you can redistribute it and/or modify
12# it under the terms of the GNU General Public License as published by
13# the Free Software Foundation; either version 2 of the License, or
14# (at your option) any later version.
15#
16# This program is distributed in the hope that it will be useful,
17# but WITHOUT ANY WARRANTY; without even the implied warranty of
18# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
19# GNU General Public License for more details.
20#
21# You should have received a copy of the GNU General Public License
22# along with this program; if not, write to the Free Software
23# Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
24#
25###########################################################################
26
27# PagedImagePlugin
28# processes sequences of images, with optional OCR text
29#
30# This plugin takes *.item files, which contain metadata and lists of image
31# files, and produces a document containing sections, one for each page.
32# The files should be named something.item, then you can have more than one
33# book in a directory. You will need to create these files, one for each
34# document/book.
35#
36#There are two formats for the item files: a plain text format, and an xml
37#format. You can use either format, and can have both formats in the same
38#collection if you like. If you use the plain format, you must not start the
39#file off with <PagedDocument>
40
41#### PLAIN FORMAT
42# The format of the xxx.item file is as follows:
43# The first lines contain any metadata for the whole document
44# <metadata-name>metadata-value
45# eg.
46# <Title>Snail farming
47# <Date>19230102
48# Then comes a list of pages, one page per line, each line has the format
49#
50# pagenum:imagefile:textfile:r
51#
52# page num and imagefile are required. pagenum is used for the Title
53# of the section, and in the display is shown as page <pagenum>.
54# imagefile is the image for the page. textfile is an optional text
55# file containing the OCR (or any) text for the page - this gets added
56# as the text for the section. r is optional, and signals that the image
57# should be rotated 180deg. Eg use this if the image has been made upside down.
58# So an example item file looks like:
59# <Title>Snail farming
60# <Date>19960403
61# 1:p1.gif:p1.txt:
62# 2:p2.gif::
63# 3:p3.gif:p3.txt:
64# 3b:p3b.gif:p3b.txt:r
65# The second page has no text, the fourth page is a back page, and
66# should be rotated.
67#
68
69#### XML FORMAT
70# The xml format looks like the following
71#<PagedDocument>
72#<Metadata name="Title">The Title of the entire document</Metadata>
73#<Page pagenum="1" imgfile="xxx.jpg" txtfile="yyy.txt">
74#<Metadata name="Title">The Title of this page</Metadata>
75#</Page>
76#... more pages
77#</PagedDocument>
78#PagedDocument contains a list of Pages, Metadata and PageGroups. Any metadata
79#that is not inside another tag will belong to the document.
80#Each Page has a pagenum (not used at the moment), an imgfile and/or a txtfile.
81#These are both optional - if neither is used, the section will have no content.
82#Pages can also have metadata associated with them.
83#PageGroups can be introduced at any point - they can contain Metadata and Pages and other PageGroups. They are used to introduce hierarchical structure into the document.
84#For example
85#<PagedDocument>
86#<PageGroup>
87#<Page>
88#<Page>
89#</PageGroup>
90#<Page>
91#</PagedDocument>
92#would generate a structure like
93#X
94#--X
95# --X
96# --X
97#--X
98#PageGroup tags can also have imgfile/textfile metadata if you like - this way they get some content themselves.
99
100#Currently the XML structure doesn't work very well with the paged document type, unless you use numerical Titles for each section.
101#There is still a bit of work to do on this format:
102#* enable other text file types, eg html, pdf etc
103#* make the document paging work properly
104#* add pagenum as Title unless a Title is present?
105
106# All the supplemetary image amd text files should be in the same folder as
107# the .item file.
108#
109# To display the images instead of the document text, you can use [srcicon]
110# in the DocumentText format statement.
111# For example,
112#
113# format DocumentText "<center><table width=_pagewidth_><tr><td>[srcicon]</td></tr></table></center>"
114#
115# To have it create thumbnail size images, use the '-create_thumbnail' option.
116# To have it create medium size images for display, use the '-create_screenview'
117# option. As usual, running
118# 'perl -S pluginfo.pl PagedImagePlugin' will list all the options.
119
120# If you want the resulting documents to be presented with a table of
121# contents, use '-documenttype hierarchy', otherwise they will have
122# next and previous arrows, and a goto page X box.
123
124# If you have used -create_screenview, you can also use [screenicon] in the format
125# statement to display the smaller image. Here is an example that switches
126# between the two:
127#
128# format DocumentText "<center><table width=_pagewidth_><tr><td>{If}{_cgiargp_ eq full,<a href='_httpdocument_&d=_cgiargd_&p=small'>Switch to small version.</a>,<a href='_httpdocument_&d=_cgiargd_&p=full'>Switch to fullsize version</a>}</td></tr><tr><td>{If}{_cgiargp_ eq full,<a href='_httpdocument_&d=_cgiargd_&p=small' title='Switch to small version'>[srcicon]</a>,<a href='_httpdocument_&d=_cgiargd_&p=full' title='Switch to fullsize version'>[screenicon]</a>}</td></tr></table></center>"
129#
130# Additional metadata can be added into the .item files, alternatively you can
131# use normal metadata.xml files, with the name of the xxx.item file as the
132# FileName (only for document level metadata).
133
134package PagedImagePlugin;
135
136use ReadXMLFile;
137use ReadTextFile;
138use ImageConverter;
139
140use strict;
141no strict 'refs'; # allow filehandles to be variables and viceversa
142
143sub BEGIN {
144 @PagedImagePlugin::ISA = ('ReadXMLFile', 'ReadTextFile', 'ImageConverter');
145}
146
147my $type_list =
148 [ { 'name' => "paged",
149 'desc' => "{PagedImagePlugin.documenttype.paged}" },
150 { 'name' => "hierarchy",
151 'desc' => "{PagedImagePlugin.documenttype.hierarchy}" } ];
152
153my $arguments =
154 [ { 'name' => "process_exp",
155 'desc' => "{BasePlugin.process_exp}",
156 'type' => "string",
157 'deft' => &get_default_process_exp(),
158 'reqd' => "no" },
159 { 'name' => "block_exp",
160 'desc' => "{BasePlugin.block_exp}",
161 'type' => "string",
162 'deft' => &get_default_block_exp(),
163 'reqd' => "no" },
164 { 'name' => "title_sub",
165 'desc' => "{HTMLPlugin.title_sub}",
166 'type' => "string",
167 'deft' => "" },
168 { 'name' => "headerpage",
169 'desc' => "{PagedImagePlugin.headerpage}",
170 'type' => "flag",
171 'reqd' => "no" },
172 { 'name' => "documenttype",
173 'desc' => "{PagedImagePlugin.documenttype}",
174 'type' => "enum",
175 'list' => $type_list,
176 'deft' => "paged",
177 'reqd' => "no" } ];
178
179
180my $options = { 'name' => "PagedImagePlugin",
181 'desc' => "{PagedImagePlugin.desc}",
182 'abstract' => "no",
183 'inherits' => "yes",
184 'args' => $arguments };
185
186sub new {
187 my ($class) = shift (@_);
188 my ($pluginlist,$inputargs,$hashArgOptLists) = @_;
189 push(@$pluginlist, $class);
190
191 push(@{$hashArgOptLists->{"ArgList"}},@{$arguments});
192 push(@{$hashArgOptLists->{"OptList"}},$options);
193
194 my $imc_self = new ImageConverter($pluginlist, $inputargs, $hashArgOptLists);
195 my $rtf_self = new ReadTextFile($pluginlist, $inputargs, $hashArgOptLists, 1);
196 my $rxf_self = new ReadXMLFile($pluginlist, $inputargs, $hashArgOptLists);
197
198 my $self = BasePlugin::merge_inheritance($imc_self,$rtf_self,$rxf_self);
199
200 # Update $self used by XML::Parser so it finds callback functions
201 # such as start_document here and not in ReadXMLFile (which is what
202 # $self was when new XML::Parser was done)
203 #
204 # If the $self returned by this constructor is the same as the one
205 # used in ReadXMLFile (e.g. in the GAPlugin) then this step isn't necessary
206 #
207 # Consider embedding this type of assignment into merge_inheritance
208 # to help catch all cases?
209
210 $rxf_self->{'parser'}->{'PluginObj'} = $self;
211
212 return bless $self, $class;
213}
214
215
216sub init {
217 my $self = shift (@_);
218 my ($verbosity, $outhandle, $failhandle) = @_;
219
220 $self->SUPER::init(@_);
221 $self->ImageConverter::init();
222}
223
224sub get_default_process_exp {
225 my $self = shift (@_);
226
227 return q^\.item$^;
228}
229
230sub get_doctype {
231 my $self = shift(@_);
232
233 return "PagedDocument";
234}
235
236
237# want to block everything except the .item ones
238# but instead we will block images and txt files
239sub get_default_block_exp {
240 my $self = shift (@_);
241
242 return q^(?i)(\.jpe?g|\.gif|\.png|\.tif?f|\.te?xt|\.html?|~)$^
243### return q^(?i)(\.jpe?g|\.gif|\.png|\.tif?f|\.te?xt|\.html?|\.css|\.opd|\.pdf|~)$^
244}
245
246# we want to use BasePlugin's read, not ReadXMLFile's
247sub read
248{
249 my $self = shift (@_);
250
251 $self->BasePlugin::read(@_);
252}
253
254
255sub read_into_doc_obj {
256 my $self = shift (@_);
257 my ($pluginfo, $base_dir, $file, $block_hash, $metadata, $processor, $maxdocs, $total_count, $gli) = @_;
258 my $outhandle = $self->{'outhandle'};
259 my $verbosity = $self->{'verbosity'};
260
261 my ($filename_full_path, $filename_no_path) = &util::get_full_filenames($base_dir, $file);
262
263 print $outhandle "PagedImagePlugin processing \"$filename_full_path\"\n"
264 if $verbosity > 1;
265 print STDERR "<Processing n='$file' p='PagedImagePlugin'>\n" if ($gli);
266
267
268 # here we need to decide if we have an old text .item file, or a new xml
269 # .item file
270 my $xml_version = $self->is_xml_item_file($filename_full_path);
271
272 $self->tidy_item_file($filename_full_path);
273
274 my $doc_obj;
275 if ($xml_version) {
276 # careful checking needed here!! are we using local xml handlers or super ones
277 $self->ReadXMLFile::read($pluginfo, $base_dir, $file, $block_hash, $metadata, $processor, $maxdocs, $total_count, $gli);
278 $doc_obj = $self->{'doc_obj'};
279 } else {
280 my ($dir);
281 ($dir, $file) = $filename_full_path =~ /^(.*?)([^\/\\]*)$/;
282
283 #process the .item file
284 $doc_obj = $self->process_item($filename_full_path, $dir, $file, $processor);
285
286 }
287
288 my $section = $doc_obj->get_top_section();
289
290 $doc_obj->add_utf8_metadata($section, "Plugin", "$self->{'plugin_type'}");
291 $doc_obj->add_metadata($section, "FileFormat", "PagedImage");
292
293 # include any metadata passed in from previous plugins
294 # note that this metadata is associated with the top level section
295 $self->add_associated_files($doc_obj, $filename_full_path);
296 $self->extra_metadata ($doc_obj, $section, $metadata);
297 $self->auto_extract_metadata ($doc_obj);
298
299 # if we haven't found any Title so far, assign one
300 $self->title_fallback($doc_obj,$section,$filename_no_path);
301
302 $self->add_OID($doc_obj);
303 return (1,$doc_obj);
304}
305
306# for now, the test is if the first non-empty line is <PagedDocument>, then its xml
307sub is_xml_item_file {
308 my $self = shift(@_);
309 my ($filename) = @_;
310
311 my $xml_version = 0;
312 open (ITEMFILE, $filename) || die "couldn't open $filename\n";
313
314 my $line = "";
315 my $num = 0;
316 $line = <ITEMFILE>;
317 while ($line !~ /\w/) {
318 $line = <ITEMFILE>;
319 }
320 chomp $line;
321 if ($line =~ /<PagedDocument/) {
322 $xml_version = 1;
323 }
324 close ITEMFILE;
325 return $xml_version;
326}
327
328sub tidy_item_file {
329 my $self = shift(@_);
330 my ($filename) = @_;
331
332 open (ITEMFILE, $filename) || die "couldn't open $filename\n";
333 my $backup_filename = "backup.item";
334 open (BACKUP,">$backup_filename")|| die "couldn't write to $backup_filename\n";
335 my $line = "";
336 $line = <ITEMFILE>;
337 $line =~ s/^\xEF\xBB\xBF//; # strip BOM
338 $line =~ s/\x0B+//ig;
339 $line =~ s/&/&amp;/g;
340 print BACKUP ($line);
341 #Tidy up the item file some metadata title contains \vt-vertical tab
342 while ($line = <ITEMFILE>) {
343 $line =~ s/\x0B+//ig;
344 $line =~ s/&/&amp;/g;
345 print BACKUP ($line);
346 }
347 close ITEMFILE;
348 close BACKUP;
349 &File::Copy::copy ($backup_filename, $filename);
350 &util::rm($backup_filename);
351
352}
353
354sub rotate_image {
355 my $self = shift (@_);
356 my ($filename_full_path) = @_;
357
358 my ($this_filetype) = $filename_full_path =~ /\.([^\.]*)$/;
359 my $result = $self->convert($filename_full_path, $this_filetype, "-rotate 180", "ROTATE");
360 my ($new_filename) = ($result =~ /=>(.*\.$this_filetype)/);
361 if (-e "$new_filename") {
362 return $new_filename;
363 }
364 # somethings gone wrong
365 return $filename_full_path;
366
367}
368
369sub process_image {
370 my $self = shift(@_);
371 my ($filename_full_path, $filename_no_path, $doc_obj, $section, $rotate) = @_;
372 # do rotation
373 if ((defined $rotate) && ($rotate eq "r")) {
374 # check the filenames
375 return 0 if ($filename_no_path eq "" || !-f $filename_full_path);
376
377 # we get a new temporary file which is rotated
378 $filename_full_path = $self->rotate_image($filename_full_path);
379 }
380
381 # do generate images
382 my $result = 0;
383 if ($self->{'image_conversion_available'} == 1) {
384 # do we need to convert $filename_no_path to utf8? We are already reading in from a file, what encoding is it in???
385 $result = $self->generate_images($filename_full_path, $filename_no_path, $doc_obj, $section);
386 }
387 #overwrite one set in ImageConverter
388 $doc_obj->set_metadata_element ($section, "FileFormat", "PagedImage");
389 return $result;
390}
391
392
393sub xml_start_tag {
394 my $self = shift(@_);
395 my ($expat, $element) = @_;
396 $self->{'element'} = $element;
397
398 my $doc_obj = $self->{'doc_obj'};
399 if ($element eq "PagedDocument") {
400 $self->{'current_section'} = $doc_obj->get_top_section();
401 } elsif ($element eq "PageGroup" || $element eq "Page") {
402 # create a new section as a child
403 $self->{'current_section'} = $doc_obj->insert_section($doc_obj->get_end_child($self->{'current_section'}));
404 $self->{'num_pages'}++;
405 # assign pagenum as what??
406 my $pagenum = $_{'pagenum'}; #TODO!!
407 if (defined $pagenum) {
408 $doc_obj->set_utf8_metadata_element($self->{'current_section'}, 'PageNum', $pagenum);
409 }
410 my ($imgfile) = $_{'imgfile'};
411 if (defined $imgfile) {
412 # *****
413 # What about support for rotate image (e.g. old ':r' notation)?
414 $self->process_image($self->{'xml_file_dir'}.$imgfile, $imgfile, $doc_obj, $self->{'current_section'});
415 }
416 my ($txtfile) = $_{'txtfile'};
417 if (defined($txtfile)&& $txtfile ne "") {
418 $self->process_text ($self->{'xml_file_dir'}.$txtfile, $txtfile, $doc_obj, $self->{'current_section'});
419 } else {
420 $self->add_dummy_text($doc_obj, $self->{'current_section'});
421 }
422 } elsif ($element eq "Metadata") {
423 $self->{'metadata_name'} = $_{'name'};
424 }
425}
426
427sub xml_end_tag {
428 my $self = shift(@_);
429 my ($expat, $element) = @_;
430
431 my $doc_obj = $self->{'doc_obj'};
432 if ($element eq "Page" || $element eq "PageGroup") {
433 # if Title hasn't been assigned, set PageNum as Title
434 if (!defined $doc_obj->get_metadata_element ($self->{'current_section'}, "Title") && defined $doc_obj->get_metadata_element ($self->{'current_section'}, "PageNum" )) {
435 $doc_obj->add_utf8_metadata ($self->{'current_section'}, "Title", $doc_obj->get_metadata_element ($self->{'current_section'}, "PageNum" ));
436 }
437 # move the current section back to the parent
438 $self->{'current_section'} = $doc_obj->get_parent_section($self->{'current_section'});
439 } elsif ($element eq "Metadata") {
440
441 $doc_obj->add_utf8_metadata ($self->{'current_section'}, $self->{'metadata_name'}, $self->{'metadata_value'});
442 $self->{'metadata_name'} = "";
443 $self->{'metadata_value'} = "";
444
445 }
446 # otherwise we ignore the end tag
447}
448
449
450sub xml_text {
451 my $self = shift(@_);
452 my ($expat) = @_;
453
454 if ($self->{'element'} eq "Metadata" && $self->{'metadata_name'}) {
455 $self->{'metadata_value'} .= $_;
456 }
457}
458
459sub xml_doctype {
460}
461
462sub open_document {
463 my $self = shift(@_);
464
465 # create a new document
466 $self->{'doc_obj'} = new doc ($self->{'filename'}, "indexed_doc");
467 # TODO is file filenmae_no_path??
468 $self->set_initial_doc_fields($self->{'doc_obj'}, $self->{'file'}, $self->{'processor'});
469
470 my ($dir, $file) = $self->{'filename'} =~ /^(.*?)([^\/\\]*)$/;
471 $self->{'xml_file_dir'} = $dir;
472 $self->{'num_pages'} = 0;
473
474}
475
476sub close_document {
477 my $self = shift(@_);
478 my $doc_obj = $self->{'doc_obj'};
479
480 # add numpages metadata
481 $doc_obj->set_utf8_metadata_element ($doc_obj->get_top_section(), 'NumPages', $self->{'num_pages'});
482
483
484}
485
486
487sub set_initial_doc_fields {
488 my $self = shift(@_);
489 my ($doc_obj, $filename_no_path, $processor) = @_;
490
491 $doc_obj->set_OIDtype ($processor->{'OIDtype'}, $processor->{'OIDmetadata'});
492 my $topsection = $doc_obj->get_top_section();
493
494 if ($self->{'documenttype'} eq 'paged') {
495 # set the gsdlthistype metadata to Paged - this ensures this document will
496 # be treated as a Paged doc, even if Titles are not numeric
497 $doc_obj->set_utf8_metadata_element ($topsection, "gsdlthistype", "Paged");
498 } else {
499 $doc_obj->set_utf8_metadata_element ($topsection, "gsdlthistype", "Hierarchy");
500 }
501
502 $self->set_Source_metadata($doc_obj, $filename_no_path);
503
504 # if we want a header page, we need to add some text into the top section, otherwise this section will become invisible
505 if ($self->{'headerpage'}) {
506 $self->add_dummy_text($doc_obj, $topsection);
507 }
508
509
510}
511
512
513sub process_item {
514 my $self = shift (@_);
515 my ($filename_full_path, $dir, $filename_no_path, $processor) = @_;
516
517 my $doc_obj = new doc ($filename_full_path, "indexed_doc");
518 $self->set_initial_doc_fields($doc_obj, $filename_no_path, $processor);
519 my $topsection = $doc_obj->get_top_section();
520 open (ITEMFILE, $filename_full_path) || die "couldn't open $filename_full_path\n";
521 my $line = "";
522 my $num = 0;
523 while (defined ($line = <ITEMFILE>)) {
524 next unless $line =~ /\w/;
525 chomp $line;
526 next if $line =~ /^#/; # ignore comment lines
527 if ($line =~ /^<([^>]*)>\s*(.*?)\s*$/) {
528 $doc_obj->set_utf8_metadata_element ($topsection, $1, $2);
529 #$meta->{$1} = $2;
530 } else {
531 $num++;
532 # line should be like page:imagefilename:textfilename:r - the r is optional -> means rotate the image 180 deg
533 $line =~ s/^\s+//; #remove space at the front
534 $line =~ s/\s+$//; #remove space at the end
535 my ($pagenum, $imgname, $txtname, $rotate) = split /:/, $line;
536
537 # create a new section for each image file
538 my $cursection = $doc_obj->insert_section($doc_obj->get_end_child($topsection));
539 # the page number becomes the Title
540 $doc_obj->set_utf8_metadata_element($cursection, 'Title', $pagenum);
541
542 # process the image for this page if there is one
543 if (defined $imgname && $imgname ne "") {
544 my $result1 = $self->process_image($dir.$imgname, $imgname, $doc_obj, $cursection, $rotate);
545
546 if (!defined $result1)
547 {
548 print "PagedImagePlugin: couldn't process image \"$dir.$imgname\" for item \"$filename_full_path\"\n";
549 }
550 }
551 # process the text file if one is there
552 if (defined $txtname && $txtname ne "") {
553 my $result2 = $self->process_text ($dir.$txtname, $txtname, $doc_obj, $cursection);
554
555 if (!defined $result2) {
556 print "PagedImagePlugin: couldn't process text file \"$dir.$txtname\" for item \"$filename_full_path\"\n";
557 $self->add_dummy_text($doc_obj, $cursection);
558 }
559 } else {
560 # otherwise add in some dummy text
561 $self->add_dummy_text($doc_obj, $cursection);
562 }
563 }
564 }
565
566 close ITEMFILE;
567
568 # add numpages metadata
569 $doc_obj->set_utf8_metadata_element ($topsection, 'NumPages', "$num");
570 return $doc_obj;
571}
572
573sub process_text {
574 my $self = shift (@_);
575 my ($filename_full_path, $file, $doc_obj, $cursection) = @_;
576
577 # check that the text file exists!!
578 if (!-f $filename_full_path) {
579 print "PagedImagePlugin: ERROR: File $filename_full_path does not exist, skipping\n";
580 return 0;
581 }
582
583 # Do encoding stuff
584 my ($language, $encoding) = $self->textcat_get_language_encoding ($filename_full_path);
585
586 my $text="";
587 &ReadTextFile::read_file($self, $filename_full_path, $encoding, $language, \$text);
588 if (!length ($text)) {
589 # It's a bit unusual but not out of the question to have no text, so just give a warning
590 print "PagedImagePlugin: WARNING: $filename_full_path contains no text\n";
591 }
592
593 # we need to escape the escape character, or else mg will convert into
594 # eg literal newlines, instead of leaving the text as '\n'
595 $text =~ s/\\/\\\\/g; # macro language
596 $text =~ s/_/\\_/g; # macro language
597
598
599 if ($text =~ m/<html.*?>\s*<head.*?>.*<\/head>\s*<body.*?>(.*)<\/body>\s*<\/html>\s*$/is) {
600 # looks like HTML input
601 # no need to escape < and > or put in <pre> tags
602
603 $text = $1;
604
605 # insert preformat tags and add text to document object
606 $doc_obj->add_utf8_text($cursection, "$text");
607 }
608 else {
609 $text =~ s/</&lt;/g;
610 $text =~ s/>/&gt;/g;
611
612 # insert preformat tags and add text to document object
613 $doc_obj->add_utf8_text($cursection, "<pre>\n$text\n</pre>");
614 }
615
616
617 return 1;
618}
619
620
621sub clean_up_after_doc_obj_processing {
622 my $self = shift(@_);
623
624 $self->ImageConverter::clean_up_temporary_files();
625}
626
6271;
Note: See TracBrowser for help on using the repository browser.