source: gsdl/trunk/perllib/plugins/HTMLPlug.pm@ 14913

Last change on this file since 14913 was 14913, checked in by qq6, 15 years ago

updated by Veronica

  • Property svn:executable set to *
  • Property svn:keywords set to Author Date Id Revision
File size: 45.7 KB
Line 
1###########################################################################
2#
3# HTMLPlug.pm -- basic html plugin
4#
5# A component of the Greenstone digital library software
6# from the New Zealand Digital Library Project at the
7# University of Waikato, New Zealand.
8#
9# Copyright (C) 1999 New Zealand Digital Library Project
10#
11# This program is free software; you can redistribute it and/or modify
12# it under the terms of the GNU General Public License as published by
13# the Free Software Foundation; either version 2 of the License, or
14# (at your option) any later version.
15#
16# This program is distributed in the hope that it will be useful,
17# but WITHOUT ANY WARRANTY; without even the implied warranty of
18# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
19# GNU General Public License for more details.
20#
21# You should have received a copy of the GNU General Public License
22# along with this program; if not, write to the Free Software
23# Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
24#
25###########################################################################
26
27#
28# Note that this plugin handles frames only in a very simple way
29# i.e. each frame is treated as a separate document. This means
30# search results will contain links to individual frames rather
31# than linking to the top level frameset.
32# There may also be some problems caused by the _parent target
33# (it's removed by this plugin)
34#
35
36package HTMLPlug;
37
38use BasPlug;
39use ghtml;
40use unicode;
41use util;
42use XMLParser;
43
44use Image::Size;
45use File::Copy;
46
47sub BEGIN {
48 @HTMLPlug::ISA = ('BasPlug');
49}
50
51use strict; # every perl program should have this!
52no strict 'refs'; # make an exception so we can use variables as filehandles
53
54my $arguments =
55 [ { 'name' => "process_exp",
56 'desc' => "{BasPlug.process_exp}",
57 'type' => "regexp",
58 'deft' => &get_default_process_exp() },
59 { 'name' => "block_exp",
60 'desc' => "{BasPlug.block_exp}",
61 'type' => 'regexp',
62 'deft' => &get_default_block_exp() },
63 { 'name' => "nolinks",
64 'desc' => "{HTMLPlug.nolinks}",
65 'type' => "flag" },
66 { 'name' => "keep_head",
67 'desc' => "{HTMLPlug.keep_head}",
68 'type' => "flag" },
69 { 'name' => "no_metadata",
70 'desc' => "{HTMLPlug.no_metadata}",
71 'type' => "flag" },
72 { 'name' => "metadata_fields",
73 'desc' => "{HTMLPlug.metadata_fields}",
74 'type' => "string",
75 'deft' => "Title" },
76 { 'name' => "hunt_creator_metadata",
77 'desc' => "{HTMLPlug.hunt_creator_metadata}",
78 'type' => "flag" },
79 { 'name' => "file_is_url",
80 'desc' => "{HTMLPlug.file_is_url}",
81 'type' => "flag" },
82 { 'name' => "assoc_files",
83 'desc' => "{HTMLPlug.assoc_files}",
84 'type' => "regexp",
85 'deft' => &get_default_block_exp() },
86 { 'name' => "rename_assoc_files",
87 'desc' => "{HTMLPlug.rename_assoc_files}",
88 'type' => "flag" },
89 { 'name' => "title_sub",
90 'desc' => "{HTMLPlug.title_sub}",
91 'type' => "string",
92 'deft' => "" },
93 { 'name' => "description_tags",
94 'desc' => "{HTMLPlug.description_tags}",
95 'type' => "flag" },
96 # retain this for backward compatibility (w3mir option was replaced by
97 # file_is_url)
98 { 'name' => "w3mir",
99# 'desc' => "{HTMLPlug.w3mir}",
100 'type' => "flag",
101 'hiddengli' => "yes"},
102 { 'name' => "no_strip_metadata_html",
103 'desc' => "{HTMLPlug.no_strip_metadata_html}",
104 'type' => "string",
105 'deft' => "",
106 'reqd' => "no"},
107 { 'name' => "sectionalise_using_h_tags",
108 'desc' => "{HTMLPlug.sectionalise_using_h_tags}",
109 'type' => "flag" },
110 { 'name' => "use_realistic_book",
111 'desc' => "{HTMLPlug.tidy_html}",
112 'type' => "flag"},
113 { 'name' => "is_old_HDL_tags",
114 'desc' => "{HTMLPlug.old_style_HDL}",
115 'type' => "flag"}
116 ];
117
118my $options = { 'name' => "HTMLPlug",
119 'desc' => "{HTMLPlug.desc}",
120 'abstract' => "no",
121 'inherits' => "yes",
122 'args' => $arguments };
123
124
125sub HB_read_html_file {
126 my $self = shift (@_);
127 my ($htmlfile, $text) = @_;
128
129 # load in the file
130 if (!open (FILE, $htmlfile)) {
131 print STDERR "ERROR - could not open $htmlfile\n";
132 return;
133 }
134
135 my $foundbody = 0;
136 $self->HB_gettext (\$foundbody, $text, "FILE");
137 close FILE;
138
139 # just in case there was no <body> tag
140 if (!$foundbody) {
141 $foundbody = 1;
142 open (FILE, $htmlfile) || return;
143 $self->HB_gettext (\$foundbody, $text, "FILE");
144 close FILE;
145 }
146 # text is in utf8
147}
148
149# converts the text to utf8, as ghtml does that for &eacute; etc.
150sub HB_gettext {
151 my $self = shift (@_);
152 my ($foundbody, $text, $handle) = @_;
153
154 my $line = "";
155 while (defined ($line = <$handle>)) {
156 # look for body tag
157 if (!$$foundbody) {
158 if ($line =~ s/^.*<body[^>]*>//i) {
159 $$foundbody = 1;
160 } else {
161 next;
162 }
163 }
164
165 # check for symbol fonts
166 if ($line =~ /<font [^>]*?face\s*=\s*\"?(\w+)\"?/i) {
167 my $font = $1;
168 print STDERR "HBPlug::HB_gettext - warning removed font $font\n"
169 if ($font !~ /^arial$/i);
170 }
171
172 $$text .= $line;
173 }
174
175 if ($self->{'input_encoding'} eq "iso_8859_1") {
176 # convert to utf-8
177 $$text=&unicode::unicode2utf8(&unicode::convert2unicode("iso_8859_1", $text));
178 }
179 # convert any alphanumeric character entities to their utf-8
180 # equivalent for indexing purposes
181 #&ghtml::convertcharentities ($$text);
182
183 $$text =~ s/\s+/ /g; # remove \n's
184}
185
186sub HB_clean_section {
187 my $self = shift (@_);
188 my ($section) = @_;
189
190 # remove tags without a starting tag from the section
191 my ($tag, $tagstart);
192 while ($section =~ /<\/([^>]{1,10})>/) {
193 $tag = $1;
194 $tagstart = index($section, "<$tag");
195 last if (($tagstart >= 0) && ($tagstart < index($section, "<\/$tag")));
196 $section =~ s/<\/$tag>//;
197 }
198
199 # remove extra paragraph tags
200 while ($section =~ s/<p\b[^>]*>\s*<p\b/<p/ig) {}
201
202 # remove extra stuff at the end of the section
203 while ($section =~ s/(<u>|<i>|<b>|<p\b[^>]*>|&nbsp;|\s)$//i) {}
204
205 # add a newline at the beginning of each paragraph
206 $section =~ s/(.)\s*<p\b/$1\n\n<p/gi;
207
208 # add a newline every 80 characters at a word boundary
209 # Note: this regular expression puts a line feed before
210 # the last word in each section, even when it is not
211 # needed.
212 $section =~ s/(.{1,80})\s/$1\n/g;
213
214 # fix up the image links
215 $section =~ s/<img[^>]*?src=\"?([^\">]+)\"?[^>]*>/
216 <center><img src=\"$1\" \/><\/center><br\/>/ig;
217 $section =~ s/&lt;&lt;I&gt;&gt;\s*([^\.]+\.(png|jpg|gif))/
218 <center><img src=\"$1\" \/><\/center><br\/>/ig;
219
220 return $section;
221}
222
223# Will convert the oldHDL format to the new HDL format (using the Section tag)
224sub convert_to_newHDLformat
225{
226 my $self = shift (@_);
227 my ($file,$cnfile) = @_;
228 my $input_filename = $file;
229 my $tmp_filename = $cnfile;
230
231 # write HTML tmp file with new HDL format
232 open (PROD, ">$tmp_filename") || die("Error Writing to File: $tmp_filename $!");
233
234 # read in the file and do basic html cleaning (removing header etc)
235 my $html = "";
236 $self->HB_read_html_file ($input_filename, \$html);
237
238 # process the file one section at a time
239 my $curtoclevel = 1;
240 my $firstsection = 1;
241 my $toclevel = 0;
242 while (length ($html) > 0) {
243 if ($html =~ s/^.*?(?:<p\b[^>]*>)?((<b>|<i>|<u>|\s)*)&lt;&lt;TOC(\d+)&gt;&gt;\s*(.*?)<p\b/<p/i) {
244 $toclevel = $3;
245 my $title = $4;
246 my $sectiontext = "";
247 if ($html =~ s/^(.*?)((?:<p\b[^>]*>)?((<b>|<i>|<u>|\s)*)&lt;&lt;TOC\d+&gt;&gt;)/$2/i) {
248 $sectiontext = $1;
249 } else {
250 $sectiontext = $html;
251 $html = "";
252 }
253
254 # remove tags and extra spaces from the title
255 $title =~ s/<\/?[^>]+>//g;
256 $title =~ s/^\s+|\s+$//g;
257
258 # close any sections below the current level and
259 # create a new section (special case for the firstsection)
260 print PROD "<!--\n";
261 while (($curtoclevel > $toclevel) ||
262 (!$firstsection && $curtoclevel == $toclevel)) {
263 $curtoclevel--;
264 print PROD "</Section>\n";
265 }
266 if ($curtoclevel+1 < $toclevel) {
267 print STDERR "WARNING - jump in toc levels in $input_filename " .
268 "from $curtoclevel to $toclevel\n";
269 }
270 while ($curtoclevel < $toclevel) {
271 $curtoclevel++;
272 }
273
274 if ($curtoclevel == 1) {
275 # add the header tag
276 print PROD "-->\n";
277 print PROD "<HTML>\n<HEAD>\n<TITLE>$title</TITLE>\n</HEAD>\n<BODY>\n";
278 print PROD "<!--\n";
279 }
280
281 print PROD "<Section>\n\t<Description>\n\t\t<Metadata name=\"Title\">$title</Metadata>\n\t</Description>\n";
282
283 print PROD "-->\n";
284
285 # clean up the section html
286 $sectiontext = $self->HB_clean_section($sectiontext);
287
288 print PROD "$sectiontext\n";
289
290 } else {
291 print STDERR "WARNING - leftover text\n" , $self->shorten($html),
292 "\nin $input_filename\n";
293 last;
294 }
295 $firstsection = 0;
296 }
297
298 print PROD "<!--\n";
299 while ($curtoclevel > 0) {
300 $curtoclevel--;
301 print PROD "</Section>\n";
302 }
303 print PROD "-->\n";
304
305 close (PROD) || die("Error Closing File: $tmp_filename $!");
306
307 return $tmp_filename;
308}
309
310sub shorten {
311 my $self = shift (@_);
312 my ($text) = @_;
313
314 return "\"$text\"" if (length($text) < 100);
315
316 return "\"" . substr ($text, 0, 50) . "\" ... \"" .
317 substr ($text, length($text)-50) . "\"";
318}
319
320sub convert_tidy_or_oldHDL_file
321{
322 my $self = shift (@_);
323 my ($file) = @_;
324 my $input_filename = $file;
325
326 if (-d $input_filename)
327 {
328 return $input_filename;
329 }
330
331 # get the input filename
332 my ($tailname, $dirname, $suffix) = &File::Basename::fileparse($input_filename, "\\.[^\\.]+\$");
333 my $base_dirname = $dirname;
334 $suffix = lc($suffix);
335
336 # derive tmp filename from input filename
337 # Remove any white space from filename -- no risk of name collision, and
338 # makes later conversion by utils simpler. Leave spaces in path...
339 # tidy up the filename with space, dot, hyphen between
340 $tailname =~ s/\s+//g;
341 $tailname =~ s/\.+//g;
342 $tailname =~ s/\-+//g;
343 # convert to utf-8 otherwise we have problems with the doc.xml file
344 # later on
345 &unicode::ensure_utf8(\$tailname);
346
347 # softlink to collection tmp dir
348 my $tmp_dirname = &util::filename_cat($ENV{'GSDLCOLLECTDIR'}, "tidytmp");
349 &util::mk_dir($tmp_dirname) if (!-e $tmp_dirname);
350
351 my $test_dirname = "";
352 my $f_separator = &util::get_os_dirsep();
353
354 if ($dirname =~ /import$f_separator/)
355 {
356 $test_dirname = $';
357
358 #print STDERR "init $'\n";
359
360 while ($test_dirname =~ /[$f_separator]/)
361 {
362 my $folderdirname = $`;
363 $tmp_dirname = &util::filename_cat($tmp_dirname,$folderdirname);
364 &util::mk_dir($tmp_dirname) if (!-e $tmp_dirname);
365 $test_dirname = $';
366 }
367 }
368
369 my $tmp_filename = &util::filename_cat($tmp_dirname, "$tailname$suffix");
370
371 # tidy or convert the input file if it is a HTML-like file or it is accepted by the process_exp
372 if (($suffix eq ".htm") || ($suffix eq ".html") || ($suffix eq ".shtml"))
373 {
374 #convert the input file to a new style HDL
375 my $hdl_output_filename = $input_filename;
376 if ($self->{'old_style_HDL'})
377 {
378 $hdl_output_filename = &util::filename_cat($tmp_dirname, "$tailname$suffix");
379 $hdl_output_filename = $self->convert_to_newHDLformat($input_filename,$hdl_output_filename);
380 }
381
382 #just for checking copy all other file from the base dir to tmp dir if it is not exists
383 opendir(DIR,$base_dirname) or die "Can't open base directory : $base_dirname!";
384 my @files = grep {!/^\.+$/} readdir(DIR);
385 close(DIR);
386
387 foreach my $file (@files)
388 {
389 my $src_file = &util::filename_cat($base_dirname,$file);
390 my $dest_file = &util::filename_cat($tmp_dirname,$file);
391 if ((!-e $dest_file) && (!-d $src_file))
392 {
393 # just copy the original file back to the tmp directory
394 copy($src_file,$dest_file) or die "Can't copy file $src_file to $dest_file $!";
395 }
396 }
397
398 # tidy the input file
399 my $tidy_output_filename = $hdl_output_filename;
400 if ($self->{'tidy_html'})
401 {
402 $tidy_output_filename = &util::filename_cat($tmp_dirname, "$tailname$suffix");
403 $tidy_output_filename = $self->tmp_tidy_file($hdl_output_filename,$tidy_output_filename);
404 }
405 $tmp_filename = $tidy_output_filename;
406 }
407 else
408 {
409 if (!-e $tmp_filename)
410 {
411 # just copy the original file back to the tmp directory
412 copy($input_filename,$tmp_filename) or die "Can't copy file $input_filename to $tmp_filename $!";
413 }
414 }
415
416 return $tmp_filename;
417}
418
419
420# Will make the html input file as a proper XML file with removed font tag and
421# image size added to the img tag.
422# The tidying process takes place in a collection specific 'tmp' directory so
423# that we don't accidentally damage the input.
424sub tmp_tidy_file
425{
426 my $self = shift (@_);
427 my ($file,$cnfile) = @_;
428 my $input_filename = $file;
429 my $tmp_filename = $cnfile;
430
431 # get the input filename
432 my ($tailname, $dirname, $suffix) = &File::Basename::fileparse($input_filename, "\\.[^\\.]+\$");
433
434 require HTML::TokeParser::Simple;
435
436 # create HTML parser to decode the input file
437 my $parser = HTML::TokeParser::Simple->new($input_filename);
438
439 # write HTML tmp file without the font tag and image size are added to the img tag
440 open (PROD, ">$tmp_filename") || die("Error Writing to File: $tmp_filename $!");
441 while (my $token = $parser->get_token())
442 {
443 # is it an img tag
444 if ($token->is_start_tag('img'))
445 {
446 # get the attributes
447 my $attr = $token->return_attr;
448
449 # get the full path to the image
450 my $img_file = &util::filename_cat($dirname,$attr->{src});
451
452 # set the width and height attribute
453 ($attr->{width}, $attr->{height}) = imgsize($img_file);
454
455 # recreate the tag
456 print PROD "<img";
457 print PROD map { qq { $_="$attr->{$_}"} } keys %$attr;
458 print PROD ">";
459 }
460 # is it a font tag
461 else
462 {
463 if (($token->is_start_tag('font')) || ($token->is_end_tag('font')))
464 {
465 # remove font tag
466 print PROD "";
467 }
468 else
469 {
470 # print without changes
471 print PROD $token->as_is;
472 }
473 }
474 }
475 close (PROD) || die("Error Closing File: $tmp_filename $!");
476
477 # run html-tidy on the tmp file to make it a proper XML file
478 my $tidyfile = `tidy -utf8 -wrap 0 -asxml $tmp_filename`;
479
480 # write result back to the tmp file
481 open (PROD, ">$tmp_filename") || die("Error Writing to File: $tmp_filename $!");
482 print PROD $tidyfile;
483 close (PROD) || die("Error Closing File: $tmp_filename $!");
484
485 # return the output filename
486 return $tmp_filename;
487}
488
489sub read_into_doc_obj
490{
491 my $self = shift (@_);
492 my ($pluginfo, $base_dir, $file, $metadata, $processor, $maxdocs, $total_count, $gli) = @_;
493
494 # check the process_exp and block_exp thing
495 my ($block_status,$filename) = $self->read_block(@_);
496 return $block_status if ((!defined $block_status) || ($block_status==0));
497
498 # get the input file
499 my $input_filename = $file;
500 my ($tailname, $dirname, $suffix) = &File::Basename::fileparse($input_filename, "\\.[^\\.]+\$");
501 $suffix = lc($suffix);
502
503 if (($self->{'tidy_html'}) || ($self->{'old_style_HDL'}))
504 {
505 # because the document has to be sectionalized set the description tags
506 $self->{'description_tags'} = 1;
507
508 # set the file to be tidied
509 $input_filename = &util::filename_cat($base_dir,$file) if $base_dir =~ /\w/;
510
511 # get the tidied file
512 #my $tidy_filename = $self->tmp_tidy_file($input_filename);
513 my $tidy_filename = $self->convert_tidy_or_oldHDL_file($input_filename);
514
515 # derive tmp filename from input filename
516 my ($tailname, $dirname, $suffix) = &File::Basename::fileparse($tidy_filename, "\\.[^\\.]+\$");
517
518 # set the new input file and base_dir to be from the tidied file
519 $file = "$tailname$suffix";
520 $base_dir = $dirname;
521 }
522
523 # call the parent read_into_doc_obj
524 my ($process_status,$doc_obj) = &BasPlug::read_into_doc_obj($self,$pluginfo, $base_dir, $file, $metadata, $processor, $maxdocs, $total_count, $gli);
525
526 return ($process_status,$doc_obj);
527}
528
529sub new {
530 my ($class) = shift (@_);
531 my ($pluginlist,$inputargs,$hashArgOptLists) = @_;
532 push(@$pluginlist, $class);
533
534 if(defined $arguments){ push(@{$hashArgOptLists->{"ArgList"}},@{$arguments});}
535 if(defined $options) { push(@{$hashArgOptLists->{"OptList"}},$options)};
536
537
538 my $self = (defined $hashArgOptLists)? new BasPlug($pluginlist,$inputargs,$hashArgOptLists): new BasPlug($pluginlist,$inputargs);
539
540 if ($self->{'w3mir'}) {
541 $self->{'file_is_url'} = 1;
542 }
543 $self->{'aux_files'} = {};
544 $self->{'dir_num'} = 0;
545 $self->{'file_num'} = 0;
546
547 return bless $self, $class;
548}
549
550# may want to use (?i)\.(gif|jpe?g|jpe|png|css|js(?:@.*)?)$
551# if have eg <script language="javascript" src="img/lib.js@123">
552sub get_default_block_exp {
553 my $self = shift (@_);
554
555 return q^(?i)\.(gif|jpe?g|jpe|jpg|png|css)$^;
556}
557
558sub get_default_process_exp {
559 my $self = shift (@_);
560
561 # the last option is an attempt to encode the concept of an html query ...
562 return q^(?i)(\.html?|\.shtml|\.shm|\.asp|\.php\d?|\.cgi|.+\?.+=.*)$^;
563}
564
565sub store_block_files
566{
567 my $self =shift (@_);
568 my ($filename) = @_;
569 my $html_fname = $filename;
570 my @file_blocks;
571
572 my ($language, $encoding) = $self->textcat_get_language_encoding ($filename);
573
574 # read in file ($text will be in utf8)
575 my $text = "";
576 $self->read_file ($filename, $encoding, $language, \$text);
577 my $textref = \$text;
578 my $opencom = '(?:<!--|&lt;!(?:&mdash;|&#151;|--))';
579 my $closecom = '(?:-->|(?:&mdash;|&#151;|--)&gt;)';
580 $$textref =~ s/$opencom(.*?)$closecom//gs;
581
582 my $attval = "\\\"[^\\\"]+\\\"|[^\\s>]+";
583 my @img_matches = ($$textref =~ m/<img[^>]*?src\s*=\s*($attval)[^>]*>/igs);
584 my @usemap_matches = ($$textref =~ m/<img[^>]*?usemap\s*=\s*($attval)[^>]*>/igs);
585 my @link_matches = ($$textref =~ m/<link[^>]*?href\s*=\s*($attval)[^>]*>/igs);
586 my @embed_matches = ($$textref =~ m/<embed[^>]*?src\s*=\s*($attval)[^>]*>/igs);
587 my @tabbg_matches = ($$textref =~ m/<(?:table|tr|td)[^>]*?background\s*=\s*($attval)[^>]*>/igs);
588
589 foreach my $link (@img_matches, @usemap_matches, @link_matches, @embed_matches, @tabbg_matches) {
590
591 # remove quotes from link at start and end if necessary
592 if ($link=~/^\"/) {
593 $link=~s/^\"//;
594 $link=~s/\"$//;
595 }
596
597 $link =~ s/\#.*$//s; # remove any anchor names, e.g. foo.html#name becomes foo.html
598
599 if ($link !~ m@^/@ && $link !~ m/^([A-Z]:?)\\/) {
600 # Turn relative file path into full path
601 my $dirname = &File::Basename::dirname($filename);
602 $link = &util::filename_cat($dirname, $link);
603 }
604 $link = $self->eval_dir_dots($link);
605
606 $self->{'file_blocks'}->{$link} = 1;
607 }
608}
609
610
611# do plugin specific processing of doc_obj
612sub process {
613 my $self = shift (@_);
614 my ($textref, $pluginfo, $base_dir, $file, $metadata, $doc_obj, $gli) = @_;
615 my $outhandle = $self->{'outhandle'};
616
617 print STDERR "<Processing n='$file' p='HTMLPlug'>\n" if ($gli);
618
619 print $outhandle "HTMLPlug: processing $file\n"
620 if $self->{'verbosity'} > 1;
621
622 if ($ENV{'GSDLOS'} =~ /^windows/i) {
623 # this makes life so much easier... perl can cope with unix-style '/'s.
624 $base_dir =~ s@(\\)+@/@g;
625 $file =~ s@(\\)+@/@g;
626 }
627
628 # reset per-doc stuff...
629 $self->{'aux_files'} = {};
630 $self->{'dir_num'} = 0;
631 $self->{'file_num'} = 0;
632
633 # process an HTML file where sections are divided by headings tags (H1, H2 ...)
634 # you can also include metadata in the format (X can be any number)
635 # <hX>Title<!--gsdl-metadata
636 # <Metadata name="name1">value1</Metadata>
637 # ...
638 # <Metadata name="nameN">valueN</Metadata>
639 #--></hX>
640 if ($self->{'sectionalise_using_h_tags'}) {
641 # description_tags should allways be activated because we convert headings to description tags
642 $self->{'description_tags'} = 1;
643
644 my $arrSections = [];
645 $$textref =~ s/<h([0-9]+)[^>]*>(.*?)<\/h[0-9]+>/$self->process_heading($1, $2, $arrSections, $file)/isge;
646
647 if (scalar(@$arrSections)) {
648 my $strMetadata = $self->update_section_data($arrSections, -1);
649 if (length($strMetadata)) {
650 $strMetadata = '<!--' . $strMetadata . "\n-->\n</body>";
651 $$textref =~ s/<\/body>/$strMetadata/ig;
652 }
653 }
654 }
655
656 my $cursection = $doc_obj->get_top_section();
657
658 $self->extract_metadata ($textref, $metadata, $doc_obj, $cursection)
659 unless $self->{'no_metadata'} || $self->{'description_tags'};
660
661 # Store URL for page as metadata - this can be used for an
662 # altavista style search interface. The URL won't be valid
663 # unless the file structure contains the domain name (i.e.
664 # like when w3mir is used to download a website).
665
666 # URL metadata (even invalid ones) are used to support internal
667 # links, so even if 'file_is_url' is off, still need to store info
668
669 my $web_url = "http://$file";
670 $doc_obj->add_metadata($cursection, "URL", $web_url);
671
672 if ($self->{'file_is_url'}) {
673 $doc_obj->add_metadata($cursection, "weblink", "<a href=\"$web_url\">");
674 $doc_obj->add_metadata($cursection, "webicon", "_iconworld_");
675 $doc_obj->add_metadata($cursection, "/weblink", "</a>");
676 }
677
678 if ($self->{'description_tags'}) {
679 # remove the html header - note that doing this here means any
680 # sections defined within the header will be lost (so all <Section>
681 # tags must appear within the body of the HTML)
682 my ($head_keep) = ($$textref =~ m/^(.*?)<body[^>]*>/is);
683
684 $$textref =~ s/^.*?<body[^>]*>//is;
685 $$textref =~ s/(<\/body[^>]*>|<\/html[^>]*>)//isg;
686
687 my $opencom = '(?:<!--|&lt;!(?:&mdash;|&#151;|--))';
688 my $closecom = '(?:-->|(?:&mdash;|&#151;|--)&gt;)';
689
690 my $lt = '(?:<|&lt;)';
691 my $gt = '(?:>|&gt;)';
692 my $quot = '(?:"|&quot;|&rdquo;|&ldquo;)';
693
694 my $dont_strip = '';
695 if ($self->{'no_strip_metadata_html'}) {
696 ($dont_strip = $self->{'no_strip_metadata_html'}) =~ s{,}{|}g;
697 }
698
699 my $found_something = 0; my $top = 1;
700 while ($$textref =~ s/^(.*?)$opencom(.*?)$closecom//s) {
701 my $text = $1;
702 my $comment = $2;
703 if (defined $text) {
704 # text before a comment - note that getting to here
705 # doesn't necessarily mean there are Section tags in
706 # the document
707 $self->process_section(\$text, $base_dir, $file, $doc_obj, $cursection);
708 }
709 while ($comment =~ s/$lt(.*?)$gt//s) {
710 my $tag = $1;
711 if ($tag eq "Section") {
712 $found_something = 1;
713 $cursection = $doc_obj->insert_section($doc_obj->get_end_child($cursection)) unless $top;
714 $top = 0;
715 } elsif ($tag eq "/Section") {
716 $found_something = 1;
717 $cursection = $doc_obj->get_parent_section ($cursection);
718 } elsif ($tag =~ /^Metadata name=$quot(.*?)$quot/s) {
719 my $metaname = $1;
720 my $accumulate = $tag =~ /mode=${quot}accumulate${quot}/ ? 1 : 0;
721 $comment =~ s/^(.*?)$lt\/Metadata$gt//s;
722 my $metavalue = $1;
723 $metavalue =~ s/^\s+//;
724 $metavalue =~ s/\s+$//;
725 # assume that no metadata value intentionally includes
726 # carriage returns or HTML tags (if they're there they
727 # were probably introduced when converting to HTML from
728 # some other format).
729 # actually some people want to have html tags in their
730 # metadata.
731 $metavalue =~ s/[\cJ\cM]/ /sg;
732 $metavalue =~ s/<[^>]+>//sg
733 unless $dont_strip && ($dont_strip eq 'all' || $metaname =~ /^($dont_strip)$/);
734 $metavalue =~ s/\s+/ /sg;
735 if ($accumulate) {
736 $doc_obj->add_utf8_metadata($cursection, $metaname, $metavalue);
737 } else {
738 $doc_obj->set_utf8_metadata_element($cursection, $metaname, $metavalue);
739 }
740 } elsif ($tag eq "Description" || $tag eq "/Description") {
741 # do nothing with containing Description tags
742 } else {
743 # simple HTML tag (probably created by the conversion
744 # to HTML from some other format) - we'll ignore it and
745 # hope for the best ;-)
746 }
747 }
748 }
749 if ($cursection ne "") {
750 print $outhandle "HTMLPlug: WARNING: $file contains unmatched <Section></Section> tags\n";
751 }
752
753 $$textref =~ s/^.*?<body[^>]*>//is;
754 $$textref =~ s/(<\/body[^>]*>|<\/html[^>]*>)//isg;
755 if ($$textref =~ /\S/) {
756 if (!$found_something) {
757 if ($self->{'verbosity'} > 2) {
758 print $outhandle "HTMLPlug: WARNING: $file appears to contain no Section tags so\n";
759 print $outhandle " will be processed as a single section document\n";
760 }
761
762 # go ahead and process single-section document
763 $self->process_section($textref, $base_dir, $file, $doc_obj, $cursection);
764
765 # if document contains no Section tags we'll go ahead
766 # and extract metadata (this won't have been done
767 # above as the -description_tags option prevents it)
768 my $complete_text = $head_keep.$doc_obj->get_text($cursection);
769 $self->extract_metadata (\$complete_text, $metadata, $doc_obj, $cursection)
770 unless $self->{'no_metadata'};
771
772 } else {
773 print $outhandle "HTMLPlug: WARNING: $file contains the following text outside\n";
774 print $outhandle " of the final closing </Section> tag. This text will\n";
775 print $outhandle " be ignored.";
776
777 my ($text);
778 if (length($$textref) > 30) {
779 $text = substr($$textref, 0, 30) . "...";
780 } else {
781 $text = $$textref;
782 }
783 $text =~ s/\n/ /isg;
784 print $outhandle " ($text)\n";
785 }
786 } elsif (!$found_something) {
787
788 if ($self->{'verbosity'} > 2) {
789 # may get to here if document contained no valid Section
790 # tags but did contain some comments. The text will have
791 # been processed already but we should print the warning
792 # as above and extract metadata
793 print $outhandle "HTMLPlug: WARNING: $file appears to contain no Section tags and\n";
794 print $outhandle " is blank or empty. Metadata will be assigned if present.\n";
795 }
796
797 my $complete_text = $head_keep.$doc_obj->get_text($cursection);
798 $self->extract_metadata (\$complete_text, $metadata, $doc_obj, $cursection)
799 unless $self->{'no_metadata'};
800 }
801
802 } else {
803
804 # remove header and footer
805 if (!$self->{'keep_head'} || $self->{'description_tags'}) {
806 $$textref =~ s/^.*?<body[^>]*>//is;
807 $$textref =~ s/(<\/body[^>]*>|<\/html[^>]*>)//isg;
808 }
809
810 # single section document
811 $self->process_section($textref, $base_dir, $file, $doc_obj, $cursection);
812 }
813 return 1;
814}
815
816
817sub process_heading
818{
819 my ($self, $nHeadNo, $strHeadingText, $arrSections, $file) = @_;
820 $strHeadingText = '' if (!defined($strHeadingText));
821
822 my $strMetadata = $self->update_section_data($arrSections, int($nHeadNo));
823
824 my $strSecMetadata = '';
825 while ($strHeadingText =~ s/<!--gsdl-metadata(.*?)-->//is)
826 {
827 $strSecMetadata .= $1;
828 }
829
830 $strHeadingText =~ s/^\s+//g;
831 $strHeadingText =~ s/\s+$//g;
832 $strSecMetadata =~ s/^\s+//g;
833 $strSecMetadata =~ s/\s+$//g;
834
835 $strMetadata .= "\n<Section>\n\t<Description>\n\t\t<Metadata name=\"Title\">" . $strHeadingText . "</Metadata>\n";
836
837 if (length($strSecMetadata)) {
838 $strMetadata .= "\t\t" . $strSecMetadata . "\n";
839 }
840
841 $strMetadata .= "\t</Description>\n";
842
843 return "<!--" . $strMetadata . "-->";
844}
845
846
847sub update_section_data
848{
849 my ($self, $arrSections, $nCurTocNo) = @_;
850 my ($strBuffer, $nLast, $nSections) = ('', 0, scalar(@$arrSections));
851
852 if ($nSections == 0) {
853 push @$arrSections, $nCurTocNo;
854 return $strBuffer;
855 }
856 $nLast = $arrSections->[$nSections - 1];
857 if ($nCurTocNo > $nLast) {
858 push @$arrSections, $nCurTocNo;
859 return $strBuffer;
860 }
861 for(my $i = $nSections - 1; $i >= 0; $i--) {
862 if ($nCurTocNo <= $arrSections->[$i]) {
863 $strBuffer .= "\n</Section>";
864 pop @$arrSections;
865 }
866 }
867 push @$arrSections, $nCurTocNo;
868 return $strBuffer;
869}
870
871
872# note that process_section may be called multiple times for a single
873# section (relying on the fact that add_utf8_text appends the text to any
874# that may exist already).
875sub process_section {
876 my $self = shift (@_);
877 my ($textref, $base_dir, $file, $doc_obj, $cursection) = @_;
878 # trap links
879 if (!$self->{'nolinks'}) {
880
881 # usemap="./#index" not handled correctly => change to "#index"
882 $$textref =~ s/(<img[^>]*?usemap\s*=\s*[\"\']?)([^\"\'>\s]+)([\"\']?[^>]*>)/
883 $self->replace_usemap_links($1, $2, $3)/isge;
884
885 $$textref =~ s/(<(?:a|area|frame|link|script)\s+[^>]*?\s*(?:href|src)\s*=\s*[\"\']?)([^\"\'>\s]+)([\"\']?[^>]*>)/
886 $self->replace_href_links ($1, $2, $3, $base_dir, $file, $doc_obj, $cursection)/isge;
887 }
888
889 # trap images
890
891 # allow spaces if inside quotes - jrm21
892 $$textref =~ s/(<(?:img|embed|table|tr|td)[^>]*?(?:src|background)\s*=\s*)([\"\'][^\"\']+[\"\']|[^\s>]+)([^>]*>)/
893 $self->replace_images ($1, $2, $3, $base_dir, $file, $doc_obj, $cursection)/isge;
894
895 # add text to document object
896 # turn \ into \\ so that the rest of greenstone doesn't think there
897 # is an escape code following. (Macro parsing loses them...)
898 $$textref =~ s/\\/\\\\/go;
899
900 $doc_obj->add_utf8_text($cursection, $$textref);
901}
902
903sub replace_images {
904 my $self = shift (@_);
905 my ($front, $link, $back, $base_dir,
906 $file, $doc_obj, $section) = @_;
907
908 # remove quotes from link at start and end if necessary
909 if ($link=~/^[\"\']/) {
910 $link=~s/^[\"\']//;$link=~s/[\"\']$//;
911 $front.='"';
912 $back="\"$back";
913 }
914
915 $link =~ s/\n/ /g;
916
917 # Hack to overcome Windows wv 0.7.1 bug that causes embedded images to be broken
918 # If the Word file path has spaces in it, wv messes up and you end up with
919 # absolute paths for the images, and without the "file://" prefix
920 # So check for this special case and massage the data to be correct
921 if ($ENV{'GSDLOS'} =~ /^windows/i && $self->{'plugin_type'} eq "WordPlug" && $link =~ /^[A-Za-z]\:\\/) {
922 $link =~ s/^.*\\([^\\]+)$/$1/;
923 }
924
925 my ($href, $hash_part, $rl) = $self->format_link ($link, $base_dir, $file);
926
927 my $img_file = $self->add_file ($href, $rl, $hash_part, $base_dir, $doc_obj, $section);
928
929 my $anchor_name = $img_file;
930 #$anchor_name =~ s/^.*\///;
931 #$anchor_name = "<a name=\"$anchor_name\" ></a>";
932
933 my $image_link = $front . $img_file .$back;
934 my $anchor_link = "<a href=\"$img_file\" >".$image_link."</a>";
935
936 return $anchor_link;
937 #return $front . $img_file . $back . $anchor_name;
938}
939
940sub replace_href_links {
941 my $self = shift (@_);
942 my ($front, $link, $back, $base_dir, $file, $doc_obj, $section) = @_;
943
944 # attempt to sort out targets - frames are not handled
945 # well in this plugin and some cases will screw things
946 # up - e.g. the _parent target (so we'll just remove
947 # them all ;-)
948 $front =~ s/(target=\"?)_top(\"?)/$1_gsdltop_$2/is;
949 $back =~ s/(target=\"?)_top(\"?)/$1_gsdltop_$2/is;
950 $front =~ s/target=\"?_parent\"?//is;
951 $back =~ s/target=\"?_parent\"?//is;
952
953 return $front . $link . $back if $link =~ /^\#/s;
954 $link =~ s/\n/ /g;
955
956 my ($href, $hash_part, $rl) = $self->format_link ($link, $base_dir, $file);
957 # href may use '\'s where '/'s should be on Windows
958 $href =~ s/\\/\//g;
959
960 my ($filename) = $href =~ /^(?:.*?):(?:\/\/)?(.*)/;
961
962
963 ##### leave all these links alone (they won't be picked up by intermediate
964 ##### pages). I think that's safest when dealing with frames, targets etc.
965 ##### (at least until I think of a better way to do it). Problems occur with
966 ##### mailto links from within small frames, the intermediate page is displayed
967 ##### within that frame and can't be seen. There is still potential for this to
968 ##### happen even with html pages - the solution seems to be to somehow tell
969 ##### the browser from the server side to display the page being sent (i.e.
970 ##### the intermediate page) in the top level window - I'm not sure if that's
971 ##### possible - the following line should probably be deleted if that can be done
972 return $front . $link . $back if $href =~ /^(mailto|news|gopher|nntp|telnet|javascript):/is;
973
974
975 if (($rl == 0) || ($filename =~ /$self->{'process_exp'}/) ||
976 ($href =~ /\/$/) || ($href =~ /^(mailto|news|gopher|nntp|telnet|javascript):/i)) {
977 &ghtml::urlsafe ($href);
978 return $front . "_httpextlink_&rl=" . $rl . "&href=" . $href . $hash_part . $back;
979 } else {
980 # link is to some other type of file (eg image) so we'll
981 # need to associate that file
982 return $front . $self->add_file ($href, $rl, $hash_part, $base_dir, $doc_obj, $section) . $back;
983 }
984}
985
986sub add_file {
987 my $self = shift (@_);
988 my ($href, $rl, $hash_part, $base_dir, $doc_obj, $section) = @_;
989 my ($newname);
990
991 my $filename = $href;
992 if ($base_dir eq "") {
993 # remove http:/ thereby leaving one slash at the start
994 $filename =~ s/^[^:]*:\///;
995 }
996 else {
997 # remove http://
998 $filename =~ s/^[^:]*:\/\///;
999 }
1000
1001 $filename = &util::filename_cat($base_dir, $filename);
1002
1003 # Replace %20's in URL with a space if required. Note that the filename
1004 # may include the %20 in some situations
1005 if ($filename =~ /\%20/) {
1006 if (!-e $filename) {
1007 $filename =~ s/\%20/ /g;
1008 }
1009 }
1010
1011 my ($ext) = $filename =~ /(\.[^\.]*)$/;
1012
1013 if ($rl == 0) {
1014 if ((!defined $ext) || ($ext !~ /$self->{'assoc_files'}/)) {
1015 return "_httpextlink_&rl=0&el=prompt&href=" . $href . $hash_part;
1016 }
1017 else {
1018 return "_httpextlink_&rl=0&el=direct&href=" . $href . $hash_part;
1019 }
1020 }
1021
1022 if ((!defined $ext) || ($ext !~ /$self->{'assoc_files'}/)) {
1023 return "_httpextlink_&rl=" . $rl . "&href=" . $href . $hash_part;
1024 }
1025 if ($self->{'rename_assoc_files'}) {
1026 if (defined $self->{'aux_files'}->{$href}) {
1027 $newname = $self->{'aux_files'}->{$href}->{'dir_num'} . "/" .
1028 $self->{'aux_files'}->{$href}->{'file_num'} . $ext;
1029 } else {
1030 $newname = $self->{'dir_num'} . "/" . $self->{'file_num'} . $ext;
1031 $self->{'aux_files'}->{$href} = {'dir_num' => $self->{'dir_num'}, 'file_num' => $self->{'file_num'}};
1032 $self->inc_filecount ();
1033 }
1034 $doc_obj->associate_file($filename, $newname, undef, $section);
1035 return "_httpdocimg_/$newname";
1036 } else {
1037 ($newname) = $filename =~ /([^\/\\]*)$/;
1038 $doc_obj->associate_file($filename, $newname, undef, $section);
1039 return "_httpdocimg_/$newname";
1040 }
1041}
1042
1043
1044sub format_link {
1045 my $self = shift (@_);
1046 my ($link, $base_dir, $file) = @_;
1047
1048 my ($before_hash, $hash_part) = $link =~ /^([^\#]*)(\#?.*)$/;
1049
1050 $hash_part = "" if !defined $hash_part;
1051 if (!defined $before_hash || $before_hash !~ /[\w\.\/]/) {
1052 my $outhandle = $self->{'outhandle'};
1053 print $outhandle "HTMLPlug: ERROR - badly formatted tag ignored ($link)\n"
1054 if $self->{'verbosity'};
1055 return ($link, "", 0);
1056 }
1057
1058 if ($before_hash =~ s@^((?:http|ftp|file)://)@@i) {
1059 my $type = $1;
1060
1061 if ($link =~ /^(http|ftp):/i) {
1062 # Turn url (using /) into file name (possibly using \ on windows)
1063 my @http_dir_split = split('/', $before_hash);
1064 $before_hash = &util::filename_cat(@http_dir_split);
1065 }
1066
1067 $before_hash = $self->eval_dir_dots($before_hash);
1068
1069 my $linkfilename = &util::filename_cat ($base_dir, $before_hash);
1070
1071 my $rl = 0;
1072 $rl = 1 if (-e $linkfilename);
1073
1074 # make sure there's a slash on the end if it's a directory
1075 if ($before_hash !~ /\/$/) {
1076 $before_hash .= "/" if (-d $linkfilename);
1077 }
1078
1079 return ($type . $before_hash, $hash_part, $rl);
1080
1081 } elsif ($link !~ /^(mailto|news|gopher|nntp|telnet|javascript):/i && $link !~ /^\//) {
1082 if ($before_hash =~ s@^/@@ || $before_hash =~ /\\/) {
1083
1084 # the first directory will be the domain name if file_is_url
1085 # to generate archives, otherwise we'll assume all files are
1086 # from the same site and base_dir is the root
1087
1088 if ($self->{'file_is_url'}) {
1089 my @dirs = split /[\/\\]/, $file;
1090 my $domname = shift (@dirs);
1091 $before_hash = &util::filename_cat($domname, $before_hash);
1092 $before_hash =~ s@\\@/@g; # for windows
1093 }
1094 else
1095 {
1096 # see if link shares directory with source document
1097 # => turn into relative link if this is so!
1098
1099 if ($ENV{'GSDLOS'} =~ /^windows/i) {
1100 # too difficult doing a pattern match with embedded '\'s...
1101 my $win_before_hash=$before_hash;
1102 $win_before_hash =~ s@(\\)+@/@g;
1103 # $base_dir is already similarly "converted" on windows.
1104 if ($win_before_hash =~ s@^$base_dir/@@o) {
1105 # if this is true, we removed a prefix
1106 $before_hash=$win_before_hash;
1107 }
1108 }
1109 else {
1110 # before_hash has lost leading slash by this point,
1111 # -> add back in prior to substitution with $base_dir
1112 $before_hash = "/$before_hash";
1113
1114 $before_hash = &util::filename_cat("",$before_hash);
1115 $before_hash =~ s@^$base_dir/@@;
1116 }
1117 }
1118 } else {
1119 # Turn relative file path into full path
1120 my $dirname = &File::Basename::dirname($file);
1121 $before_hash = &util::filename_cat($dirname, $before_hash);
1122 $before_hash = $self->eval_dir_dots($before_hash);
1123 }
1124
1125 my $linkfilename = &util::filename_cat ($base_dir, $before_hash);
1126 # make sure there's a slash on the end if it's a directory
1127 if ($before_hash !~ /\/$/) {
1128 $before_hash .= "/" if (-d $linkfilename);
1129 }
1130 return ("http://" . $before_hash, $hash_part, 1);
1131 } else {
1132 # mailto, news, nntp, telnet, javascript or gopher link
1133 return ($before_hash, "", 0);
1134 }
1135}
1136
1137sub extract_first_NNNN_characters {
1138 my $self = shift (@_);
1139 my ($textref, $doc_obj, $thissection) = @_;
1140
1141 foreach my $size (split /,/, $self->{'first'}) {
1142 my $tmptext = $$textref;
1143 # skip to the body
1144 $tmptext =~ s/.*<body[^>]*>//i;
1145 # remove javascript
1146 $tmptext =~ s@<script.*?</script>@ @sig;
1147 $tmptext =~ s/<[^>]*>/ /g;
1148 $tmptext =~ s/&nbsp;/ /g;
1149 $tmptext =~ s/^\s+//;
1150 $tmptext =~ s/\s+$//;
1151 $tmptext =~ s/\s+/ /gs;
1152 $tmptext = &unicode::substr ($tmptext, 0, $size);
1153 $tmptext =~ s/\s\S*$/&#8230;/; # adds an ellipse (...)
1154 $doc_obj->add_utf8_metadata ($thissection, "First$size", $tmptext);
1155 }
1156}
1157
1158
1159sub extract_metadata {
1160 my $self = shift (@_);
1161 my ($textref, $metadata, $doc_obj, $section) = @_;
1162 my $outhandle = $self->{'outhandle'};
1163 # if we don't want metadata, we may as well not be here ...
1164 return if (!defined $self->{'metadata_fields'});
1165
1166 # metadata fields to extract/save. 'key' is the (lowercase) name of the
1167 # html meta, 'value' is the metadata name for greenstone to use
1168 my %find_fields = ();
1169
1170 my %creator_fields = (); # short-cut for lookups
1171
1172
1173 foreach my $field (split /,/, $self->{'metadata_fields'}) {
1174 $field =~ s/^\s+//; # remove leading whitespace
1175 $field =~ s/\s+$//; # remove trailing whitespace
1176
1177 # support tag<tagname>
1178 if ($field =~ /^(.*?)<(.*?)>$/) {
1179 # "$2" is the user's preferred gs metadata name
1180 $find_fields{lc($1)}=$2; # lc = lowercase
1181 } else { # no <tagname> for mapping
1182 # "$field" is the user's preferred gs metadata name
1183 $find_fields{lc($field)}=$field; # lc = lowercase
1184 }
1185 }
1186
1187 if (defined $self->{'hunt_creator_metadata'} &&
1188 $self->{'hunt_creator_metadata'} == 1 ) {
1189 my @extra_fields =
1190 (
1191 'author',
1192 'author.email',
1193 'creator',
1194 'dc.creator',
1195 'dc.creator.corporatename',
1196 );
1197
1198 # add the creator_metadata fields to search for
1199 foreach my $field (@extra_fields) {
1200 $creator_fields{$field}=0; # add to lookup hash
1201 }
1202 }
1203
1204
1205 # find the header in the html file, which has the meta tags
1206 $$textref =~ m@<head>(.*?)</head>@si;
1207
1208 my $html_header=$1;
1209
1210 # go through every <meta... tag defined in the html and see if it is
1211 # one of the tags we want to match.
1212
1213 # special case for title - we want to remember if its been found
1214 my $found_title = 0;
1215 # this assumes that ">" won't appear. (I don't think it's allowed to...)
1216 $html_header =~ /^/; # match the start of the string, for \G assertion
1217
1218 while ($html_header =~ m/\G.*?<meta(.*?)>/sig) {
1219 my $metatag=$1;
1220 my ($tag, $value);
1221
1222 # find the tag name
1223 $metatag =~ /(?:name|http-equiv)\s*=\s*([\"\'])?(.*?)\1/is;
1224 $tag=$2;
1225 # in case they're not using " or ', but they should...
1226 if (! $tag) {
1227 $metatag =~ /(?:name|http-equiv)\s*=\s*([^\s\>]+)/is;
1228 $tag=$1;
1229 }
1230
1231 if (!defined $tag) {
1232 print $outhandle "HTMLPlug: can't find NAME in \"$metatag\"\n";
1233 next;
1234 }
1235
1236 # don't need to assign this field if it was passed in from a previous
1237 # (recursive) plugin
1238 if (defined $metadata->{$tag}) {next}
1239
1240 # find the tag content
1241 $metatag =~ /content\s*=\s*([\"\'])?(.*?)\1/is;
1242 $value=$2;
1243
1244 if (! $value) {
1245 $metatag =~ /(?:name|http-equiv)\s*=\s*([^\s\>]+)/is;
1246 $value=$1;
1247 }
1248 if (!defined $value) {
1249 print $outhandle "HTMLPlug: can't find VALUE in \"$metatag\"\n";
1250 next;
1251 }
1252
1253 # clean up and add
1254 $value =~ s/\s+/ /gs;
1255 chomp($value); # remove trailing \n, if any
1256 if (exists $creator_fields{lc($tag)}) {
1257 # map this value onto greenstone's "Creator" metadata
1258 $tag='Creator';
1259 } elsif (!exists $find_fields{lc($tag)}) {
1260 next; # don't want this tag
1261 } else {
1262 # get the user's preferred capitalisation
1263 $tag = $find_fields{lc($tag)};
1264 }
1265 if (lc($tag) eq "title") {
1266 $found_title = 1;
1267 }
1268 print $outhandle " extracted \"$tag\" metadata \"$value\"\n"
1269 if ($self->{'verbosity'} > 2);
1270 if ($tag =~ /date.*/i){
1271 $tag = lc($tag);
1272 }
1273 $doc_obj->add_utf8_metadata($section, $tag, $value);
1274
1275 }
1276
1277 # TITLE: extract the document title
1278 if (exists $find_fields{'title'} && !$found_title) {
1279 # we want a title, and didn't find one in the meta tags
1280 # see if there's a <title> tag
1281 my $title;
1282 my $from = ""; # for debugging output only
1283 if ($html_header =~ /<title[^>]*>([^<]+)<\/title[^>]*>/is) {
1284 $title = $1;
1285 $from = "<title> tags";
1286 }
1287
1288 if (!defined $title) {
1289 $from = "first 100 chars";
1290 # if no title use first 100 or so characters
1291 $title = $$textref;
1292 $title =~ s/^\xFE\xFF//; # Remove unicode byte order mark
1293 $title =~ s/^.*?<body>//si;
1294 # ignore javascript!
1295 $title =~ s@<script.*?</script>@ @sig;
1296 $title =~ s/<\/([^>]+)><\1>//g; # (eg) </b><b> - no space
1297 $title =~ s/<[^>]*>/ /g; # remove all HTML tags
1298 $title = substr ($title, 0, 100);
1299 $title =~ s/\s\S*$/.../;
1300 }
1301 $title =~ s/<[^>]*>/ /g; # remove html tags
1302 $title =~ s/&nbsp;/ /g;
1303 $title =~ s/(?:&nbsp;|\xc2\xa0)/ /g; # utf-8 for nbsp...
1304 $title =~ s/\s+/ /gs; # collapse multiple spaces
1305 $title =~ s/^\s*//; # remove leading spaces
1306 $title =~ s/\s*$//; # remove trailing spaces
1307
1308 $title =~ s/^$self->{'title_sub'}// if ($self->{'title_sub'});
1309 $title =~ s/^\s+//s; # in case title_sub introduced any...
1310 $doc_obj->add_utf8_metadata ($section, 'Title', $title);
1311 print $outhandle " extracted Title metadata \"$title\" from $from\n"
1312 if ($self->{'verbosity'} > 2);
1313 }
1314
1315 # add FileFormat metadata
1316 $doc_obj->add_metadata($section,"FileFormat", "HTML");
1317
1318 # Special, for metadata names such as tagH1 - extracts
1319 # the text between the first <H1> and </H1> tags into "H1" metadata.
1320
1321 foreach my $field (keys %find_fields) {
1322 if ($field !~ /^tag([a-z0-9]+)$/i) {next}
1323 my $tag = $1;
1324 if ($$textref =~ m@<$tag[^>]*>(.*?)</$tag[^>]*>@g) {
1325 my $content = $1;
1326 $content =~ s/&nbsp;/ /g;
1327 $content =~ s/<[^>]*>/ /g;
1328 $content =~ s/^\s+//;
1329 $content =~ s/\s+$//;
1330 $content =~ s/\s+/ /gs;
1331 if ($content) {
1332 $tag=$find_fields{"tag$tag"}; # get the user's capitalisation
1333 $tag =~ s/^tag//i;
1334 $doc_obj->add_utf8_metadata ($section, $tag, $content);
1335 print $outhandle " extracted \"$tag\" metadata \"$content\"\n"
1336 if ($self->{'verbosity'} > 2);
1337 }
1338 }
1339 }
1340}
1341
1342
1343# evaluate any "../" to next directory up
1344# evaluate any "./" as here
1345sub eval_dir_dots {
1346 my $self = shift (@_);
1347 my ($filename) = @_;
1348 my $dirsep_os = &util::get_os_dirsep();
1349 my @dirsep = split(/$dirsep_os/,$filename);
1350
1351 my @eval_dirs = ();
1352 foreach my $d (@dirsep) {
1353 if ($d eq "..") {
1354 pop(@eval_dirs);
1355
1356 } elsif ($d eq ".") {
1357 # do nothing!
1358
1359 } else {
1360 push(@eval_dirs,$d);
1361 }
1362 }
1363
1364 # Need to fiddle with number of elements in @eval_dirs if the
1365 # first one is the empty string. This is because of a
1366 # modification to util::filename_cat that supresses the addition
1367 # of a leading '/' character (or \ if windows) (intended to help
1368 # filename cat with relative paths) if the first entry in the
1369 # array is the empty string. Making the array start with *two*
1370 # empty strings is a way to defeat this "smart" option.
1371 #
1372 if (scalar(@eval_dirs) > 0) {
1373 if ($eval_dirs[0] eq ""){
1374 unshift(@eval_dirs,"");
1375 }
1376 }
1377 return &util::filename_cat(@eval_dirs);
1378}
1379
1380sub replace_usemap_links {
1381 my $self = shift (@_);
1382 my ($front, $link, $back) = @_;
1383
1384 $link =~ s/^\.\///;
1385 return $front . $link . $back;
1386}
1387
1388sub inc_filecount {
1389 my $self = shift (@_);
1390
1391 if ($self->{'file_num'} == 1000) {
1392 $self->{'dir_num'} ++;
1393 $self->{'file_num'} = 0;
1394 } else {
1395 $self->{'file_num'} ++;
1396 }
1397}
1398
1399
1400# Extend the BasPlug read_file so that strings like &eacute; are
1401# converted to UTF8 internally.
1402#
1403# We don't convert &lt; or &gt; or &amp; or &quot; in case
1404# they interfere with the GML files
1405
1406sub read_file {
1407 my ($self, $filename, $encoding, $language, $textref) = @_;
1408
1409 &BasPlug::read_file($self, $filename, $encoding, $language, $textref);
1410
1411 # Convert entities to their UTF8 equivalents
1412 $$textref =~ s/&(lt|gt|amp|quot|nbsp);/&z$1;/go;
1413 $$textref =~ s/&([^;]+);/&ghtml::getcharequiv($1,1)/gseo;
1414 $$textref =~ s/&z(lt|gt|amp|quot|nbsp);/&$1;/go;
1415}
1416
14171;
Note: See TracBrowser for help on using the repository browser.