Context Navigation

source: gsdl/trunk/perllib/plugins/HTMLPlugin.pm@ 20791

Last change on this file since 20791 was 20791, checked in by kjdon, 15 years ago
new option -processing_tmp_files - if set, then don't store the assicated files using associate_source_file - we don't want to store tmp files in the archivesinf databases
Property svn:executable set to ``* Property svn:keywords set to `Author Date Id Revision`
File size: 52.1 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 ReadTextFile;
39	use HBPlugin;
40	use ghtml;
41	use unicode;
42	use util;
43	use XMLParser;
44
45	use Image::Size;
46	use File::Copy;
47
48	sub BEGIN {
49	@HTMLPlugin::ISA = ('ReadTextFile', 'HBPlugin');
50	}
51
52	use strict; # every perl program should have this!
53	no strict 'refs'; # make an exception so we can use variables as filehandles
54
55	my $arguments =
56	[ { 'name' => "process_exp",
57	'desc' => "{BasePlugin.process_exp}",
58	'type' => "regexp",
59	'deft' => &get_default_process_exp() },
60	{ 'name' => "block_exp",
61	'desc' => "{BasePlugin.block_exp}",
62	'type' => 'regexp',
63	'deft' => &get_default_block_exp() },
64	{ 'name' => "nolinks",
65	'desc' => "{HTMLPlugin.nolinks}",
66	'type' => "flag" },
67	{ 'name' => "keep_head",
68	'desc' => "{HTMLPlugin.keep_head}",
69	'type' => "flag" },
70	{ 'name' => "no_metadata",
71	'desc' => "{HTMLPlugin.no_metadata}",
72	'type' => "flag" },
73	{ 'name' => "metadata_fields",
74	'desc' => "{HTMLPlugin.metadata_fields}",
75	'type' => "string",
76	'deft' => "Title" },
77	{ 'name' => "hunt_creator_metadata",
78	'desc' => "{HTMLPlugin.hunt_creator_metadata}",
79	'type' => "flag" },
80	{ 'name' => "file_is_url",
81	'desc' => "{HTMLPlugin.file_is_url}",
82	'type' => "flag" },
83	{ 'name' => "assoc_files",
84	'desc' => "{HTMLPlugin.assoc_files}",
85	'type' => "regexp",
86	'deft' => &get_default_block_exp() },
87	{ 'name' => "rename_assoc_files",
88	'desc' => "{HTMLPlugin.rename_assoc_files}",
89	'type' => "flag" },
90	{ 'name' => "title_sub",
91	'desc' => "{HTMLPlugin.title_sub}",
92	'type' => "string",
93	'deft' => "" },
94	{ 'name' => "description_tags",
95	'desc' => "{HTMLPlugin.description_tags}",
96	'type' => "flag" },
97	# retain this for backward compatibility (w3mir option was replaced by
98	# file_is_url)
99	{ 'name' => "w3mir",
100	# 'desc' => "{HTMLPlugin.w3mir}",
101	'type' => "flag",
102	'hiddengli' => "yes"},
103	{ 'name' => "no_strip_metadata_html",
104	'desc' => "{HTMLPlugin.no_strip_metadata_html}",
105	'type' => "string",
106	'deft' => "",
107	'reqd' => "no"},
108	{ 'name' => "sectionalise_using_h_tags",
109	'desc' => "{HTMLPlugin.sectionalise_using_h_tags}",
110	'type' => "flag" },
111	{ 'name' => "use_realistic_book",
112	'desc' => "{HTMLPlugin.tidy_html}",
113	'type' => "flag"},
114	{ 'name' => "old_style_HDL",
115	'desc' => "{HTMLPlugin.old_style_HDL}",
116	'type' => "flag"},
117	{'name' => "processing_tmp_files",
118	'desc' => "{BasePlugin.processing_tmp_files}",
119	'type' => "flag",
120	'hiddengli' => "yes"}
121	];
122
123	my $options = { 'name' => "HTMLPlugin",
124	'desc' => "{HTMLPlugin.desc}",
125	'abstract' => "no",
126	'inherits' => "yes",
127	'args' => $arguments };
128
129
130	sub new {
131	my ($class) = shift (@_);
132	my ($pluginlist,$inputargs,$hashArgOptLists) = @_;
133	push(@$pluginlist, $class);
134
135	push(@{$hashArgOptLists->{"ArgList"}},@{$arguments});
136	push(@{$hashArgOptLists->{"OptList"}},$options);
137
138
139	my $self = new ReadTextFile($pluginlist,$inputargs,$hashArgOptLists);
140
141	if ($self->{'w3mir'}) {
142	$self->{'file_is_url'} = 1;
143	}
144	$self->{'aux_files'} = {};
145	$self->{'dir_num'} = 0;
146	$self->{'file_num'} = 0;
147
148	return bless $self, $class;
149	}
150
151	# may want to use (?i)\.(gif\|jpe?g\|jpe\|png\|css\|js(?:@.*)?)$
152	# if have eg <script language="javascript" src="img/lib.js@123">
153	# blocking is now done by reading through the file and recording all the
154	# images and other files
155	sub get_default_block_exp {
156	my $self = shift (@_);
157
158	#return q^(?i)\.(gif\|jpe?g\|jpe\|jpg\|png\|css)$^;
159	return "";
160	}
161
162	sub get_default_process_exp {
163	my $self = shift (@_);
164
165	# the last option is an attempt to encode the concept of an html query ...
166	return q^(?i)(\.html?\|\.shtml\|\.shm\|\.asp\|\.php\d?\|\.cgi\|.+\?.+=.*)$^;
167	}
168
169	sub store_block_files
170	{
171	my $self =shift (@_);
172	my ($filename_full_path, $block_hash) = @_;
173
174	my $html_fname = $filename_full_path;
175	my @file_blocks;
176
177	my ($language, $encoding) = $self->textcat_get_language_encoding ($filename_full_path);
178
179	# read in file ($text will be in utf8)
180	my $raw_text = "";
181	$self->read_file_no_decoding ($filename_full_path, \$raw_text);
182
183	my $textref = \$raw_text;
184	my $opencom = '(?:<!--\|<!(?:—\|\|--))';
185	my $closecom = '(?:-->\|(?:—\|\|--)>)';
186	$$textref =~ s/$opencom(.*?)$closecom//gs;
187
188	my $attval = "\\\"[^\\\"]+\\\"\|[^\\s>]+";
189	my @img_matches = ($$textref =~ m/<img[^>]?src\s=\s($attval)[^>]>/igs);
190	my @usemap_matches = ($$textref =~ m/<img[^>]?usemap\s=\s($attval)[^>]>/igs);
191	my @link_matches = ($$textref =~ m/<link[^>]?href\s=\s($attval)[^>]>/igs);
192	my @embed_matches = ($$textref =~ m/<embed[^>]?src\s=\s($attval)[^>]>/igs);
193	my @tabbg_matches = ($$textref =~ m/<(?:body\|table\|tr\|td)[^>]?background\s=\s($attval)[^>]>/igs);
194	my @script_matches = ($$textref =~ m/<script[^>]?src\s=\s($attval)[^>]>/igs);
195
196	if(!defined $self->{'utf8_to_original_filename'}) {
197	# maps from utf8 converted link name -> original filename referrred to by (possibly URL-encoded) src url
198	$self->{'utf8_to_original_filename'} = {};
199	}
200
201	foreach my $link (@img_matches, @usemap_matches, @link_matches, @embed_matches, @tabbg_matches, @script_matches) {
202
203	# remove quotes from link at start and end if necessary
204	if ($link=~/^\"/) {
205	$link=~s/^\"//;
206	$link=~s/\"$//;
207	}
208
209	$link =~ s/\#.*$//s; # remove any anchor names, e.g. foo.html#name becomes foo.html
210	# some links may just be anchor names
211	next unless ($link =~ /\S+/);
212
213	if ($link !~ m@^/@ && $link !~ m/^([A-Z]:?)\\/) {
214	# Turn relative file path into full path
215	my $dirname = &File::Basename::dirname($filename_full_path);
216	$link = &util::filename_cat($dirname, $link);
217	}
218	$link = $self->eval_dir_dots($link);
219
220	# this is the actual filename on the filesystem (that the link refers to)
221	my $url_original_filename = $self->opt_url_decode($link);
222
223	# Convert the url_original_filename into its utf8 version. Store the utf8 link along with the url_original_filename
224	my $utf8_link = "";
225	$self->decode_text($link,$encoding,$language,\$utf8_link);
226
227	$self->{'utf8_to_original_filename'}->{$utf8_link} = $url_original_filename;
228	# print STDERR "**** utf8_encoded_link to original src filename:\n\t$utf8_link\n\t".$self->{'utf8_to_original_filename'}->{$utf8_link}."\n";
229
230	if ($url_original_filename ne $utf8_link) {
231	my $outhandle = $self->{'outhandle'};
232
233	print $outhandle "URL Encoding $url_original_filename\n";
234	print $outhandle " ->$utf8_link\n";
235	}
236
237	$block_hash->{'file_blocks'}->{$url_original_filename} = 1;
238	}
239	}
240
241	# Given a filename in any encoding, will URL decode it to get back the original filename
242	# in the original encoding. Because this method is intended to work out the original
243	# filename, it does not URL decode any filename if a file by the name of the URL-encoded*
244	# string already exists in the local folder.
245	# Return the original filename corresponding to the parameter URL-encoded filename, and
246	# a decoded flag that is set to true iff URL-decoding had to be applied.
247	sub opt_url_decode {
248	my $self = shift (@_);
249	my ($link) = @_;
250
251	# Replace %XX's in URL with decoded value if required.
252	# Note that the filename may include the %XX in some situations
253	if ($link =~ m/\%[A-F0-9]{2}/i) {
254	if (!-e $link) {
255	$link = &unicode::url_decode($link);
256	}
257	}
258
259	return $link;
260	}
261
262	sub read_into_doc_obj
263	{
264	my $self = shift (@_);
265	my ($pluginfo, $base_dir, $file, $block_hash, $metadata, $processor, $maxdocs, $total_count, $gli) = @_;
266
267	# get the input file
268	my $input_filename = $file;
269	my ($tailname, $dirname, $suffix) = &File::Basename::fileparse($input_filename, "\\.[^\\.]+\$");
270	$suffix = lc($suffix);
271
272	if (($self->{'use_realistic_book'}) \|\| ($self->{'old_style_HDL'}))
273	{
274	# because the document has to be sectionalized set the description tags
275	$self->{'description_tags'} = 1;
276
277	# set the file to be tidied
278	$input_filename = &util::filename_cat($base_dir,$file) if $base_dir =~ m/\w/;
279
280	# get the tidied file
281	#my $tidy_filename = $self->tmp_tidy_file($input_filename);
282	my $tidy_filename = $self->convert_tidy_or_oldHDL_file($input_filename);
283
284	# derive tmp filename from input filename
285	my ($tailname, $dirname, $suffix) = &File::Basename::fileparse($tidy_filename, "\\.[^\\.]+\$");
286
287	# set the new input file and base_dir to be from the tidied file
288	$file = "$tailname$suffix";
289	$base_dir = $dirname;
290	}
291
292	# call the parent read_into_doc_obj
293	my ($process_status,$doc_obj) = $self->SUPER::read_into_doc_obj($pluginfo, $base_dir, $file, $block_hash, $metadata, $processor, $maxdocs, $total_count, $gli);
294
295	return ($process_status,$doc_obj);
296	}
297
298	# do plugin specific processing of doc_obj
299	sub process {
300	my $self = shift (@_);
301	my ($textref, $pluginfo, $base_dir, $file, $metadata, $doc_obj, $gli) = @_;
302	my $outhandle = $self->{'outhandle'};
303
304	if ($ENV{'GSDLOS'} =~ m/^windows/i) {
305	# this makes life so much easier... perl can cope with unix-style '/'s.
306	$base_dir =~ s@(\\)+@/@g;
307	$file =~ s@(\\)+@/@g;
308	}
309
310	# reset per-doc stuff...
311	$self->{'aux_files'} = {};
312	$self->{'dir_num'} = 0;
313	$self->{'file_num'} = 0;
314
315	# process an HTML file where sections are divided by headings tags (H1, H2 ...)
316	# you can also include metadata in the format (X can be any number)
317	# <hX>Title<!--gsdl-metadata
318	# <Metadata name="name1">value1</Metadata>
319	# ...
320	# <Metadata name="nameN">valueN</Metadata>
321	#--></hX>
322	if ($self->{'sectionalise_using_h_tags'}) {
323	# description_tags should allways be activated because we convert headings to description tags
324	$self->{'description_tags'} = 1;
325
326	my $arrSections = [];
327	$$textref =~ s/<h([0-9]+)[^>]>(.?)<\/h[0-9]+>/$self->process_heading($1, $2, $arrSections, $file)/isge;
328
329	if (scalar(@$arrSections)) {
330	my $strMetadata = $self->update_section_data($arrSections, -1);
331	if (length($strMetadata)) {
332	$strMetadata = '<!--' . $strMetadata . "\n-->\n</body>";
333	$$textref =~ s/<\/body>/$strMetadata/ig;
334	}
335	}
336	}
337
338	my $cursection = $doc_obj->get_top_section();
339
340	$self->extract_metadata ($textref, $metadata, $doc_obj, $cursection)
341	unless $self->{'no_metadata'} \|\| $self->{'description_tags'};
342
343	# Store URL for page as metadata - this can be used for an
344	# altavista style search interface. The URL won't be valid
345	# unless the file structure contains the domain name (i.e.
346	# like when w3mir is used to download a website).
347
348	# URL metadata (even invalid ones) are used to support internal
349	# links, so even if 'file_is_url' is off, still need to store info
350
351	my ($tailname,$dirname,$suffix) = &File::Basename::fileparse($file, "\\.[^\\.]+\$");
352	my $utf8_file = $self->filename_to_utf8_metadata($file);
353	$utf8_file =~ s/&\#095;/_/g;
354	my $web_url = "http://";
355	if(defined $dirname) { # local directory
356	$dirname = $self->eval_dir_dots($dirname);
357	$dirname .= &util::get_dirsep() if $dirname ne ""; # if there's a directory, it should end on "/"
358	$web_url = $web_url.$dirname.$utf8_file;
359	} else {
360	$web_url = $web_url.$utf8_file;
361	}
362	$web_url =~ s/\\/\//g;
363	$doc_obj->add_utf8_metadata($cursection, "URL", $web_url);
364
365	if ($self->{'file_is_url'}) {
366	$doc_obj->add_metadata($cursection, "weblink", "<a href=\"$web_url\">");
367	$doc_obj->add_metadata($cursection, "webicon", "_iconworld_");
368	$doc_obj->add_metadata($cursection, "/weblink", "</a>");
369	}
370
371	if ($self->{'description_tags'}) {
372	# remove the html header - note that doing this here means any
373	# sections defined within the header will be lost (so all <Section>
374	# tags must appear within the body of the HTML)
375	my ($head_keep) = ($$textref =~ m/^(.?)<body[^>]>/is);
376
377	$$textref =~ s/^.?<body[^>]>//is;
378	$$textref =~ s/(<\/body[^>]>\|<\/html[^>]>)//isg;
379
380	my $opencom = '(?:<!--\|<!(?:—\|\|--))';
381	my $closecom = '(?:-->\|(?:—\|\|--)>)';
382
383	my $lt = '(?:<\|<)';
384	my $gt = '(?:>\|>)';
385	my $quot = '(?:"\|"\|”\|“)';
386
387	my $dont_strip = '';
388	if ($self->{'no_strip_metadata_html'}) {
389	($dont_strip = $self->{'no_strip_metadata_html'}) =~ s{,}{\|}g;
390	}
391
392	my $found_something = 0; my $top = 1;
393	while ($$textref =~ s/^(.?)$opencom(.?)$closecom//s) {
394	my $text = $1;
395	my $comment = $2;
396	if (defined $text) {
397	# text before a comment - note that getting to here
398	# doesn't necessarily mean there are Section tags in
399	# the document
400	$self->process_section(\$text, $base_dir, $file, $doc_obj, $cursection);
401	}
402	while ($comment =~ s/$lt(.*?)$gt//s) {
403	my $tag = $1;
404	if ($tag eq "Section") {
405	$found_something = 1;
406	$cursection = $doc_obj->insert_section($doc_obj->get_end_child($cursection)) unless $top;
407	$top = 0;
408	} elsif ($tag eq "/Section") {
409	$found_something = 1;
410	$cursection = $doc_obj->get_parent_section ($cursection);
411	} elsif ($tag =~ m/^Metadata name=$quot(.*?)$quot/s) {
412	my $metaname = $1;
413	my $accumulate = $tag =~ m/mode=${quot}accumulate${quot}/ ? 1 : 0;
414	$comment =~ s/^(.*?)$lt\/Metadata$gt//s;
415	my $metavalue = $1;
416	$metavalue =~ s/^\s+//;
417	$metavalue =~ s/\s+$//;
418	# assume that no metadata value intentionally includes
419	# carriage returns or HTML tags (if they're there they
420	# were probably introduced when converting to HTML from
421	# some other format).
422	# actually some people want to have html tags in their
423	# metadata.
424	$metavalue =~ s/[\cJ\cM]/ /sg;
425	$metavalue =~ s/<[^>]+>//sg
426	unless $dont_strip && ($dont_strip eq 'all' \|\| $metaname =~ m/^($dont_strip)$/);
427	$metavalue =~ s/\s+/ /sg;
428	if ($accumulate) {
429	$doc_obj->add_utf8_metadata($cursection, $metaname, $metavalue);
430	} else {
431	$doc_obj->set_utf8_metadata_element($cursection, $metaname, $metavalue);
432	}
433	} elsif ($tag eq "Description" \|\| $tag eq "/Description") {
434	# do nothing with containing Description tags
435	} else {
436	# simple HTML tag (probably created by the conversion
437	# to HTML from some other format) - we'll ignore it and
438	# hope for the best ;-)
439	}
440	}
441	}
442	if ($cursection ne "") {
443	print $outhandle "HTMLPlugin: WARNING: $file contains unmatched <Section></Section> tags\n";
444	}
445
446	$$textref =~ s/^.?<body[^>]>//is;
447	$$textref =~ s/(<\/body[^>]>\|<\/html[^>]>)//isg;
448	if ($$textref =~ m/\S/) {
449	if (!$found_something) {
450	if ($self->{'verbosity'} > 2) {
451	print $outhandle "HTMLPlugin: WARNING: $file appears to contain no Section tags so\n";
452	print $outhandle " will be processed as a single section document\n";
453	}
454
455	# go ahead and process single-section document
456	$self->process_section($textref, $base_dir, $file, $doc_obj, $cursection);
457
458	# if document contains no Section tags we'll go ahead
459	# and extract metadata (this won't have been done
460	# above as the -description_tags option prevents it)
461	my $complete_text = $head_keep.$doc_obj->get_text($cursection);
462	$self->extract_metadata (\$complete_text, $metadata, $doc_obj, $cursection)
463	unless $self->{'no_metadata'};
464
465	} else {
466	print $outhandle "HTMLPlugin: WARNING: $file contains the following text outside\n";
467	print $outhandle " of the final closing </Section> tag. This text will\n";
468	print $outhandle " be ignored.";
469
470	my ($text);
471	if (length($$textref) > 30) {
472	$text = substr($$textref, 0, 30) . "...";
473	} else {
474	$text = $$textref;
475	}
476	$text =~ s/\n/ /isg;
477	print $outhandle " ($text)\n";
478	}
479	} elsif (!$found_something) {
480
481	if ($self->{'verbosity'} > 2) {
482	# may get to here if document contained no valid Section
483	# tags but did contain some comments. The text will have
484	# been processed already but we should print the warning
485	# as above and extract metadata
486	print $outhandle "HTMLPlugin: WARNING: $file appears to contain no Section tags and\n";
487	print $outhandle " is blank or empty. Metadata will be assigned if present.\n";
488	}
489
490	my $complete_text = $head_keep.$doc_obj->get_text($cursection);
491	$self->extract_metadata (\$complete_text, $metadata, $doc_obj, $cursection)
492	unless $self->{'no_metadata'};
493	}
494
495	} else {
496
497	# remove header and footer
498	if (!$self->{'keep_head'} \|\| $self->{'description_tags'}) {
499	$$textref =~ s/^.?<body[^>]>//is;
500	$$textref =~ s/(<\/body[^>]>\|<\/html[^>]>)//isg;
501	}
502
503	# single section document
504	$self->process_section($textref, $base_dir, $file, $doc_obj, $cursection);
505	}
506	return 1;
507	}
508
509
510	sub process_heading
511	{
512	my ($self, $nHeadNo, $strHeadingText, $arrSections, $file) = @_;
513	$strHeadingText = '' if (!defined($strHeadingText));
514
515	my $strMetadata = $self->update_section_data($arrSections, int($nHeadNo));
516
517	my $strSecMetadata = '';
518	while ($strHeadingText =~ s/<!--gsdl-metadata(.*?)-->//is)
519	{
520	$strSecMetadata .= $1;
521	}
522
523	$strHeadingText =~ s/^\s+//g;
524	$strHeadingText =~ s/\s+$//g;
525	$strSecMetadata =~ s/^\s+//g;
526	$strSecMetadata =~ s/\s+$//g;
527
528	$strMetadata .= "\n<Section>\n\t<Description>\n\t\t<Metadata name=\"Title\">" . $strHeadingText . "</Metadata>\n";
529
530	if (length($strSecMetadata)) {
531	$strMetadata .= "\t\t" . $strSecMetadata . "\n";
532	}
533
534	$strMetadata .= "\t</Description>\n";
535
536	return "<!--" . $strMetadata . "-->";
537	}
538
539
540	sub update_section_data
541	{
542	my ($self, $arrSections, $nCurTocNo) = @_;
543	my ($strBuffer, $nLast, $nSections) = ('', 0, scalar(@$arrSections));
544
545	if ($nSections == 0) {
546	push @$arrSections, $nCurTocNo;
547	return $strBuffer;
548	}
549	$nLast = $arrSections->[$nSections - 1];
550	if ($nCurTocNo > $nLast) {
551	push @$arrSections, $nCurTocNo;
552	return $strBuffer;
553	}
554	for(my $i = $nSections - 1; $i >= 0; $i--) {
555	if ($nCurTocNo <= $arrSections->[$i]) {
556	$strBuffer .= "\n</Section>";
557	pop @$arrSections;
558	}
559	}
560	push @$arrSections, $nCurTocNo;
561	return $strBuffer;
562	}
563
564
565	# note that process_section may be called multiple times for a single
566	# section (relying on the fact that add_utf8_text appends the text to any
567	# that may exist already).
568	sub process_section {
569	my $self = shift (@_);
570	my ($textref, $base_dir, $file, $doc_obj, $cursection) = @_;
571	# trap links
572	if (!$self->{'nolinks'}) {
573	# usemap="./#index" not handled correctly => change to "#index"
574	## $$textref =~ s/(<img[^>]?usemap\s=\s[\"\']?)([^\"\'>\s]+)([\"\']?[^>]>)/
575
576	$$textref =~ s/(<img[^>]?usemap\s=\s)((?:[\"][^\"]+[\"])\|(?:[\'][^\']+[\'])\|(?:[^\s\/>]+))([^>]>)/
577	$self->replace_usemap_links($1, $2, $3)/isge;
578
579	## $$textref =~ s/(<(?:a\|area\|frame\|link\|script)\s+[^>]?\s(?:href\|src)\s=\s[\"\']?)([^\"\'>\s]+)([\"\']?[^>]*>)/
580
581	$$textref =~ s/(<(?:a\|area\|frame\|link\|script)\s+[^>]?\s(?:href\|src)\s=\s)((?:[\"][^\"]+[\"])\|(?:[\'][^\']+[\'])\|(?:[^\s\/>]+))([^>]*>)/
582	$self->replace_href_links ($1, $2, $3, $base_dir, $file, $doc_obj, $cursection)/isge;
583	}
584
585	# trap images
586
587	# Previously, by default, HTMLPlugin would embed <img> tags inside anchor tags
588	# i.e. <a href="image><img src="image"></a> in order to overcome a problem that
589	# turned regular text succeeding images into links. That is, by embedding <imgs>
590	# inside <a href=""></a>, the text following images were no longer misbehaving.
591	# However, there would be many occasions whereby images were not meant to link
592	# to their source images but where the images would link to another web page.
593	# To allow this, the no_image_links option was introduced: it would prevent
594	# the behaviour of embedding images into links that referenced the source images.
595
596	# Somewhere along the line, the problem of normal text turning into links when
597	# such text followed images which were not embedded in <a href=""></a> ceased
598	# to occur. This is why the following lines have been commented out (as well as
599	# two lines in replace_images). They appear to no longer apply.
600
601	# If at any time, there is a need for having images embedded in <a> anchor tags,
602	# then it might be better to turn that into an HTMLPlugin option rather than make
603	# it the default behaviour. Also, eventually, no_image_links needs to become
604	# a deprecated option for HTMLPlugin as it has now become the default behaviour.
605
606	#if(!$self->{'no_image_links'}){
607	$$textref =~ s/(<(?:img\|embed\|table\|tr\|td)[^>]?(?:src\|background)\s=\s)((?:[\"][^\"]+[\"])\|(?:[\'][^\']+[\'])\|(?:[^\s\/>]+))([^>]>)/
608	$self->replace_images ($1, $2, $3, $base_dir, $file, $doc_obj, $cursection)/isge;
609	#}
610
611	# add text to document object
612	# turn \ into \\ so that the rest of greenstone doesn't think there
613	# is an escape code following. (Macro parsing loses them...)
614	$$textref =~ s/\\/\\\\/go;
615
616	$doc_obj->add_utf8_text($cursection, $$textref);
617	}
618
619	sub replace_images {
620	my $self = shift (@_);
621	my ($front, $link, $back, $base_dir,
622	$file, $doc_obj, $section) = @_;
623
624	# remove quotes from link at start and end if necessary
625	if ($link=~/^[\"\']/) {
626	$link=~s/^[\"\']//;
627	$link=~s/[\"\']$//;
628	$front.='"';
629	$back="\"$back";
630	}
631
632	$link =~ s/\n/ /g;
633
634	# Hack to overcome Windows wv 0.7.1 bug that causes embedded images to be broken
635	# If the Word file path has spaces in it, wv messes up and you end up with
636	# absolute paths for the images, and without the "file://" prefix
637	# So check for this special case and massage the data to be correct
638	if ($ENV{'GSDLOS'} =~ m/^windows/i && $self->{'plugin_type'} eq "WordPlug" && $link =~ m/^[A-Za-z]\:\\/) {
639	$link =~ s/^.*\\([^\\]+)$/$1/;
640	}
641
642	my ($href, $hash_part, $rl) = $self->format_link ($link, $base_dir, $file);
643
644	my $img_file = $self->add_file ($href, $rl, $hash_part, $base_dir, $doc_obj, $section);
645
646	# print STDERR "** link = $link\n href = $href\n** img_file = $img_file, rl = $rl\n";
647
648	my $anchor_name = $img_file;
649	#$anchor_name =~ s/^.*\///;
650	#$anchor_name = "<a name=\"$anchor_name\" ></a>";
651
652	my $image_link = $front . $img_file .$back;
653	return $image_link;
654
655	# The reasons for why the following two lines are no longer necessary can be
656	# found in subroutine process_section
657	#my $anchor_link = "<a href=\"$img_file\" >".$image_link."</a>";
658	#return $anchor_link;
659
660	#return $front . $img_file . $back . $anchor_name;
661	}
662
663	sub replace_href_links {
664	my $self = shift (@_);
665	my ($front, $link, $back, $base_dir, $file, $doc_obj, $section) = @_;
666
667	# remove quotes from link at start and end if necessary
668	if ($link=~/^[\"\']/) {
669	$link=~s/^[\"\']//;
670	$link=~s/[\"\']$//;
671	$front.='"';
672	$back="\"$back";
673	}
674
675	# attempt to sort out targets - frames are not handled
676	# well in this plugin and some cases will screw things
677	# up - e.g. the _parent target (so we'll just remove
678	# them all ;-)
679	$front =~ s/(target=\"?)_top(\"?)/$1_gsdltop_$2/is;
680	$back =~ s/(target=\"?)_top(\"?)/$1_gsdltop_$2/is;
681	$front =~ s/target=\"?_parent\"?//is;
682	$back =~ s/target=\"?_parent\"?//is;
683
684	return $front . $link . $back if $link =~ m/^\#/s;
685	$link =~ s/\n/ /g;
686
687	# Find file referred to by $link on file system
688	# This is more complicated than it sounds when char encodings
689	# is taken in to account
690	my ($href, $hash_part, $rl) = $self->format_link ($link, $base_dir, $file);
691
692	# href may use '\'s where '/'s should be on Windows
693	$href =~ s/\\/\//g;
694	my ($filename) = $href =~ m/^(?:.?):(?:\/\/)?(.)/;
695
696
697	##### leave all these links alone (they won't be picked up by intermediate
698	##### pages). I think that's safest when dealing with frames, targets etc.
699	##### (at least until I think of a better way to do it). Problems occur with
700	##### mailto links from within small frames, the intermediate page is displayed
701	##### within that frame and can't be seen. There is still potential for this to
702	##### happen even with html pages - the solution seems to be to somehow tell
703	##### the browser from the server side to display the page being sent (i.e.
704	##### the intermediate page) in the top level window - I'm not sure if that's
705	##### possible - the following line should probably be deleted if that can be done
706	return $front . $link . $back if $href =~ m/^(mailto\|news\|gopher\|nntp\|telnet\|javascript):/is;
707
708
709	if (($rl == 0) \|\| ($filename =~ m/$self->{'process_exp'}/) \|\|
710	($href =~ m/\/$/) \|\| ($href =~ m/^(mailto\|news\|gopher\|nntp\|telnet\|javascript):/i)) {
711	&ghtml::urlsafe ($href);
712	return $front . "_httpextlink_&rl=" . $rl . "&href=" . $href . $hash_part . $back;
713	} else {
714	# link is to some other type of file (eg image) so we'll
715	# need to associate that file
716	return $front . $self->add_file ($href, $rl, $hash_part, $base_dir, $doc_obj, $section) . $back;
717	}
718	}
719
720	sub add_file {
721	my $self = shift (@_);
722	my ($href, $rl, $hash_part, $base_dir, $doc_obj, $section) = @_;
723	my ($newname);
724
725	my $filename = $href;
726	if ($base_dir eq "") {
727	# remove http:/ thereby leaving one slash at the start
728	$filename =~ s/^[^:]*:\///;
729	}
730	else {
731	# remove http://
732	$filename =~ s/^[^:]*:\/\///;
733	}
734
735	$filename = &util::filename_cat($base_dir, $filename);
736	# Replace %XX's in URL with decoded value if required. Note that the filename may include the %XX in some
737	# situations. If the original file's name was in URL encoding, the following method will not decode it.
738	my $utf8_filename = $filename;
739	$filename = $self->opt_url_decode($utf8_filename);
740
741	# some special processing if the intended filename was converted to utf8, but
742	# the actual file still needs to be renamed
743	if (!-e $filename) {
744	# try the original filename stored in map
745	my $original_filename = $self->{'utf8_to_original_filename'}->{$filename};
746	if (defined $original_filename && -e $original_filename) {
747	$filename = $original_filename;
748	}
749	}
750
751	my ($ext) = $filename =~ m/(\.[^\.]*)$/;
752
753	if ($rl == 0) {
754	if ((!defined $ext) \|\| ($ext !~ m/$self->{'assoc_files'}/)) {
755	return "_httpextlink_&rl=0&el=prompt&href=" . $href . $hash_part;
756	}
757	else {
758	return "_httpextlink_&rl=0&el=direct&href=" . $href . $hash_part;
759	}
760	}
761
762	if ((!defined $ext) \|\| ($ext !~ m/$self->{'assoc_files'}/)) {
763	return "_httpextlink_&rl=" . $rl . "&href=" . $href . $hash_part;
764	}
765	# add the original image file as a source file
766	if (!$self->{'processing_tmp_files'} ) {
767	$doc_obj->associate_source_file($filename);
768	}
769	if ($self->{'rename_assoc_files'}) {
770	if (defined $self->{'aux_files'}->{$href}) {
771	$newname = $self->{'aux_files'}->{$href}->{'dir_num'} . "/" .
772	$self->{'aux_files'}->{$href}->{'file_num'} . $ext;
773	} else {
774	$newname = $self->{'dir_num'} . "/" . $self->{'file_num'} . $ext;
775	$self->{'aux_files'}->{$href} = {'dir_num' => $self->{'dir_num'}, 'file_num' => $self->{'file_num'}};
776	$self->inc_filecount ();
777	}
778	$doc_obj->associate_file($filename, $newname, undef, $section);
779	return "_httpdocimg_/$newname";
780	} else {
781	if(&unicode::is_url_encoded($utf8_filename)) {
782	# use the possibly-decoded filename instead to avoid double URL encoding
783	($newname) = $filename =~ m/([^\/\\]*)$/;
784	} else {
785	($newname) = $utf8_filename =~ m/([^\/\\]*)$/;
786	}
787
788	# Make sure this name uses only ASCII characters.
789	# We use either base64 or URL encoding, as these preserve original encoding
790	$newname = &util::rename_file($newname, $self->{'file_rename_method'});
791
792	$doc_obj->associate_file($filename, $newname, undef, $section);
793
794	# Since the generated image will be URL-encoded to avoid file-system/browser mess-ups
795	# of filenames, URL-encode the additional percent signs of the URL-encoded filename
796	my $newname_url = $newname;
797	$newname_url = &unicode::filename_to_url($newname_url);
798	return "_httpdocimg_/$newname_url";
799	}
800	}
801
802
803	sub format_link {
804	my $self = shift (@_);
805	my ($link, $base_dir, $file) = @_;
806
807	my ($before_hash, $hash_part) = $link =~ m/^([^\#])(\#?.)$/;
808
809	$hash_part = "" if !defined $hash_part;
810	if (!defined $before_hash \|\| $before_hash !~ m/[\w\.\/]/) {
811	my $outhandle = $self->{'outhandle'};
812	print $outhandle "HTMLPlugin: ERROR - badly formatted tag ignored ($link)\n"
813	if $self->{'verbosity'};
814	return ($link, "", 0);
815	}
816
817	if ($before_hash =~ s@^((?:http\|https\|ftp\|file\|mms)://)@@i) {
818	my $type = $1;
819
820	if ($link =~ m/^(http\|ftp):/i) {
821	# Turn url (using /) into file name (possibly using \ on windows)
822	my @http_dir_split = split('/', $before_hash);
823	$before_hash = &util::filename_cat(@http_dir_split);
824	}
825
826	$before_hash = $self->eval_dir_dots($before_hash);
827
828	my $linkfilename = &util::filename_cat ($base_dir, $before_hash);
829
830	my $rl = 0;
831	$rl = 1 if (-e $linkfilename);
832
833	# make sure there's a slash on the end if it's a directory
834	if ($before_hash !~ m/\/$/) {
835	$before_hash .= "/" if (-d $linkfilename);
836	}
837	return ($type . $before_hash, $hash_part, $rl);
838
839	} elsif ($link !~ m/^(mailto\|news\|gopher\|nntp\|telnet\|javascript):/i && $link !~ m/^\//) {
840
841	if ($before_hash =~ s@^/@@ \|\| $before_hash =~ m/\\/) {
842
843	# the first directory will be the domain name if file_is_url
844	# to generate archives, otherwise we'll assume all files are
845	# from the same site and base_dir is the root
846
847	if ($self->{'file_is_url'}) {
848	my @dirs = split /[\/\\]/, $file;
849	my $domname = shift (@dirs);
850	$before_hash = &util::filename_cat($domname, $before_hash);
851	$before_hash =~ s@\\@/@g; # for windows
852	}
853	else
854	{
855	# see if link shares directory with source document
856	# => turn into relative link if this is so!
857
858	if ($ENV{'GSDLOS'} =~ m/^windows/i) {
859	# too difficult doing a pattern match with embedded '\'s...
860	my $win_before_hash=$before_hash;
861	$win_before_hash =~ s@(\\)+@/@g;
862	# $base_dir is already similarly "converted" on windows.
863	if ($win_before_hash =~ s@^$base_dir/@@o) {
864	# if this is true, we removed a prefix
865	$before_hash=$win_before_hash;
866	}
867	}
868	else {
869	# before_hash has lost leading slash by this point,
870	# -> add back in prior to substitution with $base_dir
871	$before_hash = "/$before_hash";
872
873	$before_hash = &util::filename_cat("",$before_hash);
874	$before_hash =~ s@^$base_dir/@@;
875	}
876	}
877	} else {
878	# Turn relative file path into full path
879	my $dirname = &File::Basename::dirname($file);
880	$before_hash = &util::filename_cat($dirname, $before_hash);
881	$before_hash = $self->eval_dir_dots($before_hash);
882	}
883
884	my $linkfilename = &util::filename_cat ($base_dir, $before_hash);
885	# make sure there's a slash on the end if it's a directory
886	if ($before_hash !~ m/\/$/) {
887	$before_hash .= "/" if (-d $linkfilename);
888	}
889	return ("http://" . $before_hash, $hash_part, 1);
890	} else {
891	# mailto, news, nntp, telnet, javascript or gopher link
892	return ($before_hash, "", 0);
893	}
894	}
895
896	sub extract_first_NNNN_characters {
897	my $self = shift (@_);
898	my ($textref, $doc_obj, $thissection) = @_;
899
900	foreach my $size (split /,/, $self->{'first'}) {
901	my $tmptext = $$textref;
902	# skip to the body
903	$tmptext =~ s/.<body[^>]>//i;
904	# remove javascript
905	$tmptext =~ s@<script.*?</script>@ @sig;
906	$tmptext =~ s/<[^>]*>/ /g;
907	$tmptext =~ s/ / /g;
908	$tmptext =~ s/^\s+//;
909	$tmptext =~ s/\s+$//;
910	$tmptext =~ s/\s+/ /gs;
911	$tmptext = &unicode::substr ($tmptext, 0, $size);
912	$tmptext =~ s/\s\S*$/…/; # adds an ellipse (...)
913	$doc_obj->add_utf8_metadata ($thissection, "First$size", $tmptext);
914	}
915	}
916
917
918	sub extract_metadata {
919	my $self = shift (@_);
920	my ($textref, $metadata, $doc_obj, $section) = @_;
921	my $outhandle = $self->{'outhandle'};
922	# if we don't want metadata, we may as well not be here ...
923	return if (!defined $self->{'metadata_fields'});
924
925	# metadata fields to extract/save. 'key' is the (lowercase) name of the
926	# html meta, 'value' is the metadata name for greenstone to use
927	my %find_fields = ();
928
929	my %creator_fields = (); # short-cut for lookups
930
931
932	foreach my $field (split /,/, $self->{'metadata_fields'}) {
933	$field =~ s/^\s+//; # remove leading whitespace
934	$field =~ s/\s+$//; # remove trailing whitespace
935
936	# support tag<tagname>
937	if ($field =~ m/^(.?)\s<(.*?)>$/) {
938	# "$2" is the user's preferred gs metadata name
939	$find_fields{lc($1)}=$2; # lc = lowercase
940	} else { # no <tagname> for mapping
941	# "$field" is the user's preferred gs metadata name
942	$find_fields{lc($field)}=$field; # lc = lowercase
943	}
944	}
945
946	if (defined $self->{'hunt_creator_metadata'} &&
947	$self->{'hunt_creator_metadata'} == 1 ) {
948	my @extra_fields =
949	(
950	'author',
951	'author.email',
952	'creator',
953	'dc.creator',
954	'dc.creator.corporatename',
955	);
956
957	# add the creator_metadata fields to search for
958	foreach my $field (@extra_fields) {
959	$creator_fields{$field}=0; # add to lookup hash
960	}
961	}
962
963
964	# find the header in the html file, which has the meta tags
965	$$textref =~ m@<head>(.*?)</head>@si;
966
967	my $html_header=$1;
968
969	# go through every <meta... tag defined in the html and see if it is
970	# one of the tags we want to match.
971
972	# special case for title - we want to remember if its been found
973	my $found_title = 0;
974	# this assumes that ">" won't appear. (I don't think it's allowed to...)
975	$html_header =~ m/^/; # match the start of the string, for \G assertion
976
977	while ($html_header =~ m/\G.?<meta(.?)>/sig) {
978	my $metatag=$1;
979	my ($tag, $value);
980
981	# find the tag name
982	$metatag =~ m/(?:name\|http-equiv)\s=\s([\"\'])?(.*?)\1/is;
983	$tag=$2;
984	# in case they're not using " or ', but they should...
985	if (! $tag) {
986	$metatag =~ m/(?:name\|http-equiv)\s=\s([^\s\>]+)/is;
987	$tag=$1;
988	}
989
990	if (!defined $tag) {
991	print $outhandle "HTMLPlugin: can't find NAME in \"$metatag\"\n";
992	next;
993	}
994
995	# don't need to assign this field if it was passed in from a previous
996	# (recursive) plugin
997	if (defined $metadata->{$tag}) {next}
998
999	# find the tag content
1000	$metatag =~ m/content\s=\s([\"\'])?(.*?)\1/is;
1001	$value=$2;
1002
1003	if (! $value) {
1004	$metatag =~ m/(?:name\|http-equiv)\s=\s([^\s\>]+)/is;
1005	$value=$1;
1006	}
1007	if (!defined $value) {
1008	print $outhandle "HTMLPlugin: can't find VALUE in \"$metatag\"\n";
1009	next;
1010	}
1011
1012	# clean up and add
1013	$value =~ s/\s+/ /gs;
1014	chomp($value); # remove trailing \n, if any
1015	if (exists $creator_fields{lc($tag)}) {
1016	# map this value onto greenstone's "Creator" metadata
1017	$tag='Creator';
1018	} elsif (!exists $find_fields{lc($tag)}) {
1019	next; # don't want this tag
1020	} else {
1021	# get the user's preferred capitalisation
1022	$tag = $find_fields{lc($tag)};
1023	}
1024	if (lc($tag) eq "title") {
1025	$found_title = 1;
1026	}
1027
1028	if ($self->{'verbosity'} > 2) {
1029	print $outhandle " extracted \"$tag\" metadata \"$value\"\n";
1030	}
1031
1032	# Do we still reply on the following? Surely there must
1033	# be a better way to go about this outside of the plugin?
1034	#
1035	#if ($tag =~ m/date.*/i){
1036	# $tag = lc($tag);
1037	#}
1038
1039	$doc_obj->add_utf8_metadata($section, $tag, $value);
1040
1041	}
1042
1043	# TITLE: extract the document title
1044	if (exists $find_fields{'title'} && !$found_title) {
1045	# we want a title, and didn't find one in the meta tags
1046	# see if there's a <title> tag
1047	my $title;
1048	my $from = ""; # for debugging output only
1049	if ($html_header =~ m/<title[^>]>([^<]+)<\/title[^>]>/is) {
1050	$title = $1;
1051	$from = "<title> tags";
1052	}
1053
1054	if (!defined $title) {
1055	$from = "first 100 chars";
1056	# if no title use first 100 or so characters
1057	$title = $$textref;
1058	$title =~ s/^\xFE\xFF//; # Remove unicode byte order mark
1059	$title =~ s/^.*?<body>//si;
1060	# ignore javascript!
1061	$title =~ s@<script.*?</script>@ @sig;
1062	$title =~ s/<\/([^>]+)><\1>//g; # (eg) </b><b> - no space
1063	$title =~ s/<[^>]*>/ /g; # remove all HTML tags
1064	$title = substr ($title, 0, 100);
1065	$title =~ s/\s\S*$/.../;
1066	}
1067	$title =~ s/<[^>]*>/ /g; # remove html tags
1068	$title =~ s/ / /g;
1069	$title =~ s/(?: \|\xc2\xa0)/ /g; # utf-8 for nbsp...
1070	$title =~ s/\s+/ /gs; # collapse multiple spaces
1071	$title =~ s/^\s*//; # remove leading spaces
1072	$title =~ s/\s*$//; # remove trailing spaces
1073
1074	$title =~ s/^$self->{'title_sub'}// if ($self->{'title_sub'});
1075	$title =~ s/^\s+//s; # in case title_sub introduced any...
1076	$doc_obj->add_utf8_metadata ($section, 'Title', $title);
1077	print $outhandle " extracted Title metadata \"$title\" from $from\n"
1078	if ($self->{'verbosity'} > 2);
1079	}
1080
1081	# add FileFormat metadata
1082	$doc_obj->add_metadata($section,"FileFormat", "HTML");
1083
1084	# Special, for metadata names such as tagH1 - extracts
1085	# the text between the first <H1> and </H1> tags into "H1" metadata.
1086
1087	foreach my $field (keys %find_fields) {
1088	if ($field !~ m/^tag([a-z0-9]+)$/i) {next}
1089	my $tag = $1;
1090	if ($$textref =~ m@<$tag[^>]>(.?)</$tag[^>]*>@g) {
1091	my $content = $1;
1092	$content =~ s/ / /g;
1093	$content =~ s/<[^>]*>/ /g;
1094	$content =~ s/^\s+//;
1095	$content =~ s/\s+$//;
1096	$content =~ s/\s+/ /gs;
1097	if ($content) {
1098	$tag=$find_fields{"tag$tag"}; # get the user's capitalisation
1099	$tag =~ s/^tag//i;
1100	$doc_obj->add_utf8_metadata ($section, $tag, $content);
1101	print $outhandle " extracted \"$tag\" metadata \"$content\"\n"
1102	if ($self->{'verbosity'} > 2);
1103	}
1104	}
1105	}
1106	}
1107
1108
1109	# evaluate any "../" to next directory up
1110	# evaluate any "./" as here
1111	sub eval_dir_dots {
1112	my $self = shift (@_);
1113	my ($filename) = @_;
1114	my $dirsep_os = &util::get_os_dirsep();
1115	my @dirsep = split(/$dirsep_os/,$filename);
1116
1117	my @eval_dirs = ();
1118	foreach my $d (@dirsep) {
1119	if ($d eq "..") {
1120	pop(@eval_dirs);
1121
1122	} elsif ($d eq ".") {
1123	# do nothing!
1124
1125	} else {
1126	push(@eval_dirs,$d);
1127	}
1128	}
1129
1130	# Need to fiddle with number of elements in @eval_dirs if the
1131	# first one is the empty string. This is because of a
1132	# modification to util::filename_cat that supresses the addition
1133	# of a leading '/' character (or \ if windows) (intended to help
1134	# filename cat with relative paths) if the first entry in the
1135	# array is the empty string. Making the array start with two
1136	# empty strings is a way to defeat this "smart" option.
1137	#
1138	if (scalar(@eval_dirs) > 0) {
1139	if ($eval_dirs[0] eq ""){
1140	unshift(@eval_dirs,"");
1141	}
1142	}
1143
1144	my $evaluated_filename = (scalar @eval_dirs > 0) ? &util::filename_cat(@eval_dirs) : "";
1145	return $evaluated_filename;
1146	}
1147
1148	sub replace_usemap_links {
1149	my $self = shift (@_);
1150	my ($front, $link, $back) = @_;
1151
1152	# remove quotes from link at start and end if necessary
1153	if ($link=~/^[\"\']/) {
1154	$link=~s/^[\"\']//;
1155	$link=~s/[\"\']$//;
1156	$front.='"';
1157	$back="\"$back";
1158	}
1159
1160	$link =~ s/^\.\///;
1161	return $front . $link . $back;
1162	}
1163
1164	sub inc_filecount {
1165	my $self = shift (@_);
1166
1167	if ($self->{'file_num'} == 1000) {
1168	$self->{'dir_num'} ++;
1169	$self->{'file_num'} = 0;
1170	} else {
1171	$self->{'file_num'} ++;
1172	}
1173	}
1174
1175
1176	# Extend read_file so that strings like é are
1177	# converted to UTF8 internally.
1178	#
1179	# We don't convert < or > or & or " in case
1180	# they interfere with the GML files
1181
1182	sub read_file {
1183	my $self = shift(@_);
1184	my ($filename, $encoding, $language, $textref) = @_;
1185
1186	$self->SUPER::read_file($filename, $encoding, $language, $textref);
1187
1188	# Convert entities to their UTF8 equivalents
1189	$$textref =~ s/&(lt\|gt\|amp\|quot\|nbsp);/&z$1;/go;
1190	$$textref =~ s/&([^;]+);/&ghtml::getcharequiv($1,1)/gseo;
1191	$$textref =~ s/&z(lt\|gt\|amp\|quot\|nbsp);/&$1;/go;
1192	}
1193
1194	sub HB_read_html_file {
1195	my $self = shift (@_);
1196	my ($htmlfile, $text) = @_;
1197
1198	# load in the file
1199	if (!open (FILE, $htmlfile)) {
1200	print STDERR "ERROR - could not open $htmlfile\n";
1201	return;
1202	}
1203
1204	my $foundbody = 0;
1205	$self->HB_gettext (\$foundbody, $text, "FILE");
1206	close FILE;
1207
1208	# just in case there was no <body> tag
1209	if (!$foundbody) {
1210	$foundbody = 1;
1211	open (FILE, $htmlfile) \|\| return;
1212	$self->HB_gettext (\$foundbody, $text, "FILE");
1213	close FILE;
1214	}
1215	# text is in utf8
1216	}
1217
1218	# converts the text to utf8, as ghtml does that for é etc.
1219	sub HB_gettext {
1220	my $self = shift (@_);
1221	my ($foundbody, $text, $handle) = @_;
1222
1223	my $line = "";
1224	while (defined ($line = <$handle>)) {
1225	# look for body tag
1226	if (!$$foundbody) {
1227	if ($line =~ s/^.<body[^>]>//i) {
1228	$$foundbody = 1;
1229	} else {
1230	next;
1231	}
1232	}
1233
1234	# check for symbol fonts
1235	if ($line =~ m/<font [^>]?face\s=\s*\"?(\w+)\"?/i) {
1236	my $font = $1;
1237	print STDERR "HBPlug::HB_gettext - warning removed font $font\n"
1238	if ($font !~ m/^arial$/i);
1239	}
1240
1241	$$text .= $line;
1242	}
1243
1244	if ($self->{'input_encoding'} eq "iso_8859_1") {
1245	# convert to utf-8
1246	$$text=&unicode::unicode2utf8(&unicode::convert2unicode("iso_8859_1", $text));
1247	}
1248	# convert any alphanumeric character entities to their utf-8
1249	# equivalent for indexing purposes
1250	#&ghtml::convertcharentities ($$text);
1251
1252	$$text =~ s/\s+/ /g; # remove \n's
1253	}
1254
1255	sub HB_clean_section {
1256	my $self = shift (@_);
1257	my ($section) = @_;
1258
1259	# remove tags without a starting tag from the section
1260	my ($tag, $tagstart);
1261	while ($section =~ m/<\/([^>]{1,10})>/) {
1262	$tag = $1;
1263	$tagstart = index($section, "<$tag");
1264	last if (($tagstart >= 0) && ($tagstart < index($section, "<\/$tag")));
1265	$section =~ s/<\/$tag>//;
1266	}
1267
1268	# remove extra paragraph tags
1269	while ($section =~ s/<p\b[^>]>\s<p\b/<p/ig) {}
1270
1271	# remove extra stuff at the end of the section
1272	while ($section =~ s/(<u>\|<i>\|<b>\|<p\b[^>]*>\| \|\s)$//i) {}
1273
1274	# add a newline at the beginning of each paragraph
1275	$section =~ s/(.)\s*<p\b/$1\n\n<p/gi;
1276
1277	# add a newline every 80 characters at a word boundary
1278	# Note: this regular expression puts a line feed before
1279	# the last word in each section, even when it is not
1280	# needed.
1281	$section =~ s/(.{1,80})\s/$1\n/g;
1282
1283	# fix up the image links
1284	$section =~ s/<img[^>]?src=\"?([^\">]+)\"?[^>]>/
1285	<center><img src=\"$1\" \/><\/center><br\/>/ig;
1286	$section =~ s/<<I>>\s*([^\.]+\.(png\|jpg\|gif))/
1287	<center><img src=\"$1\" \/><\/center><br\/>/ig;
1288
1289	return $section;
1290	}
1291
1292	# Will convert the oldHDL format to the new HDL format (using the Section tag)
1293	sub convert_to_newHDLformat
1294	{
1295	my $self = shift (@_);
1296	my ($file,$cnfile) = @_;
1297	my $input_filename = $file;
1298	my $tmp_filename = $cnfile;
1299
1300	# write HTML tmp file with new HDL format
1301	open (PROD, ">$tmp_filename") \|\| die("Error Writing to File: $tmp_filename $!");
1302
1303	# read in the file and do basic html cleaning (removing header etc)
1304	my $html = "";
1305	$self->HB_read_html_file ($input_filename, \$html);
1306
1307	# process the file one section at a time
1308	my $curtoclevel = 1;
1309	my $firstsection = 1;
1310	my $toclevel = 0;
1311	while (length ($html) > 0) {
1312	if ($html =~ s/^.?(?:<p\b[^>]>)?((<b>\|<i>\|<u>\|\s))<<TOC(\d+)>>\s(.*?)<p\b/<p/i) {
1313	$toclevel = $3;
1314	my $title = $4;
1315	my $sectiontext = "";
1316	if ($html =~ s/^(.?)((?:<p\b[^>]>)?((<b>\|<i>\|<u>\|\s)*)<<TOC\d+>>)/$2/i) {
1317	$sectiontext = $1;
1318	} else {
1319	$sectiontext = $html;
1320	$html = "";
1321	}
1322
1323	# remove tags and extra spaces from the title
1324	$title =~ s/<\/?[^>]+>//g;
1325	$title =~ s/^\s+\|\s+$//g;
1326
1327	# close any sections below the current level and
1328	# create a new section (special case for the firstsection)
1329	print PROD "<!--\n";
1330	while (($curtoclevel > $toclevel) \|\|
1331	(!$firstsection && $curtoclevel == $toclevel)) {
1332	$curtoclevel--;
1333	print PROD "</Section>\n";
1334	}
1335	if ($curtoclevel+1 < $toclevel) {
1336	print STDERR "WARNING - jump in toc levels in $input_filename " .
1337	"from $curtoclevel to $toclevel\n";
1338	}
1339	while ($curtoclevel < $toclevel) {
1340	$curtoclevel++;
1341	}
1342
1343	if ($curtoclevel == 1) {
1344	# add the header tag
1345	print PROD "-->\n";
1346	print PROD "<HTML>\n<HEAD>\n<TITLE>$title</TITLE>\n</HEAD>\n<BODY>\n";
1347	print PROD "<!--\n";
1348	}
1349
1350	print PROD "<Section>\n\t<Description>\n\t\t<Metadata name=\"Title\">$title</Metadata>\n\t</Description>\n";
1351
1352	print PROD "-->\n";
1353
1354	# clean up the section html
1355	$sectiontext = $self->HB_clean_section($sectiontext);
1356
1357	print PROD "$sectiontext\n";
1358
1359	} else {
1360	print STDERR "WARNING - leftover text\n" , $self->shorten($html),
1361	"\nin $input_filename\n";
1362	last;
1363	}
1364	$firstsection = 0;
1365	}
1366
1367	print PROD "<!--\n";
1368	while ($curtoclevel > 0) {
1369	$curtoclevel--;
1370	print PROD "</Section>\n";
1371	}
1372	print PROD "-->\n";
1373
1374	close (PROD) \|\| die("Error Closing File: $tmp_filename $!");
1375
1376	return $tmp_filename;
1377	}
1378
1379	sub shorten {
1380	my $self = shift (@_);
1381	my ($text) = @_;
1382
1383	return "\"$text\"" if (length($text) < 100);
1384
1385	return "\"" . substr ($text, 0, 50) . "\" ... \"" .
1386	substr ($text, length($text)-50) . "\"";
1387	}
1388
1389	sub convert_tidy_or_oldHDL_file
1390	{
1391	my $self = shift (@_);
1392	my ($file) = @_;
1393	my $input_filename = $file;
1394
1395	if (-d $input_filename)
1396	{
1397	return $input_filename;
1398	}
1399
1400	# get the input filename
1401	my ($tailname, $dirname, $suffix) = &File::Basename::fileparse($input_filename, "\\.[^\\.]+\$");
1402	my $base_dirname = $dirname;
1403	$suffix = lc($suffix);
1404
1405	# derive tmp filename from input filename
1406	# Remove any white space from filename -- no risk of name collision, and
1407	# makes later conversion by utils simpler. Leave spaces in path...
1408	# tidy up the filename with space, dot, hyphen between
1409	$tailname =~ s/\s+//g;
1410	$tailname =~ s/\.+//g;
1411	$tailname =~ s/\-+//g;
1412	# convert to utf-8 otherwise we have problems with the doc.xml file
1413	# later on
1414	&unicode::ensure_utf8(\$tailname);
1415
1416	# softlink to collection tmp dir
1417	my $tmp_dirname = &util::filename_cat($ENV{'GSDLCOLLECTDIR'}, "tidytmp");
1418	&util::mk_dir($tmp_dirname) if (!-e $tmp_dirname);
1419
1420	my $test_dirname = "";
1421	my $f_separator = &util::get_os_dirsep();
1422
1423	if ($dirname =~ m/import$f_separator/)
1424	{
1425	$test_dirname = $'; #'
1426
1427	#print STDERR "init $'\n";
1428
1429	while ($test_dirname =~ m/[$f_separator]/)
1430	{
1431	my $folderdirname = $`;
1432	$tmp_dirname = &util::filename_cat($tmp_dirname,$folderdirname);
1433	&util::mk_dir($tmp_dirname) if (!-e $tmp_dirname);
1434	$test_dirname = $'; #'
1435	}
1436	}
1437
1438	my $tmp_filename = &util::filename_cat($tmp_dirname, "$tailname$suffix");
1439
1440	# tidy or convert the input file if it is a HTML-like file or it is accepted by the process_exp
1441	if (($suffix eq ".htm") \|\| ($suffix eq ".html") \|\| ($suffix eq ".shtml"))
1442	{
1443	#convert the input file to a new style HDL
1444	my $hdl_output_filename = $input_filename;
1445	if ($self->{'old_style_HDL'})
1446	{
1447	$hdl_output_filename = &util::filename_cat($tmp_dirname, "$tailname$suffix");
1448	$hdl_output_filename = $self->convert_to_newHDLformat($input_filename,$hdl_output_filename);
1449	}
1450
1451	#just for checking copy all other file from the base dir to tmp dir if it is not exists
1452	opendir(DIR,$base_dirname) or die "Can't open base directory : $base_dirname!";
1453	my @files = grep {!/^\.+$/} readdir(DIR);
1454	close(DIR);
1455
1456	foreach my $file (@files)
1457	{
1458	my $src_file = &util::filename_cat($base_dirname,$file);
1459	my $dest_file = &util::filename_cat($tmp_dirname,$file);
1460	if ((!-e $dest_file) && (!-d $src_file))
1461	{
1462	# just copy the original file back to the tmp directory
1463	copy($src_file,$dest_file) or die "Can't copy file $src_file to $dest_file $!";
1464	}
1465	}
1466
1467	# tidy the input file
1468	my $tidy_output_filename = $hdl_output_filename;
1469	if ($self->{'use_realistic_book'})
1470	{
1471	$tidy_output_filename = &util::filename_cat($tmp_dirname, "$tailname$suffix");
1472	$tidy_output_filename = $self->tmp_tidy_file($hdl_output_filename,$tidy_output_filename);
1473	}
1474	$tmp_filename = $tidy_output_filename;
1475	}
1476	else
1477	{
1478	if (!-e $tmp_filename)
1479	{
1480	# just copy the original file back to the tmp directory
1481	copy($input_filename,$tmp_filename) or die "Can't copy file $input_filename to $tmp_filename $!";
1482	}
1483	}
1484
1485	return $tmp_filename;
1486	}
1487
1488
1489	# Will make the html input file as a proper XML file with removed font tag and
1490	# image size added to the img tag.
1491	# The tidying process takes place in a collection specific 'tmp' directory so
1492	# that we don't accidentally damage the input.
1493	sub tmp_tidy_file
1494	{
1495	my $self = shift (@_);
1496	my ($file,$cnfile) = @_;
1497	my $input_filename = $file;
1498	my $tmp_filename = $cnfile;
1499
1500	# get the input filename
1501	my ($tailname, $dirname, $suffix) = &File::Basename::fileparse($input_filename, "\\.[^\\.]+\$");
1502
1503	require HTML::TokeParser::Simple;
1504
1505	# create HTML parser to decode the input file
1506	my $parser = HTML::TokeParser::Simple->new($input_filename);
1507
1508	# write HTML tmp file without the font tag and image size are added to the img tag
1509	open (PROD, ">$tmp_filename") \|\| die("Error Writing to File: $tmp_filename $!");
1510	while (my $token = $parser->get_token())
1511	{
1512	# is it an img tag
1513	if ($token->is_start_tag('img'))
1514	{
1515	# get the attributes
1516	my $attr = $token->return_attr;
1517
1518	# get the full path to the image
1519	my $img_file = &util::filename_cat($dirname,$attr->{src});
1520
1521	# set the width and height attribute
1522	($attr->{width}, $attr->{height}) = imgsize($img_file);
1523
1524	# recreate the tag
1525	print PROD "<img";
1526	print PROD map { qq { $_="$attr->{$_}"} } keys %$attr;
1527	print PROD ">";
1528	}
1529	# is it a font tag
1530	else
1531	{
1532	if (($token->is_start_tag('font')) \|\| ($token->is_end_tag('font')))
1533	{
1534	# remove font tag
1535	print PROD "";
1536	}
1537	else
1538	{
1539	# print without changes
1540	print PROD $token->as_is;
1541	}
1542	}
1543	}
1544	close (PROD) \|\| die("Error Closing File: $tmp_filename $!");
1545
1546	# run html-tidy on the tmp file to make it a proper XML file
1547	my $tidyfile = `tidy -utf8 -wrap 0 -asxml "$tmp_filename"`;
1548
1549	# write result back to the tmp file
1550	open (PROD, ">$tmp_filename") \|\| die("Error Writing to File: $tmp_filename $!");
1551	print PROD $tidyfile;
1552	close (PROD) \|\| die("Error Closing File: $tmp_filename $!");
1553
1554	# return the output filename
1555	return $tmp_filename;
1556	}
1557
1558	1;

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

Download in other formats: