source: main/trunk/greenstone2/perllib/plugins/HTMLPlugin.pm@ 32325

Last change on this file since 32325 was 32325, checked in by ak19, 6 years ago

Dr Bainbridge worked out the solution to HTMLPlugin not handling embedded base64 encoded images. Fixed!
  • Property svn:executable set to *
  • Property svn:keywords set to Author Date Id Revision
File size: 70.9 KB
Line 
1###########################################################################
2#
3# HTMLPlugin.pm -- basic html plugin
4#
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#
28# Note that this plugin handles frames only in a very simple way
29# i.e. each frame is treated as a separate document. This means
30# search results will contain links to individual frames rather
31# than linking to the top level frameset.
32# There may also be some problems caused by the _parent target
33# (it's removed by this plugin)
34#
35
36package HTMLPlugin;
37
38use Encode;
39use Unicode::Normalize 'normalize';
40
41use ReadTextFile;
42use HBPlugin;
43use ghtml;
44use unicode;
45use util;
46use FileUtils;
47use XMLParser;
48
49use File::Copy;
50
51sub BEGIN {
52 @HTMLPlugin::ISA = ('ReadTextFile', 'HBPlugin');
53 die "GSDLHOME not set\n" unless defined $ENV{'GSDLHOME'};
54 unshift (@INC, "$ENV{'GSDLHOME'}/perllib/cpan"); # for Image/Size.pm
55}
56
57use Image::Size;
58
59use strict; # every perl program should have this!
60no strict 'refs'; # make an exception so we can use variables as filehandles
61
62my $arguments =
63 [ { 'name' => "process_exp",
64 'desc' => "{BaseImporter.process_exp}",
65 'type' => "regexp",
66 'deft' => &get_default_process_exp() },
67 { 'name' => "block_exp",
68 'desc' => "{CommonUtil.block_exp}",
69 'type' => 'regexp',
70 'deft' => &get_default_block_exp() },
71 { 'name' => "nolinks",
72 'desc' => "{HTMLPlugin.nolinks}",
73 'type' => "flag" },
74 { 'name' => "keep_head",
75 'desc' => "{HTMLPlugin.keep_head}",
76 'type' => "flag" },
77 { 'name' => "no_metadata",
78 'desc' => "{HTMLPlugin.no_metadata}",
79 'type' => "flag" },
80 { 'name' => "metadata_fields",
81 'desc' => "{HTMLPlugin.metadata_fields}",
82 'type' => "string",
83 'deft' => "Title" },
84 { 'name' => "metadata_field_separator",
85 'desc' => "{HTMLPlugin.metadata_field_separator}",
86 'type' => "string",
87 'deft' => "" },
88 { 'name' => "hunt_creator_metadata",
89 'desc' => "{HTMLPlugin.hunt_creator_metadata}",
90 'type' => "flag" },
91 { 'name' => "file_is_url",
92 'desc' => "{HTMLPlugin.file_is_url}",
93 'type' => "flag" },
94 { 'name' => "assoc_files",
95 'desc' => "{HTMLPlugin.assoc_files}",
96 'type' => "regexp",
97 'deft' => &get_default_block_exp() },
98 { 'name' => "rename_assoc_files",
99 'desc' => "{HTMLPlugin.rename_assoc_files}",
100 'type' => "flag" },
101 { 'name' => "title_sub",
102 'desc' => "{HTMLPlugin.title_sub}",
103 'type' => "string",
104 'deft' => "" },
105 { 'name' => "description_tags",
106 'desc' => "{HTMLPlugin.description_tags}",
107 'type' => "flag" },
108 # retain this for backward compatibility (w3mir option was replaced by
109 # file_is_url)
110 { 'name' => "w3mir",
111# 'desc' => "{HTMLPlugin.w3mir}",
112 'type' => "flag",
113 'hiddengli' => "yes"},
114 { 'name' => "no_strip_metadata_html",
115 'desc' => "{HTMLPlugin.no_strip_metadata_html}",
116 'type' => "string",
117 'deft' => "",
118 'reqd' => "no"},
119 { 'name' => "sectionalise_using_h_tags",
120 'desc' => "{HTMLPlugin.sectionalise_using_h_tags}",
121 'type' => "flag" },
122 { 'name' => "use_realistic_book",
123 'desc' => "{HTMLPlugin.tidy_html}",
124 'type' => "flag"},
125 { 'name' => "old_style_HDL",
126 'desc' => "{HTMLPlugin.old_style_HDL}",
127 'type' => "flag"},
128 {'name' => "processing_tmp_files",
129 'desc' => "{BaseImporter.processing_tmp_files}",
130 'type' => "flag",
131 'hiddengli' => "yes"}
132 ];
133
134my $options = { 'name' => "HTMLPlugin",
135 'desc' => "{HTMLPlugin.desc}",
136 'abstract' => "no",
137 'inherits' => "yes",
138 'args' => $arguments };
139
140
141sub new {
142 my ($class) = shift (@_);
143 my ($pluginlist,$inputargs,$hashArgOptLists) = @_;
144 push(@$pluginlist, $class);
145
146 push(@{$hashArgOptLists->{"ArgList"}},@{$arguments});
147 push(@{$hashArgOptLists->{"OptList"}},$options);
148
149
150 my $self = new ReadTextFile($pluginlist,$inputargs,$hashArgOptLists);
151
152 if ($self->{'w3mir'}) {
153 $self->{'file_is_url'} = 1;
154 }
155 $self->{'aux_files'} = {};
156 $self->{'dir_num'} = 0;
157 $self->{'file_num'} = 0;
158
159 return bless $self, $class;
160}
161
162# may want to use (?i)\.(gif|jpe?g|jpe|png|css|js(?:@.*)?)$
163# if have eg <script language="javascript" src="img/lib.js@123">
164# blocking is now done by reading through the file and recording all the
165# images and other files
166sub get_default_block_exp {
167 my $self = shift (@_);
168
169 #return q^(?i)\.(gif|jpe?g|jpe|jpg|png|css)$^;
170 return "";
171}
172
173sub get_default_process_exp {
174 my $self = shift (@_);
175
176 # the last option is an attempt to encode the concept of an html query ...
177 return q^(?i)(\.html?|\.shtml|\.shm|\.asp|\.php\d?|\.cgi|.+\?.+=.*)$^;
178}
179
180sub store_block_files
181{
182 my $self =shift (@_);
183 my ($filename_full_path, $block_hash) = @_;
184
185 my $html_fname = $filename_full_path;
186
187 my ($language, $content_encoding) = $self->textcat_get_language_encoding ($filename_full_path);
188 $self->{'store_content_encoding'}->{$filename_full_path} = [$content_encoding, $language];
189
190
191 # read in file ($text will be in the filesystem encoding)
192 my $raw_text = "";
193 $self->read_file_no_decoding($filename_full_path, \$raw_text);
194
195 my $textref = \$raw_text;
196 my $opencom = '(?:<!--|&lt;!(?:&mdash;|&#151;|--))';
197 my $closecom = '(?:-->|(?:&mdash;|&#151;|--)&gt;)';
198 $$textref =~ s/$opencom(.*?)$closecom//gs;
199
200 # Convert entities to their UTF8 equivalents
201 $$textref =~ s/&(lt|gt|amp|quot|nbsp);/&z$1;/go;
202 $$textref =~ s/&([^;]+);/&ghtml::getcharequiv($1,1,0)/gseo; # on this occassion, want it left as utf8
203 $$textref =~ s/&z(lt|gt|amp|quot|nbsp);/&$1;/go;
204
205 my $attval = "\\\"[^\\\"]+\\\"|[^\\s>]+";
206 my @img_matches = ($$textref =~ m/<img[^>]*?src\s*=\s*($attval)[^>]*>/igs);
207 my @usemap_matches = ($$textref =~ m/<img[^>]*?usemap\s*=\s*($attval)[^>]*>/igs);
208 my @link_matches = ($$textref =~ m/<link[^>]*?href\s*=\s*($attval)[^>]*>/igs);
209 my @embed_matches = ($$textref =~ m/<embed[^>]*?src\s*=\s*($attval)[^>]*>/igs);
210 my @tabbg_matches = ($$textref =~ m/<(?:body|table|tr|td)[^>]*?background\s*=\s*($attval)[^>]*>/igs);
211 my @script_matches = ($$textref =~ m/<script[^>]*?src\s*=\s*($attval)[^>]*>/igs);
212
213 if(!defined $self->{'unicode_to_original_filename'}) {
214 # maps from utf8 converted link name -> original filename referrred to by (possibly URL-encoded) src url
215 $self->{'unicode_to_original_filename'} = {};
216 }
217
218 foreach my $raw_link (@img_matches, @usemap_matches, @link_matches, @embed_matches, @tabbg_matches, @script_matches) {
219
220 # remove quotes from link at start and end if necessary
221 if ($raw_link =~ m/^\"/) {
222 $raw_link =~ s/^\"//;
223 $raw_link =~ s/\"$//;
224 }
225
226 # remove any anchor names, e.g. foo.html#name becomes foo.html
227 # but watch out for any #'s that are part of entities, such as &#x3B1;
228 $raw_link =~ s/([^&])\#.*$/$1/s;
229
230 # some links may just be anchor names
231 next unless ($raw_link =~ /\S+/);
232 # don't block embedded images, which start with src="data:image..."
233 next if($raw_link =~ m@data\:image@);
234
235 if ($raw_link !~ m@^/@ && $raw_link !~ m/^([A-Z]:?)\\/i) {
236 # Turn relative file path into full path
237 my $dirname = &File::Basename::dirname($filename_full_path);
238 $raw_link = &FileUtils::filenameConcatenate($dirname, $raw_link);
239 }
240 $raw_link = $self->eval_dir_dots($raw_link);
241
242 # this is the actual filename on the filesystem (that the link refers to)
243 my $url_original_filename = $self->opt_url_decode($raw_link);
244
245 my ($uses_bytecodes,$exceeds_bytecodes) = &unicode::analyze_raw_string($url_original_filename);
246
247 if ($exceeds_bytecodes) {
248 # We have a link to a file name that is more complicated than a raw byte filename
249 # What we do next depends on the operating system we are on
250
251 if ($ENV{'GSDLOS'} =~ /^(linux|solaris)$/i) {
252 # Assume we're dealing with a UTF-8 encoded filename
253 $url_original_filename = encode("utf8", $url_original_filename);
254 }
255 elsif ($ENV{'GSDLOS'} =~ /^darwin$/i) {
256 # HFS+ is UTF8 with decompostion
257 $url_original_filename = encode("utf8", $url_original_filename);
258 $url_original_filename = normalize('D', $url_original_filename); # Normalization Form D (decomposition)
259 }
260 elsif ($ENV{'GSDLOS'} =~ /^windows$/i) {
261 # Don't need to do anything as later code maps Windows
262 # unicode filenames to DOS short filenames when needed
263 }
264 else {
265 my $outhandle = $self->{'outhandle'};
266 print $outhandle "Warning: Unrecognized operating system ", $ENV{'GSDLOS'}, "\n";
267 print $outhandle " in raw file system encoding of: $raw_link\n";
268 print $outhandle " Assuming filesystem is UTF-8 based.\n";
269 $url_original_filename = encode("utf8", $url_original_filename);
270 }
271 }
272
273 # Convert the (currently raw) link into its Unicode version.
274 # Store the Unicode link along with the url_original_filename
275 my $unicode_url_original_filename = "";
276 $self->decode_text($raw_link,$content_encoding,$language,\$unicode_url_original_filename);
277
278
279 $self->{'unicode_to_original_filename'}->{$unicode_url_original_filename} = $url_original_filename;
280
281
282 if ($url_original_filename ne $unicode_url_original_filename) {
283 my $outhandle = $self->{'outhandle'};
284
285 print $outhandle "URL Encoding $url_original_filename\n";
286 print $outhandle " ->$unicode_url_original_filename\n";
287
288 # make sure not to block the file itself, as happens when an html file links to itself
289 # e.g. if the current file is mary-boleyn/index.html and contains <link rel="canonical" href="index.html" />
290 my $unicode_html_fname = "";
291 $self->decode_text($html_fname,$content_encoding,$language,\$unicode_html_fname);
292 if($unicode_url_original_filename ne $unicode_html_fname) {
293 # Allow for possibility of raw byte version and Unicode versions of file
294 $self->block_filename($block_hash,$unicode_url_original_filename);
295 }
296 }
297
298 # $url_original_filename = &util::upgrade_if_dos_filename($url_original_filename);
299 # TODO now use unicode files in block, do we need this??
300 $self->block_raw_filename($block_hash,$url_original_filename) if $url_original_filename ne $html_fname;
301
302 # but only add the linked file to the blocklist if the current html file does not link to itself
303
304 }
305}
306
307# Given a filename in any encoding, will URL decode it to get back the original filename
308# in the original encoding. Because this method is intended to work out the *original*
309# filename*, it does not URL decode any filename if a file by the name of the *URL-encoded*
310# string already exists in the local folder.
311#
312sub opt_url_decode {
313 my $self = shift (@_);
314 my ($raw_link) = @_;
315
316
317 # Replace %XX's in URL with decoded value if required.
318 # Note that the filename may include the %XX in some situations
319
320## if ($raw_link =~ m/\%[A-F0-9]{2}/i) {
321
322 if (($raw_link =~ m/\%[A-F0-9]{2}/i) || ($raw_link =~ m/\&\#x[0-9A-F]+;/i) || ($raw_link =~ m/\&\#[0-9]+;/i)) {
323 if (!-e $raw_link) {
324 $raw_link = &unicode::url_decode($raw_link,1);
325 }
326 }
327
328 return $raw_link;
329}
330
331sub read_into_doc_obj
332{
333 my $self = shift (@_);
334 my ($pluginfo, $base_dir, $file, $block_hash, $metadata, $processor, $maxdocs, $total_count, $gli) = @_;
335
336 my ($filename_full_path, $filename_no_path) = &util::get_full_filenames($base_dir, $file);
337
338 # Lookup content_encoding and language worked out in file_block pass for this file
339 # Store them under the local names they are nice and easy to access
340 $self->{'content_encoding'} = $self->{'store_content_encoding'}->{$filename_full_path}[0];
341 $self->{'language'} = $self->{'store_content_encoding'}->{$filename_full_path}[1];
342
343 # get the input file
344 my $input_filename = $file;
345 my ($tailname, $dirname, $suffix) = &File::Basename::fileparse($input_filename, "\\.[^\\.]+\$");
346 $suffix = lc($suffix);
347 my $tidy_filename;
348 if (($self->{'use_realistic_book'}) || ($self->{'old_style_HDL'}))
349 {
350 # because the document has to be sectionalized set the description tags
351 $self->{'description_tags'} = 1;
352
353 # set the file to be tidied
354 $input_filename = &FileUtils::filenameConcatenate($base_dir,$file) if $base_dir =~ m/\w/;
355
356 # get the tidied file
357 #my $tidy_filename = $self->tmp_tidy_file($input_filename);
358 $tidy_filename = $self->convert_tidy_or_oldHDL_file($input_filename);
359
360 # derive tmp filename from input filename
361 my ($tailname, $dirname, $suffix) = &File::Basename::fileparse($tidy_filename, "\\.[^\\.]+\$");
362
363 # set the new input file and base_dir to be from the tidied file
364 $file = "$tailname$suffix";
365 $base_dir = $dirname;
366 }
367
368 # call the parent read_into_doc_obj
369 my ($process_status,$doc_obj) = $self->SUPER::read_into_doc_obj($pluginfo, $base_dir, $file, $block_hash, $metadata, $processor, $maxdocs, $total_count, $gli);
370 if (($self->{'use_realistic_book'}) || ($self->{'old_style_HDL'}))
371 {
372 # now we need to reset the filenames in the doc obj so that the converted filenames are not used
373 my $collect_file = &util::filename_within_collection($filename_full_path);
374 $doc_obj->set_source_filename ($collect_file, $self->{'file_rename_method'});
375 ## set_source_filename does not set the doc_obj source_path which is used in archives dbs for incremental
376 # build. So set it manually.
377 $doc_obj->set_source_path($filename_full_path);
378 my $collect_conv_file = &util::filename_within_collection($tidy_filename);
379 $doc_obj->set_converted_filename($collect_conv_file);
380
381 my $plugin_filename_encoding = $self->{'filename_encoding'};
382 my $filename_encoding = $self->deduce_filename_encoding($file,$metadata,$plugin_filename_encoding);
383 $self->set_Source_metadata($doc_obj, $filename_full_path, $filename_encoding);
384 }
385
386 delete $self->{'store_content_encoding'}->{$filename_full_path};
387 $self->{'content_encoding'} = undef;
388
389 return ($process_status,$doc_obj);
390}
391
392# do plugin specific processing of doc_obj
393sub process {
394 my $self = shift (@_);
395 my ($textref, $pluginfo, $base_dir, $file, $metadata, $doc_obj, $gli) = @_;
396 my $outhandle = $self->{'outhandle'};
397
398 if ($ENV{'GSDLOS'} =~ m/^windows/i) {
399 # this makes life so much easier... perl can cope with unix-style '/'s.
400 $base_dir =~ s@(\\)+@/@g;
401 $file =~ s@(\\)+@/@g;
402 }
403
404 my $filename = &FileUtils::filenameConcatenate($base_dir,$file);
405 my $upgraded_base_dir = &util::upgrade_if_dos_filename($base_dir);
406 my $upgraded_filename = &util::upgrade_if_dos_filename($filename);
407
408 if ($ENV{'GSDLOS'} =~ m/^windows/i) {
409 # And again
410 $upgraded_base_dir =~ s@(\\)+@/@g;
411 $upgraded_filename =~ s@(\\)+@/@g;
412
413 # Need to make sure there is a '/' on the end of upgraded_base_dir
414 if (($upgraded_base_dir ne "") && ($upgraded_base_dir !~ m/\/$/)) {
415 $upgraded_base_dir .= "/";
416 }
417 }
418 my $upgraded_file = &util::filename_within_directory($upgraded_filename,$upgraded_base_dir);
419
420 # reset per-doc stuff...
421 $self->{'aux_files'} = {};
422 $self->{'dir_num'} = 0;
423 $self->{'file_num'} = 0;
424
425 # process an HTML file where sections are divided by headings tags (H1, H2 ...)
426 # you can also include metadata in the format (X can be any number)
427 # <hX>Title<!--gsdl-metadata
428 # <Metadata name="name1">value1</Metadata>
429 # ...
430 # <Metadata name="nameN">valueN</Metadata>
431 #--></hX>
432 if ($self->{'sectionalise_using_h_tags'}) {
433 # description_tags should allways be activated because we convert headings to description tags
434 $self->{'description_tags'} = 1;
435
436 my $arrSections = [];
437 $$textref =~ s/<h([0-9]+)[^>]*>(.*?)<\/h[0-9]+>/$self->process_heading($1, $2, $arrSections, $upgraded_file)/isge;
438
439 if (scalar(@$arrSections)) {
440 my $strMetadata = $self->update_section_data($arrSections, -1);
441 if (length($strMetadata)) {
442 $strMetadata = '<!--' . $strMetadata . "\n-->\n</body>";
443 $$textref =~ s/<\/body>/$strMetadata/ig;
444 }
445 }
446 }
447
448 my $cursection = $doc_obj->get_top_section();
449
450 $self->extract_metadata ($textref, $metadata, $doc_obj, $cursection)
451 unless $self->{'no_metadata'} || $self->{'description_tags'};
452
453 # Store URL for page as metadata - this can be used for an
454 # altavista style search interface. The URL won't be valid
455 # unless the file structure contains the domain name (i.e.
456 # like when w3mir is used to download a website).
457
458 # URL metadata (even invalid ones) are used to support internal
459 # links, so even if 'file_is_url' is off, still need to store info
460
461 my ($tailname,$dirname) = &File::Basename::fileparse($upgraded_file);
462
463# my $utf8_file = $self->filename_to_utf8_metadata($file);
464# $utf8_file =~ s/&\#095;/_/g;
465# variable below used to be utf8_file
466
467 my $url_encoded_file = &unicode::raw_filename_to_url_encoded($tailname);
468 my $utf8_url_encoded_file = &unicode::raw_filename_to_utf8_url_encoded($tailname);
469
470 my $web_url = "http://";
471 my $utf8_web_url = "http://";
472
473 if(defined $dirname) { # local directory
474
475 # Check for "ftp" in the domain name of the directory
476 # structure to determine if this URL should be a ftp:// URL
477 # This check is not infallible, but better than omitting the
478 # check, which would cause all files downloaded from ftp sites
479 # via mirroring with wget to have potentially erroneous http:// URLs
480 # assigned in their metadata
481 if ($dirname =~ /^[^\/]*ftp/i)
482 {
483 $web_url = "ftp://";
484 $utf8_web_url = "ftp://";
485 }
486 $dirname = $self->eval_dir_dots($dirname);
487 $dirname .= &util::get_dirsep() if $dirname ne ""; # if there's a directory, it should end on "/"
488
489 # this local directory in import may need to be URL encoded like the file
490 my $url_encoded_dir = &unicode::raw_filename_to_url_encoded($dirname);
491 my $utf8_url_encoded_dir = &unicode::raw_filename_to_utf8_url_encoded($dirname);
492
493 # changed here
494 $web_url = $web_url.$url_encoded_dir.$url_encoded_file;
495 $utf8_web_url = $utf8_web_url.$utf8_url_encoded_dir.$utf8_url_encoded_file;
496 } else {
497 $web_url = $web_url.$url_encoded_file;
498 $utf8_web_url = $utf8_web_url.$utf8_url_encoded_file;
499 }
500 $web_url =~ s/\\/\//g;
501 $utf8_web_url =~ s/\\/\//g;
502
503 if ((defined $ENV{"DEBUG_UNICODE"}) && ($ENV{"DEBUG_UNICODE"})) {
504 print STDERR "*******DEBUG: upgraded_file: $upgraded_file\n";
505 print STDERR "*******DEBUG: adding URL metadata: $utf8_url_encoded_file\n";
506 print STDERR "*******DEBUG: web url: $web_url\n";
507 print STDERR "*******DEBUG: utf8 web url: $utf8_web_url\n";
508 }
509
510
511 $doc_obj->add_utf8_metadata($cursection, "URL", $web_url);
512 $doc_obj->add_utf8_metadata($cursection, "UTF8URL", $utf8_web_url);
513
514 if ($self->{'file_is_url'}) {
515 $doc_obj->add_metadata($cursection, "weblink", "<a href=\"$web_url\">");
516 $doc_obj->add_metadata($cursection, "webicon", "_iconworld_");
517 $doc_obj->add_metadata($cursection, "/weblink", "</a>");
518 }
519
520 if ($self->{'description_tags'}) {
521 # remove the html header - note that doing this here means any
522 # sections defined within the header will be lost (so all <Section>
523 # tags must appear within the body of the HTML)
524 my ($head_keep) = ($$textref =~ m/^(.*?)<body[^>]*>/is);
525
526 $$textref =~ s/^.*?<body[^>]*>//is;
527 $$textref =~ s/(<\/body[^>]*>|<\/html[^>]*>)//isg;
528
529 my $opencom = '(?:<!--|&lt;!(?:&mdash;|&#151;|--))';
530 my $closecom = '(?:-->|(?:&mdash;|&#151;|--)&gt;)';
531
532 my $lt = '(?:<|&lt;)';
533 my $gt = '(?:>|&gt;)';
534 my $quot = '(?:"|&quot;|&rdquo;|&ldquo;)';
535
536 my $dont_strip = '';
537 if ($self->{'no_strip_metadata_html'}) {
538 ($dont_strip = $self->{'no_strip_metadata_html'}) =~ s{,}{|}g;
539 }
540
541 my $found_something = 0; my $top = 1;
542 while ($$textref =~ s/^(.*?)$opencom(.*?)$closecom//s) {
543 my $text = $1;
544 my $comment = $2;
545 if (defined $text) {
546 # text before a comment - note that getting to here
547 # doesn't necessarily mean there are Section tags in
548 # the document
549 $self->process_section(\$text, $upgraded_base_dir, $upgraded_file, $doc_obj, $cursection);
550 }
551 while ($comment =~ s/$lt(.*?)$gt//s) {
552 my $tag = $1;
553 if ($tag eq "Section") {
554 $found_something = 1;
555 $cursection = $doc_obj->insert_section($doc_obj->get_end_child($cursection)) unless $top;
556 $top = 0;
557 } elsif ($tag eq "/Section") {
558 $found_something = 1;
559 $cursection = $doc_obj->get_parent_section ($cursection);
560 } elsif ($tag =~ m/^Metadata name=$quot(.*?)$quot/s) {
561 my $metaname = $1;
562 my $accumulate = $tag =~ m/mode=${quot}accumulate${quot}/ ? 1 : 0;
563 $comment =~ s/^(.*?)$lt\/Metadata$gt//s;
564 my $metavalue = $1;
565 $metavalue =~ s/^\s+//;
566 $metavalue =~ s/\s+$//;
567 # assume that no metadata value intentionally includes
568 # carriage returns or HTML tags (if they're there they
569 # were probably introduced when converting to HTML from
570 # some other format).
571 # actually some people want to have html tags in their
572 # metadata.
573 $metavalue =~ s/[\cJ\cM]/ /sg;
574 $metavalue =~ s/<[^>]+>//sg
575 unless $dont_strip && ($dont_strip eq 'all' || $metaname =~ m/^($dont_strip)$/);
576 $metavalue =~ s/\s+/ /sg;
577 if ($metaname =~ /\./) { # has a namespace
578 $metaname = "ex.$metaname";
579 }
580 if ($accumulate) {
581 $doc_obj->add_utf8_metadata($cursection, $metaname, $metavalue);
582 } else {
583 $doc_obj->set_utf8_metadata_element($cursection, $metaname, $metavalue);
584 }
585 } elsif ($tag eq "Description" || $tag eq "/Description") {
586 # do nothing with containing Description tags
587 } else {
588 # simple HTML tag (probably created by the conversion
589 # to HTML from some other format) - we'll ignore it and
590 # hope for the best ;-)
591 }
592 }
593 }
594 if ($cursection ne "") {
595 print $outhandle "HTMLPlugin: WARNING: $upgraded_file contains unmatched <Section></Section> tags\n";
596 }
597
598 $$textref =~ s/^.*?<body[^>]*>//is;
599 $$textref =~ s/(<\/body[^>]*>|<\/html[^>]*>)//isg;
600 if ($$textref =~ m/\S/) {
601 if (!$found_something) {
602 if ($self->{'verbosity'} > 2) {
603 print $outhandle "HTMLPlugin: WARNING: $upgraded_file appears to contain no Section tags so\n";
604 print $outhandle " will be processed as a single section document\n";
605 }
606
607 # go ahead and process single-section document
608 $self->process_section($textref, $upgraded_base_dir, $upgraded_file, $doc_obj, $cursection);
609
610 # if document contains no Section tags we'll go ahead
611 # and extract metadata (this won't have been done
612 # above as the -description_tags option prevents it)
613 my $complete_text = $head_keep.$doc_obj->get_text($cursection);
614 $self->extract_metadata (\$complete_text, $metadata, $doc_obj, $cursection)
615 unless $self->{'no_metadata'};
616
617 } else {
618 print $outhandle "HTMLPlugin: WARNING: $upgraded_file contains the following text outside\n";
619 print $outhandle " of the final closing </Section> tag. This text will\n";
620 print $outhandle " be ignored.";
621
622 my ($text);
623 if (length($$textref) > 30) {
624 $text = substr($$textref, 0, 30) . "...";
625 } else {
626 $text = $$textref;
627 }
628 $text =~ s/\n/ /isg;
629 print $outhandle " ($text)\n";
630 }
631 } elsif (!$found_something) {
632
633 if ($self->{'verbosity'} > 2) {
634 # may get to here if document contained no valid Section
635 # tags but did contain some comments. The text will have
636 # been processed already but we should print the warning
637 # as above and extract metadata
638 print $outhandle "HTMLPlugin: WARNING: $upgraded_file appears to contain no Section tags and\n";
639 print $outhandle " is blank or empty. Metadata will be assigned if present.\n";
640 }
641
642 my $complete_text = $head_keep.$doc_obj->get_text($cursection);
643 $self->extract_metadata (\$complete_text, $metadata, $doc_obj, $cursection)
644 unless $self->{'no_metadata'};
645 }
646 $self->replace_section_links($doc_obj);
647 } else {
648
649 # remove header and footer
650 if (!$self->{'keep_head'} || $self->{'description_tags'}) {
651 $$textref =~ s/^.*?<body[^>]*>//is;
652 $$textref =~ s/(<\/body[^>]*>|<\/html[^>]*>)//isg;
653 }
654
655 $self->{'css_assoc_files'} = {};
656
657 # single section document
658 $self->process_section($textref, $upgraded_base_dir, $upgraded_file, $doc_obj, $cursection);
659
660 #my $upgraded_filename_dirname = &File::Basename::dirname($upgraded_filename);
661
662 $self->acquire_css_associated_files($doc_obj, $cursection);
663
664 $self->{'css_assoc_files'} = {};
665 }
666
667 return 1;
668}
669
670
671sub process_heading
672{
673 my ($self, $nHeadNo, $strHeadingText, $arrSections, $file) = @_;
674 $strHeadingText = '' if (!defined($strHeadingText));
675
676 my $strMetadata = $self->update_section_data($arrSections, int($nHeadNo));
677
678 my $strSecMetadata = '';
679 while ($strHeadingText =~ s/<!--gsdl-metadata(.*?)-->//is)
680 {
681 $strSecMetadata .= $1;
682 }
683
684 $strHeadingText =~ s/^\s+//g;
685 $strHeadingText =~ s/\s+$//g;
686 $strSecMetadata =~ s/^\s+//g;
687 $strSecMetadata =~ s/\s+$//g;
688
689 $strMetadata .= "\n<Section>\n\t<Description>\n\t\t<Metadata name=\"Title\">" . $strHeadingText . "</Metadata>\n";
690
691 if (length($strSecMetadata)) {
692 $strMetadata .= "\t\t" . $strSecMetadata . "\n";
693 }
694
695 $strMetadata .= "\t</Description>\n";
696
697 return "<!--" . $strMetadata . "-->";
698}
699
700
701sub update_section_data
702{
703 my ($self, $arrSections, $nCurTocNo) = @_;
704 my ($strBuffer, $nLast, $nSections) = ('', 0, scalar(@$arrSections));
705
706 if ($nSections == 0) {
707 push @$arrSections, $nCurTocNo;
708 return $strBuffer;
709 }
710 $nLast = $arrSections->[$nSections - 1];
711 if ($nCurTocNo > $nLast) {
712 push @$arrSections, $nCurTocNo;
713 return $strBuffer;
714 }
715 for(my $i = $nSections - 1; $i >= 0; $i--) {
716 if ($nCurTocNo <= $arrSections->[$i]) {
717 $strBuffer .= "\n</Section>";
718 pop @$arrSections;
719 }
720 }
721 push @$arrSections, $nCurTocNo;
722 return $strBuffer;
723}
724
725
726# note that process_section may be called multiple times for a single
727# section (relying on the fact that add_utf8_text appends the text to any
728# that may exist already).
729sub process_section {
730 my $self = shift (@_);
731 my ($textref, $base_dir, $file, $doc_obj, $cursection) = @_;
732
733 my @styleTagsText = ($$textref =~ m/<style[^>]*>([^<]*)<\/style>/sg);
734 if(scalar(@styleTagsText) > 0)
735 {
736 my $css_filename_dirname = &File::Basename::dirname(&FileUtils::filenameConcatenate($base_dir, $file));
737 foreach my $styleText (@styleTagsText)
738 {
739 $self->acquire_css_associated_files_from_text_block($styleText, $css_filename_dirname);
740 }
741 }
742
743 # trap links
744 if (!$self->{'nolinks'}) {
745 # usemap="./#index" not handled correctly => change to "#index"
746## $$textref =~ s/(<img[^>]*?usemap\s*=\s*[\"\']?)([^\"\'>\s]+)([\"\']?[^>]*>)/
747
748## my $opencom = '(?:<!--|&lt;!(?:&mdash;|&#151;|--))';
749## my $closecom = '(?:-->|(?:&mdash;|&#151;|--)&gt;)';
750
751 $$textref =~ s/(<img[^>]*?usemap\s*=\s*)((?:[\"][^\"]+[\"])|(?:[\'][^\']+[\'])|(?:[^\s\/>]+))([^>]*>)/
752 $self->replace_usemap_links($1, $2, $3)/isge;
753
754 $$textref =~ s/(<(?:a|area|frame|link|script)\s+[^>]*?\s*(?:href|src)\s*=\s*)((?:[\"][^\"]+[\"])|(?:[\'][^\']+[\'])|(?:[^\s\/>]+))([^>]*>)/
755 $self->replace_href_links ($1, $2, $3, $base_dir, $file, $doc_obj, $cursection)/isge;
756
757## $$textref =~ s/($opencom.*?)?+(<(?:a|area|frame|link|script)\s+[^>]*?\s*(?:href|src)\s*=\s*)((?:[\"][^\"]+[\"])|(?:[\'][^\']+[\'])|(?:[^\s\/>]+))([^>]*>)(.*?$closecom)?+/
758# $self->replace_href_links ($1, $2, $3, $4, $5, $base_dir, $file, $doc_obj, $cursection)/isge;
759 }
760
761 # trap images
762
763 # Previously, by default, HTMLPlugin would embed <img> tags inside anchor tags
764 # i.e. <a href="image><img src="image"></a> in order to overcome a problem that
765 # turned regular text succeeding images into links. That is, by embedding <imgs>
766 # inside <a href=""></a>, the text following images were no longer misbehaving.
767 # However, there would be many occasions whereby images were not meant to link
768 # to their source images but where the images would link to another web page.
769 # To allow this, the no_image_links option was introduced: it would prevent
770 # the behaviour of embedding images into links that referenced the source images.
771
772 # Somewhere along the line, the problem of normal text turning into links when
773 # such text followed images which were not embedded in <a href=""></a> ceased
774 # to occur. This is why the following lines have been commented out (as well as
775 # two lines in replace_images). They appear to no longer apply.
776
777 # If at any time, there is a need for having images embedded in <a> anchor tags,
778 # then it might be better to turn that into an HTMLPlugin option rather than make
779 # it the default behaviour. Also, eventually, no_image_links needs to become
780 # a deprecated option for HTMLPlugin as it has now become the default behaviour.
781
782 #if(!$self->{'no_image_links'}){
783 $$textref =~ s/(<(?:img|embed|table|tr|td)[^>]*?(?:src|background)\s*=\s*)((?:[\"][^\"]+[\"])|(?:[\'][^\']+[\'])|(?:[^\s\/>]+))([^>]*>)/
784 $self->replace_images ($1, $2, $3, $base_dir, $file, $doc_obj, $cursection)/isge;
785 #}
786
787 # add text to document object
788 # turn \ into \\ so that the rest of greenstone doesn't think there
789 # is an escape code following. (Macro parsing loses them...)
790 $$textref =~ s/\\/\\\\/go;
791
792 $doc_obj->add_utf8_text($cursection, $$textref);
793}
794
795sub replace_images {
796 my $self = shift (@_);
797 my ($front, $link, $back, $base_dir,
798 $file, $doc_obj, $section) = @_;
799
800 # don't modify embedded images. These start with data:image and are base-64 encoded. Return them as is
801 # If we let this method proceed, it will not just mark embedded images as "external links", but will wrongly
802 # clean up double slashes in the base-64 encoded part as single slash, which clobbers the encoding/image.
803 return ($front . $link .$back) if($link =~ m@^[\"\']?data\:image@);
804
805
806 # remove quotes from link at start and end if necessary
807 if ($link=~/^[\"\']/) {
808 $link=~s/^[\"\']//;
809 $link=~s/[\"\']$//;
810 $front.='"';
811 $back="\"$back";
812 }
813
814 $link =~ s/\n/ /g;
815
816 # Hack to overcome Windows wv 0.7.1 bug that causes embedded images to be broken
817 # If the Word file path has spaces in it, wv messes up and you end up with
818 # absolute paths for the images, and without the "file://" prefix
819 # So check for this special case and massage the data to be correct
820 if ($ENV{'GSDLOS'} =~ m/^windows/i && $self->{'plugin_type'} eq "WordPlug" && $link =~ m/^[A-Za-z]\:\\/) {
821 $link =~ s/^.*\\([^\\]+)$/$1/;
822 }
823
824 my ($href, $hash_part, $rl) = $self->format_link ($link, $base_dir, $file);
825
826 my $img_file = $self->add_file ($href, $rl, $hash_part, $base_dir, $doc_obj, $section);
827
828# print STDERR "**** link = $link\n**** href = $href\n**** img_file = $img_file, rl = $rl\n\n";
829
830 my $anchor_name = $img_file;
831 #$anchor_name =~ s/^.*\///;
832 #$anchor_name = "<a name=\"$anchor_name\" ></a>";
833
834 my $image_link = $front . $img_file .$back;
835 return $image_link;
836
837 # The reasons for why the following two lines are no longer necessary can be
838 # found in subroutine process_section
839 #my $anchor_link = "<a href=\"$img_file\" >".$image_link."</a>";
840 #return $anchor_link;
841
842 #return $front . $img_file . $back . $anchor_name;
843}
844
845sub replace_href_links {
846 my $self = shift (@_);
847 my ($front, $link, $back, $base_dir, $file, $doc_obj, $section) = @_;
848
849 if($front =~ m/^<link / && $link =~ m/\.css"$/)
850 {
851 my $actual_link = $link;
852 $actual_link =~ s/^"(.*)"$/$1/;
853
854 my $directory = &File::Basename::dirname($file);
855
856 my $css_filename = &FileUtils::filenameConcatenate($base_dir, $directory, $actual_link);
857 $self->retrieve_css_associated_files($css_filename);
858 }
859
860 # remove quotes from link at start and end if necessary
861 if ($link=~/^[\"\']/) {
862 $link=~s/^[\"\']//;
863 $link=~s/[\"\']$//;
864 $front.='"';
865 $back="\"$back";
866 }
867
868 # can't remember adding this :-( must have had a reason though...
869 if ($link =~ /^\_http/ || $link =~ /^\_libraryname\_/) {
870 # assume it is a greenstone one and leave alone
871 return $front . $link . $back;
872 }
873
874 # attempt to sort out targets - frames are not handled
875 # well in this plugin and some cases will screw things
876 # up - e.g. the _parent target (so we'll just remove
877 # them all ;-)
878 $front =~ s/(target=\"?)_top(\"?)/$1_gsdltop_$2/is;
879 $back =~ s/(target=\"?)_top(\"?)/$1_gsdltop_$2/is;
880 $front =~ s/target=\"?_parent\"?//is;
881 $back =~ s/target=\"?_parent\"?//is;
882
883 if($link =~ m/^\#/s)
884 {
885 return $front . "_httpsamepagelink_" . $link . $back;
886 }
887
888 $link =~ s/\n/ /g;
889
890 # Find file referred to by $link on file system
891 # This is more complicated than it sounds when char encodings
892 # is taken in to account
893 my ($href, $hash_part, $rl) = $self->format_link ($link, $base_dir, $file);
894
895 # href may use '\'s where '/'s should be on Windows
896 $href =~ s/\\/\//g;
897 my ($filename) = $href =~ m/^(?:.*?):(?:\/\/)?(.*)/;
898
899 ##### leave all these links alone (they won't be picked up by intermediate
900 ##### pages). I think that's safest when dealing with frames, targets etc.
901 ##### (at least until I think of a better way to do it). Problems occur with
902 ##### mailto links from within small frames, the intermediate page is displayed
903 ##### within that frame and can't be seen. There is still potential for this to
904 ##### happen even with html pages - the solution seems to be to somehow tell
905 ##### the browser from the server side to display the page being sent (i.e.
906 ##### the intermediate page) in the top level window - I'm not sure if that's
907 ##### possible - the following line should probably be deleted if that can be done
908 return $front . $link . $back if $href =~ m/^(mailto|news|gopher|nntp|telnet|javascript):/is;
909
910 if (($rl == 0) || ($filename =~ m/$self->{'process_exp'}/) ||
911 ($href =~ m/\/$/) || ($href =~ m/^(mailto|news|gopher|nntp|telnet|javascript):/i)) {
912
913 if ($ENV{'GSDLOS'} =~ m/^windows$/) {
914
915 # Don't do any encoding for now, as not clear what
916 # the right thing to do is to support filename
917 # encoding on Windows when they are not UTF16
918 #
919 }
920 else {
921 # => Unix-based system
922
923 # If web page didn't give encoding, then default to utf8
924 my $content_encoding= $self->{'content_encoding'} || "utf8";
925
926 if ((defined $ENV{"DEBUG_UNICODE"}) && ($ENV{"DEBUG_UNICODE"})) {
927 print STDERR "**** Encoding with '$content_encoding', href: $href\n";
928 }
929
930 # on Darwin, the unicode filenames are stored on the file
931 # system in decomposed form, so any href link (including when
932 # URL-encoded) should refer to the decomposed name of the file
933 if ($ENV{'GSDLOS'} =~ /^darwin$/i) {
934 $href = normalize('D', $href); # Normalization Form D (decomposition)
935 }
936
937 $href = encode($content_encoding,$href);
938 }
939
940 $href = &unicode::raw_filename_to_utf8_url_encoded($href);
941 $href = &unicode::filename_to_url($href);
942
943 &ghtml::urlsafe ($href);
944
945 if ((defined $ENV{"DEBUG_UNICODE"}) && ($ENV{"DEBUG_UNICODE"})) {
946 print STDERR "******DEBUG: href=$href\n";
947 }
948
949 #TODO here
950# if ($rl ==1) {
951 # have a relative link, we need to do URL encoding etc so it matches what has happened for that file
952 #$href = &util::rename_file($href, $self->{'file_rename_method'});
953# $href = &unicode::raw_filename_to_url_encoded($href);
954 # then, this might be url encoded, so we replace % with %25
955# $href = &unicode::filename_to_url($href);
956# print STDERR "DEBUG: url encoded href = $href\n";
957# }
958
959 return $front . "_httpextlink_&amp;rl=" . $rl . "&amp;href=" . $href . $hash_part . $back;
960 } else {
961 # link is to some other type of file (e.g., an image) so we'll
962 # need to associate that file
963 return $front . $self->add_file ($href, $rl, $hash_part, $base_dir, $doc_obj, $section) . $back;
964 }
965}
966
967sub retrieve_css_associated_files {
968 my $self = shift (@_);
969 my ($css_filename) = @_;
970
971 my $css_filename_dirname = &File::Basename::dirname($css_filename);
972
973 # The following file reading section is a candidate to use FileUtils::readUTF8File()
974 # in place of calling sysread() directly. But only if we can reason we'd be working with UTF8
975 open (CSSFILE, $css_filename) || return;
976 sysread (CSSFILE, my $file_string, -s CSSFILE);
977
978 $self->acquire_css_associated_files_from_text_block($file_string, $css_filename_dirname) unless !defined $file_string;
979
980 close CSSFILE;
981}
982
983sub acquire_css_associated_files_from_text_block {
984 my $self = shift (@_);
985 my ($text, $css_filename_dirname) = @_;
986
987 my @image_urls = ($text =~ m/background-image:\s*url[^;]*;/sg);
988 foreach my $img_url (@image_urls)
989 {
990 $img_url =~ s/^.*url.*\((.*)\).*$/$1/;
991 $img_url =~ s/^\s*"?([^"]*)"?\s*$/$1/;
992
993 $self->{'css_assoc_files'}->{&FileUtils::filenameConcatenate($css_filename_dirname, $img_url)} = $img_url;
994 }
995}
996
997sub acquire_css_associated_files {
998 my $self = shift(@_);
999
1000 my ($doc_obj, $section) = @_;
1001
1002 foreach my $image_filename (keys %{$self->{'css_assoc_files'}})
1003 {
1004 $doc_obj->associate_file($image_filename, $self->{'css_assoc_files'}->{$image_filename}, undef, $section);
1005 }
1006}
1007
1008sub add_file {
1009 my $self = shift (@_);
1010 my ($href, $rl, $hash_part, $base_dir, $doc_obj, $section) = @_;
1011 my ($newname);
1012
1013 my $filename = $href;
1014 if ($base_dir eq "") {
1015 if ($ENV{'GSDLOS'} =~ m/^windows$/i) {
1016 # remove http://
1017 $filename =~ s/^[^:]*:\/\///;
1018 }
1019 else {
1020 # remove http:/ thereby leaving one slash at the start as
1021 # part of full pathname
1022 $filename =~ s/^[^:]*:\///;
1023 }
1024 }
1025 else {
1026 # remove http://
1027 $filename =~ s/^[^:]*:\/\///;
1028 }
1029
1030 if ($ENV{'GSDLOS'} =~ m/^windows$/i) {
1031 $filename =~ s@\/@\\@g;
1032 }
1033
1034 $filename = &FileUtils::filenameConcatenate($base_dir, $filename);
1035
1036 if (($self->{'use_realistic_book'}) || ($self->{'old_style_HDL'})) {
1037 # we are processing a tidytmp file - want paths to be in import
1038 $filename =~ s/([\\\/])tidytmp([\\\/])/$1import$2/;
1039 }
1040
1041 # Replace %XX's in URL with decoded value if required. Note that the
1042 # filename may include the %XX in some situations. If the *original*
1043 # file's name was in URL encoding, the following method will not decode
1044 # it.
1045 my $unicode_filename = $filename;
1046 my $opt_decode_unicode_filename = $self->opt_url_decode($unicode_filename);
1047
1048 # wvWare can generate <img src="StrangeNoGraphicData"> tags, but with no
1049 # (it seems) accompanying file
1050 if ($opt_decode_unicode_filename =~ m/StrangeNoGraphicData$/) { return ""; }
1051
1052 my $content_encoding= $self->{'content_encoding'} || "utf8";
1053
1054 if ($ENV{'GSDLOS'} =~ /^(linux|solaris)$/i) {
1055 # The filenames that come through the HTML file have been decoded
1056 # into Unicode aware Perl strings. Need to convert them back
1057 # to their initial raw-byte encoding to match the file that
1058 # exists on the file system
1059 $filename = encode($content_encoding, $opt_decode_unicode_filename);
1060
1061 }
1062 elsif ($ENV{'GSDLOS'} =~ /^darwin$/i) {
1063 # HFS+ is UTF8 with decompostion
1064 $filename = encode($content_encoding, $opt_decode_unicode_filename);
1065 $filename = normalize('D', $filename); # Normalization Form D (decomposition)
1066
1067 }
1068 elsif ($ENV{'GSDLOS'} =~ /^windows$/i) {
1069 my $long_filename = Win32::GetLongPathName($opt_decode_unicode_filename);
1070
1071 if (defined $long_filename) {
1072 my $short_filename = Win32::GetLongPathName($long_filename);
1073 $filename = $short_filename;
1074 }
1075# else {
1076# print STDERR "***** failed to map href to real file:\n";
1077# print STDERR "****** $href -> $opt_decode_unicode_filename\n";
1078# }
1079 }
1080 else {
1081 my $outhandle = $self->{'outhandle'};
1082 print $outhandle "Warning: Unrecognized operating system ", $ENV{'GSDLOS'}, "\n";
1083 print $outhandle " in file system encoding of href: $href\n";
1084 print $outhandle " No character encoding done.\n";
1085 }
1086
1087
1088 # some special processing if the intended filename was converted to utf8, but
1089 # the actual file still needs to be renamed
1090 if (!&FileUtils::fileExists($filename)) {
1091 # try the original filename stored in map
1092 if ((defined $ENV{"DEBUG_UNICODE"}) && ($ENV{"DEBUG_UNICODE"})) {
1093 print STDERR "******!! orig filename did not exist: $filename\n";
1094 }
1095
1096## print STDERR "**** trying to look up unicode_filename: $unicode_filename\n";
1097
1098 my $original_filename = $self->{'unicode_to_original_filename'}->{$unicode_filename};
1099
1100 if ((defined $ENV{"DEBUG_UNICODE"}) && ($ENV{"DEBUG_UNICODE"})) {
1101 print STDERR "****** From lookup unicode_filename, now trying for: $original_filename\n";
1102 }
1103
1104 if (defined $original_filename && -e $original_filename) {
1105 if ((defined $ENV{"DEBUG_UNICODE"}) && ($ENV{"DEBUG_UNICODE"})) {
1106 print STDERR "****** Found match!\n";
1107 }
1108 $filename = $original_filename;
1109 }
1110 }
1111
1112 my ($ext) = $filename =~ m/(\.[^\.]*)$/;
1113
1114 if ($rl == 0) {
1115 if ((!defined $ext) || ($ext !~ m/$self->{'assoc_files'}/)) {
1116 return "_httpextlink_&amp;rl=0&amp;el=prompt&amp;href=" . $href . $hash_part;
1117 }
1118 else {
1119 return "_httpextlink_&amp;rl=0&amp;el=direct&amp;href=" . $href . $hash_part;
1120 }
1121 }
1122
1123 if ((!defined $ext) || ($ext !~ m/$self->{'assoc_files'}/)) {
1124 return "_httpextlink_&amp;rl=" . $rl . "&amp;href=" . $href . $hash_part;
1125 }
1126 # add the original image file as a source file
1127 if (!$self->{'processing_tmp_files'} ) {
1128 $doc_obj->associate_source_file($filename);
1129 }
1130 if ($self->{'rename_assoc_files'}) {
1131 if (defined $self->{'aux_files'}->{$href}) {
1132 $newname = $self->{'aux_files'}->{$href}->{'dir_num'} . "/" .
1133 $self->{'aux_files'}->{$href}->{'file_num'} . $ext;
1134 } else {
1135 $newname = $self->{'dir_num'} . "/" . $self->{'file_num'} . $ext;
1136 $self->{'aux_files'}->{$href} = {'dir_num' => $self->{'dir_num'}, 'file_num' => $self->{'file_num'}};
1137 $self->inc_filecount ();
1138 }
1139 $doc_obj->associate_file($filename, $newname, undef, $section);
1140 return "_httpdocimg_/$newname";
1141 } else {
1142 if(&unicode::is_url_encoded($unicode_filename)) {
1143 # use the possibly-decoded filename instead to avoid double URL encoding
1144 ($newname) = $filename =~ m/([^\/\\]*)$/;
1145 } else {
1146 ($newname) = $unicode_filename =~ m/([^\/\\]*)$/;
1147 }
1148
1149 # Make sure this name uses only ASCII characters.
1150 # We use either base64 or URL encoding, as these preserve original encoding
1151 $newname = &util::rename_file($newname, $self->{'file_rename_method'});
1152
1153### print STDERR "***** associating $filename (raw-byte/utf8)-> $newname\n";
1154 $doc_obj->associate_file($filename, $newname, undef, $section);
1155
1156 # Since the generated image will be URL-encoded to avoid file-system/browser mess-ups
1157 # of filenames, URL-encode the additional percent signs of the URL-encoded filename
1158 my $newname_url = $newname;
1159 $newname_url = &unicode::filename_to_url($newname_url);
1160 return "_httpdocimg_/$newname_url";
1161 }
1162}
1163
1164sub replace_section_links {
1165 my $self = shift(@_);
1166 my ($doc_obj) = @_;
1167 my %anchors;
1168 my $top_section = $doc_obj->get_top_section();
1169 my $thissection = $doc_obj->get_next_section($top_section);
1170 while ( defined $thissection ) {
1171 my $text = $doc_obj->get_text($thissection);
1172 while ( $text =~ /(?:(?:id|name)\s*=\s*[\'\"])([^\'\"]+)/gi ) {
1173 $anchors{$1} = $thissection;
1174 }
1175 $thissection = $doc_obj->get_next_section($thissection);
1176 }
1177 $thissection = $top_section;
1178 while (defined $thissection) {
1179 my $text = $doc_obj->get_text($thissection);
1180 $text =~ s/(href\s*=\s*[\"\'])(_httpsamepagelink_#)([^\'\"]+)/$self->replace_link_to_anchor($1,$2,$3,$thissection,$anchors{$3})/ige;
1181 $doc_obj->delete_text( $thissection);
1182 $doc_obj->add_utf8_text( $thissection, $text );
1183 $thissection = $doc_obj->get_next_section ($thissection);
1184 }
1185}
1186sub replace_link_to_anchor {
1187 my $self = shift(@_);
1188 my ($href_part,$old_link,$identifier,$current_section,$target_section) = @_;
1189 if (length $target_section){
1190 return $href_part . "javascript:goToAnchor(\'" . $target_section . "\',\'" . $identifier . "\');" ;
1191 }
1192 return $href_part . $old_link . $identifier ;
1193}
1194
1195sub format_link {
1196 my $self = shift (@_);
1197 my ($link, $base_dir, $file) = @_;
1198
1199 # strip off hash part, e.g. #foo, but watch out for any entities, e.g. &#x3B1;
1200 my ($before_hash, $hash_part) = $link =~ m/^(.*?[^&])(\#.*)?$/;
1201
1202 $hash_part = "" if !defined $hash_part;
1203 if (!defined $before_hash || $before_hash !~ m/[\w\.\/]/) {
1204 my $outhandle = $self->{'outhandle'};
1205 print $outhandle "HTMLPlugin: ERROR - badly formatted tag ignored ($link)\n"
1206 if $self->{'verbosity'};
1207 return ($link, "", 0);
1208 }
1209
1210# my $dirname;
1211 if ($before_hash =~ s@^((?:http|https|ftp|file|mms)://)@@i) {
1212 my $type = $1;
1213 my $before_hash_file = $before_hash;
1214
1215 if ($link =~ m/^(http|ftp):/i) {
1216
1217 # Turn url (using /) into file name (possibly using \ on windows)
1218 my @http_dir_split = split('/', $before_hash_file);
1219 $before_hash_file = &FileUtils::filenameConcatenate(@http_dir_split);
1220 }
1221
1222 # want to maintain two version of "before_hash": one representing the URL, the other using filesystem specific directory separator
1223 $before_hash_file = $self->eval_dir_dots($before_hash_file);
1224 my $before_hash_url = $before_hash_file;
1225 if ($ENV{'GSDLOS'} =~ /^windows$/i) {
1226 $before_hash_url =~ s@\\@\/@g;
1227 }
1228
1229 ######## TODO need to check this for encoding stufff
1230 my $linkfilename = &FileUtils::filenameConcatenate($base_dir, $before_hash_file);
1231 print STDERR "checking for existence whether relative link or not $linkfilename\n";
1232 my $rl = 0;
1233 $rl = 1 if (-e $linkfilename);
1234 if (-e $linkfilename) {
1235
1236 print STDERR "DOES exist $linkfilename\n";
1237 } else {
1238 print STDERR "DOESN'T exist $linkfilename\n";
1239 }
1240 # make sure there's a slash on the end if it's a directory
1241 if ($before_hash_url !~ m/\/$/) {
1242 $before_hash_url .= "/" if (-d $linkfilename);
1243 }
1244 return ($type . $before_hash_url, $hash_part, $rl);
1245
1246 } elsif ($link !~ m/^(mailto|news|gopher|nntp|telnet|javascript):/i && $link !~ m/^\//) {
1247
1248 #### TODO whst is this test doing???
1249 if ($before_hash =~ s@^/@@ || $before_hash =~ m/\\/) {
1250
1251 # the first directory will be the domain name if file_is_url
1252 # to generate archives, otherwise we'll assume all files are
1253 # from the same site and base_dir is the root
1254
1255 if ($self->{'file_is_url'}) {
1256 my @dirs = split /[\/\\]/, $file;
1257 my $domname = shift (@dirs);
1258 $before_hash = &FileUtils::filenameConcatenate($domname, $before_hash);
1259 $before_hash =~ s@\\@/@g; # for windows
1260 }
1261 else
1262 {
1263 # see if link shares directory with source document
1264 # => turn into relative link if this is so!
1265
1266 if ($ENV{'GSDLOS'} =~ m/^windows/i) {
1267 # too difficult doing a pattern match with embedded '\'s...
1268 my $win_before_hash=$before_hash;
1269 $win_before_hash =~ s@(\\)+@/@g;
1270 # $base_dir is already similarly "converted" on windows.
1271 if ($win_before_hash =~ s@^$base_dir/@@o) {
1272 # if this is true, we removed a prefix
1273 $before_hash=$win_before_hash;
1274 }
1275 }
1276 else {
1277 # before_hash has lost leading slash by this point,
1278 # -> add back in prior to substitution with $base_dir
1279 $before_hash = "/$before_hash";
1280
1281 $before_hash = &FileUtils::filenameConcatenate("",$before_hash);
1282 $before_hash =~ s@^$base_dir/@@;
1283 }
1284 }
1285 } else {
1286
1287 # Turn relative file path into full path (inside import dir)
1288 my $dirname = &File::Basename::dirname($file);
1289
1290 # we want to add dirname (which is raw filesystem path) onto $before_hash, (which is perl unicode aware string). Convert dirname to perl string
1291
1292 my $unicode_dirname ="";
1293 # we need to turn raw filesystem filename into perl unicode aware string
1294 my $filename_encoding = $self->guess_filesystem_encoding();
1295
1296 # copied this from set_Source_metadata in BaseImporter
1297 if ((defined $filename_encoding) && ($filename_encoding ne "ascii")) {
1298 # Use filename_encoding to map raw filename to a Perl unicode-aware string
1299 $unicode_dirname = decode($filename_encoding,$dirname);
1300 }
1301 else {
1302 # otherwise generate %xx encoded version of filename for char > 127
1303 $unicode_dirname = &unicode::raw_filename_to_url_encoded($dirname);
1304 }
1305 $before_hash = &FileUtils::filenameConcatenate($unicode_dirname, $before_hash);
1306 $before_hash = $self->eval_dir_dots($before_hash);
1307 $before_hash =~ s@\\@/@g; # for windows
1308 }
1309
1310 my $linkfilename = &FileUtils::filenameConcatenate($base_dir, $before_hash);
1311
1312 # make sure there's a slash on the end if it's a directory
1313 if ($before_hash !~ m/\/$/) {
1314 $before_hash .= "/" if (-d $linkfilename);
1315 }
1316 return ("http://" . $before_hash, $hash_part, 1);
1317 } else {
1318 # mailto, news, nntp, telnet, javascript or gopher link
1319 return ($before_hash, "", 0);
1320 }
1321}
1322
1323sub extract_first_NNNN_characters {
1324 my $self = shift (@_);
1325 my ($textref, $doc_obj, $thissection) = @_;
1326
1327 foreach my $size (split /,/, $self->{'first'}) {
1328 my $tmptext = $$textref;
1329 # skip to the body
1330 $tmptext =~ s/.*<body[^>]*>//i;
1331 # remove javascript
1332 $tmptext =~ s@<script.*?</script>@ @sig;
1333 $tmptext =~ s/<[^>]*>/ /g;
1334 $tmptext =~ s/&nbsp;/ /g;
1335 $tmptext =~ s/^\s+//;
1336 $tmptext =~ s/\s+$//;
1337 $tmptext =~ s/\s+/ /gs;
1338 $tmptext = &unicode::substr ($tmptext, 0, $size);
1339 $tmptext =~ s/\s\S*$/&#8230;/; # adds an ellipse (...)
1340 $doc_obj->add_utf8_metadata ($thissection, "First$size", $tmptext);
1341 }
1342}
1343
1344
1345sub extract_metadata {
1346 my $self = shift (@_);
1347 my ($textref, $metadata, $doc_obj, $section) = @_;
1348 my $outhandle = $self->{'outhandle'};
1349 # if we don't want metadata, we may as well not be here ...
1350 return if (!defined $self->{'metadata_fields'});
1351 my $separator = $self->{'metadata_field_separator'};
1352 if ($separator eq "") {
1353 undef $separator;
1354 }
1355
1356 # metadata fields to extract/save. 'key' is the (lowercase) name of the
1357 # html meta, 'value' is the metadata name for greenstone to use
1358 my %find_fields = ();
1359
1360 my %creator_fields = (); # short-cut for lookups
1361
1362
1363 foreach my $field (split /,/, $self->{'metadata_fields'}) {
1364 $field =~ s/^\s+//; # remove leading whitespace
1365 $field =~ s/\s+$//; # remove trailing whitespace
1366
1367 # support tag<tagname>
1368 if ($field =~ m/^(.*?)\s*<(.*?)>$/) {
1369 # "$2" is the user's preferred gs metadata name
1370 $find_fields{lc($1)}=$2; # lc = lowercase
1371 } else { # no <tagname> for mapping
1372 # "$field" is the user's preferred gs metadata name
1373 $find_fields{lc($field)}=$field; # lc = lowercase
1374 }
1375 }
1376
1377 if (defined $self->{'hunt_creator_metadata'} &&
1378 $self->{'hunt_creator_metadata'} == 1 ) {
1379 my @extra_fields =
1380 (
1381 'author',
1382 'author.email',
1383 'creator',
1384 'dc.creator',
1385 'dc.creator.corporatename',
1386 );
1387
1388 # add the creator_metadata fields to search for
1389 foreach my $field (@extra_fields) {
1390 $creator_fields{$field}=0; # add to lookup hash
1391 }
1392 }
1393
1394
1395 # find the header in the html file, which has the meta tags
1396 $$textref =~ m@<head>(.*?)</head>@si;
1397
1398 my $html_header=$1;
1399
1400 # go through every <meta... tag defined in the html and see if it is
1401 # one of the tags we want to match.
1402
1403 # special case for title - we want to remember if its been found
1404 my $found_title = 0;
1405 # this assumes that ">" won't appear. (I don't think it's allowed to...)
1406 $html_header =~ m/^/; # match the start of the string, for \G assertion
1407
1408 while ($html_header =~ m/\G.*?<meta(.*?)>/sig) {
1409 my $metatag=$1;
1410 my ($tag, $value);
1411
1412 # find the tag name
1413 $metatag =~ m/(?:name|http-equiv)\s*=\s*([\"\'])?(.*?)\1/is;
1414 $tag=$2;
1415 # in case they're not using " or ', but they should...
1416 if (! $tag) {
1417 $metatag =~ m/(?:name|http-equiv)\s*=\s*([^\s\>]+)/is;
1418 $tag=$1;
1419 }
1420
1421 if (!defined $tag) {
1422 print $outhandle "HTMLPlugin: can't find NAME in \"$metatag\"\n";
1423 next;
1424 }
1425
1426 # don't need to assign this field if it was passed in from a previous
1427 # (recursive) plugin
1428 if (defined $metadata->{$tag}) {next}
1429
1430 # find the tag content
1431 $metatag =~ m/content\s*=\s*([\"\'])?(.*?)\1/is;
1432 $value=$2;
1433
1434 # The following code assigns the metaname to value if value is
1435 # empty. Why would we do this?
1436 #if (! $value) {
1437 # $metatag =~ m/(?:name|http-equiv)\s*=\s*([^\s\>]+)/is;
1438 # $value=$1;
1439 #}
1440 if (!defined $value || $value eq "") {
1441 print $outhandle "HTMLPlugin: can't find VALUE in <meta $metatag >\n" if ($self->{'verbosity'} > 2);
1442 next;
1443 }
1444
1445 # clean up and add
1446 $value =~ s/\s+/ /gs;
1447 chomp($value); # remove trailing \n, if any
1448 if (exists $creator_fields{lc($tag)}) {
1449 # map this value onto greenstone's "Creator" metadata
1450 $tag='Creator';
1451 } elsif (!exists $find_fields{lc($tag)}) {
1452 next; # don't want this tag
1453 } else {
1454 # get the user's preferred capitalisation
1455 $tag = $find_fields{lc($tag)};
1456 }
1457 if (lc($tag) eq "title") {
1458 $found_title = 1;
1459 }
1460
1461 if ($self->{'verbosity'} > 2) {
1462 print $outhandle " extracted \"$tag\" metadata \"$value\"\n";
1463 }
1464
1465 if ($tag =~ /\./) {
1466 # there is a . so has a namespace, add ex.
1467 $tag = "ex.$tag";
1468 }
1469 if (defined $separator) {
1470 my @values = split($separator, $value);
1471 foreach my $v (@values) {
1472 $doc_obj->add_utf8_metadata($section, $tag, $v) if $v =~ /\S/;
1473 }
1474 }
1475 else {
1476 $doc_obj->add_utf8_metadata($section, $tag, $value);
1477 }
1478 }
1479
1480 # TITLE: extract the document title
1481 if (exists $find_fields{'title'} && !$found_title) {
1482 # we want a title, and didn't find one in the meta tags
1483 # see if there's a <title> tag
1484 my $title;
1485 my $from = ""; # for debugging output only
1486 if ($html_header =~ m/<title[^>]*>([^<]+)<\/title[^>]*>/is) {
1487 $title = $1;
1488 $from = "<title> tags";
1489 }
1490
1491 if (!defined $title) {
1492 $from = "first 100 chars";
1493 # if no title use first 100 or so characters
1494 $title = $$textref;
1495 $title =~ s/^\xFE\xFF//; # Remove unicode byte order mark
1496 $title =~ s/^.*?<body>//si;
1497 # ignore javascript!
1498 $title =~ s@<script.*?</script>@ @sig;
1499 $title =~ s/<\/([^>]+)><\1>//g; # (eg) </b><b> - no space
1500 $title =~ s/<[^>]*>/ /g; # remove all HTML tags
1501 $title =~ s@\r@@g; # remove Windows carriage returns to ensure that titles of pdftohtml docs are consistent (the same 100 chars) across windows and linux
1502 $title = substr ($title, 0, 100);
1503 $title =~ s/\s\S*$/.../;
1504 }
1505 $title =~ s/<[^>]*>/ /g; # remove html tags
1506 $title =~ s/&nbsp;/ /g;
1507 $title =~ s/(?:&nbsp;|\xc2\xa0)/ /g; # utf-8 for nbsp...
1508 $title =~ s/\s+/ /gs; # collapse multiple spaces
1509 $title =~ s/^\s*//; # remove leading spaces
1510 $title =~ s/\s*$//; # remove trailing spaces
1511
1512 $title =~ s/^$self->{'title_sub'}// if ($self->{'title_sub'});
1513 $title =~ s/^\s+//s; # in case title_sub introduced any...
1514 $doc_obj->add_utf8_metadata ($section, "Title", $title);
1515 print $outhandle " extracted Title metadata \"$title\" from $from\n"
1516 if ($self->{'verbosity'} > 2);
1517 }
1518
1519 # add FileFormat metadata
1520 $doc_obj->add_metadata($section,"FileFormat", "HTML");
1521
1522 # Special, for metadata names such as tagH1 - extracts
1523 # the text between the first <H1> and </H1> tags into "H1" metadata.
1524
1525 foreach my $field (keys %find_fields) {
1526 if ($field !~ m/^tag([a-z0-9]+)$/i) {next}
1527 my $tag = $1;
1528 if ($$textref =~ m@<$tag[^>]*>(.*?)</$tag[^>]*>@g) {
1529 my $content = $1;
1530 $content =~ s/&nbsp;/ /g;
1531 $content =~ s/<[^>]*>/ /g;
1532 $content =~ s/^\s+//;
1533 $content =~ s/\s+$//;
1534 $content =~ s/\s+/ /gs;
1535 if ($content) {
1536 $tag=$find_fields{"tag$tag"}; # get the user's capitalisation
1537 $tag =~ s/^tag//i;
1538 $doc_obj->add_utf8_metadata ($section, $tag, $content);
1539 print $outhandle " extracted \"$tag\" metadata \"$content\"\n"
1540 if ($self->{'verbosity'} > 2);
1541 }
1542 }
1543 }
1544}
1545
1546
1547# evaluate any "../" to next directory up
1548# evaluate any "./" as here
1549sub eval_dir_dots {
1550 my $self = shift (@_);
1551 my ($filename) = @_;
1552 my $dirsep_os = &util::get_os_dirsep();
1553 my @dirsep = split(/$dirsep_os/,$filename);
1554
1555 my @eval_dirs = ();
1556 foreach my $d (@dirsep) {
1557 if ($d eq "..") {
1558 pop(@eval_dirs);
1559
1560 } elsif ($d eq ".") {
1561 # do nothing!
1562
1563 } else {
1564 push(@eval_dirs,$d);
1565 }
1566 }
1567
1568 # Need to fiddle with number of elements in @eval_dirs if the
1569 # first one is the empty string. This is because of a
1570 # modification to FileUtils::filenameConcatenate that supresses the addition
1571 # of a leading '/' character (or \ if windows) (intended to help
1572 # filename cat with relative paths) if the first entry in the
1573 # array is the empty string. Making the array start with *two*
1574 # empty strings is a way to defeat this "smart" option.
1575 #
1576 if (scalar(@eval_dirs) > 0) {
1577 if ($eval_dirs[0] eq ""){
1578 unshift(@eval_dirs,"");
1579 }
1580 }
1581
1582 my $evaluated_filename = (scalar @eval_dirs > 0) ? &FileUtils::filenameConcatenate(@eval_dirs) : "";
1583 return $evaluated_filename;
1584}
1585
1586sub replace_usemap_links {
1587 my $self = shift (@_);
1588 my ($front, $link, $back) = @_;
1589
1590 # remove quotes from link at start and end if necessary
1591 if ($link=~/^[\"\']/) {
1592 $link=~s/^[\"\']//;
1593 $link=~s/[\"\']$//;
1594 $front.='"';
1595 $back="\"$back";
1596 }
1597
1598 $link =~ s/^\.\///;
1599 return $front . $link . $back;
1600}
1601
1602sub inc_filecount {
1603 my $self = shift (@_);
1604
1605 if ($self->{'file_num'} == 1000) {
1606 $self->{'dir_num'} ++;
1607 $self->{'file_num'} = 0;
1608 } else {
1609 $self->{'file_num'} ++;
1610 }
1611}
1612
1613
1614# Extend read_file so that strings like &eacute; are
1615# converted to UTF8 internally.
1616#
1617# We don't convert &lt; or &gt; or &amp; or &quot; in case
1618# they interfere with the GML files
1619
1620sub read_file {
1621 my $self = shift(@_);
1622 my ($filename, $encoding, $language, $textref) = @_;
1623
1624 $self->SUPER::read_file($filename, $encoding, $language, $textref);
1625
1626 # Convert entities to their Unicode code-point equivalents
1627 $$textref =~ s/&(lt|gt|amp|quot|nbsp);/&z$1;/go;
1628 $$textref =~ s/&([^;]+);/&ghtml::getcharequiv($1,1,1)/gseo;
1629 $$textref =~ s/&z(lt|gt|amp|quot|nbsp);/&$1;/go;
1630
1631}
1632
1633sub HB_read_html_file {
1634 my $self = shift (@_);
1635 my ($htmlfile, $text) = @_;
1636
1637 # load in the file
1638 if (!open (FILE, $htmlfile)) {
1639 print STDERR "ERROR - could not open $htmlfile\n";
1640 return;
1641 }
1642
1643 my $foundbody = 0;
1644 $self->HB_gettext (\$foundbody, $text, "FILE");
1645 close FILE;
1646
1647 # just in case there was no <body> tag
1648 if (!$foundbody) {
1649 $foundbody = 1;
1650 open (FILE, $htmlfile) || return;
1651 $self->HB_gettext (\$foundbody, $text, "FILE");
1652 close FILE;
1653 }
1654 # text is in utf8
1655}
1656
1657# converts the text to utf8, as ghtml does that for &eacute; etc.
1658sub HB_gettext {
1659 my $self = shift (@_);
1660 my ($foundbody, $text, $handle) = @_;
1661
1662 my $line = "";
1663 while (defined ($line = <$handle>)) {
1664 # look for body tag
1665 if (!$$foundbody) {
1666 if ($line =~ s/^.*<body[^>]*>//i) {
1667 $$foundbody = 1;
1668 } else {
1669 next;
1670 }
1671 }
1672
1673 # check for symbol fonts
1674 if ($line =~ m/<font [^>]*?face\s*=\s*\"?(\w+)\"?/i) {
1675 my $font = $1;
1676 print STDERR "HBPlug::HB_gettext - warning removed font $font\n"
1677 if ($font !~ m/^arial$/i);
1678 }
1679
1680 $$text .= $line;
1681 }
1682
1683 if ($self->{'input_encoding'} eq "iso_8859_1") {
1684 # convert to utf-8
1685 $$text=&unicode::unicode2utf8(&unicode::convert2unicode("iso_8859_1", $text));
1686 }
1687 # convert any alphanumeric character entities to their utf-8
1688 # equivalent for indexing purposes
1689 #&ghtml::convertcharentities ($$text);
1690
1691 $$text =~ s/\s+/ /g; # remove \n's
1692
1693 # At this point $$text is a binary byte string
1694 # => turn it into a Unicode aware string, so full
1695 # Unicode aware pattern matching can be used.
1696 # For instance: 's/\x{0101}//g' or '[[:upper:]]'
1697 #
1698
1699 $$text = decode("utf8",$$text);
1700}
1701
1702sub HB_clean_section {
1703 my $self = shift (@_);
1704 my ($section) = @_;
1705
1706 # remove tags without a starting tag from the section
1707 my ($tag, $tagstart);
1708 while ($section =~ m/<\/([^>]{1,10})>/) {
1709 $tag = $1;
1710 $tagstart = index($section, "<$tag");
1711 last if (($tagstart >= 0) && ($tagstart < index($section, "<\/$tag")));
1712 $section =~ s/<\/$tag>//;
1713 }
1714
1715 # remove extra paragraph tags
1716 while ($section =~ s/<p\b[^>]*>\s*<p\b/<p/ig) {}
1717
1718 # remove extra stuff at the end of the section
1719 while ($section =~ s/(<u>|<i>|<b>|<p\b[^>]*>|&nbsp;|\s)$//i) {}
1720
1721 # add a newline at the beginning of each paragraph
1722 $section =~ s/(.)\s*<p\b/$1\n\n<p/gi;
1723
1724 # add a newline every 80 characters at a word boundary
1725 # Note: this regular expression puts a line feed before
1726 # the last word in each section, even when it is not
1727 # needed.
1728 $section =~ s/(.{1,80})\s/$1\n/g;
1729
1730 # fix up the image links
1731 $section =~ s/<img[^>]*?src=\"?([^\">]+)\"?[^>]*>/
1732 <center><img src=\"$1\" \/><\/center><br\/>/ig;
1733 $section =~ s/&lt;&lt;I&gt;&gt;\s*([^\.]+\.(png|jpg|gif))/
1734 <center><img src=\"$1\" \/><\/center><br\/>/ig;
1735
1736 return $section;
1737}
1738
1739# Will convert the oldHDL format to the new HDL format (using the Section tag)
1740sub convert_to_newHDLformat
1741{
1742 my $self = shift (@_);
1743 my ($file,$cnfile) = @_;
1744 my $input_filename = $file;
1745 my $tmp_filename = $cnfile;
1746
1747 # write HTML tmp file with new HDL format
1748 open (PROD, ">$tmp_filename") || die("Error Writing to File: $tmp_filename $!");
1749
1750 # read in the file and do basic html cleaning (removing header etc)
1751 my $html = "";
1752 $self->HB_read_html_file ($input_filename, \$html);
1753
1754 # process the file one section at a time
1755 my $curtoclevel = 1;
1756 my $firstsection = 1;
1757 my $toclevel = 0;
1758 while (length ($html) > 0) {
1759 if ($html =~ s/^.*?(?:<p\b[^>]*>)?((<b>|<i>|<u>|\s)*)&lt;&lt;TOC(\d+)&gt;&gt;\s*(.*?)<p\b/<p/i) {
1760 $toclevel = $3;
1761 my $title = $4;
1762 my $sectiontext = "";
1763 if ($html =~ s/^(.*?)((?:<p\b[^>]*>)?((<b>|<i>|<u>|\s)*)&lt;&lt;TOC\d+&gt;&gt;)/$2/i) {
1764 $sectiontext = $1;
1765 } else {
1766 $sectiontext = $html;
1767 $html = "";
1768 }
1769
1770 # remove tags and extra spaces from the title
1771 $title =~ s/<\/?[^>]+>//g;
1772 $title =~ s/^\s+|\s+$//g;
1773
1774 # close any sections below the current level and
1775 # create a new section (special case for the firstsection)
1776 print PROD "<!--\n";
1777 while (($curtoclevel > $toclevel) ||
1778 (!$firstsection && $curtoclevel == $toclevel)) {
1779 $curtoclevel--;
1780 print PROD "</Section>\n";
1781 }
1782 if ($curtoclevel+1 < $toclevel) {
1783 print STDERR "WARNING - jump in toc levels in $input_filename " .
1784 "from $curtoclevel to $toclevel\n";
1785 }
1786 while ($curtoclevel < $toclevel) {
1787 $curtoclevel++;
1788 }
1789
1790 if ($curtoclevel == 1) {
1791 # add the header tag
1792 print PROD "-->\n";
1793 print PROD "<HTML>\n<HEAD>\n<TITLE>$title</TITLE>\n</HEAD>\n<BODY>\n";
1794 print PROD "<!--\n";
1795 }
1796
1797 print PROD "<Section>\n\t<Description>\n\t\t<Metadata name=\"Title\">$title</Metadata>\n\t</Description>\n";
1798
1799 print PROD "-->\n";
1800
1801 # clean up the section html
1802 $sectiontext = $self->HB_clean_section($sectiontext);
1803
1804 print PROD "$sectiontext\n";
1805
1806 } else {
1807 print STDERR "WARNING - leftover text\n" , $self->shorten($html),
1808 "\nin $input_filename\n";
1809 last;
1810 }
1811 $firstsection = 0;
1812 }
1813
1814 print PROD "<!--\n";
1815 while ($curtoclevel > 0) {
1816 $curtoclevel--;
1817 print PROD "</Section>\n";
1818 }
1819 print PROD "-->\n";
1820
1821 close (PROD) || die("Error Closing File: $tmp_filename $!");
1822
1823 return $tmp_filename;
1824}
1825
1826sub shorten {
1827 my $self = shift (@_);
1828 my ($text) = @_;
1829
1830 return "\"$text\"" if (length($text) < 100);
1831
1832 return "\"" . substr ($text, 0, 50) . "\" ... \"" .
1833 substr ($text, length($text)-50) . "\"";
1834}
1835
1836sub convert_tidy_or_oldHDL_file
1837{
1838 my $self = shift (@_);
1839 my ($file) = @_;
1840 my $input_filename = $file;
1841
1842 if (-d $input_filename)
1843 {
1844 return $input_filename;
1845 }
1846
1847 # get the input filename
1848 my ($tailname, $dirname, $suffix) = &File::Basename::fileparse($input_filename, "\\.[^\\.]+\$");
1849 my $base_dirname = $dirname;
1850 $suffix = lc($suffix);
1851
1852 # derive tmp filename from input filename
1853 # Remove any white space from filename -- no risk of name collision, and
1854 # makes later conversion by utils simpler. Leave spaces in path...
1855 # tidy up the filename with space, dot, hyphen between
1856 $tailname =~ s/\s+//g;
1857 $tailname =~ s/\.+//g;
1858 $tailname =~ s/\-+//g;
1859 # convert to utf-8 otherwise we have problems with the doc.xml file
1860 # later on
1861 &unicode::ensure_utf8(\$tailname);
1862
1863 # softlink to collection tmp dir
1864 my $tmp_dirname = &FileUtils::filenameConcatenate($ENV{'GSDLCOLLECTDIR'}, "tidytmp");
1865 &FileUtils::makeDirectory($tmp_dirname) if (!-e $tmp_dirname);
1866
1867 my $test_dirname = "";
1868 my $f_separator = &util::get_os_dirsep();
1869
1870 if ($dirname =~ m/import$f_separator/)
1871 {
1872 $test_dirname = $'; #'
1873
1874 #print STDERR "init $'\n";
1875
1876 while ($test_dirname =~ m/[$f_separator]/)
1877 {
1878 my $folderdirname = $`;
1879 $tmp_dirname = &FileUtils::filenameConcatenate($tmp_dirname,$folderdirname);
1880 &FileUtils::makeDirectory($tmp_dirname) if (!-e $tmp_dirname);
1881 $test_dirname = $'; #'
1882 }
1883 }
1884
1885 my $tmp_filename = &FileUtils::filenameConcatenate($tmp_dirname, "$tailname$suffix");
1886
1887 # tidy or convert the input file if it is a HTML-like file or it is accepted by the process_exp
1888 if (($suffix eq ".htm") || ($suffix eq ".html") || ($suffix eq ".shtml"))
1889 {
1890 #convert the input file to a new style HDL
1891 my $hdl_output_filename = $input_filename;
1892 if ($self->{'old_style_HDL'})
1893 {
1894 $hdl_output_filename = &FileUtils::filenameConcatenate($tmp_dirname, "$tailname$suffix");
1895 $hdl_output_filename = $self->convert_to_newHDLformat($input_filename,$hdl_output_filename);
1896 }
1897
1898 #just for checking copy all other file from the base dir to tmp dir if it is not exists
1899 opendir(DIR,$base_dirname) or die "Can't open base directory : $base_dirname!";
1900 my @files = grep {!/^\.+$/} readdir(DIR);
1901 close(DIR);
1902
1903 foreach my $file (@files)
1904 {
1905 my $src_file = &FileUtils::filenameConcatenate($base_dirname,$file);
1906 my $dest_file = &FileUtils::filenameConcatenate($tmp_dirname,$file);
1907 if ((!-e $dest_file) && (!-d $src_file))
1908 {
1909 # just copy the original file back to the tmp directory
1910 copy($src_file,$dest_file) or die "Can't copy file $src_file to $dest_file $!";
1911 }
1912 }
1913
1914 # tidy the input file
1915 my $tidy_output_filename = $hdl_output_filename;
1916 if ($self->{'use_realistic_book'})
1917 {
1918 $tidy_output_filename = &FileUtils::filenameConcatenate($tmp_dirname, "$tailname$suffix");
1919 $tidy_output_filename = $self->tmp_tidy_file($hdl_output_filename,$tidy_output_filename);
1920 }
1921 $tmp_filename = $tidy_output_filename;
1922 }
1923 else
1924 {
1925 if (!-e $tmp_filename)
1926 {
1927 # just copy the original file back to the tmp directory
1928 copy($input_filename,$tmp_filename) or die "Can't copy file $input_filename to $tmp_filename $!";
1929 }
1930 }
1931
1932 return $tmp_filename;
1933}
1934
1935
1936# Will make the html input file as a proper XML file with removed font tag and
1937# image size added to the img tag.
1938# The tidying process takes place in a collection specific 'tmp' directory so
1939# that we don't accidentally damage the input.
1940sub tmp_tidy_file
1941{
1942 my $self = shift (@_);
1943 my ($file,$cnfile) = @_;
1944 my $input_filename = $file;
1945 my $tmp_filename = $cnfile;
1946
1947 # get the input filename
1948 my ($tailname, $dirname, $suffix) = &File::Basename::fileparse($input_filename, "\\.[^\\.]+\$");
1949
1950 require HTML::TokeParser::Simple;
1951
1952 # create HTML parser to decode the input file
1953 my $parser = HTML::TokeParser::Simple->new($input_filename);
1954
1955 # write HTML tmp file without the font tag and image size are added to the img tag
1956 open (PROD, ">$tmp_filename") || die("Error Writing to File: $tmp_filename $!");
1957 while (my $token = $parser->get_token())
1958 {
1959 # is it an img tag
1960 if ($token->is_start_tag('img'))
1961 {
1962 # get the attributes
1963 my $attr = $token->return_attr;
1964
1965 # get the full path to the image
1966 my $img_file = &FileUtils::filenameConcatenate($dirname,$attr->{src});
1967
1968 # set the width and height attribute
1969 ($attr->{width}, $attr->{height}) = imgsize($img_file);
1970
1971 # recreate the tag
1972 print PROD "<img";
1973 print PROD map { qq { $_="$attr->{$_}"} } keys %$attr;
1974 print PROD ">";
1975 }
1976 # is it a font tag
1977 else
1978 {
1979 if (($token->is_start_tag('font')) || ($token->is_end_tag('font')))
1980 {
1981 # remove font tag
1982 print PROD "";
1983 }
1984 else
1985 {
1986 # print without changes
1987 print PROD $token->as_is;
1988 }
1989 }
1990 }
1991 close (PROD) || die("Error Closing File: $tmp_filename $!");
1992
1993 # run html-tidy on the tmp file to make it a proper XML file
1994
1995 my $outhandle = $self->{'outhandle'};
1996 print $outhandle "Converting HTML to be XML compliant:\n";
1997
1998 my $tidy_cmd = "tidy";
1999 $tidy_cmd .= " -q" if ($self->{'verbosity'} <= 2);
2000 $tidy_cmd .= " -raw -wrap 0 -asxml \"$tmp_filename\"";
2001 if ($self->{'verbosity'} <= 2) {
2002 if ($ENV{'GSDLOS'} =~ m/^windows/i) {
2003 $tidy_cmd .= " 2>nul";
2004 }
2005 else {
2006 $tidy_cmd .= " 2>/dev/null";
2007 }
2008 print $outhandle " => $tidy_cmd\n";
2009 }
2010
2011 my $tidyfile = `$tidy_cmd`;
2012
2013 # write result back to the tmp file
2014 open (PROD, ">$tmp_filename") || die("Error Writing to File: $tmp_filename $!");
2015 print PROD $tidyfile;
2016 close (PROD) || die("Error Closing File: $tmp_filename $!");
2017
2018 # return the output filename
2019 return $tmp_filename;
2020}
2021
2022sub associate_cover_image
2023{
2024 my $self = shift(@_);
2025 my ($doc_obj, $filename) = @_;
2026 if (($self->{'use_realistic_book'}) || ($self->{'old_style_HDL'}))
2027 {
2028 # we will have cover image in tidytmp, but want it from import
2029 $filename =~ s/([\\\/])tidytmp([\\\/])/$1import$2/;
2030 }
2031 $self->SUPER::associate_cover_image($doc_obj, $filename);
2032}
2033
2034
20351;
Note: See TracBrowser for help on using the repository browser.