Context Navigation

source: main/trunk/greenstone2/perllib/plugins/HTMLPlugin.pm@ 23418

Last change on this file since 23418 was 23418, checked in by davidb, 13 years ago
A few further additions to help windows keep track of c\... and C:\... filenames
Property svn:executable set to ``* Property svn:keywords set to `Author Date Id Revision`
File size: 63.4 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
36	package HTMLPlugin;
37
38	use Encode;
39	use Unicode::Normalize 'normalize';
40
41	use ReadTextFile;
42	use HBPlugin;
43	use ghtml;
44	use unicode;
45	use util;
46	use XMLParser;
47
48	use Image::Size;
49	use File::Copy;
50
51	sub BEGIN {
52	@HTMLPlugin::ISA = ('ReadTextFile', 'HBPlugin');
53	}
54
55	use strict; # every perl program should have this!
56	no strict 'refs'; # make an exception so we can use variables as filehandles
57
58	my $arguments =
59	[ { 'name' => "process_exp",
60	'desc' => "{BasePlugin.process_exp}",
61	'type' => "regexp",
62	'deft' => &get_default_process_exp() },
63	{ 'name' => "block_exp",
64	'desc' => "{BasePlugin.block_exp}",
65	'type' => 'regexp',
66	'deft' => &get_default_block_exp() },
67	{ 'name' => "nolinks",
68	'desc' => "{HTMLPlugin.nolinks}",
69	'type' => "flag" },
70	{ 'name' => "keep_head",
71	'desc' => "{HTMLPlugin.keep_head}",
72	'type' => "flag" },
73	{ 'name' => "no_metadata",
74	'desc' => "{HTMLPlugin.no_metadata}",
75	'type' => "flag" },
76	{ 'name' => "metadata_fields",
77	'desc' => "{HTMLPlugin.metadata_fields}",
78	'type' => "string",
79	'deft' => "Title" },
80	{ 'name' => "metadata_field_separator",
81	'desc' => "{HTMLPlugin.metadata_field_separator}",
82	'type' => "string",
83	'deft' => "" },
84	{ 'name' => "hunt_creator_metadata",
85	'desc' => "{HTMLPlugin.hunt_creator_metadata}",
86	'type' => "flag" },
87	{ 'name' => "file_is_url",
88	'desc' => "{HTMLPlugin.file_is_url}",
89	'type' => "flag" },
90	{ 'name' => "assoc_files",
91	'desc' => "{HTMLPlugin.assoc_files}",
92	'type' => "regexp",
93	'deft' => &get_default_block_exp() },
94	{ 'name' => "rename_assoc_files",
95	'desc' => "{HTMLPlugin.rename_assoc_files}",
96	'type' => "flag" },
97	{ 'name' => "title_sub",
98	'desc' => "{HTMLPlugin.title_sub}",
99	'type' => "string",
100	'deft' => "" },
101	{ 'name' => "description_tags",
102	'desc' => "{HTMLPlugin.description_tags}",
103	'type' => "flag" },
104	# retain this for backward compatibility (w3mir option was replaced by
105	# file_is_url)
106	{ 'name' => "w3mir",
107	# 'desc' => "{HTMLPlugin.w3mir}",
108	'type' => "flag",
109	'hiddengli' => "yes"},
110	{ 'name' => "no_strip_metadata_html",
111	'desc' => "{HTMLPlugin.no_strip_metadata_html}",
112	'type' => "string",
113	'deft' => "",
114	'reqd' => "no"},
115	{ 'name' => "sectionalise_using_h_tags",
116	'desc' => "{HTMLPlugin.sectionalise_using_h_tags}",
117	'type' => "flag" },
118	{ 'name' => "use_realistic_book",
119	'desc' => "{HTMLPlugin.tidy_html}",
120	'type' => "flag"},
121	{ 'name' => "old_style_HDL",
122	'desc' => "{HTMLPlugin.old_style_HDL}",
123	'type' => "flag"},
124	{'name' => "processing_tmp_files",
125	'desc' => "{BasePlugin.processing_tmp_files}",
126	'type' => "flag",
127	'hiddengli' => "yes"}
128	];
129
130	my $options = { 'name' => "HTMLPlugin",
131	'desc' => "{HTMLPlugin.desc}",
132	'abstract' => "no",
133	'inherits' => "yes",
134	'args' => $arguments };
135
136
137	sub new {
138	my ($class) = shift (@_);
139	my ($pluginlist,$inputargs,$hashArgOptLists) = @_;
140	push(@$pluginlist, $class);
141
142	push(@{$hashArgOptLists->{"ArgList"}},@{$arguments});
143	push(@{$hashArgOptLists->{"OptList"}},$options);
144
145
146	my $self = new ReadTextFile($pluginlist,$inputargs,$hashArgOptLists);
147
148	if ($self->{'w3mir'}) {
149	$self->{'file_is_url'} = 1;
150	}
151	$self->{'aux_files'} = {};
152	$self->{'dir_num'} = 0;
153	$self->{'file_num'} = 0;
154
155	return bless $self, $class;
156	}
157
158	# may want to use (?i)\.(gif\|jpe?g\|jpe\|png\|css\|js(?:@.*)?)$
159	# if have eg <script language="javascript" src="img/lib.js@123">
160	# blocking is now done by reading through the file and recording all the
161	# images and other files
162	sub get_default_block_exp {
163	my $self = shift (@_);
164
165	#return q^(?i)\.(gif\|jpe?g\|jpe\|jpg\|png\|css)$^;
166	return "";
167	}
168
169	sub get_default_process_exp {
170	my $self = shift (@_);
171
172	# the last option is an attempt to encode the concept of an html query ...
173	return q^(?i)(\.html?\|\.shtml\|\.shm\|\.asp\|\.php\d?\|\.cgi\|.+\?.+=.*)$^;
174	}
175
176	sub store_block_files
177	{
178	my $self =shift (@_);
179	my ($filename_full_path, $block_hash) = @_;
180
181	my $html_fname = $filename_full_path;
182	my @file_blocks;
183
184	my ($language, $content_encoding) = $self->textcat_get_language_encoding ($filename_full_path);
185	$self->{'store_content_encoding'}->{$filename_full_path} = $content_encoding;
186
187	# read in file ($text will be in utf8)
188	my $raw_text = "";
189	$self->read_file_no_decoding($filename_full_path, \$raw_text);
190
191	my $textref = \$raw_text;
192	my $opencom = '(?:<!--\|<!(?:—\|\|--))';
193	my $closecom = '(?:-->\|(?:—\|\|--)>)';
194	$$textref =~ s/$opencom(.*?)$closecom//gs;
195
196	# Convert entities to their UTF8 equivalents
197	$$textref =~ s/&(lt\|gt\|amp\|quot\|nbsp);/&z$1;/go;
198	$$textref =~ s/&([^;]+);/&ghtml::getcharequiv($1,1,0)/gseo; # on this occassion, want it left as utf8
199	$$textref =~ s/&z(lt\|gt\|amp\|quot\|nbsp);/&$1;/go;
200
201	my $attval = "\\\"[^\\\"]+\\\"\|[^\\s>]+";
202	my @img_matches = ($$textref =~ m/<img[^>]?src\s=\s($attval)[^>]>/igs);
203	my @usemap_matches = ($$textref =~ m/<img[^>]?usemap\s=\s($attval)[^>]>/igs);
204	my @link_matches = ($$textref =~ m/<link[^>]?href\s=\s($attval)[^>]>/igs);
205	my @embed_matches = ($$textref =~ m/<embed[^>]?src\s=\s($attval)[^>]>/igs);
206	my @tabbg_matches = ($$textref =~ m/<(?:body\|table\|tr\|td)[^>]?background\s=\s($attval)[^>]>/igs);
207	my @script_matches = ($$textref =~ m/<script[^>]?src\s=\s($attval)[^>]>/igs);
208
209	if(!defined $self->{'unicode_to_original_filename'}) {
210	# maps from utf8 converted link name -> original filename referrred to by (possibly URL-encoded) src url
211	$self->{'unicode_to_original_filename'} = {};
212	}
213
214	foreach my $raw_link (@img_matches, @usemap_matches, @link_matches, @embed_matches, @tabbg_matches, @script_matches) {
215
216	# remove quotes from link at start and end if necessary
217	if ($raw_link =~ m/^\"/) {
218	$raw_link =~ s/^\"//;
219	$raw_link =~ s/\"$//;
220	}
221
222	# remove any anchor names, e.g. foo.html#name becomes foo.html
223	# but watch out for any #'s that are part of entities, such as α
224	$raw_link =~ s/([^&])\#.*$/$1/s;
225
226	# some links may just be anchor names
227	next unless ($raw_link =~ /\S+/);
228
229	if ($raw_link !~ m@^/@ && $raw_link !~ m/^([A-Z]:?)\\/i) {
230	# Turn relative file path into full path
231	my $dirname = &File::Basename::dirname($filename_full_path);
232	$raw_link = &util::filename_cat($dirname, $raw_link);
233	}
234	$raw_link = $self->eval_dir_dots($raw_link);
235
236	# this is the actual filename on the filesystem (that the link refers to)
237	my $url_original_filename = $self->opt_url_decode($raw_link);
238
239	my ($uses_bytecodes,$exceeds_bytecodes) = &unicode::analyze_raw_string($url_original_filename);
240
241	if ($exceeds_bytecodes) {
242	# We have a link to a file name that is more complicated than a raw byte filename
243	# What we do next depends on the operating system we are on
244
245	if ($ENV{'GSDLOS'} =~ /^(linux\|solaris)$/i) {
246	# Assume we're dealing with a UTF-8 encoded filename
247	$url_original_filename = encode("utf8", $url_original_filename);
248	}
249	elsif ($ENV{'GSDLOS'} =~ /^darwin$/i) {
250	# HFS+ is UTF8 with decompostion
251	$url_original_filename = encode("utf8", $url_original_filename);
252	$url_original_filename = normalize('D', $url_original_filename); # Normalization Form D (decomposition)
253	}
254	elsif ($ENV{'GSDLOS'} =~ /^windows$/i) {
255	# Don't need to do anything as later code maps Windows
256	# unicode filenames to DOS short filenames when needed
257	}
258	else {
259	my $outhandle = $self->{'outhandle'};
260	print $outhandle "Warning: Unrecognized operating system ", $ENV{'GSDLOS'}, "\n";
261	print $outhandle " in raw file system encoding of: $raw_link\n";
262	print $outhandle " Assuming filesystem is UTF-8 based.\n";
263	$url_original_filename = encode("utf8", $url_original_filename);
264	}
265	}
266
267	# Convert the (currently raw) link into its Unicode version.
268	# Store the Unicode link along with the url_original_filename
269	my $unicode_url_original_filename = "";
270	$self->decode_text($raw_link,$content_encoding,$language,\$unicode_url_original_filename);
271
272
273	$self->{'unicode_to_original_filename'}->{$unicode_url_original_filename} = $url_original_filename;
274
275
276	if ($url_original_filename ne $unicode_url_original_filename) {
277	my $outhandle = $self->{'outhandle'};
278
279	print $outhandle "URL Encoding $url_original_filename\n";
280	print $outhandle " ->$unicode_url_original_filename\n";
281
282	# Allow for possibility of raw byte version and Unicode versions of file
283	$block_hash->{'file_blocks'}->{$unicode_url_original_filename} = 1;
284
285	}
286
287	# $url_original_filename = &util::upgrade_if_dos_filename($url_original_filename);
288
289
290	$block_hash->{'file_blocks'}->{$url_original_filename} = 1;
291	if ($ENV{'GSDLOS'} =~ m/^windows$/) {
292
293	my $lower_drive = $url_original_filename;
294	$lower_drive =~ s/^([A-Z]):/\l$1:/i;
295	my $upper_drive = $url_original_filename;
296
297	$upper_drive =~ s/^([A-Z]):/\u$1:/i;
298
299	## print STDERR "**** HTML plugin, blocking: \"$lower_drive\"\n";
300	## print STDERR "**** HTML plugin, blocking: \"$upper_drive\"\n";
301
302	$block_hash->{'file_blocks'}->{$lower_drive} = 1;
303	$block_hash->{'file_blocks'}->{$upper_drive} = 1;
304	}
305	else {
306	## print STDERR "**** HTML plugin, blocking: $url_original_filename\n";
307	$block_hash->{'file_blocks'}->{$url_original_filename} = 1;
308	}
309
310	}
311	}
312
313	# Given a filename in any encoding, will URL decode it to get back the original filename
314	# in the original encoding. Because this method is intended to work out the original
315	# filename, it does not URL decode any filename if a file by the name of the URL-encoded*
316	# string already exists in the local folder.
317	#
318	sub opt_url_decode {
319	my $self = shift (@_);
320	my ($raw_link) = @_;
321
322
323	# Replace %XX's in URL with decoded value if required.
324	# Note that the filename may include the %XX in some situations
325
326	## if ($raw_link =~ m/\%[A-F0-9]{2}/i) {
327
328	if (($raw_link =~ m/\%[A-F0-9]{2}/i) \|\| ($raw_link =~ m/\&\#x[0-9A-F]+;/i) \|\| ($raw_link =~ m/\&\#[0-9]+;/i)) {
329	if (!-e $raw_link) {
330	$raw_link = &unicode::url_decode($raw_link,1);
331	}
332	}
333
334	return $raw_link;
335	}
336
337	sub read_into_doc_obj
338	{
339	my $self = shift (@_);
340	my ($pluginfo, $base_dir, $file, $block_hash, $metadata, $processor, $maxdocs, $total_count, $gli) = @_;
341
342	my ($filename_full_path, $filename_no_path) = &util::get_full_filenames($base_dir, $file);
343
344	# Lookup content_encoding worked out in file_block pass for this file
345	# Store it under the local name 'content_encoding' so its nice and
346	# easy to access
347	$self->{'content_encoding'} = $self->{'store_content_encoding'}->{$filename_full_path};
348
349	# get the input file
350	my $input_filename = $file;
351	my ($tailname, $dirname, $suffix) = &File::Basename::fileparse($input_filename, "\\.[^\\.]+\$");
352	$suffix = lc($suffix);
353	my $tidy_filename;
354	if (($self->{'use_realistic_book'}) \|\| ($self->{'old_style_HDL'}))
355	{
356	# because the document has to be sectionalized set the description tags
357	$self->{'description_tags'} = 1;
358
359	# set the file to be tidied
360	$input_filename = &util::filename_cat($base_dir,$file) if $base_dir =~ m/\w/;
361
362	# get the tidied file
363	#my $tidy_filename = $self->tmp_tidy_file($input_filename);
364	$tidy_filename = $self->convert_tidy_or_oldHDL_file($input_filename);
365
366	# derive tmp filename from input filename
367	my ($tailname, $dirname, $suffix) = &File::Basename::fileparse($tidy_filename, "\\.[^\\.]+\$");
368
369	# set the new input file and base_dir to be from the tidied file
370	$file = "$tailname$suffix";
371	$base_dir = $dirname;
372	}
373
374	# call the parent read_into_doc_obj
375	my ($process_status,$doc_obj) = $self->SUPER::read_into_doc_obj($pluginfo, $base_dir, $file, $block_hash, $metadata, $processor, $maxdocs, $total_count, $gli);
376	if (($self->{'use_realistic_book'}) \|\| ($self->{'old_style_HDL'}))
377	{
378	# now we need to reset the filenames in the doc obj so that the converted filenames are not used
379	my $collect_file = &util::filename_within_collection($filename_full_path);
380	$doc_obj->set_source_filename ($collect_file, $self->{'file_rename_method'});
381	## set_source_filename does not set the doc_obj source_path which is used in archives dbs for incremental
382	# build. So set it manually.
383	$doc_obj->set_source_path($filename_full_path);
384	my $collect_conv_file = &util::filename_within_collection($tidy_filename);
385	$doc_obj->set_converted_filename($collect_conv_file);
386
387	my $plugin_filename_encoding = $self->{'filename_encoding'};
388	my $filename_encoding = $self->deduce_filename_encoding($file,$metadata,$plugin_filename_encoding);
389	$self->set_Source_metadata($doc_obj, $filename_full_path, $filename_encoding);
390	}
391
392	delete $self->{'store_content_encoding'}->{$filename_full_path};
393	$self->{'content_encoding'} = undef;
394
395	return ($process_status,$doc_obj);
396	}
397
398	# do plugin specific processing of doc_obj
399	sub process {
400	my $self = shift (@_);
401	my ($textref, $pluginfo, $base_dir, $file, $metadata, $doc_obj, $gli) = @_;
402	my $outhandle = $self->{'outhandle'};
403
404	if ($ENV{'GSDLOS'} =~ m/^windows/i) {
405	# this makes life so much easier... perl can cope with unix-style '/'s.
406	$base_dir =~ s@(\\)+@/@g;
407	$file =~ s@(\\)+@/@g;
408	}
409
410	my $filename = &util::filename_cat($base_dir,$file);
411	my $upgraded_base_dir = &util::upgrade_if_dos_filename($base_dir);
412	my $upgraded_filename = &util::upgrade_if_dos_filename($filename);
413
414	if ($ENV{'GSDLOS'} =~ m/^windows/i) {
415	# And again
416	$upgraded_base_dir =~ s@(\\)+@/@g;
417	$upgraded_filename =~ s@(\\)+@/@g;
418
419	# Need to make sure there is a '/' on the end of upgraded_base_dir
420	if (($upgraded_base_dir ne "") && ($upgraded_base_dir !~ m/\/$/)) {
421	$upgraded_base_dir .= "/";
422	}
423	}
424	my $upgraded_file = &util::filename_within_directory($upgraded_filename,$upgraded_base_dir);
425
426	# reset per-doc stuff...
427	$self->{'aux_files'} = {};
428	$self->{'dir_num'} = 0;
429	$self->{'file_num'} = 0;
430
431	# process an HTML file where sections are divided by headings tags (H1, H2 ...)
432	# you can also include metadata in the format (X can be any number)
433	# <hX>Title<!--gsdl-metadata
434	# <Metadata name="name1">value1</Metadata>
435	# ...
436	# <Metadata name="nameN">valueN</Metadata>
437	#--></hX>
438	if ($self->{'sectionalise_using_h_tags'}) {
439	# description_tags should allways be activated because we convert headings to description tags
440	$self->{'description_tags'} = 1;
441
442	my $arrSections = [];
443	$$textref =~ s/<h([0-9]+)[^>]>(.?)<\/h[0-9]+>/$self->process_heading($1, $2, $arrSections, $upgraded_file)/isge;
444
445	if (scalar(@$arrSections)) {
446	my $strMetadata = $self->update_section_data($arrSections, -1);
447	if (length($strMetadata)) {
448	$strMetadata = '<!--' . $strMetadata . "\n-->\n</body>";
449	$$textref =~ s/<\/body>/$strMetadata/ig;
450	}
451	}
452	}
453
454	my $cursection = $doc_obj->get_top_section();
455
456	$self->extract_metadata ($textref, $metadata, $doc_obj, $cursection)
457	unless $self->{'no_metadata'} \|\| $self->{'description_tags'};
458
459	# Store URL for page as metadata - this can be used for an
460	# altavista style search interface. The URL won't be valid
461	# unless the file structure contains the domain name (i.e.
462	# like when w3mir is used to download a website).
463
464	# URL metadata (even invalid ones) are used to support internal
465	# links, so even if 'file_is_url' is off, still need to store info
466
467	my ($tailname,$dirname) = &File::Basename::fileparse($upgraded_file);
468
469	# my $utf8_file = $self->filename_to_utf8_metadata($file);
470	# $utf8_file =~ s/&\#095;/_/g;
471	# variable below used to be utf8_file
472	my $url_encoded_file = &unicode::raw_filename_to_url_encoded($tailname);
473	my $utf8_url_encoded_file = &unicode::raw_filename_to_utf8_url_encoded($tailname);
474
475	my $web_url = "http://";
476	my $utf8_web_url = "http://";
477	if(defined $dirname) { # local directory
478	# Check for "ftp" in the domain name of the directory
479	# structure to determine if this URL should be a ftp:// URL
480	# This check is not infallible, but better than omitting the
481	# check, which would cause all files downloaded from ftp sites
482	# via mirroring with wget to have potentially erroneous http:// URLs
483	# assigned in their metadata
484	if ($dirname =~ /^[^\/]*ftp/i)
485	{
486	$web_url = "ftp://";
487	$utf8_web_url = "ftp://";
488	}
489	$dirname = $self->eval_dir_dots($dirname);
490	$dirname .= &util::get_dirsep() if $dirname ne ""; # if there's a directory, it should end on "/"
491
492	$web_url = $web_url.$dirname.$url_encoded_file;
493	$utf8_web_url = $utf8_web_url.$dirname.$utf8_url_encoded_file;
494	} else {
495	$web_url = $web_url.$url_encoded_file;
496	$utf8_web_url = $utf8_web_url.$utf8_url_encoded_file;
497	}
498	$web_url =~ s/\\/\//g;
499	$utf8_web_url =~ s/\\/\//g;
500
501	if ((defined $ENV{"DEBUG_UNICODE"}) && ($ENV{"DEBUG_UNICODE"})) {
502	print STDERR "*******DEBUG: upgraded_file: $upgraded_file\n";
503	print STDERR "*******DEBUG: adding URL metadata: $utf8_url_encoded_file\n";
504	}
505
506
507	$doc_obj->add_utf8_metadata($cursection, "URL", $web_url);
508	$doc_obj->add_utf8_metadata($cursection, "UTF8URL", $utf8_web_url);
509
510	if ($self->{'file_is_url'}) {
511	$doc_obj->add_metadata($cursection, "weblink", "<a href=\"$web_url\">");
512	$doc_obj->add_metadata($cursection, "webicon", "_iconworld_");
513	$doc_obj->add_metadata($cursection, "/weblink", "</a>");
514	}
515
516	if ($self->{'description_tags'}) {
517	# remove the html header - note that doing this here means any
518	# sections defined within the header will be lost (so all <Section>
519	# tags must appear within the body of the HTML)
520	my ($head_keep) = ($$textref =~ m/^(.?)<body[^>]>/is);
521
522	$$textref =~ s/^.?<body[^>]>//is;
523	$$textref =~ s/(<\/body[^>]>\|<\/html[^>]>)//isg;
524
525	my $opencom = '(?:<!--\|<!(?:—\|\|--))';
526	my $closecom = '(?:-->\|(?:—\|\|--)>)';
527
528	my $lt = '(?:<\|<)';
529	my $gt = '(?:>\|>)';
530	my $quot = '(?:"\|"\|”\|“)';
531
532	my $dont_strip = '';
533	if ($self->{'no_strip_metadata_html'}) {
534	($dont_strip = $self->{'no_strip_metadata_html'}) =~ s{,}{\|}g;
535	}
536
537	my $found_something = 0; my $top = 1;
538	while ($$textref =~ s/^(.?)$opencom(.?)$closecom//s) {
539	my $text = $1;
540	my $comment = $2;
541	if (defined $text) {
542	# text before a comment - note that getting to here
543	# doesn't necessarily mean there are Section tags in
544	# the document
545	$self->process_section(\$text, $upgraded_base_dir, $upgraded_file, $doc_obj, $cursection);
546	}
547	while ($comment =~ s/$lt(.*?)$gt//s) {
548	my $tag = $1;
549	if ($tag eq "Section") {
550	$found_something = 1;
551	$cursection = $doc_obj->insert_section($doc_obj->get_end_child($cursection)) unless $top;
552	$top = 0;
553	} elsif ($tag eq "/Section") {
554	$found_something = 1;
555	$cursection = $doc_obj->get_parent_section ($cursection);
556	} elsif ($tag =~ m/^Metadata name=$quot(.*?)$quot/s) {
557	my $metaname = $1;
558	my $accumulate = $tag =~ m/mode=${quot}accumulate${quot}/ ? 1 : 0;
559	$comment =~ s/^(.*?)$lt\/Metadata$gt//s;
560	my $metavalue = $1;
561	$metavalue =~ s/^\s+//;
562	$metavalue =~ s/\s+$//;
563	# assume that no metadata value intentionally includes
564	# carriage returns or HTML tags (if they're there they
565	# were probably introduced when converting to HTML from
566	# some other format).
567	# actually some people want to have html tags in their
568	# metadata.
569	$metavalue =~ s/[\cJ\cM]/ /sg;
570	$metavalue =~ s/<[^>]+>//sg
571	unless $dont_strip && ($dont_strip eq 'all' \|\| $metaname =~ m/^($dont_strip)$/);
572	$metavalue =~ s/\s+/ /sg;
573	if ($metaname =~ /\./) { # has a namespace
574	$metaname = "ex.$metaname";
575	}
576	if ($accumulate) {
577	$doc_obj->add_utf8_metadata($cursection, $metaname, $metavalue);
578	} else {
579	$doc_obj->set_utf8_metadata_element($cursection, $metaname, $metavalue);
580	}
581	} elsif ($tag eq "Description" \|\| $tag eq "/Description") {
582	# do nothing with containing Description tags
583	} else {
584	# simple HTML tag (probably created by the conversion
585	# to HTML from some other format) - we'll ignore it and
586	# hope for the best ;-)
587	}
588	}
589	}
590	if ($cursection ne "") {
591	print $outhandle "HTMLPlugin: WARNING: $upgraded_file contains unmatched <Section></Section> tags\n";
592	}
593
594	$$textref =~ s/^.?<body[^>]>//is;
595	$$textref =~ s/(<\/body[^>]>\|<\/html[^>]>)//isg;
596	if ($$textref =~ m/\S/) {
597	if (!$found_something) {
598	if ($self->{'verbosity'} > 2) {
599	print $outhandle "HTMLPlugin: WARNING: $upgraded_file appears to contain no Section tags so\n";
600	print $outhandle " will be processed as a single section document\n";
601	}
602
603	# go ahead and process single-section document
604	$self->process_section($textref, $upgraded_base_dir, $upgraded_file, $doc_obj, $cursection);
605
606	# if document contains no Section tags we'll go ahead
607	# and extract metadata (this won't have been done
608	# above as the -description_tags option prevents it)
609	my $complete_text = $head_keep.$doc_obj->get_text($cursection);
610	$self->extract_metadata (\$complete_text, $metadata, $doc_obj, $cursection)
611	unless $self->{'no_metadata'};
612
613	} else {
614	print $outhandle "HTMLPlugin: WARNING: $upgraded_file contains the following text outside\n";
615	print $outhandle " of the final closing </Section> tag. This text will\n";
616	print $outhandle " be ignored.";
617
618	my ($text);
619	if (length($$textref) > 30) {
620	$text = substr($$textref, 0, 30) . "...";
621	} else {
622	$text = $$textref;
623	}
624	$text =~ s/\n/ /isg;
625	print $outhandle " ($text)\n";
626	}
627	} elsif (!$found_something) {
628
629	if ($self->{'verbosity'} > 2) {
630	# may get to here if document contained no valid Section
631	# tags but did contain some comments. The text will have
632	# been processed already but we should print the warning
633	# as above and extract metadata
634	print $outhandle "HTMLPlugin: WARNING: $upgraded_file appears to contain no Section tags and\n";
635	print $outhandle " is blank or empty. Metadata will be assigned if present.\n";
636	}
637
638	my $complete_text = $head_keep.$doc_obj->get_text($cursection);
639	$self->extract_metadata (\$complete_text, $metadata, $doc_obj, $cursection)
640	unless $self->{'no_metadata'};
641	}
642
643	} else {
644
645	# remove header and footer
646	if (!$self->{'keep_head'} \|\| $self->{'description_tags'}) {
647	$$textref =~ s/^.?<body[^>]>//is;
648	$$textref =~ s/(<\/body[^>]>\|<\/html[^>]>)//isg;
649	}
650
651
652	# single section document
653	$self->process_section($textref, $upgraded_base_dir, $upgraded_file, $doc_obj, $cursection);
654	}
655
656	return 1;
657	}
658
659
660	sub process_heading
661	{
662	my ($self, $nHeadNo, $strHeadingText, $arrSections, $file) = @_;
663	$strHeadingText = '' if (!defined($strHeadingText));
664
665	my $strMetadata = $self->update_section_data($arrSections, int($nHeadNo));
666
667	my $strSecMetadata = '';
668	while ($strHeadingText =~ s/<!--gsdl-metadata(.*?)-->//is)
669	{
670	$strSecMetadata .= $1;
671	}
672
673	$strHeadingText =~ s/^\s+//g;
674	$strHeadingText =~ s/\s+$//g;
675	$strSecMetadata =~ s/^\s+//g;
676	$strSecMetadata =~ s/\s+$//g;
677
678	$strMetadata .= "\n<Section>\n\t<Description>\n\t\t<Metadata name=\"Title\">" . $strHeadingText . "</Metadata>\n";
679
680	if (length($strSecMetadata)) {
681	$strMetadata .= "\t\t" . $strSecMetadata . "\n";
682	}
683
684	$strMetadata .= "\t</Description>\n";
685
686	return "<!--" . $strMetadata . "-->";
687	}
688
689
690	sub update_section_data
691	{
692	my ($self, $arrSections, $nCurTocNo) = @_;
693	my ($strBuffer, $nLast, $nSections) = ('', 0, scalar(@$arrSections));
694
695	if ($nSections == 0) {
696	push @$arrSections, $nCurTocNo;
697	return $strBuffer;
698	}
699	$nLast = $arrSections->[$nSections - 1];
700	if ($nCurTocNo > $nLast) {
701	push @$arrSections, $nCurTocNo;
702	return $strBuffer;
703	}
704	for(my $i = $nSections - 1; $i >= 0; $i--) {
705	if ($nCurTocNo <= $arrSections->[$i]) {
706	$strBuffer .= "\n</Section>";
707	pop @$arrSections;
708	}
709	}
710	push @$arrSections, $nCurTocNo;
711	return $strBuffer;
712	}
713
714
715	# note that process_section may be called multiple times for a single
716	# section (relying on the fact that add_utf8_text appends the text to any
717	# that may exist already).
718	sub process_section {
719	my $self = shift (@_);
720	my ($textref, $base_dir, $file, $doc_obj, $cursection) = @_;
721	# trap links
722	if (!$self->{'nolinks'}) {
723	# usemap="./#index" not handled correctly => change to "#index"
724	## $$textref =~ s/(<img[^>]?usemap\s=\s[\"\']?)([^\"\'>\s]+)([\"\']?[^>]>)/
725
726	## my $opencom = '(?:<!--\|<!(?:—\|\|--))';
727	## my $closecom = '(?:-->\|(?:—\|\|--)>)';
728
729	$$textref =~ s/(<img[^>]?usemap\s=\s)((?:[\"][^\"]+[\"])\|(?:[\'][^\']+[\'])\|(?:[^\s\/>]+))([^>]>)/
730	$self->replace_usemap_links($1, $2, $3)/isge;
731
732	$$textref =~ s/(<(?:a\|area\|frame\|link\|script)\s+[^>]?\s(?:href\|src)\s=\s[\"\']?)([^\"\'>\s]+)([\"\']?[^>]*>)/
733	$self->replace_href_links ($1, $2, $3, $base_dir, $file, $doc_obj, $cursection)/isge;
734
735	## $$textref =~ s/($opencom.?)?+(<(?:a\|area\|frame\|link\|script)\s+[^>]?\s(?:href\|src)\s=\s)((?:[\"][^\"]+[\"])\|(?:[\'][^\']+[\'])\|(?:[^\s\/>]+))([^>]>)(.*?$closecom)?+/
736	# $self->replace_href_links ($1, $2, $3, $4, $5, $base_dir, $file, $doc_obj, $cursection)/isge;
737	}
738
739	# trap images
740
741	# Previously, by default, HTMLPlugin would embed <img> tags inside anchor tags
742	# i.e. <a href="image><img src="image"></a> in order to overcome a problem that
743	# turned regular text succeeding images into links. That is, by embedding <imgs>
744	# inside <a href=""></a>, the text following images were no longer misbehaving.
745	# However, there would be many occasions whereby images were not meant to link
746	# to their source images but where the images would link to another web page.
747	# To allow this, the no_image_links option was introduced: it would prevent
748	# the behaviour of embedding images into links that referenced the source images.
749
750	# Somewhere along the line, the problem of normal text turning into links when
751	# such text followed images which were not embedded in <a href=""></a> ceased
752	# to occur. This is why the following lines have been commented out (as well as
753	# two lines in replace_images). They appear to no longer apply.
754
755	# If at any time, there is a need for having images embedded in <a> anchor tags,
756	# then it might be better to turn that into an HTMLPlugin option rather than make
757	# it the default behaviour. Also, eventually, no_image_links needs to become
758	# a deprecated option for HTMLPlugin as it has now become the default behaviour.
759
760	#if(!$self->{'no_image_links'}){
761	$$textref =~ s/(<(?:img\|embed\|table\|tr\|td)[^>]?(?:src\|background)\s=\s)((?:[\"][^\"]+[\"])\|(?:[\'][^\']+[\'])\|(?:[^\s\/>]+))([^>]>)/
762	$self->replace_images ($1, $2, $3, $base_dir, $file, $doc_obj, $cursection)/isge;
763	#}
764
765	# add text to document object
766	# turn \ into \\ so that the rest of greenstone doesn't think there
767	# is an escape code following. (Macro parsing loses them...)
768	$$textref =~ s/\\/\\\\/go;
769
770	$doc_obj->add_utf8_text($cursection, $$textref);
771	}
772
773	sub replace_images {
774	my $self = shift (@_);
775	my ($front, $link, $back, $base_dir,
776	$file, $doc_obj, $section) = @_;
777
778	# remove quotes from link at start and end if necessary
779	if ($link=~/^[\"\']/) {
780	$link=~s/^[\"\']//;
781	$link=~s/[\"\']$//;
782	$front.='"';
783	$back="\"$back";
784	}
785
786	$link =~ s/\n/ /g;
787
788	# Hack to overcome Windows wv 0.7.1 bug that causes embedded images to be broken
789	# If the Word file path has spaces in it, wv messes up and you end up with
790	# absolute paths for the images, and without the "file://" prefix
791	# So check for this special case and massage the data to be correct
792	if ($ENV{'GSDLOS'} =~ m/^windows/i && $self->{'plugin_type'} eq "WordPlug" && $link =~ m/^[A-Za-z]\:\\/) {
793	$link =~ s/^.*\\([^\\]+)$/$1/;
794	}
795
796	my ($href, $hash_part, $rl) = $self->format_link ($link, $base_dir, $file);
797
798	my $img_file = $self->add_file ($href, $rl, $hash_part, $base_dir, $doc_obj, $section);
799
800	# print STDERR "** link = $link\n href = $href\n** img_file = $img_file, rl = $rl\n";
801
802	my $anchor_name = $img_file;
803	#$anchor_name =~ s/^.*\///;
804	#$anchor_name = "<a name=\"$anchor_name\" ></a>";
805
806	my $image_link = $front . $img_file .$back;
807	return $image_link;
808
809	# The reasons for why the following two lines are no longer necessary can be
810	# found in subroutine process_section
811	#my $anchor_link = "<a href=\"$img_file\" >".$image_link."</a>";
812	#return $anchor_link;
813
814	#return $front . $img_file . $back . $anchor_name;
815	}
816
817	sub replace_href_links {
818	my $self = shift (@_);
819	my ($front, $link, $back, $base_dir, $file, $doc_obj, $section) = @_;
820
821	# remove quotes from link at start and end if necessary
822	if ($link=~/^[\"\']/) {
823	$link=~s/^[\"\']//;
824	$link=~s/[\"\']$//;
825	$front.='"';
826	$back="\"$back";
827	}
828
829	# attempt to sort out targets - frames are not handled
830	# well in this plugin and some cases will screw things
831	# up - e.g. the _parent target (so we'll just remove
832	# them all ;-)
833	$front =~ s/(target=\"?)_top(\"?)/$1_gsdltop_$2/is;
834	$back =~ s/(target=\"?)_top(\"?)/$1_gsdltop_$2/is;
835	$front =~ s/target=\"?_parent\"?//is;
836	$back =~ s/target=\"?_parent\"?//is;
837
838	return $front . $link . $back if $link =~ m/^\#/s;
839	$link =~ s/\n/ /g;
840
841	# Find file referred to by $link on file system
842	# This is more complicated than it sounds when char encodings
843	# is taken in to account
844	my ($href, $hash_part, $rl) = $self->format_link ($link, $base_dir, $file);
845
846	# href may use '\'s where '/'s should be on Windows
847	$href =~ s/\\/\//g;
848	my ($filename) = $href =~ m/^(?:.?):(?:\/\/)?(.)/;
849
850
851	##### leave all these links alone (they won't be picked up by intermediate
852	##### pages). I think that's safest when dealing with frames, targets etc.
853	##### (at least until I think of a better way to do it). Problems occur with
854	##### mailto links from within small frames, the intermediate page is displayed
855	##### within that frame and can't be seen. There is still potential for this to
856	##### happen even with html pages - the solution seems to be to somehow tell
857	##### the browser from the server side to display the page being sent (i.e.
858	##### the intermediate page) in the top level window - I'm not sure if that's
859	##### possible - the following line should probably be deleted if that can be done
860	return $front . $link . $back if $href =~ m/^(mailto\|news\|gopher\|nntp\|telnet\|javascript):/is;
861
862	if (($rl == 0) \|\| ($filename =~ m/$self->{'process_exp'}/) \|\|
863	($href =~ m/\/$/) \|\| ($href =~ m/^(mailto\|news\|gopher\|nntp\|telnet\|javascript):/i)) {
864
865	if ($ENV{'GSDLOS'} =~ m/^windows$/) {
866
867	# Don't do any encoding for now, as not clear what
868	# the right thing to do is to support filename
869	# encoding on Windows when they are not UTF16
870	#
871	}
872	else {
873	# => Unix-based system
874
875	# If web page didn't give encoding, then default to utf8
876	my $content_encoding= $self->{'content_encoding'} \|\| "utf8";
877
878	if ((defined $ENV{"DEBUG_UNICODE"}) && ($ENV{"DEBUG_UNICODE"})) {
879	print STDERR "**** Encoding with '$content_encoding', href: $href\n";
880	}
881
882	$href = encode($content_encoding,$href);
883	}
884
885	$href = &unicode::raw_filename_to_utf8_url_encoded($href);
886	$href = &unicode::filename_to_url($href);
887
888	&ghtml::urlsafe ($href);
889
890	if ((defined $ENV{"DEBUG_UNICODE"}) && ($ENV{"DEBUG_UNICODE"})) {
891	print STDERR "******DEBUG: href=$href\n";
892	}
893
894
895	return $front . "_httpextlink_&rl=" . $rl . "&href=" . $href . $hash_part . $back;
896	} else {
897	# link is to some other type of file (e.g., an image) so we'll
898	# need to associate that file
899	return $front . $self->add_file ($href, $rl, $hash_part, $base_dir, $doc_obj, $section) . $back;
900	}
901	}
902
903	sub add_file {
904	my $self = shift (@_);
905	my ($href, $rl, $hash_part, $base_dir, $doc_obj, $section) = @_;
906	my ($newname);
907
908	my $filename = $href;
909	if ($base_dir eq "") {
910	if ($ENV{'GSDLOS'} =~ m/^windows$/i) {
911	# remove http://
912	$filename =~ s/^[^:]*:\/\///;
913	}
914	else {
915	# remove http:/ thereby leaving one slash at the start as
916	# part of full pathname
917	$filename =~ s/^[^:]*:\///;
918	}
919	}
920	else {
921	# remove http://
922	$filename =~ s/^[^:]*:\/\///;
923	}
924
925	$filename = &util::filename_cat($base_dir, $filename);
926
927	if (($self->{'use_realistic_book'}) \|\| ($self->{'old_style_HDL'})) {
928	# we are processing a tidytmp file - want paths to be in import
929	$filename =~ s/([\\\/])tidytmp([\\\/])/$1import$2/;
930	}
931
932	# Replace %XX's in URL with decoded value if required. Note that the
933	# filename may include the %XX in some situations. If the original
934	# file's name was in URL encoding, the following method will not decode
935	# it.
936	my $unicode_filename = $filename;
937	my $opt_decode_unicode_filename = $self->opt_url_decode($unicode_filename);
938
939	# wvWare can generate <img src="StrangeNoGraphicData"> tags, but with no
940	# (it seems) accompanying file
941	if ($opt_decode_unicode_filename =~ m/StrangeNoGraphicData$/) { return ""; }
942
943	my $content_encoding= $self->{'content_encoding'} \|\| "utf8";
944
945	if ($ENV{'GSDLOS'} =~ /^(linux\|solaris)$/i) {
946	# The filenames that come through the HTML file have been decoded
947	# into Unicode aware Perl strings. Need to convert them back
948	# to their initial raw-byte encoding to match the file that
949	# exists on the file system
950	$filename = encode($content_encoding, $opt_decode_unicode_filename);
951	}
952	elsif ($ENV{'GSDLOS'} =~ /^darwin$/i) {
953	# HFS+ is UTF8 with decompostion
954	$filename = encode($content_encoding, $opt_decode_unicode_filename);
955	$filename = normalize('D', $filename); # Normalization Form D (decomposition)
956
957	}
958	elsif ($ENV{'GSDLOS'} =~ /^windows$/i) {
959	my $long_filename = Win32::GetLongPathName($opt_decode_unicode_filename);
960
961	if (defined $long_filename) {
962	my $short_filename = Win32::GetLongPathName($long_filename);
963	$filename = $short_filename;
964	}
965	# else {
966	# print STDERR "***** failed to map href to real file:\n";
967	# print STDERR "****** $href -> $opt_decode_unicode_filename\n";
968	# }
969	}
970	else {
971	my $outhandle = $self->{'outhandle'};
972	print $outhandle "Warning: Unrecognized operating system ", $ENV{'GSDLOS'}, "\n";
973	print $outhandle " in file system encoding of href: $href\n";
974	print $outhandle " No character encoding done.\n";
975	}
976
977
978	# some special processing if the intended filename was converted to utf8, but
979	# the actual file still needs to be renamed
980	if (!&util::fd_exists($filename)) {
981	# try the original filename stored in map
982	if ((defined $ENV{"DEBUG_UNICODE"}) && ($ENV{"DEBUG_UNICODE"})) {
983	print STDERR "******!! orig filename did not exist: $filename\n";
984	}
985
986	## print STDERR "**** trying to look up unicode_filename: $unicode_filename\n";
987
988	my $original_filename = $self->{'unicode_to_original_filename'}->{$unicode_filename};
989
990	if ((defined $ENV{"DEBUG_UNICODE"}) && ($ENV{"DEBUG_UNICODE"})) {
991	print STDERR "****** From lookup unicode_filename, now trying for: $original_filename\n";
992	}
993
994	if (defined $original_filename && -e $original_filename) {
995	if ((defined $ENV{"DEBUG_UNICODE"}) && ($ENV{"DEBUG_UNICODE"})) {
996	print STDERR "****** Found match!\n";
997	}
998	$filename = $original_filename;
999	}
1000	}
1001
1002	my ($ext) = $filename =~ m/(\.[^\.]*)$/;
1003
1004	if ($rl == 0) {
1005	if ((!defined $ext) \|\| ($ext !~ m/$self->{'assoc_files'}/)) {
1006	return "_httpextlink_&rl=0&el=prompt&href=" . $href . $hash_part;
1007	}
1008	else {
1009	return "_httpextlink_&rl=0&el=direct&href=" . $href . $hash_part;
1010	}
1011	}
1012
1013	if ((!defined $ext) \|\| ($ext !~ m/$self->{'assoc_files'}/)) {
1014	return "_httpextlink_&rl=" . $rl . "&href=" . $href . $hash_part;
1015	}
1016	# add the original image file as a source file
1017	if (!$self->{'processing_tmp_files'} ) {
1018	$doc_obj->associate_source_file($filename);
1019	}
1020	if ($self->{'rename_assoc_files'}) {
1021	if (defined $self->{'aux_files'}->{$href}) {
1022	$newname = $self->{'aux_files'}->{$href}->{'dir_num'} . "/" .
1023	$self->{'aux_files'}->{$href}->{'file_num'} . $ext;
1024	} else {
1025	$newname = $self->{'dir_num'} . "/" . $self->{'file_num'} . $ext;
1026	$self->{'aux_files'}->{$href} = {'dir_num' => $self->{'dir_num'}, 'file_num' => $self->{'file_num'}};
1027	$self->inc_filecount ();
1028	}
1029	$doc_obj->associate_file($filename, $newname, undef, $section);
1030	return "_httpdocimg_/$newname";
1031	} else {
1032	if(&unicode::is_url_encoded($unicode_filename)) {
1033	# use the possibly-decoded filename instead to avoid double URL encoding
1034	($newname) = $filename =~ m/([^\/\\]*)$/;
1035	} else {
1036	($newname) = $unicode_filename =~ m/([^\/\\]*)$/;
1037	}
1038
1039	# Make sure this name uses only ASCII characters.
1040	# We use either base64 or URL encoding, as these preserve original encoding
1041	$newname = &util::rename_file($newname, $self->{'file_rename_method'});
1042
1043	### print STDERR "***** associating $filename (raw-byte/utf8)-> $newname\n";
1044	$doc_obj->associate_file($filename, $newname, undef, $section);
1045
1046	# Since the generated image will be URL-encoded to avoid file-system/browser mess-ups
1047	# of filenames, URL-encode the additional percent signs of the URL-encoded filename
1048	my $newname_url = $newname;
1049	$newname_url = &unicode::filename_to_url($newname_url);
1050	return "_httpdocimg_/$newname_url";
1051	}
1052	}
1053
1054
1055	sub format_link {
1056	my $self = shift (@_);
1057	my ($link, $base_dir, $file) = @_;
1058
1059	# strip off hash part, e.g. #foo, but watch out for any entities, e.g. α
1060	my ($before_hash, $hash_part) = $link =~ m/^(.?[^&])(\#.)?$/;
1061
1062	$hash_part = "" if !defined $hash_part;
1063	if (!defined $before_hash \|\| $before_hash !~ m/[\w\.\/]/) {
1064	my $outhandle = $self->{'outhandle'};
1065	print $outhandle "HTMLPlugin: ERROR - badly formatted tag ignored ($link)\n"
1066	if $self->{'verbosity'};
1067	return ($link, "", 0);
1068	}
1069
1070	if ($before_hash =~ s@^((?:http\|https\|ftp\|file\|mms)://)@@i) {
1071	my $type = $1;
1072
1073	if ($link =~ m/^(http\|ftp):/i) {
1074	# Turn url (using /) into file name (possibly using \ on windows)
1075	my @http_dir_split = split('/', $before_hash);
1076	$before_hash = &util::filename_cat(@http_dir_split);
1077	}
1078
1079	$before_hash = $self->eval_dir_dots($before_hash);
1080
1081	my $linkfilename = &util::filename_cat ($base_dir, $before_hash);
1082
1083	my $rl = 0;
1084	$rl = 1 if (-e $linkfilename);
1085
1086	# make sure there's a slash on the end if it's a directory
1087	if ($before_hash !~ m/\/$/) {
1088	$before_hash .= "/" if (-d $linkfilename);
1089	}
1090	return ($type . $before_hash, $hash_part, $rl);
1091
1092	} elsif ($link !~ m/^(mailto\|news\|gopher\|nntp\|telnet\|javascript):/i && $link !~ m/^\//) {
1093
1094	if ($before_hash =~ s@^/@@ \|\| $before_hash =~ m/\\/) {
1095
1096	# the first directory will be the domain name if file_is_url
1097	# to generate archives, otherwise we'll assume all files are
1098	# from the same site and base_dir is the root
1099
1100	if ($self->{'file_is_url'}) {
1101	my @dirs = split /[\/\\]/, $file;
1102	my $domname = shift (@dirs);
1103	$before_hash = &util::filename_cat($domname, $before_hash);
1104	$before_hash =~ s@\\@/@g; # for windows
1105	}
1106	else
1107	{
1108	# see if link shares directory with source document
1109	# => turn into relative link if this is so!
1110
1111	if ($ENV{'GSDLOS'} =~ m/^windows/i) {
1112	# too difficult doing a pattern match with embedded '\'s...
1113	my $win_before_hash=$before_hash;
1114	$win_before_hash =~ s@(\\)+@/@g;
1115	# $base_dir is already similarly "converted" on windows.
1116	if ($win_before_hash =~ s@^$base_dir/@@o) {
1117	# if this is true, we removed a prefix
1118	$before_hash=$win_before_hash;
1119	}
1120	}
1121	else {
1122	# before_hash has lost leading slash by this point,
1123	# -> add back in prior to substitution with $base_dir
1124	$before_hash = "/$before_hash";
1125
1126	$before_hash = &util::filename_cat("",$before_hash);
1127	$before_hash =~ s@^$base_dir/@@;
1128	}
1129	}
1130	} else {
1131	# Turn relative file path into full path
1132	my $dirname = &File::Basename::dirname($file);
1133	$before_hash = &util::filename_cat($dirname, $before_hash);
1134	$before_hash = $self->eval_dir_dots($before_hash);
1135	}
1136
1137	my $linkfilename = &util::filename_cat ($base_dir, $before_hash);
1138
1139
1140	# print STDERR "**** linkfilename = $linkfilename\n";
1141	# if (!&util::fd_exists($linkfilename)) {
1142	# print STDERR "***** Warning: Could not find $linkfilename\n";
1143	# }
1144
1145
1146	# make sure there's a slash on the end if it's a directory
1147	if ($before_hash !~ m/\/$/) {
1148	$before_hash .= "/" if (-d $linkfilename);
1149	}
1150
1151	# print STDERR "*** returning: $before_hash\n";
1152
1153	return ("http://" . $before_hash, $hash_part, 1);
1154	} else {
1155	# mailto, news, nntp, telnet, javascript or gopher link
1156	return ($before_hash, "", 0);
1157	}
1158	}
1159
1160	sub extract_first_NNNN_characters {
1161	my $self = shift (@_);
1162	my ($textref, $doc_obj, $thissection) = @_;
1163
1164	foreach my $size (split /,/, $self->{'first'}) {
1165	my $tmptext = $$textref;
1166	# skip to the body
1167	$tmptext =~ s/.<body[^>]>//i;
1168	# remove javascript
1169	$tmptext =~ s@<script.*?</script>@ @sig;
1170	$tmptext =~ s/<[^>]*>/ /g;
1171	$tmptext =~ s/ / /g;
1172	$tmptext =~ s/^\s+//;
1173	$tmptext =~ s/\s+$//;
1174	$tmptext =~ s/\s+/ /gs;
1175	$tmptext = &unicode::substr ($tmptext, 0, $size);
1176	$tmptext =~ s/\s\S*$/…/; # adds an ellipse (...)
1177	$doc_obj->add_utf8_metadata ($thissection, "First$size", $tmptext);
1178	}
1179	}
1180
1181
1182	sub extract_metadata {
1183	my $self = shift (@_);
1184	my ($textref, $metadata, $doc_obj, $section) = @_;
1185	my $outhandle = $self->{'outhandle'};
1186	# if we don't want metadata, we may as well not be here ...
1187	return if (!defined $self->{'metadata_fields'});
1188
1189	my $separator = $self->{'metadata_field_separator'};
1190	if ($separator eq "") {
1191	undef $separator;
1192	}
1193
1194	# metadata fields to extract/save. 'key' is the (lowercase) name of the
1195	# html meta, 'value' is the metadata name for greenstone to use
1196	my %find_fields = ();
1197
1198	my %creator_fields = (); # short-cut for lookups
1199
1200
1201	foreach my $field (split /,/, $self->{'metadata_fields'}) {
1202	$field =~ s/^\s+//; # remove leading whitespace
1203	$field =~ s/\s+$//; # remove trailing whitespace
1204
1205	# support tag<tagname>
1206	if ($field =~ m/^(.?)\s<(.*?)>$/) {
1207	# "$2" is the user's preferred gs metadata name
1208	$find_fields{lc($1)}=$2; # lc = lowercase
1209	} else { # no <tagname> for mapping
1210	# "$field" is the user's preferred gs metadata name
1211	$find_fields{lc($field)}=$field; # lc = lowercase
1212	}
1213	}
1214
1215	if (defined $self->{'hunt_creator_metadata'} &&
1216	$self->{'hunt_creator_metadata'} == 1 ) {
1217	my @extra_fields =
1218	(
1219	'author',
1220	'author.email',
1221	'creator',
1222	'dc.creator',
1223	'dc.creator.corporatename',
1224	);
1225
1226	# add the creator_metadata fields to search for
1227	foreach my $field (@extra_fields) {
1228	$creator_fields{$field}=0; # add to lookup hash
1229	}
1230	}
1231
1232
1233	# find the header in the html file, which has the meta tags
1234	$$textref =~ m@<head>(.*?)</head>@si;
1235
1236	my $html_header=$1;
1237
1238	# go through every <meta... tag defined in the html and see if it is
1239	# one of the tags we want to match.
1240
1241	# special case for title - we want to remember if its been found
1242	my $found_title = 0;
1243	# this assumes that ">" won't appear. (I don't think it's allowed to...)
1244	$html_header =~ m/^/; # match the start of the string, for \G assertion
1245
1246	while ($html_header =~ m/\G.?<meta(.?)>/sig) {
1247	my $metatag=$1;
1248	my ($tag, $value);
1249
1250	# find the tag name
1251	$metatag =~ m/(?:name\|http-equiv)\s=\s([\"\'])?(.*?)\1/is;
1252	$tag=$2;
1253	# in case they're not using " or ', but they should...
1254	if (! $tag) {
1255	$metatag =~ m/(?:name\|http-equiv)\s=\s([^\s\>]+)/is;
1256	$tag=$1;
1257	}
1258
1259	if (!defined $tag) {
1260	print $outhandle "HTMLPlugin: can't find NAME in \"$metatag\"\n";
1261	next;
1262	}
1263
1264	# don't need to assign this field if it was passed in from a previous
1265	# (recursive) plugin
1266	if (defined $metadata->{$tag}) {next}
1267
1268	# find the tag content
1269	$metatag =~ m/content\s=\s([\"\'])?(.*?)\1/is;
1270	$value=$2;
1271
1272	if (! $value) {
1273	$metatag =~ m/(?:name\|http-equiv)\s=\s([^\s\>]+)/is;
1274	$value=$1;
1275	}
1276	if (!defined $value) {
1277	print $outhandle "HTMLPlugin: can't find VALUE in \"$metatag\"\n";
1278	next;
1279	}
1280
1281	# clean up and add
1282	$value =~ s/\s+/ /gs;
1283	chomp($value); # remove trailing \n, if any
1284	if (exists $creator_fields{lc($tag)}) {
1285	# map this value onto greenstone's "Creator" metadata
1286	$tag='Creator';
1287	} elsif (!exists $find_fields{lc($tag)}) {
1288	next; # don't want this tag
1289	} else {
1290	# get the user's preferred capitalisation
1291	$tag = $find_fields{lc($tag)};
1292	}
1293	if (lc($tag) eq "title") {
1294	$found_title = 1;
1295	}
1296
1297	if ($self->{'verbosity'} > 2) {
1298	print $outhandle " extracted \"$tag\" metadata \"$value\"\n";
1299	}
1300
1301	if ($tag =~ /\./) {
1302	# there is a . so has a namespace, add ex.
1303	$tag = "ex.$tag";
1304	}
1305	if (defined $separator) {
1306	my @values = split($separator, $value);
1307	foreach my $v (@values) {
1308	$doc_obj->add_utf8_metadata($section, $tag, $v) if $v =~ /\S/;
1309	}
1310	}
1311	else {
1312	$doc_obj->add_utf8_metadata($section, $tag, $value);
1313	}
1314	}
1315
1316	# TITLE: extract the document title
1317	if (exists $find_fields{'title'} && !$found_title) {
1318	# we want a title, and didn't find one in the meta tags
1319	# see if there's a <title> tag
1320	my $title;
1321	my $from = ""; # for debugging output only
1322	if ($html_header =~ m/<title[^>]>([^<]+)<\/title[^>]>/is) {
1323	$title = $1;
1324	$from = "<title> tags";
1325	}
1326
1327	if (!defined $title) {
1328	$from = "first 100 chars";
1329	# if no title use first 100 or so characters
1330	$title = $$textref;
1331	$title =~ s/^\xFE\xFF//; # Remove unicode byte order mark
1332	$title =~ s/^.*?<body>//si;
1333	# ignore javascript!
1334	$title =~ s@<script.*?</script>@ @sig;
1335	$title =~ s/<\/([^>]+)><\1>//g; # (eg) </b><b> - no space
1336	$title =~ s/<[^>]*>/ /g; # remove all HTML tags
1337	$title = substr ($title, 0, 100);
1338	$title =~ s/\s\S*$/.../;
1339	}
1340	$title =~ s/<[^>]*>/ /g; # remove html tags
1341	$title =~ s/ / /g;
1342	$title =~ s/(?: \|\xc2\xa0)/ /g; # utf-8 for nbsp...
1343	$title =~ s/\s+/ /gs; # collapse multiple spaces
1344	$title =~ s/^\s*//; # remove leading spaces
1345	$title =~ s/\s*$//; # remove trailing spaces
1346
1347	$title =~ s/^$self->{'title_sub'}// if ($self->{'title_sub'});
1348	$title =~ s/^\s+//s; # in case title_sub introduced any...
1349	$doc_obj->add_utf8_metadata ($section, "Title", $title);
1350	print $outhandle " extracted Title metadata \"$title\" from $from\n"
1351	if ($self->{'verbosity'} > 2);
1352	}
1353
1354	# add FileFormat metadata
1355	$doc_obj->add_metadata($section,"FileFormat", "HTML");
1356
1357	# Special, for metadata names such as tagH1 - extracts
1358	# the text between the first <H1> and </H1> tags into "H1" metadata.
1359
1360	foreach my $field (keys %find_fields) {
1361	if ($field !~ m/^tag([a-z0-9]+)$/i) {next}
1362	my $tag = $1;
1363	if ($$textref =~ m@<$tag[^>]>(.?)</$tag[^>]*>@g) {
1364	my $content = $1;
1365	$content =~ s/ / /g;
1366	$content =~ s/<[^>]*>/ /g;
1367	$content =~ s/^\s+//;
1368	$content =~ s/\s+$//;
1369	$content =~ s/\s+/ /gs;
1370	if ($content) {
1371	$tag=$find_fields{"tag$tag"}; # get the user's capitalisation
1372	$tag =~ s/^tag//i;
1373	$doc_obj->add_utf8_metadata ($section, $tag, $content);
1374	print $outhandle " extracted \"$tag\" metadata \"$content\"\n"
1375	if ($self->{'verbosity'} > 2);
1376	}
1377	}
1378	}
1379	}
1380
1381
1382	# evaluate any "../" to next directory up
1383	# evaluate any "./" as here
1384	sub eval_dir_dots {
1385	my $self = shift (@_);
1386	my ($filename) = @_;
1387	my $dirsep_os = &util::get_os_dirsep();
1388	my @dirsep = split(/$dirsep_os/,$filename);
1389
1390	my @eval_dirs = ();
1391	foreach my $d (@dirsep) {
1392	if ($d eq "..") {
1393	pop(@eval_dirs);
1394
1395	} elsif ($d eq ".") {
1396	# do nothing!
1397
1398	} else {
1399	push(@eval_dirs,$d);
1400	}
1401	}
1402
1403	# Need to fiddle with number of elements in @eval_dirs if the
1404	# first one is the empty string. This is because of a
1405	# modification to util::filename_cat that supresses the addition
1406	# of a leading '/' character (or \ if windows) (intended to help
1407	# filename cat with relative paths) if the first entry in the
1408	# array is the empty string. Making the array start with two
1409	# empty strings is a way to defeat this "smart" option.
1410	#
1411	if (scalar(@eval_dirs) > 0) {
1412	if ($eval_dirs[0] eq ""){
1413	unshift(@eval_dirs,"");
1414	}
1415	}
1416
1417	my $evaluated_filename = (scalar @eval_dirs > 0) ? &util::filename_cat(@eval_dirs) : "";
1418	return $evaluated_filename;
1419	}
1420
1421	sub replace_usemap_links {
1422	my $self = shift (@_);
1423	my ($front, $link, $back) = @_;
1424
1425	# remove quotes from link at start and end if necessary
1426	if ($link=~/^[\"\']/) {
1427	$link=~s/^[\"\']//;
1428	$link=~s/[\"\']$//;
1429	$front.='"';
1430	$back="\"$back";
1431	}
1432
1433	$link =~ s/^\.\///;
1434	return $front . $link . $back;
1435	}
1436
1437	sub inc_filecount {
1438	my $self = shift (@_);
1439
1440	if ($self->{'file_num'} == 1000) {
1441	$self->{'dir_num'} ++;
1442	$self->{'file_num'} = 0;
1443	} else {
1444	$self->{'file_num'} ++;
1445	}
1446	}
1447
1448
1449	# Extend read_file so that strings like é are
1450	# converted to UTF8 internally.
1451	#
1452	# We don't convert < or > or & or " in case
1453	# they interfere with the GML files
1454
1455	sub read_file {
1456	my $self = shift(@_);
1457	my ($filename, $encoding, $language, $textref) = @_;
1458
1459	$self->SUPER::read_file($filename, $encoding, $language, $textref);
1460
1461	# Convert entities to their Unicode code-point equivalents
1462	$$textref =~ s/&(lt\|gt\|amp\|quot\|nbsp);/&z$1;/go;
1463	$$textref =~ s/&([^;]+);/&ghtml::getcharequiv($1,1,1)/gseo;
1464	$$textref =~ s/&z(lt\|gt\|amp\|quot\|nbsp);/&$1;/go;
1465
1466	}
1467
1468	sub HB_read_html_file {
1469	my $self = shift (@_);
1470	my ($htmlfile, $text) = @_;
1471
1472	# load in the file
1473	if (!open (FILE, $htmlfile)) {
1474	print STDERR "ERROR - could not open $htmlfile\n";
1475	return;
1476	}
1477
1478	my $foundbody = 0;
1479	$self->HB_gettext (\$foundbody, $text, "FILE");
1480	close FILE;
1481
1482	# just in case there was no <body> tag
1483	if (!$foundbody) {
1484	$foundbody = 1;
1485	open (FILE, $htmlfile) \|\| return;
1486	$self->HB_gettext (\$foundbody, $text, "FILE");
1487	close FILE;
1488	}
1489	# text is in utf8
1490	}
1491
1492	# converts the text to utf8, as ghtml does that for é etc.
1493	sub HB_gettext {
1494	my $self = shift (@_);
1495	my ($foundbody, $text, $handle) = @_;
1496
1497	my $line = "";
1498	while (defined ($line = <$handle>)) {
1499	# look for body tag
1500	if (!$$foundbody) {
1501	if ($line =~ s/^.<body[^>]>//i) {
1502	$$foundbody = 1;
1503	} else {
1504	next;
1505	}
1506	}
1507
1508	# check for symbol fonts
1509	if ($line =~ m/<font [^>]?face\s=\s*\"?(\w+)\"?/i) {
1510	my $font = $1;
1511	print STDERR "HBPlug::HB_gettext - warning removed font $font\n"
1512	if ($font !~ m/^arial$/i);
1513	}
1514
1515	$$text .= $line;
1516	}
1517
1518	if ($self->{'input_encoding'} eq "iso_8859_1") {
1519	# convert to utf-8
1520	$$text=&unicode::unicode2utf8(&unicode::convert2unicode("iso_8859_1", $text));
1521	}
1522	# convert any alphanumeric character entities to their utf-8
1523	# equivalent for indexing purposes
1524	#&ghtml::convertcharentities ($$text);
1525
1526	$$text =~ s/\s+/ /g; # remove \n's
1527
1528	# At this point $$text is a binary byte string
1529	# => turn it into a Unicode aware string, so full
1530	# Unicode aware pattern matching can be used.
1531	# For instance: 's/\x{0101}//g' or '[[:upper:]]'
1532	#
1533
1534	$$text = decode("utf8",$$text);
1535	}
1536
1537	sub HB_clean_section {
1538	my $self = shift (@_);
1539	my ($section) = @_;
1540
1541	# remove tags without a starting tag from the section
1542	my ($tag, $tagstart);
1543	while ($section =~ m/<\/([^>]{1,10})>/) {
1544	$tag = $1;
1545	$tagstart = index($section, "<$tag");
1546	last if (($tagstart >= 0) && ($tagstart < index($section, "<\/$tag")));
1547	$section =~ s/<\/$tag>//;
1548	}
1549
1550	# remove extra paragraph tags
1551	while ($section =~ s/<p\b[^>]>\s<p\b/<p/ig) {}
1552
1553	# remove extra stuff at the end of the section
1554	while ($section =~ s/(<u>\|<i>\|<b>\|<p\b[^>]*>\| \|\s)$//i) {}
1555
1556	# add a newline at the beginning of each paragraph
1557	$section =~ s/(.)\s*<p\b/$1\n\n<p/gi;
1558
1559	# add a newline every 80 characters at a word boundary
1560	# Note: this regular expression puts a line feed before
1561	# the last word in each section, even when it is not
1562	# needed.
1563	$section =~ s/(.{1,80})\s/$1\n/g;
1564
1565	# fix up the image links
1566	$section =~ s/<img[^>]?src=\"?([^\">]+)\"?[^>]>/
1567	<center><img src=\"$1\" \/><\/center><br\/>/ig;
1568	$section =~ s/<<I>>\s*([^\.]+\.(png\|jpg\|gif))/
1569	<center><img src=\"$1\" \/><\/center><br\/>/ig;
1570
1571	return $section;
1572	}
1573
1574	# Will convert the oldHDL format to the new HDL format (using the Section tag)
1575	sub convert_to_newHDLformat
1576	{
1577	my $self = shift (@_);
1578	my ($file,$cnfile) = @_;
1579	my $input_filename = $file;
1580	my $tmp_filename = $cnfile;
1581
1582	# write HTML tmp file with new HDL format
1583	open (PROD, ">$tmp_filename") \|\| die("Error Writing to File: $tmp_filename $!");
1584
1585	# read in the file and do basic html cleaning (removing header etc)
1586	my $html = "";
1587	$self->HB_read_html_file ($input_filename, \$html);
1588
1589	# process the file one section at a time
1590	my $curtoclevel = 1;
1591	my $firstsection = 1;
1592	my $toclevel = 0;
1593	while (length ($html) > 0) {
1594	if ($html =~ s/^.?(?:<p\b[^>]>)?((<b>\|<i>\|<u>\|\s))<<TOC(\d+)>>\s(.*?)<p\b/<p/i) {
1595	$toclevel = $3;
1596	my $title = $4;
1597	my $sectiontext = "";
1598	if ($html =~ s/^(.?)((?:<p\b[^>]>)?((<b>\|<i>\|<u>\|\s)*)<<TOC\d+>>)/$2/i) {
1599	$sectiontext = $1;
1600	} else {
1601	$sectiontext = $html;
1602	$html = "";
1603	}
1604
1605	# remove tags and extra spaces from the title
1606	$title =~ s/<\/?[^>]+>//g;
1607	$title =~ s/^\s+\|\s+$//g;
1608
1609	# close any sections below the current level and
1610	# create a new section (special case for the firstsection)
1611	print PROD "<!--\n";
1612	while (($curtoclevel > $toclevel) \|\|
1613	(!$firstsection && $curtoclevel == $toclevel)) {
1614	$curtoclevel--;
1615	print PROD "</Section>\n";
1616	}
1617	if ($curtoclevel+1 < $toclevel) {
1618	print STDERR "WARNING - jump in toc levels in $input_filename " .
1619	"from $curtoclevel to $toclevel\n";
1620	}
1621	while ($curtoclevel < $toclevel) {
1622	$curtoclevel++;
1623	}
1624
1625	if ($curtoclevel == 1) {
1626	# add the header tag
1627	print PROD "-->\n";
1628	print PROD "<HTML>\n<HEAD>\n<TITLE>$title</TITLE>\n</HEAD>\n<BODY>\n";
1629	print PROD "<!--\n";
1630	}
1631
1632	print PROD "<Section>\n\t<Description>\n\t\t<Metadata name=\"Title\">$title</Metadata>\n\t</Description>\n";
1633
1634	print PROD "-->\n";
1635
1636	# clean up the section html
1637	$sectiontext = $self->HB_clean_section($sectiontext);
1638
1639	print PROD "$sectiontext\n";
1640
1641	} else {
1642	print STDERR "WARNING - leftover text\n" , $self->shorten($html),
1643	"\nin $input_filename\n";
1644	last;
1645	}
1646	$firstsection = 0;
1647	}
1648
1649	print PROD "<!--\n";
1650	while ($curtoclevel > 0) {
1651	$curtoclevel--;
1652	print PROD "</Section>\n";
1653	}
1654	print PROD "-->\n";
1655
1656	close (PROD) \|\| die("Error Closing File: $tmp_filename $!");
1657
1658	return $tmp_filename;
1659	}
1660
1661	sub shorten {
1662	my $self = shift (@_);
1663	my ($text) = @_;
1664
1665	return "\"$text\"" if (length($text) < 100);
1666
1667	return "\"" . substr ($text, 0, 50) . "\" ... \"" .
1668	substr ($text, length($text)-50) . "\"";
1669	}
1670
1671	sub convert_tidy_or_oldHDL_file
1672	{
1673	my $self = shift (@_);
1674	my ($file) = @_;
1675	my $input_filename = $file;
1676
1677	if (-d $input_filename)
1678	{
1679	return $input_filename;
1680	}
1681
1682	# get the input filename
1683	my ($tailname, $dirname, $suffix) = &File::Basename::fileparse($input_filename, "\\.[^\\.]+\$");
1684	my $base_dirname = $dirname;
1685	$suffix = lc($suffix);
1686
1687	# derive tmp filename from input filename
1688	# Remove any white space from filename -- no risk of name collision, and
1689	# makes later conversion by utils simpler. Leave spaces in path...
1690	# tidy up the filename with space, dot, hyphen between
1691	$tailname =~ s/\s+//g;
1692	$tailname =~ s/\.+//g;
1693	$tailname =~ s/\-+//g;
1694	# convert to utf-8 otherwise we have problems with the doc.xml file
1695	# later on
1696	&unicode::ensure_utf8(\$tailname);
1697
1698	# softlink to collection tmp dir
1699	my $tmp_dirname = &util::filename_cat($ENV{'GSDLCOLLECTDIR'}, "tidytmp");
1700	&util::mk_dir($tmp_dirname) if (!-e $tmp_dirname);
1701
1702	my $test_dirname = "";
1703	my $f_separator = &util::get_os_dirsep();
1704
1705	if ($dirname =~ m/import$f_separator/)
1706	{
1707	$test_dirname = $'; #'
1708
1709	#print STDERR "init $'\n";
1710
1711	while ($test_dirname =~ m/[$f_separator]/)
1712	{
1713	my $folderdirname = $`;
1714	$tmp_dirname = &util::filename_cat($tmp_dirname,$folderdirname);
1715	&util::mk_dir($tmp_dirname) if (!-e $tmp_dirname);
1716	$test_dirname = $'; #'
1717	}
1718	}
1719
1720	my $tmp_filename = &util::filename_cat($tmp_dirname, "$tailname$suffix");
1721
1722	# tidy or convert the input file if it is a HTML-like file or it is accepted by the process_exp
1723	if (($suffix eq ".htm") \|\| ($suffix eq ".html") \|\| ($suffix eq ".shtml"))
1724	{
1725	#convert the input file to a new style HDL
1726	my $hdl_output_filename = $input_filename;
1727	if ($self->{'old_style_HDL'})
1728	{
1729	$hdl_output_filename = &util::filename_cat($tmp_dirname, "$tailname$suffix");
1730	$hdl_output_filename = $self->convert_to_newHDLformat($input_filename,$hdl_output_filename);
1731	}
1732
1733	#just for checking copy all other file from the base dir to tmp dir if it is not exists
1734	opendir(DIR,$base_dirname) or die "Can't open base directory : $base_dirname!";
1735	my @files = grep {!/^\.+$/} readdir(DIR);
1736	close(DIR);
1737
1738	foreach my $file (@files)
1739	{
1740	my $src_file = &util::filename_cat($base_dirname,$file);
1741	my $dest_file = &util::filename_cat($tmp_dirname,$file);
1742	if ((!-e $dest_file) && (!-d $src_file))
1743	{
1744	# just copy the original file back to the tmp directory
1745	copy($src_file,$dest_file) or die "Can't copy file $src_file to $dest_file $!";
1746	}
1747	}
1748
1749	# tidy the input file
1750	my $tidy_output_filename = $hdl_output_filename;
1751	if ($self->{'use_realistic_book'})
1752	{
1753	$tidy_output_filename = &util::filename_cat($tmp_dirname, "$tailname$suffix");
1754	$tidy_output_filename = $self->tmp_tidy_file($hdl_output_filename,$tidy_output_filename);
1755	}
1756	$tmp_filename = $tidy_output_filename;
1757	}
1758	else
1759	{
1760	if (!-e $tmp_filename)
1761	{
1762	# just copy the original file back to the tmp directory
1763	copy($input_filename,$tmp_filename) or die "Can't copy file $input_filename to $tmp_filename $!";
1764	}
1765	}
1766
1767	return $tmp_filename;
1768	}
1769
1770
1771	# Will make the html input file as a proper XML file with removed font tag and
1772	# image size added to the img tag.
1773	# The tidying process takes place in a collection specific 'tmp' directory so
1774	# that we don't accidentally damage the input.
1775	sub tmp_tidy_file
1776	{
1777	my $self = shift (@_);
1778	my ($file,$cnfile) = @_;
1779	my $input_filename = $file;
1780	my $tmp_filename = $cnfile;
1781
1782	# get the input filename
1783	my ($tailname, $dirname, $suffix) = &File::Basename::fileparse($input_filename, "\\.[^\\.]+\$");
1784
1785	require HTML::TokeParser::Simple;
1786
1787	# create HTML parser to decode the input file
1788	my $parser = HTML::TokeParser::Simple->new($input_filename);
1789
1790	# write HTML tmp file without the font tag and image size are added to the img tag
1791	open (PROD, ">$tmp_filename") \|\| die("Error Writing to File: $tmp_filename $!");
1792	while (my $token = $parser->get_token())
1793	{
1794	# is it an img tag
1795	if ($token->is_start_tag('img'))
1796	{
1797	# get the attributes
1798	my $attr = $token->return_attr;
1799
1800	# get the full path to the image
1801	my $img_file = &util::filename_cat($dirname,$attr->{src});
1802
1803	# set the width and height attribute
1804	($attr->{width}, $attr->{height}) = imgsize($img_file);
1805
1806	# recreate the tag
1807	print PROD "<img";
1808	print PROD map { qq { $_="$attr->{$_}"} } keys %$attr;
1809	print PROD ">";
1810	}
1811	# is it a font tag
1812	else
1813	{
1814	if (($token->is_start_tag('font')) \|\| ($token->is_end_tag('font')))
1815	{
1816	# remove font tag
1817	print PROD "";
1818	}
1819	else
1820	{
1821	# print without changes
1822	print PROD $token->as_is;
1823	}
1824	}
1825	}
1826	close (PROD) \|\| die("Error Closing File: $tmp_filename $!");
1827
1828	# run html-tidy on the tmp file to make it a proper XML file
1829
1830	my $outhandle = $self->{'outhandle'};
1831	print $outhandle "Converting HTML to be XML compliant:\n";
1832
1833	my $tidy_cmd = "tidy";
1834	$tidy_cmd .= " -q" if ($self->{'verbosity'} <= 2);
1835	$tidy_cmd .= " -raw -wrap 0 -asxml \"$tmp_filename\"";
1836	if ($self->{'verbosity'} <= 2) {
1837	if ($ENV{'GSDLOS'} =~ m/^windows/i) {
1838	$tidy_cmd .= " 2>nul";
1839	}
1840	else {
1841	$tidy_cmd .= " 2>/dev/null";
1842	}
1843	print $outhandle " => $tidy_cmd\n";
1844	}
1845
1846	my $tidyfile = `$tidy_cmd`;
1847
1848	# write result back to the tmp file
1849	open (PROD, ">$tmp_filename") \|\| die("Error Writing to File: $tmp_filename $!");
1850	print PROD $tidyfile;
1851	close (PROD) \|\| die("Error Closing File: $tmp_filename $!");
1852
1853	# return the output filename
1854	return $tmp_filename;
1855	}
1856
1857	sub associate_cover_image
1858	{
1859	my $self = shift(@_);
1860	my ($doc_obj, $filename) = @_;
1861	if (($self->{'use_realistic_book'}) \|\| ($self->{'old_style_HDL'}))
1862	{
1863	# we will have cover image in tidytmp, but want it from import
1864	$filename =~ s/([\\\/])tidytmp([\\\/])/$1import$2/;
1865	}
1866	$self->SUPER::associate_cover_image($doc_obj, $filename);
1867	}
1868
1869
1870	1;

Note: See TracBrowser for help on using the repository browser.

Download in other formats: