root/main/trunk/greenstone2/perllib/plugins/HTMLPlugin.pm @ 28319

Revision 28319, 66.3 KB (checked in by ak19, 7 years ago)

The replace-with-src-doc feature had stopped working. It needed Image/Size.pm, so INC needs to be updated in HTMLPlugin to include perllib/cpan for this.

  • 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 Encode;
39use Unicode::Normalize 'normalize';
40
41use ReadTextFile;
42use HBPlugin;
43use ghtml;
44use unicode;
45use util;
46use FileUtils;
47use XMLParser;
48
49use File::Copy;
50
51sub BEGIN {
52    @HTMLPlugin::ISA = ('ReadTextFile', 'HBPlugin');
53    die "GSDLHOME not set\n" unless defined $ENV{'GSDLHOME'};
54    unshift (@INC, "$ENV{'GSDLHOME'}/perllib/cpan"); # for Image/Size.pm
55}
56
57use Image::Size;
58
59use strict; # every perl program should have this!
60no strict 'refs'; # make an exception so we can use variables as filehandles
61
62my $arguments =
63    [ { 'name' => "process_exp",
64    'desc' => "{BasePlugin.process_exp}",
65    'type' => "regexp",
66    'deft' =>  &get_default_process_exp() },
67      { 'name' => "block_exp",
68    'desc' => "{BasePlugin.block_exp}",
69    'type' => 'regexp',
70    'deft' =>  &get_default_block_exp() },
71      { 'name' => "nolinks",
72    'desc' => "{HTMLPlugin.nolinks}",
73    'type' => "flag" },
74      { 'name' => "keep_head",
75    'desc' => "{HTMLPlugin.keep_head}",
76    'type' => "flag" },
77      { 'name' => "no_metadata",
78    'desc' => "{HTMLPlugin.no_metadata}",
79    'type' => "flag" },
80      { 'name' => "metadata_fields",
81    'desc' => "{HTMLPlugin.metadata_fields}",
82    'type' => "string",
83    'deft' => "Title" },
84      { 'name' => "metadata_field_separator",
85    'desc' => "{HTMLPlugin.metadata_field_separator}",
86    'type' => "string",
87    'deft' => "" },
88      { 'name' => "hunt_creator_metadata",
89    'desc' => "{HTMLPlugin.hunt_creator_metadata}",
90    'type' => "flag" },
91      { 'name' => "file_is_url",
92    'desc' => "{HTMLPlugin.file_is_url}",
93    'type' => "flag" },
94      { 'name' => "assoc_files",
95    'desc' => "{HTMLPlugin.assoc_files}",
96    'type' => "regexp",
97    'deft' => &get_default_block_exp() },
98      { 'name' => "rename_assoc_files",
99    'desc' => "{HTMLPlugin.rename_assoc_files}",
100    'type' => "flag" },
101      { 'name' => "title_sub",
102    'desc' => "{HTMLPlugin.title_sub}",
103    'type' => "string",
104    'deft' => "" },
105      { 'name' => "description_tags",
106    'desc' => "{HTMLPlugin.description_tags}",
107    'type' => "flag" },
108      # retain this for backward compatibility (w3mir option was replaced by
109      # file_is_url)
110      { 'name' => "w3mir",
111#   'desc' => "{HTMLPlugin.w3mir}",
112    'type' => "flag",
113    'hiddengli' => "yes"},
114      { 'name' => "no_strip_metadata_html",
115    'desc' => "{HTMLPlugin.no_strip_metadata_html}",
116    'type' => "string",
117    'deft' => "",
118    'reqd' => "no"},
119      { 'name' => "sectionalise_using_h_tags",
120    'desc' => "{HTMLPlugin.sectionalise_using_h_tags}",
121    'type' => "flag" },
122      { 'name' => "use_realistic_book",
123        'desc' => "{HTMLPlugin.tidy_html}",
124    'type' => "flag"},
125      { 'name' => "old_style_HDL",
126        'desc' => "{HTMLPlugin.old_style_HDL}",
127    'type' => "flag"},
128      {'name' => "processing_tmp_files",
129       'desc' => "{BasePlugin.processing_tmp_files}",
130       'type' => "flag",
131       'hiddengli' => "yes"}
132      ];
133
134my $options = { 'name'     => "HTMLPlugin",
135        'desc'     => "{HTMLPlugin.desc}",
136        'abstract' => "no",
137        'inherits' => "yes",
138        'args'     => $arguments };
139
140
141sub new {
142    my ($class) = shift (@_);
143    my ($pluginlist,$inputargs,$hashArgOptLists) = @_;
144    push(@$pluginlist, $class);
145   
146    push(@{$hashArgOptLists->{"ArgList"}},@{$arguments});
147    push(@{$hashArgOptLists->{"OptList"}},$options);
148   
149
150    my $self = new ReadTextFile($pluginlist,$inputargs,$hashArgOptLists);
151   
152    if ($self->{'w3mir'}) {
153    $self->{'file_is_url'} = 1;
154    }
155    $self->{'aux_files'} = {};
156    $self->{'dir_num'} = 0;
157    $self->{'file_num'} = 0;
158   
159    return bless $self, $class;
160}
161
162# may want to use (?i)\.(gif|jpe?g|jpe|png|css|js(?:@.*)?)$
163# if have eg <script language="javascript" src="img/lib.js@123">
164# blocking is now done by reading through the file and recording all the
165# images and other files
166sub get_default_block_exp {
167    my $self = shift (@_);
168   
169    #return q^(?i)\.(gif|jpe?g|jpe|jpg|png|css)$^;
170    return "";
171}
172
173sub get_default_process_exp {
174    my $self = shift (@_);
175   
176    # the last option is an attempt to encode the concept of an html query ...
177    return q^(?i)(\.html?|\.shtml|\.shm|\.asp|\.php\d?|\.cgi|.+\?.+=.*)$^;
178}
179
180sub store_block_files
181{
182    my $self =shift (@_);
183    my ($filename_full_path, $block_hash) = @_;
184
185    my $html_fname = $filename_full_path;
186   
187    my ($language, $content_encoding) = $self->textcat_get_language_encoding ($filename_full_path);
188    $self->{'store_content_encoding'}->{$filename_full_path} = $content_encoding;
189
190    # read in file ($text will be in utf8)
191    my $raw_text = "";
192    $self->read_file_no_decoding($filename_full_path, \$raw_text);
193
194    my $textref = \$raw_text;
195    my $opencom = '(?:<!--|&lt;!(?:&mdash;|&#151;|--))';
196    my $closecom = '(?:-->|(?:&mdash;|&#151;|--)&gt;)';
197    $$textref =~ s/$opencom(.*?)$closecom//gs;
198
199    # Convert entities to their UTF8 equivalents
200    $$textref =~ s/&(lt|gt|amp|quot|nbsp);/&z$1;/go;
201    $$textref =~ s/&([^;]+);/&ghtml::getcharequiv($1,1,0)/gseo; # on this occassion, want it left as utf8
202    $$textref =~ s/&z(lt|gt|amp|quot|nbsp);/&$1;/go;
203
204    my $attval = "\\\"[^\\\"]+\\\"|[^\\s>]+";
205    my @img_matches = ($$textref =~ m/<img[^>]*?src\s*=\s*($attval)[^>]*>/igs);
206    my @usemap_matches = ($$textref =~ m/<img[^>]*?usemap\s*=\s*($attval)[^>]*>/igs);
207    my @link_matches = ($$textref =~ m/<link[^>]*?href\s*=\s*($attval)[^>]*>/igs);
208    my @embed_matches = ($$textref =~ m/<embed[^>]*?src\s*=\s*($attval)[^>]*>/igs);
209    my @tabbg_matches = ($$textref =~ m/<(?:body|table|tr|td)[^>]*?background\s*=\s*($attval)[^>]*>/igs);
210    my @script_matches = ($$textref =~ m/<script[^>]*?src\s*=\s*($attval)[^>]*>/igs);
211
212    if(!defined $self->{'unicode_to_original_filename'}) {
213    # maps from utf8 converted link name -> original filename referrred to by (possibly URL-encoded) src url
214    $self->{'unicode_to_original_filename'} = {};
215    }
216
217    foreach my $raw_link (@img_matches, @usemap_matches, @link_matches, @embed_matches, @tabbg_matches, @script_matches) {
218
219    # remove quotes from link at start and end if necessary
220    if ($raw_link =~ m/^\"/) {
221        $raw_link =~ s/^\"//;
222        $raw_link =~ s/\"$//;
223    }
224
225    # remove any anchor names, e.g. foo.html#name becomes foo.html
226    # but watch out for any #'s that are part of entities, such as &#x3B1;
227    $raw_link =~ s/([^&])\#.*$/$1/s;
228
229    # some links may just be anchor names
230    next unless ($raw_link =~ /\S+/);
231
232    if ($raw_link !~ m@^/@ && $raw_link !~ m/^([A-Z]:?)\\/i) {
233        # Turn relative file path into full path
234        my $dirname = &File::Basename::dirname($filename_full_path);
235        $raw_link = &FileUtils::filenameConcatenate($dirname, $raw_link);
236    }
237    $raw_link = $self->eval_dir_dots($raw_link);
238
239    # this is the actual filename on the filesystem (that the link refers to)
240    my $url_original_filename = $self->opt_url_decode($raw_link);
241
242    my ($uses_bytecodes,$exceeds_bytecodes) = &unicode::analyze_raw_string($url_original_filename);
243
244    if ($exceeds_bytecodes) {
245        # We have a link to a file name that is more complicated than a raw byte filename
246        # What we do next depends on the operating system we are on
247
248        if ($ENV{'GSDLOS'} =~ /^(linux|solaris)$/i) {
249        # Assume we're dealing with a UTF-8 encoded filename
250        $url_original_filename = encode("utf8", $url_original_filename);
251        }
252        elsif ($ENV{'GSDLOS'} =~ /^darwin$/i) {
253        # HFS+ is UTF8 with decompostion
254        $url_original_filename = encode("utf8", $url_original_filename);
255        $url_original_filename = normalize('D', $url_original_filename); # Normalization Form D (decomposition)     
256        }
257        elsif ($ENV{'GSDLOS'} =~ /^windows$/i) {
258        # Don't need to do anything as later code maps Windows
259        # unicode filenames to DOS short filenames when needed     
260        }
261        else {
262        my $outhandle = $self->{'outhandle'};
263        print $outhandle "Warning: Unrecognized operating system ", $ENV{'GSDLOS'}, "\n";
264        print $outhandle "         in raw file system encoding of: $raw_link\n";
265        print $outhandle "         Assuming filesystem is UTF-8 based.\n";
266        $url_original_filename = encode("utf8", $url_original_filename);
267        }
268    }
269
270    # Convert the (currently raw) link into its Unicode version.
271    # Store the Unicode link along with the url_original_filename
272    my $unicode_url_original_filename = "";
273    $self->decode_text($raw_link,$content_encoding,$language,\$unicode_url_original_filename);
274
275
276    $self->{'unicode_to_original_filename'}->{$unicode_url_original_filename} = $url_original_filename;
277
278
279    if ($url_original_filename ne $unicode_url_original_filename) {
280        my $outhandle = $self->{'outhandle'};
281       
282        print $outhandle "URL Encoding $url_original_filename\n";
283        print $outhandle " ->$unicode_url_original_filename\n";
284
285        # Allow for possibility of raw byte version and Unicode versions of file
286        &util::block_filename($block_hash,$unicode_url_original_filename);
287    }
288
289    # $url_original_filename = &util::upgrade_if_dos_filename($url_original_filename);
290    &util::block_filename($block_hash,$url_original_filename);
291       
292    }
293}
294
295# Given a filename in any encoding, will URL decode it to get back the original filename
296# in the original encoding. Because this method is intended to work out the *original*
297# filename*, it does not URL decode any filename if a file by the name of the *URL-encoded*
298# string already exists in the local folder.
299#
300sub opt_url_decode {
301    my $self = shift (@_);
302    my ($raw_link) = @_;
303
304
305    # Replace %XX's in URL with decoded value if required.
306    # Note that the filename may include the %XX in some situations
307
308##    if ($raw_link =~ m/\%[A-F0-9]{2}/i) {
309
310    if (($raw_link =~ m/\%[A-F0-9]{2}/i) || ($raw_link =~ m/\&\#x[0-9A-F]+;/i) || ($raw_link =~ m/\&\#[0-9]+;/i)) {
311    if (!-e $raw_link) {
312        $raw_link = &unicode::url_decode($raw_link,1);
313    }
314    }
315   
316    return $raw_link;
317}
318
319sub read_into_doc_obj
320{
321    my $self = shift (@_); 
322    my ($pluginfo, $base_dir, $file, $block_hash, $metadata, $processor, $maxdocs, $total_count, $gli) = @_;
323       
324    my ($filename_full_path, $filename_no_path) = &util::get_full_filenames($base_dir, $file);
325
326    # Lookup content_encoding worked out in file_block pass for this file
327    # Store it under the local name 'content_encoding' so its nice and
328    # easy to access
329    $self->{'content_encoding'} = $self->{'store_content_encoding'}->{$filename_full_path};
330
331    # get the input file
332    my $input_filename = $file;
333    my ($tailname, $dirname, $suffix) = &File::Basename::fileparse($input_filename, "\\.[^\\.]+\$");
334    $suffix = lc($suffix);   
335    my $tidy_filename;
336    if (($self->{'use_realistic_book'}) || ($self->{'old_style_HDL'}))
337    {
338    # because the document has to be sectionalized set the description tags
339    $self->{'description_tags'} = 1;
340   
341    # set the file to be tidied
342    $input_filename = &FileUtils::filenameConcatenate($base_dir,$file) if $base_dir =~ m/\w/;
343   
344    # get the tidied file
345    #my $tidy_filename = $self->tmp_tidy_file($input_filename);
346    $tidy_filename = $self->convert_tidy_or_oldHDL_file($input_filename);
347   
348    # derive tmp filename from input filename
349    my ($tailname, $dirname, $suffix) = &File::Basename::fileparse($tidy_filename, "\\.[^\\.]+\$");
350   
351    # set the new input file and base_dir to be from the tidied file
352    $file = "$tailname$suffix";
353    $base_dir = $dirname;
354    }
355   
356    # call the parent read_into_doc_obj
357    my ($process_status,$doc_obj) = $self->SUPER::read_into_doc_obj($pluginfo, $base_dir, $file, $block_hash, $metadata, $processor, $maxdocs, $total_count, $gli);
358    if (($self->{'use_realistic_book'}) || ($self->{'old_style_HDL'}))
359    {
360    # now we need to reset the filenames in the doc obj so that the converted filenames are not used
361    my $collect_file = &util::filename_within_collection($filename_full_path);
362    $doc_obj->set_source_filename ($collect_file, $self->{'file_rename_method'});
363    ## set_source_filename does not set the doc_obj source_path which is used in archives dbs for incremental
364    # build. So set it manually.
365    $doc_obj->set_source_path($filename_full_path);
366    my $collect_conv_file = &util::filename_within_collection($tidy_filename);
367    $doc_obj->set_converted_filename($collect_conv_file);
368
369    my $plugin_filename_encoding = $self->{'filename_encoding'};
370    my $filename_encoding = $self->deduce_filename_encoding($file,$metadata,$plugin_filename_encoding);
371    $self->set_Source_metadata($doc_obj, $filename_full_path, $filename_encoding);
372    }
373
374    delete $self->{'store_content_encoding'}->{$filename_full_path};
375    $self->{'content_encoding'} = undef;
376
377    return ($process_status,$doc_obj);
378}
379
380# do plugin specific processing of doc_obj
381sub process {
382    my $self = shift (@_);
383    my ($textref, $pluginfo, $base_dir, $file, $metadata, $doc_obj, $gli) = @_;
384    my $outhandle = $self->{'outhandle'};
385
386    if ($ENV{'GSDLOS'} =~ m/^windows/i) {
387    # this makes life so much easier... perl can cope with unix-style '/'s.
388    $base_dir =~ s@(\\)+@/@g;
389    $file =~ s@(\\)+@/@g;     
390    }
391
392    my $filename = &FileUtils::filenameConcatenate($base_dir,$file);
393    my $upgraded_base_dir = &util::upgrade_if_dos_filename($base_dir);
394    my $upgraded_filename = &util::upgrade_if_dos_filename($filename);
395
396    if ($ENV{'GSDLOS'} =~ m/^windows/i) {
397    # And again
398    $upgraded_base_dir =~ s@(\\)+@/@g; 
399    $upgraded_filename =~ s@(\\)+@/@g;
400   
401    # Need to make sure there is a '/' on the end of upgraded_base_dir
402    if (($upgraded_base_dir ne "") && ($upgraded_base_dir !~ m/\/$/)) {
403        $upgraded_base_dir .= "/";
404    }
405    }
406    my $upgraded_file = &util::filename_within_directory($upgraded_filename,$upgraded_base_dir);
407   
408    # reset per-doc stuff...
409    $self->{'aux_files'} = {};
410    $self->{'dir_num'} = 0;
411    $self->{'file_num'} = 0;
412
413    # process an HTML file where sections are divided by headings tags (H1, H2 ...)
414    # you can also include metadata in the format (X can be any number)
415    # <hX>Title<!--gsdl-metadata
416    #   <Metadata name="name1">value1</Metadata>
417    #   ...
418    #   <Metadata name="nameN">valueN</Metadata>
419    #--></hX>
420    if ($self->{'sectionalise_using_h_tags'}) {
421    # description_tags should allways be activated because we convert headings to description tags
422    $self->{'description_tags'} = 1;
423
424    my $arrSections = [];
425    $$textref =~ s/<h([0-9]+)[^>]*>(.*?)<\/h[0-9]+>/$self->process_heading($1, $2, $arrSections, $upgraded_file)/isge;
426
427    if (scalar(@$arrSections)) {
428        my $strMetadata = $self->update_section_data($arrSections, -1);
429        if (length($strMetadata)) {
430        $strMetadata = '<!--' . $strMetadata . "\n-->\n</body>";
431        $$textref =~ s/<\/body>/$strMetadata/ig;
432        }
433    }
434    }
435
436    my $cursection = $doc_obj->get_top_section();
437
438    $self->extract_metadata ($textref, $metadata, $doc_obj, $cursection)
439    unless $self->{'no_metadata'} || $self->{'description_tags'};
440
441    # Store URL for page as metadata - this can be used for an
442    # altavista style search interface. The URL won't be valid
443    # unless the file structure contains the domain name (i.e.
444    # like when w3mir is used to download a website).
445
446    # URL metadata (even invalid ones) are used to support internal
447    # links, so even if 'file_is_url' is off, still need to store info
448
449    my ($tailname,$dirname) = &File::Basename::fileparse($upgraded_file);
450
451#    my $utf8_file = $self->filename_to_utf8_metadata($file);
452#    $utf8_file =~ s/&\#095;/_/g;
453#    variable below used to be utf8_file
454
455    my $url_encoded_file = &unicode::raw_filename_to_url_encoded($tailname);
456    my $utf8_url_encoded_file = &unicode::raw_filename_to_utf8_url_encoded($tailname);
457
458    my $web_url = "http://";
459    my $utf8_web_url = "http://";
460    if(defined $dirname) { # local directory
461        # Check for "ftp" in the domain name of the directory
462        #  structure to determine if this URL should be a ftp:// URL
463        # This check is not infallible, but better than omitting the
464        #  check, which would cause all files downloaded from ftp sites
465        #  via mirroring with wget to have potentially erroneous http:// URLs
466        #  assigned in their metadata
467        if ($dirname =~ /^[^\/]*ftp/i)
468    {
469      $web_url = "ftp://";
470      $utf8_web_url = "ftp://";
471    }
472    $dirname = $self->eval_dir_dots($dirname);
473    $dirname .= &util::get_dirsep() if $dirname ne ""; # if there's a directory, it should end on "/"
474
475    $web_url = $web_url.$dirname.$url_encoded_file;
476    $utf8_web_url = $utf8_web_url.$dirname.$utf8_url_encoded_file;
477    } else {
478    $web_url = $web_url.$url_encoded_file;
479    $utf8_web_url = $utf8_web_url.$utf8_url_encoded_file;
480    }
481    $web_url =~ s/\\/\//g;
482    $utf8_web_url =~ s/\\/\//g;
483
484    if ((defined $ENV{"DEBUG_UNICODE"}) && ($ENV{"DEBUG_UNICODE"})) {
485    print STDERR "*******DEBUG: upgraded_file:       $upgraded_file\n";
486    print STDERR "*******DEBUG: adding URL metadata: $utf8_url_encoded_file\n";
487    }
488
489
490    $doc_obj->add_utf8_metadata($cursection, "URL", $web_url);
491    $doc_obj->add_utf8_metadata($cursection, "UTF8URL", $utf8_web_url);
492
493    if ($self->{'file_is_url'}) {
494    $doc_obj->add_metadata($cursection, "weblink", "<a href=\"$web_url\">");
495    $doc_obj->add_metadata($cursection, "webicon", "_iconworld_");
496    $doc_obj->add_metadata($cursection, "/weblink", "</a>");
497    }
498
499    if ($self->{'description_tags'}) {
500    # remove the html header - note that doing this here means any
501    # sections defined within the header will be lost (so all <Section>
502    # tags must appear within the body of the HTML)
503    my ($head_keep) = ($$textref =~ m/^(.*?)<body[^>]*>/is);
504
505    $$textref =~ s/^.*?<body[^>]*>//is;
506    $$textref =~ s/(<\/body[^>]*>|<\/html[^>]*>)//isg;
507
508    my $opencom = '(?:<!--|&lt;!(?:&mdash;|&#151;|--))';
509    my $closecom = '(?:-->|(?:&mdash;|&#151;|--)&gt;)';
510
511    my $lt = '(?:<|&lt;)';
512    my $gt = '(?:>|&gt;)';
513    my $quot = '(?:"|&quot;|&rdquo;|&ldquo;)';
514
515    my $dont_strip = '';
516    if ($self->{'no_strip_metadata_html'}) {
517        ($dont_strip = $self->{'no_strip_metadata_html'}) =~ s{,}{|}g;
518    }
519
520    my $found_something = 0; my $top = 1;
521    while ($$textref =~ s/^(.*?)$opencom(.*?)$closecom//s) {
522        my $text = $1;
523        my $comment = $2;
524        if (defined $text) {
525        # text before a comment - note that getting to here
526        # doesn't necessarily mean there are Section tags in
527        # the document
528        $self->process_section(\$text, $upgraded_base_dir, $upgraded_file, $doc_obj, $cursection);
529        }
530        while ($comment =~ s/$lt(.*?)$gt//s) {
531        my $tag = $1;
532        if ($tag eq "Section") {
533            $found_something = 1;
534            $cursection = $doc_obj->insert_section($doc_obj->get_end_child($cursection)) unless $top;
535            $top = 0;
536        } elsif ($tag eq "/Section") {
537            $found_something = 1;
538            $cursection = $doc_obj->get_parent_section ($cursection);
539        } elsif ($tag =~ m/^Metadata name=$quot(.*?)$quot/s) {
540            my $metaname = $1;
541            my $accumulate = $tag =~ m/mode=${quot}accumulate${quot}/ ? 1 : 0;
542            $comment =~ s/^(.*?)$lt\/Metadata$gt//s;
543            my $metavalue = $1;
544            $metavalue =~ s/^\s+//;
545            $metavalue =~ s/\s+$//;
546                    # assume that no metadata value intentionally includes
547                    # carriage returns or HTML tags (if they're there they
548                    # were probably introduced when converting to HTML from
549                    # some other format).
550            # actually some people want to have html tags in their
551            # metadata.
552            $metavalue =~ s/[\cJ\cM]/ /sg;
553            $metavalue =~ s/<[^>]+>//sg
554            unless $dont_strip && ($dont_strip eq 'all' || $metaname =~ m/^($dont_strip)$/);
555            $metavalue =~ s/\s+/ /sg;
556            if ($metaname =~ /\./) { # has a namespace
557            $metaname = "ex.$metaname";
558            }
559            if ($accumulate) {
560            $doc_obj->add_utf8_metadata($cursection, $metaname, $metavalue);
561            } else {
562            $doc_obj->set_utf8_metadata_element($cursection, $metaname, $metavalue);   
563            }
564        } elsif ($tag eq "Description" || $tag eq "/Description") {
565            # do nothing with containing Description tags
566        } else {
567            # simple HTML tag (probably created by the conversion
568            # to HTML from some other format) - we'll ignore it and
569            # hope for the best ;-)
570        }
571        }
572    }
573    if ($cursection ne "") {
574        print $outhandle "HTMLPlugin: WARNING: $upgraded_file contains unmatched <Section></Section> tags\n";
575    }
576
577    $$textref =~ s/^.*?<body[^>]*>//is;
578    $$textref =~ s/(<\/body[^>]*>|<\/html[^>]*>)//isg;
579    if ($$textref =~ m/\S/) {
580        if (!$found_something) {
581        if ($self->{'verbosity'} > 2) {
582            print $outhandle "HTMLPlugin: WARNING: $upgraded_file appears to contain no Section tags so\n";
583            print $outhandle "          will be processed as a single section document\n";
584        }
585
586        # go ahead and process single-section document
587        $self->process_section($textref, $upgraded_base_dir, $upgraded_file, $doc_obj, $cursection);
588
589        # if document contains no Section tags we'll go ahead
590        # and extract metadata (this won't have been done
591        # above as the -description_tags option prevents it)
592        my $complete_text = $head_keep.$doc_obj->get_text($cursection);
593        $self->extract_metadata (\$complete_text, $metadata, $doc_obj, $cursection)
594            unless $self->{'no_metadata'};
595
596        } else {
597        print $outhandle "HTMLPlugin: WARNING: $upgraded_file contains the following text outside\n";
598        print $outhandle "          of the final closing </Section> tag. This text will\n";
599        print $outhandle "          be ignored.";
600
601        my ($text);
602        if (length($$textref) > 30) {
603            $text = substr($$textref, 0, 30) . "...";
604        } else {
605            $text = $$textref;
606        }
607        $text =~ s/\n/ /isg;
608        print $outhandle " ($text)\n";
609        }
610    } elsif (!$found_something) {
611
612        if ($self->{'verbosity'} > 2) {
613        # may get to here if document contained no valid Section
614        # tags but did contain some comments. The text will have
615        # been processed already but we should print the warning
616        # as above and extract metadata
617        print $outhandle "HTMLPlugin: WARNING: $upgraded_file appears to contain no Section tags and\n";
618        print $outhandle "          is blank or empty.  Metadata will be assigned if present.\n";
619        }
620
621        my $complete_text = $head_keep.$doc_obj->get_text($cursection);
622        $self->extract_metadata (\$complete_text, $metadata, $doc_obj, $cursection)
623        unless $self->{'no_metadata'};
624    }
625   
626    } else {
627
628    # remove header and footer
629    if (!$self->{'keep_head'} || $self->{'description_tags'}) {
630        $$textref =~ s/^.*?<body[^>]*>//is;
631        $$textref =~ s/(<\/body[^>]*>|<\/html[^>]*>)//isg;
632    }
633
634    $self->{'css_assoc_files'} = {};
635   
636    # single section document
637    $self->process_section($textref, $upgraded_base_dir, $upgraded_file, $doc_obj, $cursection);
638   
639    #my $upgraded_filename_dirname = &File::Basename::dirname($upgraded_filename);
640       
641    $self->acquire_css_associated_files($doc_obj, $cursection);
642   
643    $self->{'css_assoc_files'} = {};
644    }
645
646    return 1;
647}
648
649
650sub process_heading
651{
652    my ($self, $nHeadNo, $strHeadingText, $arrSections, $file) = @_;
653    $strHeadingText = '' if (!defined($strHeadingText));
654
655    my $strMetadata = $self->update_section_data($arrSections, int($nHeadNo));
656
657    my $strSecMetadata = '';
658    while ($strHeadingText =~ s/<!--gsdl-metadata(.*?)-->//is)
659    {
660    $strSecMetadata .= $1;
661    }
662
663    $strHeadingText =~ s/^\s+//g;
664    $strHeadingText =~ s/\s+$//g;
665    $strSecMetadata =~ s/^\s+//g;
666    $strSecMetadata =~ s/\s+$//g;
667
668    $strMetadata .= "\n<Section>\n\t<Description>\n\t\t<Metadata name=\"Title\">" . $strHeadingText . "</Metadata>\n";
669
670    if (length($strSecMetadata)) {
671    $strMetadata .= "\t\t" . $strSecMetadata . "\n";
672    }
673
674    $strMetadata .= "\t</Description>\n";
675
676    return "<!--" . $strMetadata . "-->";
677}
678
679
680sub update_section_data
681{
682    my ($self, $arrSections, $nCurTocNo) = @_;
683    my ($strBuffer, $nLast, $nSections) = ('', 0, scalar(@$arrSections));
684
685    if ($nSections == 0) {
686    push @$arrSections, $nCurTocNo;
687    return $strBuffer;
688    }
689    $nLast = $arrSections->[$nSections - 1];
690    if ($nCurTocNo > $nLast) {
691    push @$arrSections, $nCurTocNo;
692    return $strBuffer;
693    }
694    for(my $i = $nSections - 1; $i >= 0; $i--) {
695    if ($nCurTocNo <= $arrSections->[$i]) {
696        $strBuffer .= "\n</Section>";
697        pop @$arrSections;
698    }
699    }
700    push @$arrSections, $nCurTocNo;
701    return $strBuffer;
702}
703
704
705# note that process_section may be called multiple times for a single
706# section (relying on the fact that add_utf8_text appends the text to any
707# that may exist already).
708sub process_section {
709    my $self = shift (@_);
710    my ($textref, $base_dir, $file, $doc_obj, $cursection) = @_;
711
712    my @styleTagsText = ($$textref =~ m/<style[^>]*>([^<]*)<\/style>/sg);
713    if(scalar(@styleTagsText) > 0)
714    {
715        my $css_filename_dirname = &File::Basename::dirname(&FileUtils::filenameConcatenate($base_dir, $file));
716        foreach my $styleText (@styleTagsText)
717        {
718            $self->acquire_css_associated_files_from_text_block($styleText, $css_filename_dirname);
719        }
720    }
721
722    # trap links
723    if (!$self->{'nolinks'}) {
724    # usemap="./#index" not handled correctly => change to "#index"
725##  $$textref =~ s/(<img[^>]*?usemap\s*=\s*[\"\']?)([^\"\'>\s]+)([\"\']?[^>]*>)/
726
727##  my $opencom = '(?:<!--|&lt;!(?:&mdash;|&#151;|--))';
728##  my $closecom = '(?:-->|(?:&mdash;|&#151;|--)&gt;)';
729
730    $$textref =~ s/(<img[^>]*?usemap\s*=\s*)((?:[\"][^\"]+[\"])|(?:[\'][^\']+[\'])|(?:[^\s\/>]+))([^>]*>)/
731        $self->replace_usemap_links($1, $2, $3)/isge;
732
733    $$textref =~ s/(<(?:a|area|frame|link|script)\s+[^>]*?\s*(?:href|src)\s*=\s*)((?:[\"][^\"]+[\"])|(?:[\'][^\']+[\'])|(?:[^\s\/>]+))([^>]*>)/
734        $self->replace_href_links ($1, $2, $3, $base_dir, $file, $doc_obj, $cursection)/isge;
735   
736##  $$textref =~ s/($opencom.*?)?+(<(?:a|area|frame|link|script)\s+[^>]*?\s*(?:href|src)\s*=\s*)((?:[\"][^\"]+[\"])|(?:[\'][^\']+[\'])|(?:[^\s\/>]+))([^>]*>)(.*?$closecom)?+/
737#       $self->replace_href_links ($1, $2, $3, $4, $5, $base_dir, $file, $doc_obj, $cursection)/isge;
738    }
739
740    # trap images
741
742    # Previously, by default, HTMLPlugin would embed <img> tags inside anchor tags
743    # i.e. <a href="image><img src="image"></a> in order to overcome a problem that
744    # turned regular text succeeding images into links. That is, by embedding <imgs>
745    # inside <a href=""></a>, the text following images were no longer misbehaving.
746    # However, there would be many occasions whereby images were not meant to link
747    # to their source images but where the images would link to another web page.
748    # To allow this, the no_image_links option was introduced: it would prevent
749    # the behaviour of embedding images into links that referenced the source images.
750
751    # Somewhere along the line, the problem of normal text turning into links when
752    # such text followed images which were not embedded in <a href=""></a> ceased
753    # to occur. This is why the following lines have been commented out (as well as
754    # two lines in replace_images). They appear to no longer apply.
755
756    # If at any time, there is a need for having images embedded in <a> anchor tags,
757    # then it might be better to turn that into an HTMLPlugin option rather than make
758    # it the default behaviour. Also, eventually, no_image_links needs to become
759    # a deprecated option for HTMLPlugin as it has now become the default behaviour.
760
761    #if(!$self->{'no_image_links'}){
762    $$textref =~ s/(<(?:img|embed|table|tr|td)[^>]*?(?:src|background)\s*=\s*)((?:[\"][^\"]+[\"])|(?:[\'][^\']+[\'])|(?:[^\s\/>]+))([^>]*>)/
763    $self->replace_images ($1, $2, $3, $base_dir, $file, $doc_obj, $cursection)/isge;
764    #}
765
766    # add text to document object
767    # turn \ into \\ so that the rest of greenstone doesn't think there
768    # is an escape code following. (Macro parsing loses them...)
769    $$textref =~ s/\\/\\\\/go;
770   
771    $doc_obj->add_utf8_text($cursection, $$textref);
772}
773
774sub replace_images {
775    my $self = shift (@_);
776    my ($front, $link, $back, $base_dir,
777    $file, $doc_obj, $section) = @_;
778
779    # remove quotes from link at start and end if necessary
780    if ($link=~/^[\"\']/) {
781    $link=~s/^[\"\']//;
782    $link=~s/[\"\']$//;
783    $front.='"';
784    $back="\"$back";
785    }
786
787    $link =~ s/\n/ /g;
788
789    # Hack to overcome Windows wv 0.7.1 bug that causes embedded images to be broken
790    # If the Word file path has spaces in it, wv messes up and you end up with
791    #   absolute paths for the images, and without the "file://" prefix
792    # So check for this special case and massage the data to be correct
793    if ($ENV{'GSDLOS'} =~ m/^windows/i && $self->{'plugin_type'} eq "WordPlug" && $link =~ m/^[A-Za-z]\:\\/) {
794    $link =~ s/^.*\\([^\\]+)$/$1/;
795    }
796   
797    my ($href, $hash_part, $rl) = $self->format_link ($link, $base_dir, $file);
798
799    my $img_file =  $self->add_file ($href, $rl, $hash_part, $base_dir, $doc_obj, $section);
800
801#    print STDERR "**** link = $link\n**** href = $href\n**** img_file = $img_file, rl = $rl\n\n";
802
803    my $anchor_name = $img_file;
804    #$anchor_name =~ s/^.*\///;
805    #$anchor_name = "<a name=\"$anchor_name\" ></a>";
806
807    my $image_link = $front . $img_file .$back;
808    return $image_link;
809
810    # The reasons for why the following two lines are no longer necessary can be
811    # found in subroutine process_section
812    #my $anchor_link = "<a href=\"$img_file\" >".$image_link."</a>"; 
813    #return $anchor_link;   
814   
815    #return $front . $img_file . $back . $anchor_name;
816}
817
818sub replace_href_links {
819    my $self = shift (@_);
820    my ($front, $link, $back, $base_dir, $file, $doc_obj, $section) = @_;
821   
822    if($front =~ m/^<link / && $link =~ m/\.css"$/)
823    {
824        my $actual_link = $link;
825        $actual_link =~ s/^"(.*)"$/$1/;
826       
827        my $directory = &File::Basename::dirname($file);
828       
829        my $css_filename = &FileUtils::filenameConcatenate($base_dir, $directory, $actual_link);   
830        $self->retrieve_css_associated_files($css_filename);
831    }
832   
833    # remove quotes from link at start and end if necessary
834    if ($link=~/^[\"\']/) {
835    $link=~s/^[\"\']//;
836    $link=~s/[\"\']$//;
837    $front.='"';
838    $back="\"$back";
839    }
840
841    # attempt to sort out targets - frames are not handled
842    # well in this plugin and some cases will screw things
843    # up - e.g. the _parent target (so we'll just remove
844    # them all ;-)
845    $front =~ s/(target=\"?)_top(\"?)/$1_gsdltop_$2/is;
846    $back =~ s/(target=\"?)_top(\"?)/$1_gsdltop_$2/is;
847    $front =~ s/target=\"?_parent\"?//is;
848    $back =~ s/target=\"?_parent\"?//is;
849
850    if($link =~ m/^\#/s)
851    {
852        return $front . "_httpsamepagelink_" . $link . $back;   
853    }
854   
855    $link =~ s/\n/ /g;
856
857    # Find file referred to by $link on file system
858    # This is more complicated than it sounds when char encodings
859    # is taken in to account
860    my ($href, $hash_part, $rl) = $self->format_link ($link, $base_dir, $file);
861
862    # href may use '\'s where '/'s should be on Windows
863    $href =~ s/\\/\//g;
864    my ($filename) = $href =~ m/^(?:.*?):(?:\/\/)?(.*)/;
865
866    ##### leave all these links alone (they won't be picked up by intermediate
867    ##### pages). I think that's safest when dealing with frames, targets etc.
868    ##### (at least until I think of a better way to do it). Problems occur with
869    ##### mailto links from within small frames, the intermediate page is displayed
870    ##### within that frame and can't be seen. There is still potential for this to
871    ##### happen even with html pages - the solution seems to be to somehow tell
872    ##### the browser from the server side to display the page being sent (i.e.
873    ##### the intermediate page) in the top level window - I'm not sure if that's
874    ##### possible - the following line should probably be deleted if that can be done
875    return $front . $link . $back if $href =~ m/^(mailto|news|gopher|nntp|telnet|javascript):/is;
876
877    if (($rl == 0) || ($filename =~ m/$self->{'process_exp'}/) ||
878    ($href =~ m/\/$/) || ($href =~ m/^(mailto|news|gopher|nntp|telnet|javascript):/i)) {
879
880    if ($ENV{'GSDLOS'} =~ m/^windows$/) {
881
882        # Don't do any encoding for now, as not clear what
883        # the right thing to do is to support filename
884        # encoding on Windows when they are not UTF16
885        #
886    }
887    else {
888        # => Unix-based system
889
890        # If web page didn't give encoding, then default to utf8
891        my $content_encoding= $self->{'content_encoding'} || "utf8";
892
893        if ((defined $ENV{"DEBUG_UNICODE"}) && ($ENV{"DEBUG_UNICODE"})) {
894        print STDERR "**** Encoding with '$content_encoding', href: $href\n";
895        }
896
897        # on Darwin, the unicode filenames are stored on the file
898        # system in decomposed form, so any href link (including when
899        # URL-encoded) should refer to the decomposed name of the file
900        if ($ENV{'GSDLOS'} =~ /^darwin$/i) {
901        $href = normalize('D', $href); # Normalization Form D (decomposition)
902        }
903
904        $href = encode($content_encoding,$href);
905    }
906
907    $href = &unicode::raw_filename_to_utf8_url_encoded($href); 
908    $href = &unicode::filename_to_url($href);
909
910    &ghtml::urlsafe ($href);
911
912    if ((defined $ENV{"DEBUG_UNICODE"}) && ($ENV{"DEBUG_UNICODE"})) {
913        print STDERR "******DEBUG: href=$href\n";   
914    }
915
916
917    return $front . "_httpextlink_&amp;rl=" . $rl . "&amp;href=" . $href . $hash_part . $back;
918    } else {
919    # link is to some other type of file (e.g., an image) so we'll
920    # need to associate that file
921    return $front . $self->add_file ($href, $rl, $hash_part, $base_dir, $doc_obj, $section) . $back;
922    }
923}
924
925sub retrieve_css_associated_files {
926    my $self = shift (@_);
927    my ($css_filename) = @_;
928   
929    my $css_filename_dirname = &File::Basename::dirname($css_filename);
930   
931    open (CSSFILE, $css_filename) || return;
932    sysread (CSSFILE, my $file_string, -s CSSFILE);
933   
934    $self->acquire_css_associated_files_from_text_block($file_string, $css_filename_dirname) unless !defined $file_string;
935   
936    close CSSFILE;
937}
938
939sub acquire_css_associated_files_from_text_block {
940    my $self = shift (@_);
941    my ($text, $css_filename_dirname) = @_;
942   
943    my @image_urls = ($text =~ m/background-image:\s*url[^;]*;/sg);
944    foreach my $img_url (@image_urls)
945    {
946        $img_url =~ s/^.*url.*\((.*)\).*$/$1/;
947        $img_url =~ s/^\s*"?([^"]*)"?\s*$/$1/;
948
949        $self->{'css_assoc_files'}->{&FileUtils::filenameConcatenate($css_filename_dirname, $img_url)} = $img_url;
950    }
951}
952
953sub acquire_css_associated_files {
954    my $self = shift(@_);
955   
956    my ($doc_obj, $section) = @_;
957   
958    foreach my $image_filename (keys %{$self->{'css_assoc_files'}})
959    {
960        $doc_obj->associate_file($image_filename, $self->{'css_assoc_files'}->{$image_filename}, undef, $section); 
961    }
962}
963
964sub add_file {
965    my $self = shift (@_);
966    my ($href, $rl, $hash_part, $base_dir, $doc_obj, $section) = @_;
967    my ($newname);
968
969    my $filename = $href;
970    if ($base_dir eq "") {
971    if ($ENV{'GSDLOS'} =~ m/^windows$/i) {
972        # remove http://
973        $filename =~ s/^[^:]*:\/\///;
974    }
975    else {
976        # remove http:/ thereby leaving one slash at the start as
977        # part of full pathname
978        $filename =~ s/^[^:]*:\///;
979    }
980    }
981    else {
982    # remove http://
983    $filename =~ s/^[^:]*:\/\///;
984    }
985   
986    if ($ENV{'GSDLOS'} =~ m/^windows$/i) {
987        $filename =~ s@\/@\\@g;
988    }
989   
990    $filename = &FileUtils::filenameConcatenate($base_dir, $filename);
991
992    if (($self->{'use_realistic_book'}) || ($self->{'old_style_HDL'})) {
993    # we are processing a tidytmp file - want paths to be in import
994    $filename =~ s/([\\\/])tidytmp([\\\/])/$1import$2/;
995    }
996
997    # Replace %XX's in URL with decoded value if required. Note that the
998    # filename may include the %XX in some situations. If the *original*
999    # file's name was in URL encoding, the following method will not decode
1000    # it.
1001    my $unicode_filename = $filename;
1002    my $opt_decode_unicode_filename = $self->opt_url_decode($unicode_filename);
1003
1004    # wvWare can generate <img src="StrangeNoGraphicData"> tags, but with no
1005    # (it seems) accompanying file
1006    if ($opt_decode_unicode_filename =~ m/StrangeNoGraphicData$/) { return ""; }
1007
1008    my $content_encoding= $self->{'content_encoding'} || "utf8";
1009
1010    if ($ENV{'GSDLOS'} =~ /^(linux|solaris)$/i) {
1011    # The filenames that come through the HTML file have been decoded
1012    # into Unicode aware Perl strings.  Need to convert them back
1013    # to their initial raw-byte encoding to match the file that
1014    # exists on the file system
1015    $filename = encode($content_encoding, $opt_decode_unicode_filename);
1016    }
1017    elsif ($ENV{'GSDLOS'} =~ /^darwin$/i) {
1018    # HFS+ is UTF8 with decompostion
1019    $filename = encode($content_encoding, $opt_decode_unicode_filename);
1020    $filename = normalize('D', $filename); # Normalization Form D (decomposition)
1021
1022    }
1023    elsif ($ENV{'GSDLOS'} =~ /^windows$/i) {
1024    my $long_filename = Win32::GetLongPathName($opt_decode_unicode_filename);
1025
1026    if (defined $long_filename) {
1027        my $short_filename = Win32::GetLongPathName($long_filename);
1028        $filename = $short_filename;
1029    }
1030#   else {
1031#       print STDERR "***** failed to map href to real file:\n";
1032#       print STDERR "****** $href -> $opt_decode_unicode_filename\n";
1033#   }
1034    }
1035    else {
1036    my $outhandle = $self->{'outhandle'};
1037    print $outhandle "Warning: Unrecognized operating system ", $ENV{'GSDLOS'}, "\n";
1038    print $outhandle "         in file system encoding of href: $href\n";
1039    print $outhandle "         No character encoding done.\n";
1040    }
1041
1042
1043    # some special processing if the intended filename was converted to utf8, but
1044    # the actual file still needs to be renamed
1045    #if (!&util::fd_exists($filename)) {
1046    if (!&FileUtils::fileExists($filename)) {
1047    # try the original filename stored in map
1048    if ((defined $ENV{"DEBUG_UNICODE"}) && ($ENV{"DEBUG_UNICODE"})) {
1049        print STDERR "******!! orig filename did not exist: $filename\n";
1050    }
1051
1052##  print STDERR "**** trying to look up unicode_filename: $unicode_filename\n";
1053
1054    my $original_filename = $self->{'unicode_to_original_filename'}->{$unicode_filename};
1055
1056    if ((defined $ENV{"DEBUG_UNICODE"}) && ($ENV{"DEBUG_UNICODE"})) {
1057        print STDERR "******   From lookup unicode_filename, now trying for: $original_filename\n";
1058    }
1059
1060    if (defined $original_filename && -e $original_filename) {
1061        if ((defined $ENV{"DEBUG_UNICODE"}) && ($ENV{"DEBUG_UNICODE"})) {
1062            print STDERR "******   Found match!\n";
1063        }
1064        $filename = $original_filename;
1065    }
1066    }
1067   
1068    my ($ext) = $filename =~ m/(\.[^\.]*)$/;
1069
1070    if ($rl == 0) {
1071    if ((!defined $ext) || ($ext !~ m/$self->{'assoc_files'}/)) {
1072        return "_httpextlink_&amp;rl=0&amp;el=prompt&amp;href=" . $href . $hash_part;
1073    }
1074    else {
1075        return "_httpextlink_&amp;rl=0&amp;el=direct&amp;href=" . $href . $hash_part;
1076    }
1077    }
1078
1079    if ((!defined $ext) || ($ext !~ m/$self->{'assoc_files'}/)) {
1080    return "_httpextlink_&amp;rl=" . $rl . "&amp;href=" . $href . $hash_part;
1081    }
1082    # add the original image file as a source file
1083    if (!$self->{'processing_tmp_files'} ) {
1084    $doc_obj->associate_source_file($filename);
1085    }
1086    if ($self->{'rename_assoc_files'}) {
1087    if (defined $self->{'aux_files'}->{$href}) {
1088        $newname = $self->{'aux_files'}->{$href}->{'dir_num'} . "/" .
1089        $self->{'aux_files'}->{$href}->{'file_num'} . $ext;
1090    } else {
1091        $newname = $self->{'dir_num'} . "/" . $self->{'file_num'} . $ext;
1092        $self->{'aux_files'}->{$href} = {'dir_num' => $self->{'dir_num'}, 'file_num' => $self->{'file_num'}};
1093        $self->inc_filecount ();
1094    }
1095    $doc_obj->associate_file($filename, $newname, undef, $section);
1096    return "_httpdocimg_/$newname";
1097    } else {
1098    if(&unicode::is_url_encoded($unicode_filename)) {
1099        # use the possibly-decoded filename instead to avoid double URL encoding
1100        ($newname) = $filename =~ m/([^\/\\]*)$/;
1101    } else {
1102        ($newname) = $unicode_filename =~ m/([^\/\\]*)$/;
1103    }
1104
1105    # Make sure this name uses only ASCII characters.
1106    # We use either base64 or URL encoding, as these preserve original encoding
1107    $newname = &util::rename_file($newname, $self->{'file_rename_method'});
1108
1109### print STDERR "***** associating $filename (raw-byte/utf8)-> $newname\n";
1110    $doc_obj->associate_file($filename, $newname, undef, $section);
1111
1112    # Since the generated image will be URL-encoded to avoid file-system/browser mess-ups
1113    # of filenames, URL-encode the additional percent signs of the URL-encoded filename
1114    my $newname_url = $newname;
1115    $newname_url = &unicode::filename_to_url($newname_url);
1116    return "_httpdocimg_/$newname_url";
1117    }
1118}
1119
1120
1121sub format_link {
1122    my $self = shift (@_);
1123    my ($link, $base_dir, $file) = @_;
1124
1125    # strip off hash part, e.g. #foo, but watch out for any entities, e.g. &#x3B1;
1126    my ($before_hash, $hash_part) = $link =~ m/^(.*?[^&])(\#.*)?$/;
1127   
1128    $hash_part = "" if !defined $hash_part;
1129    if (!defined $before_hash || $before_hash !~ m/[\w\.\/]/) {
1130        my $outhandle = $self->{'outhandle'};
1131        print $outhandle "HTMLPlugin: ERROR - badly formatted tag ignored ($link)\n"
1132            if $self->{'verbosity'};
1133        return ($link, "", 0);
1134    }
1135   
1136    if ($before_hash =~ s@^((?:http|https|ftp|file|mms)://)@@i) {
1137        my $type = $1;
1138        my $before_hash_file = $before_hash;
1139       
1140    if ($link =~ m/^(http|ftp):/i) {
1141       
1142        # Turn url (using /) into file name (possibly using \ on windows)
1143        my @http_dir_split = split('/', $before_hash_file);
1144        $before_hash_file = &FileUtils::filenameConcatenate(@http_dir_split);       
1145    }
1146   
1147    # want to maintain two version of "before_hash": one representing the URL, the other using filesystem specific directory separator
1148    $before_hash_file = $self->eval_dir_dots($before_hash_file);
1149    my $before_hash_url = $before_hash_file;
1150    if ($ENV{'GSDLOS'} =~ /^windows$/i) {
1151        $before_hash_url =~ s@\\@\/@g;
1152    }
1153   
1154    my $linkfilename = &FileUtils::filenameConcatenate($base_dir, $before_hash_file);
1155   
1156    my $rl = 0;
1157    $rl = 1 if (-e $linkfilename);
1158
1159    # make sure there's a slash on the end if it's a directory
1160    if ($before_hash_url !~ m/\/$/) {
1161        $before_hash_url .= "/" if (-d $linkfilename);
1162    }
1163    return ($type . $before_hash_url, $hash_part, $rl);
1164   
1165    } elsif ($link !~ m/^(mailto|news|gopher|nntp|telnet|javascript):/i && $link !~ m/^\//) {
1166
1167    if ($before_hash =~ s@^/@@ || $before_hash =~ m/\\/) {
1168
1169        # the first directory will be the domain name if file_is_url
1170        # to generate archives, otherwise we'll assume all files are
1171        # from the same site and base_dir is the root
1172
1173        if ($self->{'file_is_url'}) {
1174        my @dirs = split /[\/\\]/, $file;
1175        my $domname = shift (@dirs);
1176        $before_hash = &FileUtils::filenameConcatenate($domname, $before_hash);
1177        $before_hash =~ s@\\@/@g; # for windows
1178        }
1179        else
1180        {
1181        # see if link shares directory with source document
1182        # => turn into relative link if this is so!
1183       
1184        if ($ENV{'GSDLOS'} =~ m/^windows/i) {
1185            # too difficult doing a pattern match with embedded '\'s...
1186            my $win_before_hash=$before_hash;
1187            $win_before_hash =~ s@(\\)+@/@g;
1188            # $base_dir is already similarly "converted" on windows.
1189            if ($win_before_hash =~ s@^$base_dir/@@o) {
1190            # if this is true, we removed a prefix
1191            $before_hash=$win_before_hash;
1192            }
1193        }
1194        else {
1195            # before_hash has lost leading slash by this point,
1196            # -> add back in prior to substitution with $base_dir
1197            $before_hash = "/$before_hash";
1198
1199            $before_hash = &FileUtils::filenameConcatenate("",$before_hash);
1200            $before_hash =~ s@^$base_dir/@@;
1201        }
1202        }
1203    } else {
1204        # Turn relative file path into full path
1205        my $dirname = &File::Basename::dirname($file);
1206        $before_hash = &FileUtils::filenameConcatenate($dirname, $before_hash);
1207        $before_hash = $self->eval_dir_dots($before_hash);   
1208        $before_hash =~ s@\\@/@g; # for windows         
1209    }
1210
1211    my $linkfilename = &FileUtils::filenameConcatenate($base_dir, $before_hash);
1212
1213
1214#   print STDERR "**** linkfilename = $linkfilename\n";
1215#   if (!&util::fd_exists($linkfilename)) {
1216#       print STDERR "***** Warning: Could not find $linkfilename\n";
1217#   }
1218
1219
1220    # make sure there's a slash on the end if it's a directory
1221    if ($before_hash !~ m/\/$/) {
1222        $before_hash .= "/" if (-d $linkfilename);
1223    }
1224
1225#   print STDERR "*** returning: $before_hash\n";
1226
1227    return ("http://" . $before_hash, $hash_part, 1);
1228    } else {
1229    # mailto, news, nntp, telnet, javascript or gopher link
1230    return ($before_hash, "", 0);
1231    }
1232}
1233
1234sub extract_first_NNNN_characters {
1235    my $self = shift (@_);
1236    my ($textref, $doc_obj, $thissection) = @_;
1237   
1238    foreach my $size (split /,/, $self->{'first'}) {
1239    my $tmptext =  $$textref;
1240    # skip to the body
1241    $tmptext =~ s/.*<body[^>]*>//i;
1242    # remove javascript
1243    $tmptext =~ s@<script.*?</script>@ @sig;
1244    $tmptext =~ s/<[^>]*>/ /g;
1245    $tmptext =~ s/&nbsp;/ /g;
1246    $tmptext =~ s/^\s+//;
1247    $tmptext =~ s/\s+$//;
1248    $tmptext =~ s/\s+/ /gs;
1249    $tmptext = &unicode::substr ($tmptext, 0, $size);
1250    $tmptext =~ s/\s\S*$/&#8230;/; # adds an ellipse (...)
1251    $doc_obj->add_utf8_metadata ($thissection, "First$size", $tmptext);
1252    }
1253}
1254
1255
1256sub extract_metadata {
1257    my $self = shift (@_);
1258    my ($textref, $metadata, $doc_obj, $section) = @_;
1259    my $outhandle = $self->{'outhandle'};
1260    # if we don't want metadata, we may as well not be here ...
1261    return if (!defined $self->{'metadata_fields'});
1262   
1263    my $separator = $self->{'metadata_field_separator'};
1264    if ($separator eq "") {
1265    undef $separator;
1266    }
1267
1268    # metadata fields to extract/save. 'key' is the (lowercase) name of the
1269    # html meta, 'value' is the metadata name for greenstone to use
1270    my %find_fields = ();
1271
1272    my %creator_fields = (); # short-cut for lookups
1273
1274
1275    foreach my $field (split /,/, $self->{'metadata_fields'}) {
1276        $field =~ s/^\s+//; # remove leading whitespace
1277        $field =~ s/\s+$//; # remove trailing whitespace
1278   
1279    # support tag<tagname>
1280    if ($field =~ m/^(.*?)\s*<(.*?)>$/) {
1281        # "$2" is the user's preferred gs metadata name
1282        $find_fields{lc($1)}=$2; # lc = lowercase
1283    } else { # no <tagname> for mapping
1284        # "$field" is the user's preferred gs metadata name
1285        $find_fields{lc($field)}=$field; # lc = lowercase
1286    }
1287    }
1288
1289    if (defined $self->{'hunt_creator_metadata'} &&
1290    $self->{'hunt_creator_metadata'} == 1 ) {
1291    my @extra_fields =
1292        (
1293         'author',
1294         'author.email',
1295         'creator',
1296         'dc.creator',
1297         'dc.creator.corporatename',
1298         );
1299
1300    # add the creator_metadata fields to search for
1301    foreach my $field (@extra_fields) {
1302        $creator_fields{$field}=0; # add to lookup hash
1303    }
1304    }
1305
1306
1307    # find the header in the html file, which has the meta tags
1308    $$textref =~ m@<head>(.*?)</head>@si;
1309
1310    my $html_header=$1;
1311
1312    # go through every <meta... tag defined in the html and see if it is
1313    # one of the tags we want to match.
1314   
1315    # special case for title - we want to remember if its been found
1316    my $found_title = 0;
1317    # this assumes that ">" won't appear. (I don't think it's allowed to...)
1318    $html_header =~ m/^/; # match the start of the string, for \G assertion
1319   
1320    while ($html_header =~ m/\G.*?<meta(.*?)>/sig) {
1321    my $metatag=$1;
1322    my ($tag, $value);
1323
1324    # find the tag name
1325    $metatag =~ m/(?:name|http-equiv)\s*=\s*([\"\'])?(.*?)\1/is;
1326    $tag=$2;
1327    # in case they're not using " or ', but they should...
1328    if (! $tag) {
1329        $metatag =~ m/(?:name|http-equiv)\s*=\s*([^\s\>]+)/is;
1330        $tag=$1;
1331    }
1332
1333    if (!defined $tag) {
1334        print $outhandle "HTMLPlugin: can't find NAME in \"$metatag\"\n";
1335        next;
1336    }
1337   
1338    # don't need to assign this field if it was passed in from a previous
1339    # (recursive) plugin
1340    if (defined $metadata->{$tag}) {next}
1341
1342    # find the tag content
1343    $metatag =~ m/content\s*=\s*([\"\'])?(.*?)\1/is;
1344    $value=$2;
1345
1346    # The following code assigns the metaname to value if value is
1347    # empty. Why would we do this?
1348    #if (! $value) {
1349    #    $metatag =~ m/(?:name|http-equiv)\s*=\s*([^\s\>]+)/is;
1350    #    $value=$1;
1351    #}
1352    if (!defined $value || $value eq "") {
1353        print $outhandle "HTMLPlugin: can't find VALUE in <meta $metatag >\n" if ($self->{'verbosity'} > 2);
1354        next;
1355    }
1356   
1357    # clean up and add
1358    $value =~ s/\s+/ /gs;
1359    chomp($value); # remove trailing \n, if any
1360    if (exists $creator_fields{lc($tag)}) {
1361        # map this value onto greenstone's "Creator" metadata
1362        $tag='Creator';
1363    } elsif (!exists $find_fields{lc($tag)}) {
1364        next; # don't want this tag
1365    } else {
1366        # get the user's preferred capitalisation
1367        $tag = $find_fields{lc($tag)};
1368    }
1369    if (lc($tag) eq "title") {
1370        $found_title = 1;
1371    }
1372
1373    if ($self->{'verbosity'} > 2) {
1374        print $outhandle " extracted \"$tag\" metadata \"$value\"\n";
1375    }
1376
1377    if ($tag =~ /\./) {
1378        # there is a . so has a namespace, add ex.
1379        $tag = "ex.$tag";
1380    }
1381    if (defined $separator) {
1382        my @values = split($separator, $value);
1383        foreach my $v (@values) {
1384        $doc_obj->add_utf8_metadata($section, $tag, $v) if $v =~ /\S/;
1385        }
1386    }
1387    else {
1388        $doc_obj->add_utf8_metadata($section, $tag, $value);
1389    }
1390    }
1391   
1392    # TITLE: extract the document title
1393    if (exists $find_fields{'title'} && !$found_title) {
1394    # we want a title, and didn't find one in the meta tags
1395    # see if there's a <title> tag
1396    my $title;
1397    my $from = ""; # for debugging output only
1398    if ($html_header =~ m/<title[^>]*>([^<]+)<\/title[^>]*>/is) {
1399        $title = $1;
1400        $from = "<title> tags";
1401    }
1402
1403    if (!defined $title) {
1404        $from = "first 100 chars";
1405        # if no title use first 100 or so characters
1406        $title = $$textref;
1407        $title =~ s/^\xFE\xFF//; # Remove unicode byte order mark       
1408        $title =~ s/^.*?<body>//si;
1409        # ignore javascript!
1410        $title =~ s@<script.*?</script>@ @sig;
1411        $title =~ s/<\/([^>]+)><\1>//g; # (eg) </b><b> - no space
1412        $title =~ s/<[^>]*>/ /g; # remove all HTML tags
1413        $title =~ s@\r@@g; # remove Windows carriage returns to ensure that titles of pdftohtml docs are consistent (the same 100 chars) across windows and linux
1414        $title = substr ($title, 0, 100);
1415        $title =~ s/\s\S*$/.../;
1416    }
1417    $title =~ s/<[^>]*>/ /g; # remove html tags
1418    $title =~ s/&nbsp;/ /g;
1419    $title =~ s/(?:&nbsp;|\xc2\xa0)/ /g; # utf-8 for nbsp...
1420    $title =~ s/\s+/ /gs; # collapse multiple spaces
1421    $title =~ s/^\s*//;   # remove leading spaces
1422    $title =~ s/\s*$//;   # remove trailing spaces
1423
1424    $title =~ s/^$self->{'title_sub'}// if ($self->{'title_sub'});
1425    $title =~ s/^\s+//s; # in case title_sub introduced any...
1426    $doc_obj->add_utf8_metadata ($section, "Title", $title);
1427    print $outhandle " extracted Title metadata \"$title\" from $from\n"
1428        if ($self->{'verbosity'} > 2);
1429    }
1430   
1431    # add FileFormat metadata
1432    $doc_obj->add_metadata($section,"FileFormat", "HTML");
1433
1434    # Special, for metadata names such as tagH1 - extracts
1435    # the text between the first <H1> and </H1> tags into "H1" metadata.
1436
1437    foreach my $field (keys %find_fields) {
1438    if ($field !~ m/^tag([a-z0-9]+)$/i) {next}
1439    my $tag = $1;
1440    if ($$textref =~ m@<$tag[^>]*>(.*?)</$tag[^>]*>@g) {
1441        my $content = $1;
1442        $content =~ s/&nbsp;/ /g;
1443        $content =~ s/<[^>]*>/ /g;
1444        $content =~ s/^\s+//;
1445        $content =~ s/\s+$//;
1446        $content =~ s/\s+/ /gs;
1447        if ($content) {
1448        $tag=$find_fields{"tag$tag"}; # get the user's capitalisation
1449        $tag =~ s/^tag//i;
1450        $doc_obj->add_utf8_metadata ($section, $tag, $content);
1451        print $outhandle " extracted \"$tag\" metadata \"$content\"\n"
1452            if ($self->{'verbosity'} > 2);
1453        }
1454    }
1455    }   
1456}
1457
1458
1459# evaluate any "../" to next directory up
1460# evaluate any "./" as here
1461sub eval_dir_dots {
1462    my $self = shift (@_);
1463    my ($filename) = @_;
1464    my $dirsep_os = &util::get_os_dirsep();
1465    my @dirsep = split(/$dirsep_os/,$filename);
1466
1467    my @eval_dirs = ();
1468    foreach my $d (@dirsep) {
1469    if ($d eq "..") {
1470        pop(@eval_dirs);
1471       
1472    } elsif ($d eq ".") {
1473        # do nothing!
1474
1475    } else {
1476        push(@eval_dirs,$d);
1477    }
1478    }
1479
1480    # Need to fiddle with number of elements in @eval_dirs if the
1481    # first one is the empty string.  This is because of a
1482    # modification to FileUtils::filenameConcatenate that supresses the addition
1483    # of a leading '/' character (or \ if windows) (intended to help
1484    # filename cat with relative paths) if the first entry in the
1485    # array is the empty string.  Making the array start with *two*
1486    # empty strings is a way to defeat this "smart" option.
1487    #
1488    if (scalar(@eval_dirs) > 0) {
1489    if ($eval_dirs[0] eq ""){
1490        unshift(@eval_dirs,"");
1491    }
1492    }
1493   
1494    my $evaluated_filename = (scalar @eval_dirs > 0) ? &FileUtils::filenameConcatenate(@eval_dirs) : "";
1495    return $evaluated_filename;
1496}
1497
1498sub replace_usemap_links {
1499    my $self = shift (@_);
1500    my ($front, $link, $back) = @_;
1501
1502    # remove quotes from link at start and end if necessary
1503    if ($link=~/^[\"\']/) {
1504    $link=~s/^[\"\']//;
1505    $link=~s/[\"\']$//;
1506    $front.='"';
1507    $back="\"$back";
1508    }
1509
1510    $link =~ s/^\.\///;
1511    return $front . $link . $back;
1512}
1513
1514sub inc_filecount {
1515    my $self = shift (@_);
1516
1517    if ($self->{'file_num'} == 1000) {
1518    $self->{'dir_num'} ++;
1519    $self->{'file_num'} = 0;
1520    } else {
1521    $self->{'file_num'} ++;
1522    }
1523}
1524
1525
1526# Extend read_file so that strings like &eacute; are
1527# converted to UTF8 internally. 
1528#
1529# We don't convert &lt; or &gt; or &amp; or &quot; in case
1530# they interfere with the GML files
1531
1532sub read_file {
1533    my $self = shift(@_);
1534    my ($filename, $encoding, $language, $textref) = @_;
1535
1536    $self->SUPER::read_file($filename, $encoding, $language, $textref);
1537
1538    # Convert entities to their Unicode code-point equivalents
1539    $$textref =~ s/&(lt|gt|amp|quot|nbsp);/&z$1;/go;
1540    $$textref =~ s/&([^;]+);/&ghtml::getcharequiv($1,1,1)/gseo;
1541    $$textref =~ s/&z(lt|gt|amp|quot|nbsp);/&$1;/go;
1542
1543}
1544
1545sub HB_read_html_file {
1546    my $self = shift (@_);
1547    my ($htmlfile, $text) = @_;
1548   
1549    # load in the file
1550    if (!open (FILE, $htmlfile)) {
1551    print STDERR "ERROR - could not open $htmlfile\n";
1552    return;
1553    }
1554
1555    my $foundbody = 0;
1556    $self->HB_gettext (\$foundbody, $text, "FILE");
1557    close FILE;
1558   
1559    # just in case there was no <body> tag
1560    if (!$foundbody) {
1561    $foundbody = 1;
1562    open (FILE, $htmlfile) || return;
1563    $self->HB_gettext (\$foundbody, $text, "FILE");
1564    close FILE;
1565    }
1566    # text is in utf8
1567}       
1568
1569# converts the text to utf8, as ghtml does that for &eacute; etc.
1570sub HB_gettext {
1571    my $self = shift (@_);
1572    my ($foundbody, $text, $handle) = @_;
1573   
1574    my $line = "";
1575    while (defined ($line = <$handle>)) {
1576    # look for body tag
1577    if (!$$foundbody) {
1578        if ($line =~ s/^.*<body[^>]*>//i) {
1579        $$foundbody = 1;
1580        } else {
1581        next;
1582        }
1583    }
1584   
1585    # check for symbol fonts
1586    if ($line =~ m/<font [^>]*?face\s*=\s*\"?(\w+)\"?/i) {
1587        my $font = $1;
1588        print STDERR "HBPlug::HB_gettext - warning removed font $font\n"
1589        if ($font !~ m/^arial$/i);
1590    }
1591
1592    $$text .= $line;
1593    }
1594
1595    if ($self->{'input_encoding'} eq "iso_8859_1") {
1596    # convert to utf-8
1597    $$text=&unicode::unicode2utf8(&unicode::convert2unicode("iso_8859_1", $text));
1598    }
1599    # convert any alphanumeric character entities to their utf-8
1600    # equivalent for indexing purposes
1601    #&ghtml::convertcharentities ($$text);
1602
1603    $$text =~ s/\s+/ /g; # remove \n's
1604
1605    # At this point $$text is a binary byte string
1606    # => turn it into a Unicode aware string, so full
1607    # Unicode aware pattern matching can be used.
1608    # For instance: 's/\x{0101}//g' or '[[:upper:]]'
1609    #
1610
1611    $$text = decode("utf8",$$text);
1612}
1613
1614sub HB_clean_section {
1615    my $self = shift (@_);
1616    my ($section) = @_;
1617
1618    # remove tags without a starting tag from the section
1619    my ($tag, $tagstart);
1620    while ($section =~ m/<\/([^>]{1,10})>/) {
1621    $tag = $1;
1622    $tagstart = index($section, "<$tag");
1623    last if (($tagstart >= 0) && ($tagstart < index($section, "<\/$tag")));
1624    $section =~ s/<\/$tag>//;
1625    }
1626   
1627    # remove extra paragraph tags
1628    while ($section =~ s/<p\b[^>]*>\s*<p\b/<p/ig) {}
1629   
1630    # remove extra stuff at the end of the section
1631    while ($section =~ s/(<u>|<i>|<b>|<p\b[^>]*>|&nbsp;|\s)$//i) {}
1632   
1633    # add a newline at the beginning of each paragraph
1634    $section =~ s/(.)\s*<p\b/$1\n\n<p/gi;
1635   
1636    # add a newline every 80 characters at a word boundary
1637    # Note: this regular expression puts a line feed before
1638    # the last word in each section, even when it is not
1639    # needed.
1640    $section =~ s/(.{1,80})\s/$1\n/g;
1641   
1642    # fix up the image links
1643    $section =~ s/<img[^>]*?src=\"?([^\">]+)\"?[^>]*>/
1644    <center><img src=\"$1\" \/><\/center><br\/>/ig;
1645    $section =~ s/&lt;&lt;I&gt;&gt;\s*([^\.]+\.(png|jpg|gif))/
1646    <center><img src=\"$1\" \/><\/center><br\/>/ig;
1647
1648    return $section;
1649}
1650
1651# Will convert the oldHDL format to the new HDL format (using the Section tag) 
1652sub convert_to_newHDLformat
1653{
1654    my $self = shift (@_);
1655    my ($file,$cnfile) = @_;
1656    my $input_filename = $file;
1657    my $tmp_filename = $cnfile;
1658   
1659    # write HTML tmp file with new HDL format
1660    open (PROD, ">$tmp_filename") || die("Error Writing to File: $tmp_filename $!");
1661   
1662    # read in the file and do basic html cleaning (removing header etc)
1663    my $html = "";
1664    $self->HB_read_html_file ($input_filename, \$html);
1665   
1666    # process the file one section at a time
1667    my $curtoclevel = 1;
1668    my $firstsection = 1;
1669    my $toclevel = 0;
1670    while (length ($html) > 0) {
1671    if ($html =~ s/^.*?(?:<p\b[^>]*>)?((<b>|<i>|<u>|\s)*)&lt;&lt;TOC(\d+)&gt;&gt;\s*(.*?)<p\b/<p/i) {
1672        $toclevel = $3;
1673        my $title = $4;
1674        my $sectiontext = "";
1675        if ($html =~ s/^(.*?)((?:<p\b[^>]*>)?((<b>|<i>|<u>|\s)*)&lt;&lt;TOC\d+&gt;&gt;)/$2/i) {
1676        $sectiontext = $1;
1677        } else {
1678        $sectiontext = $html;
1679        $html = "";
1680        }
1681
1682        # remove tags and extra spaces from the title
1683        $title =~ s/<\/?[^>]+>//g;
1684        $title =~ s/^\s+|\s+$//g;
1685
1686        # close any sections below the current level and
1687        # create a new section (special case for the firstsection)
1688        print PROD "<!--\n";
1689        while (($curtoclevel > $toclevel) ||
1690           (!$firstsection && $curtoclevel == $toclevel)) {
1691        $curtoclevel--;
1692        print PROD "</Section>\n";
1693        }
1694        if ($curtoclevel+1 < $toclevel) {
1695        print STDERR "WARNING - jump in toc levels in $input_filename " .
1696            "from $curtoclevel to $toclevel\n";
1697        }
1698        while ($curtoclevel < $toclevel) {
1699        $curtoclevel++;
1700        }
1701
1702        if ($curtoclevel == 1) {
1703            # add the header tag
1704        print PROD "-->\n";
1705            print PROD "<HTML>\n<HEAD>\n<TITLE>$title</TITLE>\n</HEAD>\n<BODY>\n";
1706        print PROD "<!--\n";
1707        }
1708       
1709        print PROD "<Section>\n\t<Description>\n\t\t<Metadata name=\"Title\">$title</Metadata>\n\t</Description>\n";
1710       
1711        print PROD "-->\n";
1712       
1713        # clean up the section html
1714        $sectiontext = $self->HB_clean_section($sectiontext);
1715
1716        print PROD "$sectiontext\n";     
1717
1718    } else {
1719        print STDERR "WARNING - leftover text\n" , $self->shorten($html),
1720        "\nin $input_filename\n";
1721        last;
1722    }
1723    $firstsection = 0;
1724    }
1725   
1726    print PROD "<!--\n";
1727    while ($curtoclevel > 0) {
1728    $curtoclevel--;
1729    print PROD "</Section>\n";
1730    }
1731    print PROD "-->\n";
1732   
1733    close (PROD) || die("Error Closing File: $tmp_filename $!");
1734   
1735    return $tmp_filename;
1736}       
1737
1738sub shorten {
1739    my $self = shift (@_);
1740    my ($text) = @_;
1741
1742    return "\"$text\"" if (length($text) < 100);
1743
1744    return "\"" . substr ($text, 0, 50) . "\" ... \"" .
1745    substr ($text, length($text)-50) . "\"";
1746}
1747
1748sub convert_tidy_or_oldHDL_file
1749{
1750    my $self = shift (@_);
1751    my ($file) = @_;
1752    my $input_filename = $file;
1753   
1754    if (-d $input_filename)
1755    {
1756        return $input_filename;
1757    }
1758   
1759    # get the input filename
1760    my ($tailname, $dirname, $suffix) = &File::Basename::fileparse($input_filename, "\\.[^\\.]+\$");
1761    my $base_dirname = $dirname;
1762    $suffix = lc($suffix);   
1763   
1764    # derive tmp filename from input filename
1765    # Remove any white space from filename -- no risk of name collision, and
1766    # makes later conversion by utils simpler. Leave spaces in path...
1767    # tidy up the filename with space, dot, hyphen between
1768    $tailname =~ s/\s+//g;
1769    $tailname =~ s/\.+//g;
1770    $tailname =~ s/\-+//g;
1771    # convert to utf-8 otherwise we have problems with the doc.xml file
1772    # later on
1773    &unicode::ensure_utf8(\$tailname);
1774   
1775    # softlink to collection tmp dir
1776    my $tmp_dirname = &FileUtils::filenameConcatenate($ENV{'GSDLCOLLECTDIR'}, "tidytmp");
1777    &FileUtils::makeDirectory($tmp_dirname) if (!-e $tmp_dirname);
1778   
1779    my $test_dirname = "";
1780    my $f_separator = &util::get_os_dirsep();
1781   
1782    if ($dirname =~ m/import$f_separator/)
1783    {
1784        $test_dirname = $'; #'
1785   
1786    #print STDERR "init $'\n";
1787   
1788    while ($test_dirname =~ m/[$f_separator]/)
1789    {
1790        my $folderdirname = $`;
1791        $tmp_dirname = &FileUtils::filenameConcatenate($tmp_dirname,$folderdirname);
1792        &FileUtils::makeDirectory($tmp_dirname) if (!-e $tmp_dirname);
1793        $test_dirname = $'; #'
1794    }
1795    }
1796   
1797    my $tmp_filename = &FileUtils::filenameConcatenate($tmp_dirname, "$tailname$suffix");
1798   
1799    # tidy or convert the input file if it is a HTML-like file or it is accepted by the process_exp
1800    if (($suffix eq ".htm") || ($suffix eq ".html") || ($suffix eq ".shtml"))
1801    {   
1802        #convert the input file to a new style HDL
1803        my $hdl_output_filename = $input_filename;
1804        if ($self->{'old_style_HDL'})
1805        {
1806        $hdl_output_filename = &FileUtils::filenameConcatenate($tmp_dirname, "$tailname$suffix");
1807        $hdl_output_filename = $self->convert_to_newHDLformat($input_filename,$hdl_output_filename);
1808        }
1809   
1810    #just for checking copy all other file from the base dir to tmp dir if it is not exists
1811    opendir(DIR,$base_dirname) or die "Can't open base directory : $base_dirname!";
1812    my @files = grep {!/^\.+$/} readdir(DIR);
1813    close(DIR);
1814
1815    foreach my $file (@files)
1816    {
1817        my $src_file = &FileUtils::filenameConcatenate($base_dirname,$file);
1818        my $dest_file = &FileUtils::filenameConcatenate($tmp_dirname,$file);
1819        if ((!-e $dest_file) && (!-d $src_file))
1820        {
1821        # just copy the original file back to the tmp directory
1822        copy($src_file,$dest_file) or die "Can't copy file $src_file to $dest_file $!";
1823        }
1824    }
1825   
1826    # tidy the input file
1827    my $tidy_output_filename = $hdl_output_filename;
1828    if ($self->{'use_realistic_book'})
1829    {
1830        $tidy_output_filename = &FileUtils::filenameConcatenate($tmp_dirname, "$tailname$suffix");
1831        $tidy_output_filename = $self->tmp_tidy_file($hdl_output_filename,$tidy_output_filename);
1832    }
1833    $tmp_filename = $tidy_output_filename;
1834    }
1835    else
1836    {
1837        if (!-e $tmp_filename)
1838    {
1839        # just copy the original file back to the tmp directory
1840        copy($input_filename,$tmp_filename) or die "Can't copy file $input_filename to $tmp_filename $!";
1841    }
1842    }
1843   
1844    return $tmp_filename;
1845}
1846
1847
1848# Will make the html input file as a proper XML file with removed font tag and
1849# image size added to the img tag.
1850# The tidying process takes place in a collection specific 'tmp' directory so
1851# that we don't accidentally damage the input.
1852sub tmp_tidy_file
1853{
1854    my $self = shift (@_);
1855    my ($file,$cnfile) = @_;
1856    my $input_filename = $file;
1857    my $tmp_filename = $cnfile;
1858   
1859    # get the input filename
1860    my ($tailname, $dirname, $suffix) = &File::Basename::fileparse($input_filename, "\\.[^\\.]+\$");
1861
1862    require HTML::TokeParser::Simple;
1863   
1864    # create HTML parser to decode the input file
1865    my $parser = HTML::TokeParser::Simple->new($input_filename);
1866
1867    # write HTML tmp file without the font tag and image size are added to the img tag
1868    open (PROD, ">$tmp_filename") || die("Error Writing to File: $tmp_filename $!");
1869    while (my $token = $parser->get_token())
1870    {
1871    # is it an img tag
1872    if ($token->is_start_tag('img'))
1873    {
1874        # get the attributes
1875        my $attr = $token->return_attr;
1876
1877        # get the full path to the image
1878        my $img_file = &FileUtils::filenameConcatenate($dirname,$attr->{src});
1879       
1880        # set the width and height attribute
1881        ($attr->{width}, $attr->{height}) = imgsize($img_file);
1882
1883        # recreate the tag
1884        print PROD "<img";
1885        print PROD map { qq { $_="$attr->{$_}"} } keys %$attr;
1886        print PROD ">";
1887    }
1888    # is it a font tag
1889    else
1890    {
1891        if (($token->is_start_tag('font')) || ($token->is_end_tag('font')))
1892        {
1893        # remove font tag
1894        print PROD "";
1895        }
1896        else
1897        {
1898        # print without changes
1899        print PROD $token->as_is;
1900        }
1901    }
1902    }
1903    close (PROD) || die("Error Closing File: $tmp_filename $!");
1904
1905    # run html-tidy on the tmp file to make it a proper XML file
1906
1907    my $outhandle = $self->{'outhandle'};
1908    print $outhandle "Converting HTML to be XML compliant:\n";
1909
1910    my $tidy_cmd = "tidy";
1911    $tidy_cmd .= " -q" if ($self->{'verbosity'} <= 2);
1912    $tidy_cmd .= " -raw -wrap 0 -asxml \"$tmp_filename\"";
1913    if ($self->{'verbosity'} <= 2) {
1914    if ($ENV{'GSDLOS'} =~ m/^windows/i) {
1915        $tidy_cmd .= " 2>nul";
1916    }
1917    else {
1918        $tidy_cmd .= " 2>/dev/null";
1919    }
1920    print $outhandle "  => $tidy_cmd\n";
1921    }
1922
1923    my $tidyfile = `$tidy_cmd`;
1924
1925    # write result back to the tmp file
1926    open (PROD, ">$tmp_filename") || die("Error Writing to File: $tmp_filename $!");
1927    print PROD $tidyfile;
1928    close (PROD) || die("Error Closing File: $tmp_filename $!");
1929   
1930    # return the output filename
1931    return $tmp_filename;
1932}
1933
1934sub associate_cover_image
1935{
1936    my $self = shift(@_);
1937    my ($doc_obj, $filename) = @_;
1938    if (($self->{'use_realistic_book'}) || ($self->{'old_style_HDL'}))
1939    {
1940    # we will have cover image in tidytmp, but want it from import
1941    $filename =~ s/([\\\/])tidytmp([\\\/])/$1import$2/;
1942    }
1943    $self->SUPER::associate_cover_image($doc_obj, $filename);
1944}
1945
1946   
19471;
Note: See TracBrowser for help on using the browser.