root/gsdl/trunk/perllib/plugins/HTMLPlugin.pm @ 16632

Revision 16632, 47.4 KB (checked in by ak19, 12 years ago)

Work on supporting non-utf8 characters in filenames

  • Property svn:executable set to *
  • Property svn:keywords set to Author Date Id Revision
Line 
1###########################################################################
2#
3# HTMLPlugin.pm -- basic html plugin
4#
5# A component of the Greenstone digital library software
6# from the New Zealand Digital Library Project at the
7# University of Waikato, New Zealand.
8#
9# Copyright (C) 1999 New Zealand Digital Library Project
10#
11# This program is free software; you can redistribute it and/or modify
12# it under the terms of the GNU General Public License as published by
13# the Free Software Foundation; either version 2 of the License, or
14# (at your option) any later version.
15#
16# This program is distributed in the hope that it will be useful,
17# but WITHOUT ANY WARRANTY; without even the implied warranty of
18# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
19# GNU General Public License for more details.
20#
21# You should have received a copy of the GNU General Public License
22# along with this program; if not, write to the Free Software
23# Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
24#
25###########################################################################
26
27#
28# Note that this plugin handles frames only in a very simple way
29# i.e. each frame is treated as a separate document. This means
30# search results will contain links to individual frames rather
31# than linking to the top level frameset.
32# There may also be some problems caused by the _parent target
33# (it's removed by this plugin)
34#
35
36package HTMLPlugin;
37
38use ReadTextFile;
39use HBPlugin;
40use ghtml;
41use unicode;
42use util;
43use XMLParser;
44
45use Image::Size;
46use File::Copy;
47
48sub BEGIN {
49    @HTMLPlugin::ISA = ('ReadTextFile', 'HBPlugin');
50}
51
52use strict; # every perl program should have this!
53no strict 'refs'; # make an exception so we can use variables as filehandles
54
55my $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      ];
118
119my $options = { 'name'     => "HTMLPlugin",
120        'desc'     => "{HTMLPlugin.desc}",
121        'abstract' => "no",
122        'inherits' => "yes",
123        'args'     => $arguments };
124
125
126sub HB_read_html_file {
127    my $self = shift (@_);
128    my ($htmlfile, $text) = @_;
129   
130    # load in the file
131    if (!open (FILE, $htmlfile)) {
132    print STDERR "ERROR - could not open $htmlfile\n";
133    return;
134    }
135
136    my $foundbody = 0;
137    $self->HB_gettext (\$foundbody, $text, "FILE");
138    close FILE;
139   
140    # just in case there was no <body> tag
141    if (!$foundbody) {
142    $foundbody = 1;
143    open (FILE, $htmlfile) || return;
144    $self->HB_gettext (\$foundbody, $text, "FILE");
145    close FILE;
146    }
147    # text is in utf8
148}       
149
150# converts the text to utf8, as ghtml does that for &eacute; etc.
151sub HB_gettext {
152    my $self = shift (@_);
153    my ($foundbody, $text, $handle) = @_;
154   
155    my $line = "";
156    while (defined ($line = <$handle>)) {
157    # look for body tag
158    if (!$$foundbody) {
159        if ($line =~ s/^.*<body[^>]*>//i) {
160        $$foundbody = 1;
161        } else {
162        next;
163        }
164    }
165   
166    # check for symbol fonts
167    if ($line =~ /<font [^>]*?face\s*=\s*\"?(\w+)\"?/i) {
168        my $font = $1;
169        print STDERR "HBPlug::HB_gettext - warning removed font $font\n"
170        if ($font !~ /^arial$/i);
171    }
172
173    $$text .= $line;
174    }
175
176    if ($self->{'input_encoding'} eq "iso_8859_1") {
177    # convert to utf-8
178    $$text=&unicode::unicode2utf8(&unicode::convert2unicode("iso_8859_1", $text));
179    }
180    # convert any alphanumeric character entities to their utf-8
181    # equivalent for indexing purposes
182    #&ghtml::convertcharentities ($$text);
183
184    $$text =~ s/\s+/ /g; # remove \n's
185}
186
187sub HB_clean_section {
188    my $self = shift (@_);
189    my ($section) = @_;
190
191    # remove tags without a starting tag from the section
192    my ($tag, $tagstart);
193    while ($section =~ /<\/([^>]{1,10})>/) {
194    $tag = $1;
195    $tagstart = index($section, "<$tag");
196    last if (($tagstart >= 0) && ($tagstart < index($section, "<\/$tag")));
197    $section =~ s/<\/$tag>//;
198    }
199   
200    # remove extra paragraph tags
201    while ($section =~ s/<p\b[^>]*>\s*<p\b/<p/ig) {}
202   
203    # remove extra stuff at the end of the section
204    while ($section =~ s/(<u>|<i>|<b>|<p\b[^>]*>|&nbsp;|\s)$//i) {}
205   
206    # add a newline at the beginning of each paragraph
207    $section =~ s/(.)\s*<p\b/$1\n\n<p/gi;
208   
209    # add a newline every 80 characters at a word boundary
210    # Note: this regular expression puts a line feed before
211    # the last word in each section, even when it is not
212    # needed.
213    $section =~ s/(.{1,80})\s/$1\n/g;
214   
215    # fix up the image links
216    $section =~ s/<img[^>]*?src=\"?([^\">]+)\"?[^>]*>/
217    <center><img src=\"$1\" \/><\/center><br\/>/ig;
218    $section =~ s/&lt;&lt;I&gt;&gt;\s*([^\.]+\.(png|jpg|gif))/
219    <center><img src=\"$1\" \/><\/center><br\/>/ig;
220
221    return $section;
222}
223
224# Will convert the oldHDL format to the new HDL format (using the Section tag) 
225sub convert_to_newHDLformat
226{
227    my $self = shift (@_);
228    my ($file,$cnfile) = @_;
229    my $input_filename = $file;
230    my $tmp_filename = $cnfile;
231   
232    # write HTML tmp file with new HDL format
233    open (PROD, ">$tmp_filename") || die("Error Writing to File: $tmp_filename $!");
234   
235    # read in the file and do basic html cleaning (removing header etc)
236    my $html = "";
237    $self->HB_read_html_file ($input_filename, \$html);
238   
239    # process the file one section at a time
240    my $curtoclevel = 1;
241    my $firstsection = 1;
242    my $toclevel = 0;
243    while (length ($html) > 0) {
244    if ($html =~ s/^.*?(?:<p\b[^>]*>)?((<b>|<i>|<u>|\s)*)&lt;&lt;TOC(\d+)&gt;&gt;\s*(.*?)<p\b/<p/i) {
245        $toclevel = $3;
246        my $title = $4;
247        my $sectiontext = "";
248        if ($html =~ s/^(.*?)((?:<p\b[^>]*>)?((<b>|<i>|<u>|\s)*)&lt;&lt;TOC\d+&gt;&gt;)/$2/i) {
249        $sectiontext = $1;
250        } else {
251        $sectiontext = $html;
252        $html = "";
253        }
254
255        # remove tags and extra spaces from the title
256        $title =~ s/<\/?[^>]+>//g;
257        $title =~ s/^\s+|\s+$//g;
258
259        # close any sections below the current level and
260        # create a new section (special case for the firstsection)
261        print PROD "<!--\n";
262        while (($curtoclevel > $toclevel) ||
263           (!$firstsection && $curtoclevel == $toclevel)) {
264        $curtoclevel--;
265        print PROD "</Section>\n";
266        }
267        if ($curtoclevel+1 < $toclevel) {
268        print STDERR "WARNING - jump in toc levels in $input_filename " .
269            "from $curtoclevel to $toclevel\n";
270        }
271        while ($curtoclevel < $toclevel) {
272        $curtoclevel++;
273        }
274
275        if ($curtoclevel == 1) {
276            # add the header tag
277        print PROD "-->\n";
278            print PROD "<HTML>\n<HEAD>\n<TITLE>$title</TITLE>\n</HEAD>\n<BODY>\n";
279        print PROD "<!--\n";
280        }
281       
282        print PROD "<Section>\n\t<Description>\n\t\t<Metadata name=\"Title\">$title</Metadata>\n\t</Description>\n";
283       
284        print PROD "-->\n";
285       
286        # clean up the section html
287        $sectiontext = $self->HB_clean_section($sectiontext);
288
289        print PROD "$sectiontext\n";     
290
291    } else {
292        print STDERR "WARNING - leftover text\n" , $self->shorten($html),
293        "\nin $input_filename\n";
294        last;
295    }
296    $firstsection = 0;
297    }
298   
299    print PROD "<!--\n";
300    while ($curtoclevel > 0) {
301    $curtoclevel--;
302    print PROD "</Section>\n";
303    }
304    print PROD "-->\n";
305   
306    close (PROD) || die("Error Closing File: $tmp_filename $!");
307   
308    return $tmp_filename;
309}       
310
311sub shorten {
312    my $self = shift (@_);
313    my ($text) = @_;
314
315    return "\"$text\"" if (length($text) < 100);
316
317    return "\"" . substr ($text, 0, 50) . "\" ... \"" .
318    substr ($text, length($text)-50) . "\"";
319}
320
321sub convert_tidy_or_oldHDL_file
322{
323    my $self = shift (@_);
324    my ($file) = @_;
325    my $input_filename = $file;
326   
327    if (-d $input_filename)
328    {
329        return $input_filename;
330    }
331   
332    # get the input filename
333    my ($tailname, $dirname, $suffix) = &File::Basename::fileparse($input_filename, "\\.[^\\.]+\$");
334    my $base_dirname = $dirname;
335    $suffix = lc($suffix);   
336   
337    # derive tmp filename from input filename
338    # Remove any white space from filename -- no risk of name collision, and
339    # makes later conversion by utils simpler. Leave spaces in path...
340    # tidy up the filename with space, dot, hyphen between
341    $tailname =~ s/\s+//g;
342    $tailname =~ s/\.+//g;
343    $tailname =~ s/\-+//g;
344    # convert to utf-8 otherwise we have problems with the doc.xml file
345    # later on
346    &unicode::ensure_utf8(\$tailname);
347   
348    # softlink to collection tmp dir
349    my $tmp_dirname = &util::filename_cat($ENV{'GSDLCOLLECTDIR'}, "tidytmp");
350    &util::mk_dir($tmp_dirname) if (!-e $tmp_dirname);
351   
352    my $test_dirname = "";
353    my $f_separator = &util::get_os_dirsep();
354   
355    if ($dirname =~ /import$f_separator/)
356    {
357        $test_dirname = $'; #'
358   
359    #print STDERR "init $'\n";
360   
361    while ($test_dirname =~ /[$f_separator]/)
362    {
363        my $folderdirname = $`;
364        $tmp_dirname = &util::filename_cat($tmp_dirname,$folderdirname);
365        &util::mk_dir($tmp_dirname) if (!-e $tmp_dirname);
366        $test_dirname = $'; #'
367    }
368    }
369   
370    my $tmp_filename = &util::filename_cat($tmp_dirname, "$tailname$suffix");
371   
372    # tidy or convert the input file if it is a HTML-like file or it is accepted by the process_exp
373    if (($suffix eq ".htm") || ($suffix eq ".html") || ($suffix eq ".shtml"))
374    {   
375        #convert the input file to a new style HDL
376        my $hdl_output_filename = $input_filename;
377        if ($self->{'old_style_HDL'})
378        {
379        $hdl_output_filename = &util::filename_cat($tmp_dirname, "$tailname$suffix");
380        $hdl_output_filename = $self->convert_to_newHDLformat($input_filename,$hdl_output_filename);
381        }
382   
383    #just for checking copy all other file from the base dir to tmp dir if it is not exists
384    opendir(DIR,$base_dirname) or die "Can't open base directory : $base_dirname!";
385    my @files = grep {!/^\.+$/} readdir(DIR);
386    close(DIR);
387
388    foreach my $file (@files)
389    {
390        my $src_file = &util::filename_cat($base_dirname,$file);
391        my $dest_file = &util::filename_cat($tmp_dirname,$file);
392        if ((!-e $dest_file) && (!-d $src_file))
393        {
394        # just copy the original file back to the tmp directory
395        copy($src_file,$dest_file) or die "Can't copy file $src_file to $dest_file $!";
396        }
397    }
398   
399    # tidy the input file
400    my $tidy_output_filename = $hdl_output_filename;
401    if ($self->{'tidy_html'})
402    {
403        $tidy_output_filename = &util::filename_cat($tmp_dirname, "$tailname$suffix");
404        $tidy_output_filename = $self->tmp_tidy_file($hdl_output_filename,$tidy_output_filename);
405    }
406    $tmp_filename = $tidy_output_filename;
407    }
408    else
409    {
410        if (!-e $tmp_filename)
411    {
412        # just copy the original file back to the tmp directory
413        copy($input_filename,$tmp_filename) or die "Can't copy file $input_filename to $tmp_filename $!";
414    }
415    }
416   
417    return $tmp_filename;
418}
419
420
421# Will make the html input file as a proper XML file with removed font tag and
422# image size added to the img tag.
423# The tidying process takes place in a collection specific 'tmp' directory so
424# that we don't accidentally damage the input.
425sub tmp_tidy_file
426{
427    my $self = shift (@_);
428    my ($file,$cnfile) = @_;
429    my $input_filename = $file;
430    my $tmp_filename = $cnfile;
431   
432    # get the input filename
433    my ($tailname, $dirname, $suffix) = &File::Basename::fileparse($input_filename, "\\.[^\\.]+\$");
434
435    require HTML::TokeParser::Simple;
436   
437    # create HTML parser to decode the input file
438    my $parser = HTML::TokeParser::Simple->new($input_filename);
439
440    # write HTML tmp file without the font tag and image size are added to the img tag
441    open (PROD, ">$tmp_filename") || die("Error Writing to File: $tmp_filename $!");
442    while (my $token = $parser->get_token())
443    {
444    # is it an img tag
445    if ($token->is_start_tag('img'))
446    {
447        # get the attributes
448        my $attr = $token->return_attr;
449
450        # get the full path to the image
451        my $img_file = &util::filename_cat($dirname,$attr->{src});
452       
453        # set the width and height attribute
454        ($attr->{width}, $attr->{height}) = imgsize($img_file);
455
456        # recreate the tag
457        print PROD "<img";
458        print PROD map { qq { $_="$attr->{$_}"} } keys %$attr;
459        print PROD ">";
460    }
461    # is it a font tag
462    else
463    {
464        if (($token->is_start_tag('font')) || ($token->is_end_tag('font')))
465        {
466        # remove font tag
467        print PROD "";
468        }
469        else
470        {
471        # print without changes
472        print PROD $token->as_is;
473        }
474    }
475    }
476    close (PROD) || die("Error Closing File: $tmp_filename $!");
477
478    # run html-tidy on the tmp file to make it a proper XML file
479    my $tidyfile = `tidy -utf8 -wrap 0 -asxml $tmp_filename`;
480
481    # write result back to the tmp file
482    open (PROD, ">$tmp_filename") || die("Error Writing to File: $tmp_filename $!");
483    print PROD $tidyfile;
484    close (PROD) || die("Error Closing File: $tmp_filename $!");
485   
486    # return the output filename
487    return $tmp_filename;
488}
489
490sub read_into_doc_obj
491{
492    my $self = shift (@_); 
493    my ($pluginfo, $base_dir, $file, $block_hash, $metadata, $processor, $maxdocs, $total_count, $gli) = @_;
494       
495    # get the input file
496    my $input_filename = $file;
497    my ($tailname, $dirname, $suffix) = &File::Basename::fileparse($input_filename, "\\.[^\\.]+\$");
498    $suffix = lc($suffix);   
499   
500    if (($self->{'tidy_html'}) || ($self->{'old_style_HDL'}))
501    {
502    # because the document has to be sectionalized set the description tags
503    $self->{'description_tags'} = 1;
504   
505    # set the file to be tidied
506    $input_filename = &util::filename_cat($base_dir,$file) if $base_dir =~ /\w/;
507   
508    # get the tidied file
509    #my $tidy_filename = $self->tmp_tidy_file($input_filename);
510    my $tidy_filename = $self->convert_tidy_or_oldHDL_file($input_filename);
511   
512    # derive tmp filename from input filename
513    my ($tailname, $dirname, $suffix) = &File::Basename::fileparse($tidy_filename, "\\.[^\\.]+\$");
514   
515    # set the new input file and base_dir to be from the tidied file
516    $file = "$tailname$suffix";
517    $base_dir = $dirname;
518    }
519   
520    # call the parent read_into_doc_obj
521    my ($process_status,$doc_obj) = $self->SUPER::read_into_doc_obj($pluginfo, $base_dir, $file, $block_hash, $metadata, $processor, $maxdocs, $total_count, $gli);
522   
523    return ($process_status,$doc_obj);
524}
525
526sub new {
527    my ($class) = shift (@_);
528    my ($pluginlist,$inputargs,$hashArgOptLists) = @_;
529    push(@$pluginlist, $class);
530   
531    push(@{$hashArgOptLists->{"ArgList"}},@{$arguments});
532    push(@{$hashArgOptLists->{"OptList"}},$options);
533   
534
535    my $self = new ReadTextFile($pluginlist,$inputargs,$hashArgOptLists);
536   
537    if ($self->{'w3mir'}) {
538    $self->{'file_is_url'} = 1;
539    }
540    $self->{'aux_files'} = {};
541    $self->{'dir_num'} = 0;
542    $self->{'file_num'} = 0;
543   
544    return bless $self, $class;
545}
546
547# may want to use (?i)\.(gif|jpe?g|jpe|png|css|js(?:@.*)?)$
548# if have eg <script language="javascript" src="img/lib.js@123">
549sub get_default_block_exp {
550    my $self = shift (@_);
551   
552    #return q^(?i)\.(gif|jpe?g|jpe|jpg|png|css)$^;
553    return "";
554}
555
556sub get_default_process_exp {
557    my $self = shift (@_);
558   
559    # the last option is an attempt to encode the concept of an html query ...
560    return q^(?i)(\.html?|\.shtml|\.shm|\.asp|\.php\d?|\.cgi|.+\?.+=.*)$^;
561}
562
563sub store_block_files
564{
565    my $self =shift (@_);
566    my ($filename_full_path, $block_hash) = @_;
567
568    my $html_fname = $filename_full_path;
569    my @file_blocks;
570   
571    my ($language, $encoding) = $self->textcat_get_language_encoding ($filename_full_path);
572
573    # read in file ($text will be in utf8)
574    my $text = "";
575    $self->read_file ($filename_full_path, $encoding, $language, \$text);
576    my $textref = \$text;
577    my $opencom = '(?:<!--|&lt;!(?:&mdash;|&#151;|--))';
578    my $closecom = '(?:-->|(?:&mdash;|&#151;|--)&gt;)';
579    $$textref =~ s/$opencom(.*?)$closecom//gs;
580
581    my $attval = "\\\"[^\\\"]+\\\"|[^\\s>]+";
582    my @img_matches = ($$textref =~ m/<img[^>]*?src\s*=\s*($attval)[^>]*>/igs);
583    my @usemap_matches = ($$textref =~ m/<img[^>]*?usemap\s*=\s*($attval)[^>]*>/igs);
584    my @link_matches = ($$textref =~ m/<link[^>]*?href\s*=\s*($attval)[^>]*>/igs);
585    my @embed_matches = ($$textref =~ m/<embed[^>]*?src\s*=\s*($attval)[^>]*>/igs);
586    my @tabbg_matches = ($$textref =~ m/<(?:table|tr|td)[^>]*?background\s*=\s*($attval)[^>]*>/igs);
587
588    foreach my $link (@img_matches, @usemap_matches, @link_matches, @embed_matches, @tabbg_matches) {
589
590    # remove quotes from link at start and end if necessary
591    if ($link=~/^\"/) {
592        $link=~s/^\"//;
593        $link=~s/\"$//;
594    }
595
596    $link =~ s/\#.*$//s; # remove any anchor names, e.g. foo.html#name becomes foo.html
597
598    if ($link !~ m@^/@ && $link !~ m/^([A-Z]:?)\\/) {
599        # Turn relative file path into full path
600        my $dirname = &File::Basename::dirname($filename_full_path);
601        $link = &util::filename_cat($dirname, $link);
602    }
603    $link = $self->eval_dir_dots($link);
604   
605    $block_hash->{'file_blocks'}->{$link} = 1;
606    }
607}
608
609
610# do plugin specific processing of doc_obj
611sub process {
612    my $self = shift (@_);
613    my ($textref, $pluginfo, $base_dir, $file, $metadata, $doc_obj, $gli) = @_;
614    my $outhandle = $self->{'outhandle'};
615
616    if ($ENV{'GSDLOS'} =~ /^windows/i) {
617    # this makes life so much easier... perl can cope with unix-style '/'s.
618    $base_dir =~ s@(\\)+@/@g;
619    $file =~ s@(\\)+@/@g;
620    }
621   
622    # reset per-doc stuff...
623    $self->{'aux_files'} = {};
624    $self->{'dir_num'} = 0;
625    $self->{'file_num'} = 0;
626
627    # process an HTML file where sections are divided by headings tags (H1, H2 ...)
628    # you can also include metadata in the format (X can be any number)
629    # <hX>Title<!--gsdl-metadata
630    #   <Metadata name="name1">value1</Metadata>
631    #   ...
632    #   <Metadata name="nameN">valueN</Metadata>
633    #--></hX>
634    if ($self->{'sectionalise_using_h_tags'}) {
635    # description_tags should allways be activated because we convert headings to description tags
636    $self->{'description_tags'} = 1;
637
638    my $arrSections = [];
639    $$textref =~ s/<h([0-9]+)[^>]*>(.*?)<\/h[0-9]+>/$self->process_heading($1, $2, $arrSections, $file)/isge;
640
641    if (scalar(@$arrSections)) {
642        my $strMetadata = $self->update_section_data($arrSections, -1);
643        if (length($strMetadata)) {
644        $strMetadata = '<!--' . $strMetadata . "\n-->\n</body>";
645        $$textref =~ s/<\/body>/$strMetadata/ig;
646        }
647    }
648    }
649
650    my $cursection = $doc_obj->get_top_section();
651
652    $self->extract_metadata ($textref, $metadata, $doc_obj, $cursection)
653    unless $self->{'no_metadata'} || $self->{'description_tags'};
654
655    # Store URL for page as metadata - this can be used for an
656    # altavista style search interface. The URL won't be valid
657    # unless the file structure contains the domain name (i.e.
658    # like when w3mir is used to download a website).
659
660    # URL metadata (even invalid ones) are used to support internal
661    # links, so even if 'file_is_url' is off, still need to store info
662
663    my $utf8_file = $self->filename_to_utf8_metadata($file);
664    my $web_url = "http://$utf8_file";
665    $doc_obj->add_utf8_metadata($cursection, "URL", $web_url);
666
667    if ($self->{'file_is_url'}) {
668    $doc_obj->add_metadata($cursection, "weblink", "<a href=\"$web_url\">");
669    $doc_obj->add_metadata($cursection, "webicon", "_iconworld_");
670    $doc_obj->add_metadata($cursection, "/weblink", "</a>");
671    }
672
673    if ($self->{'description_tags'}) {
674    # remove the html header - note that doing this here means any
675    # sections defined within the header will be lost (so all <Section>
676    # tags must appear within the body of the HTML)
677    my ($head_keep) = ($$textref =~ m/^(.*?)<body[^>]*>/is);
678
679    $$textref =~ s/^.*?<body[^>]*>//is;
680    $$textref =~ s/(<\/body[^>]*>|<\/html[^>]*>)//isg;
681
682    my $opencom = '(?:<!--|&lt;!(?:&mdash;|&#151;|--))';
683    my $closecom = '(?:-->|(?:&mdash;|&#151;|--)&gt;)';
684
685    my $lt = '(?:<|&lt;)';
686    my $gt = '(?:>|&gt;)';
687    my $quot = '(?:"|&quot;|&rdquo;|&ldquo;)';
688
689    my $dont_strip = '';
690    if ($self->{'no_strip_metadata_html'}) {
691        ($dont_strip = $self->{'no_strip_metadata_html'}) =~ s{,}{|}g;
692    }
693
694    my $found_something = 0; my $top = 1;
695    while ($$textref =~ s/^(.*?)$opencom(.*?)$closecom//s) {
696        my $text = $1;
697        my $comment = $2;
698        if (defined $text) {
699        # text before a comment - note that getting to here
700        # doesn't necessarily mean there are Section tags in
701        # the document
702        $self->process_section(\$text, $base_dir, $file, $doc_obj, $cursection);
703        }
704        while ($comment =~ s/$lt(.*?)$gt//s) {
705        my $tag = $1;
706        if ($tag eq "Section") {
707            $found_something = 1;
708            $cursection = $doc_obj->insert_section($doc_obj->get_end_child($cursection)) unless $top;
709            $top = 0;
710        } elsif ($tag eq "/Section") {
711            $found_something = 1;
712            $cursection = $doc_obj->get_parent_section ($cursection);
713        } elsif ($tag =~ /^Metadata name=$quot(.*?)$quot/s) {
714            my $metaname = $1;
715            my $accumulate = $tag =~ /mode=${quot}accumulate${quot}/ ? 1 : 0;
716            $comment =~ s/^(.*?)$lt\/Metadata$gt//s;
717            my $metavalue = $1;
718            $metavalue =~ s/^\s+//;
719            $metavalue =~ s/\s+$//;
720                    # assume that no metadata value intentionally includes
721                    # carriage returns or HTML tags (if they're there they
722                    # were probably introduced when converting to HTML from
723                    # some other format).
724            # actually some people want to have html tags in their
725            # metadata.
726            $metavalue =~ s/[\cJ\cM]/ /sg;
727            $metavalue =~ s/<[^>]+>//sg
728            unless $dont_strip && ($dont_strip eq 'all' || $metaname =~ /^($dont_strip)$/);
729            $metavalue =~ s/\s+/ /sg;
730            if ($accumulate) {
731            $doc_obj->add_utf8_metadata($cursection, $metaname, $metavalue);
732            } else {
733            $doc_obj->set_utf8_metadata_element($cursection, $metaname, $metavalue);   
734            }
735        } elsif ($tag eq "Description" || $tag eq "/Description") {
736            # do nothing with containing Description tags
737        } else {
738            # simple HTML tag (probably created by the conversion
739            # to HTML from some other format) - we'll ignore it and
740            # hope for the best ;-)
741        }
742        }
743    }
744    if ($cursection ne "") {
745        print $outhandle "HTMLPlugin: WARNING: $file contains unmatched <Section></Section> tags\n";
746    }
747
748    $$textref =~ s/^.*?<body[^>]*>//is;
749    $$textref =~ s/(<\/body[^>]*>|<\/html[^>]*>)//isg;
750    if ($$textref =~ /\S/) {
751        if (!$found_something) {
752        if ($self->{'verbosity'} > 2) {
753            print $outhandle "HTMLPlugin: WARNING: $file appears to contain no Section tags so\n";
754            print $outhandle "          will be processed as a single section document\n";
755        }
756
757        # go ahead and process single-section document
758        $self->process_section($textref, $base_dir, $file, $doc_obj, $cursection);
759
760        # if document contains no Section tags we'll go ahead
761        # and extract metadata (this won't have been done
762        # above as the -description_tags option prevents it)
763        my $complete_text = $head_keep.$doc_obj->get_text($cursection);
764        $self->extract_metadata (\$complete_text, $metadata, $doc_obj, $cursection)
765            unless $self->{'no_metadata'};
766
767        } else {
768        print $outhandle "HTMLPlugin: WARNING: $file contains the following text outside\n";
769        print $outhandle "          of the final closing </Section> tag. This text will\n";
770        print $outhandle "          be ignored.";
771
772        my ($text);
773        if (length($$textref) > 30) {
774            $text = substr($$textref, 0, 30) . "...";
775        } else {
776            $text = $$textref;
777        }
778        $text =~ s/\n/ /isg;
779        print $outhandle " ($text)\n";
780        }
781    } elsif (!$found_something) {
782
783        if ($self->{'verbosity'} > 2) {
784        # may get to here if document contained no valid Section
785        # tags but did contain some comments. The text will have
786        # been processed already but we should print the warning
787        # as above and extract metadata
788        print $outhandle "HTMLPlugin: WARNING: $file appears to contain no Section tags and\n";
789        print $outhandle "          is blank or empty.  Metadata will be assigned if present.\n";
790        }
791
792        my $complete_text = $head_keep.$doc_obj->get_text($cursection);
793        $self->extract_metadata (\$complete_text, $metadata, $doc_obj, $cursection)
794        unless $self->{'no_metadata'};
795    }
796   
797    } else {
798
799    # remove header and footer
800    if (!$self->{'keep_head'} || $self->{'description_tags'}) {
801        $$textref =~ s/^.*?<body[^>]*>//is;
802        $$textref =~ s/(<\/body[^>]*>|<\/html[^>]*>)//isg;
803    }
804
805    # single section document
806    $self->process_section($textref, $base_dir, $file, $doc_obj, $cursection);
807    }
808    return 1;
809}
810
811
812sub process_heading
813{
814    my ($self, $nHeadNo, $strHeadingText, $arrSections, $file) = @_;
815    $strHeadingText = '' if (!defined($strHeadingText));
816
817    my $strMetadata = $self->update_section_data($arrSections, int($nHeadNo));
818
819    my $strSecMetadata = '';
820    while ($strHeadingText =~ s/<!--gsdl-metadata(.*?)-->//is)
821    {
822    $strSecMetadata .= $1;
823    }
824
825    $strHeadingText =~ s/^\s+//g;
826    $strHeadingText =~ s/\s+$//g;
827    $strSecMetadata =~ s/^\s+//g;
828    $strSecMetadata =~ s/\s+$//g;
829
830    $strMetadata .= "\n<Section>\n\t<Description>\n\t\t<Metadata name=\"Title\">" . $strHeadingText . "</Metadata>\n";
831
832    if (length($strSecMetadata)) {
833    $strMetadata .= "\t\t" . $strSecMetadata . "\n";
834    }
835
836    $strMetadata .= "\t</Description>\n";
837
838    return "<!--" . $strMetadata . "-->";
839}
840
841
842sub update_section_data
843{
844    my ($self, $arrSections, $nCurTocNo) = @_;
845    my ($strBuffer, $nLast, $nSections) = ('', 0, scalar(@$arrSections));
846
847    if ($nSections == 0) {
848    push @$arrSections, $nCurTocNo;
849    return $strBuffer;
850    }
851    $nLast = $arrSections->[$nSections - 1];
852    if ($nCurTocNo > $nLast) {
853    push @$arrSections, $nCurTocNo;
854    return $strBuffer;
855    }
856    for(my $i = $nSections - 1; $i >= 0; $i--) {
857    if ($nCurTocNo <= $arrSections->[$i]) {
858        $strBuffer .= "\n</Section>";
859        pop @$arrSections;
860    }
861    }
862    push @$arrSections, $nCurTocNo;
863    return $strBuffer;
864}
865
866
867# note that process_section may be called multiple times for a single
868# section (relying on the fact that add_utf8_text appends the text to any
869# that may exist already).
870sub process_section {
871    my $self = shift (@_);
872    my ($textref, $base_dir, $file, $doc_obj, $cursection) = @_;
873    # trap links
874    if (!$self->{'nolinks'}) {
875    # usemap="./#index" not handled correctly => change to "#index"
876    $$textref =~ s/(<img[^>]*?usemap\s*=\s*[\"\']?)([^\"\'>\s]+)([\"\']?[^>]*>)/
877        $self->replace_usemap_links($1, $2, $3)/isge;
878
879    $$textref =~ s/(<(?:a|area|frame|link|script)\s+[^>]*?\s*(?:href|src)\s*=\s*[\"\']?)([^\"\'>\s]+)([\"\']?[^>]*>)/
880        $self->replace_href_links ($1, $2, $3, $base_dir, $file, $doc_obj, $cursection)/isge;
881    }
882
883    # trap images
884
885    # Previously, by default, HTMLPlugin would embed <img> tags inside anchor tags
886    # i.e. <a href="image><img src="image"></a> in order to overcome a problem that
887    # turned regular text succeeding images into links. That is, by embedding <imgs>
888    # inside <a href=""></a>, the text following images were no longer misbehaving.
889    # However, there would be many occasions whereby images were not meant to link
890    # to their source images but where the images would link to another web page.
891    # To allow this, the no_image_links option was introduced: it would prevent
892    # the behaviour of embedding images into links that referenced the source images.
893
894    # Somewhere along the line, the problem of normal text turning into links when
895    # such text followed images which were not embedded in <a href=""></a> ceased
896    # to occur. This is why the following lines have been commented out (as well as
897    # two lines in replace_images). They appear to no longer apply.
898
899    # If at any time, there is a need for having images embedded in <a> anchor tags,
900    # then it might be better to turn that into an HTMLPlugin option rather than make
901    # it the default behaviour. Also, eventually, no_image_links needs to become
902    # a deprecated option for HTMLPlugin as it has now become the default behaviour.
903
904    #if(!$self->{'no_image_links'}){
905    $$textref =~ s/(<(?:img|embed|table|tr|td)[^>]*?(?:src|background)\s*=\s*)((?:[\"][^\"]+[\"])|(?:[\'][^\']+[\'])|(?:[^\s\/>]+))([^>]*>)/
906    $self->replace_images ($1, $2, $3, $base_dir, $file, $doc_obj, $cursection)/isge;
907    #}
908
909    # add text to document object
910    # turn \ into \\ so that the rest of greenstone doesn't think there
911    # is an escape code following. (Macro parsing loses them...)
912    $$textref =~ s/\\/\\\\/go;
913   
914    $doc_obj->add_utf8_text($cursection, $$textref);
915}
916
917sub replace_images {
918    my $self = shift (@_);
919    my ($front, $link, $back, $base_dir,
920    $file, $doc_obj, $section) = @_;
921
922    # remove quotes from link at start and end if necessary
923    if ($link=~/^[\"\']/) {
924    $link=~s/^[\"\']//;
925    $link=~s/[\"\']$//;
926    $front.='"';
927    $back="\"$back";
928    }
929
930    $link =~ s/\n/ /g;
931
932    # Hack to overcome Windows wv 0.7.1 bug that causes embedded images to be broken
933    # If the Word file path has spaces in it, wv messes up and you end up with
934    #   absolute paths for the images, and without the "file://" prefix
935    # So check for this special case and massage the data to be correct
936    if ($ENV{'GSDLOS'} =~ /^windows/i && $self->{'plugin_type'} eq "WordPlug" && $link =~ /^[A-Za-z]\:\\/) {
937    $link =~ s/^.*\\([^\\]+)$/$1/;
938    }
939   
940    my ($href, $hash_part, $rl) = $self->format_link ($link, $base_dir, $file);
941
942    my $img_file =  $self->add_file ($href, $rl, $hash_part, $base_dir, $doc_obj, $section);
943
944    print STDERR "**** link = $link\n";
945    print STDERR "**** href = $href\n";
946    print STDERR "**** img_file = $img_file\n";
947
948    my $anchor_name = $img_file;
949    #$anchor_name =~ s/^.*\///;
950    #$anchor_name = "<a name=\"$anchor_name\" ></a>";
951
952    my $image_link = $front . $img_file .$back;
953    return $image_link;
954
955    # The reasons for why the following two lines are no longer necessary can be
956    # found in subroutine process_section
957    #my $anchor_link = "<a href=\"$img_file\" >".$image_link."</a>"; 
958    #return $anchor_link;   
959   
960    #return $front . $img_file . $back . $anchor_name;
961}
962
963sub replace_href_links {
964    my $self = shift (@_);
965    my ($front, $link, $back, $base_dir, $file, $doc_obj, $section) = @_;
966
967    # attempt to sort out targets - frames are not handled
968    # well in this plugin and some cases will screw things
969    # up - e.g. the _parent target (so we'll just remove
970    # them all ;-)
971    $front =~ s/(target=\"?)_top(\"?)/$1_gsdltop_$2/is;
972    $back =~ s/(target=\"?)_top(\"?)/$1_gsdltop_$2/is;
973    $front =~ s/target=\"?_parent\"?//is;
974    $back =~ s/target=\"?_parent\"?//is;
975
976    return $front . $link . $back if $link =~ /^\#/s;
977    $link =~ s/\n/ /g;
978
979    my ($href, $hash_part, $rl) = $self->format_link ($link, $base_dir, $file);
980    # href may use '\'s where '/'s should be on Windows
981    $href =~ s/\\/\//g;
982
983    my ($filename) = $href =~ /^(?:.*?):(?:\/\/)?(.*)/;
984   
985
986    ##### leave all these links alone (they won't be picked up by intermediate
987    ##### pages). I think that's safest when dealing with frames, targets etc.
988    ##### (at least until I think of a better way to do it). Problems occur with
989    ##### mailto links from within small frames, the intermediate page is displayed
990    ##### within that frame and can't be seen. There is still potential for this to
991    ##### happen even with html pages - the solution seems to be to somehow tell
992    ##### the browser from the server side to display the page being sent (i.e.
993    ##### the intermediate page) in the top level window - I'm not sure if that's
994    ##### possible - the following line should probably be deleted if that can be done
995    return $front . $link . $back if $href =~ /^(mailto|news|gopher|nntp|telnet|javascript):/is;
996
997
998    if (($rl == 0) || ($filename =~ /$self->{'process_exp'}/) ||
999    ($href =~ /\/$/) || ($href =~ /^(mailto|news|gopher|nntp|telnet|javascript):/i)) {
1000    &ghtml::urlsafe ($href);
1001    return $front . "_httpextlink_&rl=" . $rl . "&href=" . $href . $hash_part . $back;
1002    } else {
1003    # link is to some other type of file (eg image) so we'll
1004    # need to associate that file
1005    return $front . $self->add_file ($href, $rl, $hash_part, $base_dir, $doc_obj, $section) . $back;
1006    }
1007}
1008
1009sub add_file {
1010    my $self = shift (@_);
1011    my ($href, $rl, $hash_part, $base_dir, $doc_obj, $section) = @_;
1012    my ($newname);
1013
1014    my $filename = $href;
1015    if ($base_dir eq "") {
1016    # remove http:/ thereby leaving one slash at the start
1017    $filename =~ s/^[^:]*:\///;
1018    }
1019    else {
1020    # remove http://
1021    $filename =~ s/^[^:]*:\/\///;
1022    }
1023
1024    $filename = &util::filename_cat($base_dir, $filename);
1025   
1026    # Replace %XX's in URL with decoded value if required.
1027    # Note that the filename may include the %XX in some situations
1028    if ($filename =~ m/\%[A-F0-9]{2}/i) {
1029    if (!-e $filename) {
1030        $filename =~ s/\%([A-F0-9]{2})/pack('C', hex($1))/ige;
1031    }
1032    }
1033
1034    my ($ext) = $filename =~ /(\.[^\.]*)$/;
1035
1036    if ($rl == 0) {
1037    if ((!defined $ext) || ($ext !~ /$self->{'assoc_files'}/)) {
1038        return "_httpextlink_&rl=0&el=prompt&href=" . $href . $hash_part;
1039    }
1040    else {
1041        return "_httpextlink_&rl=0&el=direct&href=" . $href . $hash_part;
1042    }
1043    }
1044
1045    if ((!defined $ext) || ($ext !~ /$self->{'assoc_files'}/)) {
1046    return "_httpextlink_&rl=" . $rl . "&href=" . $href . $hash_part;
1047    }
1048    if ($self->{'rename_assoc_files'}) {
1049    if (defined $self->{'aux_files'}->{$href}) {
1050        $newname = $self->{'aux_files'}->{$href}->{'dir_num'} . "/" .
1051        $self->{'aux_files'}->{$href}->{'file_num'} . $ext;
1052    } else {
1053        $newname = $self->{'dir_num'} . "/" . $self->{'file_num'} . $ext;
1054        $self->{'aux_files'}->{$href} = {'dir_num' => $self->{'dir_num'}, 'file_num' => $self->{'file_num'}};
1055        $self->inc_filecount ();
1056    }
1057    $doc_obj->associate_file($filename, $newname, undef, $section);
1058    return "_httpdocimg_/$newname";
1059    } else {
1060    ($newname) = $filename =~ /([^\/\\]*)$/;
1061    # Make sure this name is a valid utf8 filename
1062    ## &unicode::ensure_utf8(\$newname);
1063    $newname =~ s/([^A-Za-z0-9])/sprintf("%%%02X", ord($1))/seg;
1064
1065    $doc_obj->associate_file($filename, $newname, undef, $section);
1066
1067    my $newname_url = $newname;
1068    $newname_url =~ s/%/%25/g;
1069    return "_httpdocimg_/$newname_url";
1070    }
1071}
1072
1073
1074
1075sub format_link {
1076    my $self = shift (@_);
1077    my ($link, $base_dir, $file) = @_;
1078
1079    my ($before_hash, $hash_part) = $link =~ /^([^\#]*)(\#?.*)$/;
1080   
1081    $hash_part = "" if !defined $hash_part;
1082    if (!defined $before_hash || $before_hash !~ /[\w\.\/]/) {
1083    my $outhandle = $self->{'outhandle'};
1084    print $outhandle "HTMLPlugin: ERROR - badly formatted tag ignored ($link)\n"
1085        if $self->{'verbosity'};
1086    return ($link, "", 0);
1087    }
1088
1089    if ($before_hash =~ s@^((?:http|ftp|file)://)@@i) {
1090    my $type = $1;
1091
1092    if ($link =~ /^(http|ftp):/i) {
1093        # Turn url (using /) into file name (possibly using \ on windows)
1094        my @http_dir_split = split('/', $before_hash);
1095        $before_hash = &util::filename_cat(@http_dir_split);
1096    }
1097
1098    $before_hash = $self->eval_dir_dots($before_hash);
1099   
1100    my $linkfilename = &util::filename_cat ($base_dir, $before_hash);
1101   
1102    my $rl = 0;
1103    $rl = 1 if (-e $linkfilename);
1104
1105    # make sure there's a slash on the end if it's a directory
1106    if ($before_hash !~ /\/$/) {
1107        $before_hash .= "/" if (-d $linkfilename);
1108    }
1109
1110    return ($type . $before_hash, $hash_part, $rl);
1111   
1112    } elsif ($link !~ /^(mailto|news|gopher|nntp|telnet|javascript):/i && $link !~ /^\//) {
1113    if ($before_hash =~ s@^/@@ || $before_hash =~ /\\/) {
1114
1115        # the first directory will be the domain name if file_is_url
1116        # to generate archives, otherwise we'll assume all files are
1117        # from the same site and base_dir is the root
1118
1119        if ($self->{'file_is_url'}) {
1120        my @dirs = split /[\/\\]/, $file;
1121        my $domname = shift (@dirs);
1122        $before_hash = &util::filename_cat($domname, $before_hash);
1123        $before_hash =~ s@\\@/@g; # for windows
1124        }
1125        else
1126        {
1127        # see if link shares directory with source document
1128        # => turn into relative link if this is so!
1129       
1130        if ($ENV{'GSDLOS'} =~ /^windows/i) {
1131            # too difficult doing a pattern match with embedded '\'s...
1132            my $win_before_hash=$before_hash;
1133            $win_before_hash =~ s@(\\)+@/@g;
1134            # $base_dir is already similarly "converted" on windows.
1135            if ($win_before_hash =~ s@^$base_dir/@@o) {
1136            # if this is true, we removed a prefix
1137            $before_hash=$win_before_hash;
1138            }
1139        }
1140        else {
1141            # before_hash has lost leading slash by this point,
1142            # -> add back in prior to substitution with $base_dir
1143            $before_hash = "/$before_hash";
1144
1145            $before_hash = &util::filename_cat("",$before_hash);
1146            $before_hash =~ s@^$base_dir/@@;
1147        }
1148        }
1149    } else {
1150        # Turn relative file path into full path
1151        my $dirname = &File::Basename::dirname($file);
1152        $before_hash = &util::filename_cat($dirname, $before_hash);
1153        $before_hash = $self->eval_dir_dots($before_hash);
1154    }
1155
1156    my $linkfilename = &util::filename_cat ($base_dir, $before_hash);
1157    # make sure there's a slash on the end if it's a directory
1158    if ($before_hash !~ /\/$/) {
1159        $before_hash .= "/" if (-d $linkfilename);
1160    }
1161    return ("http://" . $before_hash, $hash_part, 1);
1162    } else {
1163    # mailto, news, nntp, telnet, javascript or gopher link
1164    return ($before_hash, "", 0);
1165    }
1166}
1167
1168sub extract_first_NNNN_characters {
1169    my $self = shift (@_);
1170    my ($textref, $doc_obj, $thissection) = @_;
1171   
1172    foreach my $size (split /,/, $self->{'first'}) {
1173    my $tmptext =  $$textref;
1174    # skip to the body
1175    $tmptext =~ s/.*<body[^>]*>//i;
1176    # remove javascript
1177    $tmptext =~ s@<script.*?</script>@ @sig;
1178    $tmptext =~ s/<[^>]*>/ /g;
1179    $tmptext =~ s/&nbsp;/ /g;
1180    $tmptext =~ s/^\s+//;
1181    $tmptext =~ s/\s+$//;
1182    $tmptext =~ s/\s+/ /gs;
1183    $tmptext = &unicode::substr ($tmptext, 0, $size);
1184    $tmptext =~ s/\s\S*$/&#8230;/; # adds an ellipse (...)
1185    $doc_obj->add_utf8_metadata ($thissection, "First$size", $tmptext);
1186    }
1187}
1188
1189
1190sub extract_metadata {
1191    my $self = shift (@_);
1192    my ($textref, $metadata, $doc_obj, $section) = @_;
1193    my $outhandle = $self->{'outhandle'};
1194    # if we don't want metadata, we may as well not be here ...
1195    return if (!defined $self->{'metadata_fields'});
1196
1197    # metadata fields to extract/save. 'key' is the (lowercase) name of the
1198    # html meta, 'value' is the metadata name for greenstone to use
1199    my %find_fields = ();
1200
1201    my %creator_fields = (); # short-cut for lookups
1202
1203
1204    foreach my $field (split /,/, $self->{'metadata_fields'}) {
1205        $field =~ s/^\s+//; # remove leading whitespace
1206        $field =~ s/\s+$//; # remove trailing whitespace
1207
1208    # support tag<tagname>
1209    if ($field =~ /^(.*?)<(.*?)>$/) {
1210        # "$2" is the user's preferred gs metadata name
1211        $find_fields{lc($1)}=$2; # lc = lowercase
1212    } else { # no <tagname> for mapping
1213        # "$field" is the user's preferred gs metadata name
1214        $find_fields{lc($field)}=$field; # lc = lowercase
1215    }
1216    }
1217
1218    if (defined $self->{'hunt_creator_metadata'} &&
1219    $self->{'hunt_creator_metadata'} == 1 ) {
1220    my @extra_fields =
1221        (
1222         'author',
1223         'author.email',
1224         'creator',
1225         'dc.creator',
1226         'dc.creator.corporatename',
1227         );
1228
1229    # add the creator_metadata fields to search for
1230    foreach my $field (@extra_fields) {
1231        $creator_fields{$field}=0; # add to lookup hash
1232    }
1233    }
1234
1235
1236    # find the header in the html file, which has the meta tags
1237    $$textref =~ m@<head>(.*?)</head>@si;
1238
1239    my $html_header=$1;
1240
1241    # go through every <meta... tag defined in the html and see if it is
1242    # one of the tags we want to match.
1243   
1244    # special case for title - we want to remember if its been found
1245    my $found_title = 0;
1246    # this assumes that ">" won't appear. (I don't think it's allowed to...)
1247    $html_header =~ /^/; # match the start of the string, for \G assertion
1248   
1249    while ($html_header =~ m/\G.*?<meta(.*?)>/sig) {
1250    my $metatag=$1;
1251    my ($tag, $value);
1252
1253    # find the tag name
1254    $metatag =~ /(?:name|http-equiv)\s*=\s*([\"\'])?(.*?)\1/is;
1255    $tag=$2;
1256    # in case they're not using " or ', but they should...
1257    if (! $tag) {
1258        $metatag =~ /(?:name|http-equiv)\s*=\s*([^\s\>]+)/is;
1259        $tag=$1;
1260    }
1261
1262    if (!defined $tag) {
1263        print $outhandle "HTMLPlugin: can't find NAME in \"$metatag\"\n";
1264        next;
1265    }
1266   
1267    # don't need to assign this field if it was passed in from a previous
1268    # (recursive) plugin
1269    if (defined $metadata->{$tag}) {next}
1270
1271    # find the tag content
1272    $metatag =~ /content\s*=\s*([\"\'])?(.*?)\1/is;
1273    $value=$2;
1274
1275    if (! $value) {
1276        $metatag =~ /(?:name|http-equiv)\s*=\s*([^\s\>]+)/is;
1277        $value=$1;
1278    }
1279    if (!defined $value) {
1280        print $outhandle "HTMLPlugin: can't find VALUE in \"$metatag\"\n";
1281        next;
1282    }
1283
1284    # clean up and add
1285    $value =~ s/\s+/ /gs;
1286    chomp($value); # remove trailing \n, if any
1287    if (exists $creator_fields{lc($tag)}) {
1288        # map this value onto greenstone's "Creator" metadata
1289        $tag='Creator';
1290    } elsif (!exists $find_fields{lc($tag)}) {
1291        next; # don't want this tag
1292    } else {
1293        # get the user's preferred capitalisation
1294        $tag = $find_fields{lc($tag)};
1295    }
1296    if (lc($tag) eq "title") {
1297        $found_title = 1;
1298    }
1299    print $outhandle " extracted \"$tag\" metadata \"$value\"\n"
1300        if ($self->{'verbosity'} > 2);
1301    if ($tag =~ /date.*/i){
1302        $tag = lc($tag);
1303    }
1304    $doc_obj->add_utf8_metadata($section, $tag, $value);
1305
1306    }
1307   
1308    # TITLE: extract the document title
1309    if (exists $find_fields{'title'} && !$found_title) {
1310    # we want a title, and didn't find one in the meta tags
1311    # see if there's a <title> tag
1312    my $title;
1313    my $from = ""; # for debugging output only
1314    if ($html_header =~ /<title[^>]*>([^<]+)<\/title[^>]*>/is) {
1315        $title = $1;
1316        $from = "<title> tags";
1317    }
1318
1319    if (!defined $title) {
1320        $from = "first 100 chars";
1321        # if no title use first 100 or so characters
1322        $title = $$textref;
1323        $title =~ s/^\xFE\xFF//; # Remove unicode byte order mark       
1324        $title =~ s/^.*?<body>//si;
1325        # ignore javascript!
1326        $title =~ s@<script.*?</script>@ @sig;
1327        $title =~ s/<\/([^>]+)><\1>//g; # (eg) </b><b> - no space
1328        $title =~ s/<[^>]*>/ /g; # remove all HTML tags
1329        $title = substr ($title, 0, 100);
1330        $title =~ s/\s\S*$/.../;
1331    }
1332    $title =~ s/<[^>]*>/ /g; # remove html tags
1333    $title =~ s/&nbsp;/ /g;
1334    $title =~ s/(?:&nbsp;|\xc2\xa0)/ /g; # utf-8 for nbsp...
1335    $title =~ s/\s+/ /gs; # collapse multiple spaces
1336    $title =~ s/^\s*//;   # remove leading spaces
1337    $title =~ s/\s*$//;   # remove trailing spaces
1338
1339    $title =~ s/^$self->{'title_sub'}// if ($self->{'title_sub'});
1340    $title =~ s/^\s+//s; # in case title_sub introduced any...
1341    $doc_obj->add_utf8_metadata ($section, 'Title', $title);
1342    print $outhandle " extracted Title metadata \"$title\" from $from\n"
1343        if ($self->{'verbosity'} > 2);
1344    }
1345   
1346    # add FileFormat metadata
1347    $doc_obj->add_metadata($section,"FileFormat", "HTML");
1348
1349    # Special, for metadata names such as tagH1 - extracts
1350    # the text between the first <H1> and </H1> tags into "H1" metadata.
1351
1352    foreach my $field (keys %find_fields) {
1353    if ($field !~ /^tag([a-z0-9]+)$/i) {next}
1354    my $tag = $1;
1355    if ($$textref =~ m@<$tag[^>]*>(.*?)</$tag[^>]*>@g) {
1356        my $content = $1;
1357        $content =~ s/&nbsp;/ /g;
1358        $content =~ s/<[^>]*>/ /g;
1359        $content =~ s/^\s+//;
1360        $content =~ s/\s+$//;
1361        $content =~ s/\s+/ /gs;
1362        if ($content) {
1363        $tag=$find_fields{"tag$tag"}; # get the user's capitalisation
1364        $tag =~ s/^tag//i;
1365        $doc_obj->add_utf8_metadata ($section, $tag, $content);
1366        print $outhandle " extracted \"$tag\" metadata \"$content\"\n"
1367            if ($self->{'verbosity'} > 2);
1368        }
1369    }
1370    }   
1371}
1372
1373
1374# evaluate any "../" to next directory up
1375# evaluate any "./" as here
1376sub eval_dir_dots {
1377    my $self = shift (@_);
1378    my ($filename) = @_;
1379    my $dirsep_os = &util::get_os_dirsep();
1380    my @dirsep = split(/$dirsep_os/,$filename);
1381
1382    my @eval_dirs = ();
1383    foreach my $d (@dirsep) {
1384    if ($d eq "..") {
1385        pop(@eval_dirs);
1386       
1387    } elsif ($d eq ".") {
1388        # do nothing!
1389
1390    } else {
1391        push(@eval_dirs,$d);
1392    }
1393    }
1394
1395    # Need to fiddle with number of elements in @eval_dirs if the
1396    # first one is the empty string.  This is because of a
1397    # modification to util::filename_cat that supresses the addition
1398    # of a leading '/' character (or \ if windows) (intended to help
1399    # filename cat with relative paths) if the first entry in the
1400    # array is the empty string.  Making the array start with *two*
1401    # empty strings is a way to defeat this "smart" option.
1402    #
1403    if (scalar(@eval_dirs) > 0) {
1404    if ($eval_dirs[0] eq ""){
1405        unshift(@eval_dirs,"");
1406    }
1407    }
1408    return &util::filename_cat(@eval_dirs);
1409}
1410
1411sub replace_usemap_links {
1412    my $self = shift (@_);
1413    my ($front, $link, $back) = @_;
1414
1415    $link =~ s/^\.\///;
1416    return $front . $link . $back;
1417}
1418
1419sub inc_filecount {
1420    my $self = shift (@_);
1421
1422    if ($self->{'file_num'} == 1000) {
1423    $self->{'dir_num'} ++;
1424    $self->{'file_num'} = 0;
1425    } else {
1426    $self->{'file_num'} ++;
1427    }
1428}
1429
1430
1431# Extend read_file so that strings like &eacute; are
1432# converted to UTF8 internally. 
1433#
1434# We don't convert &lt; or &gt; or &amp; or &quot; in case
1435# they interfere with the GML files
1436
1437sub read_file {
1438    my $self = shift(@_);
1439    my ($filename, $encoding, $language, $textref) = @_;
1440
1441    $self->SUPER::read_file($filename, $encoding, $language, $textref);
1442
1443    # Convert entities to their UTF8 equivalents
1444    $$textref =~ s/&(lt|gt|amp|quot|nbsp);/&z$1;/go;
1445    $$textref =~ s/&([^;]+);/&ghtml::getcharequiv($1,1)/gseo;
1446    $$textref =~ s/&z(lt|gt|amp|quot|nbsp);/&$1;/go;
1447}
1448
14491;
Note: See TracBrowser for help on using the browser.