1 | ###########################################################################
|
---|
2 | #
|
---|
3 | # HTMLPlugin.pm -- basic html plugin
|
---|
4 | #
|
---|
5 | # A component of the Greenstone digital library software
|
---|
6 | # from the New Zealand Digital Library Project at the
|
---|
7 | # University of Waikato, New Zealand.
|
---|
8 | #
|
---|
9 | # Copyright (C) 1999 New Zealand Digital Library Project
|
---|
10 | #
|
---|
11 | # This program is free software; you can redistribute it and/or modify
|
---|
12 | # it under the terms of the GNU General Public License as published by
|
---|
13 | # the Free Software Foundation; either version 2 of the License, or
|
---|
14 | # (at your option) any later version.
|
---|
15 | #
|
---|
16 | # This program is distributed in the hope that it will be useful,
|
---|
17 | # but WITHOUT ANY WARRANTY; without even the implied warranty of
|
---|
18 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
---|
19 | # GNU General Public License for more details.
|
---|
20 | #
|
---|
21 | # You should have received a copy of the GNU General Public License
|
---|
22 | # along with this program; if not, write to the Free Software
|
---|
23 | # Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
|
---|
24 | #
|
---|
25 | ###########################################################################
|
---|
26 |
|
---|
27 | #
|
---|
28 | # Note that this plugin handles frames only in a very simple way
|
---|
29 | # i.e. each frame is treated as a separate document. This means
|
---|
30 | # search results will contain links to individual frames rather
|
---|
31 | # than linking to the top level frameset.
|
---|
32 | # There may also be some problems caused by the _parent target
|
---|
33 | # (it's removed by this plugin)
|
---|
34 | #
|
---|
35 |
|
---|
36 | package HTMLPlugin;
|
---|
37 |
|
---|
38 | use ReadTextFile;
|
---|
39 | use HBPlugin;
|
---|
40 | use ghtml;
|
---|
41 | use unicode;
|
---|
42 | use util;
|
---|
43 | use XMLParser;
|
---|
44 |
|
---|
45 | use Image::Size;
|
---|
46 | use File::Copy;
|
---|
47 |
|
---|
48 | sub BEGIN {
|
---|
49 | @HTMLPlugin::ISA = ('ReadTextFile', 'HBPlugin');
|
---|
50 | }
|
---|
51 |
|
---|
52 | use strict; # every perl program should have this!
|
---|
53 | no strict 'refs'; # make an exception so we can use variables as filehandles
|
---|
54 |
|
---|
55 | my $arguments =
|
---|
56 | [ { 'name' => "process_exp",
|
---|
57 | 'desc' => "{BasePlugin.process_exp}",
|
---|
58 | 'type' => "regexp",
|
---|
59 | 'deft' => &get_default_process_exp() },
|
---|
60 | { 'name' => "block_exp",
|
---|
61 | 'desc' => "{BasePlugin.block_exp}",
|
---|
62 | 'type' => 'regexp',
|
---|
63 | 'deft' => &get_default_block_exp() },
|
---|
64 | { 'name' => "nolinks",
|
---|
65 | 'desc' => "{HTMLPlugin.nolinks}",
|
---|
66 | 'type' => "flag" },
|
---|
67 | { 'name' => "keep_head",
|
---|
68 | 'desc' => "{HTMLPlugin.keep_head}",
|
---|
69 | 'type' => "flag" },
|
---|
70 | { 'name' => "no_metadata",
|
---|
71 | 'desc' => "{HTMLPlugin.no_metadata}",
|
---|
72 | 'type' => "flag" },
|
---|
73 | { 'name' => "metadata_fields",
|
---|
74 | 'desc' => "{HTMLPlugin.metadata_fields}",
|
---|
75 | 'type' => "string",
|
---|
76 | 'deft' => "Title" },
|
---|
77 | { 'name' => "metadata_field_separator",
|
---|
78 | 'desc' => "{HTMLPlugin.metadata_field_separator}",
|
---|
79 | 'type' => "string",
|
---|
80 | 'deft' => "" },
|
---|
81 | { 'name' => "hunt_creator_metadata",
|
---|
82 | 'desc' => "{HTMLPlugin.hunt_creator_metadata}",
|
---|
83 | 'type' => "flag" },
|
---|
84 | { 'name' => "file_is_url",
|
---|
85 | 'desc' => "{HTMLPlugin.file_is_url}",
|
---|
86 | 'type' => "flag" },
|
---|
87 | { 'name' => "assoc_files",
|
---|
88 | 'desc' => "{HTMLPlugin.assoc_files}",
|
---|
89 | 'type' => "regexp",
|
---|
90 | 'deft' => &get_default_block_exp() },
|
---|
91 | { 'name' => "rename_assoc_files",
|
---|
92 | 'desc' => "{HTMLPlugin.rename_assoc_files}",
|
---|
93 | 'type' => "flag" },
|
---|
94 | { 'name' => "title_sub",
|
---|
95 | 'desc' => "{HTMLPlugin.title_sub}",
|
---|
96 | 'type' => "string",
|
---|
97 | 'deft' => "" },
|
---|
98 | { 'name' => "description_tags",
|
---|
99 | 'desc' => "{HTMLPlugin.description_tags}",
|
---|
100 | 'type' => "flag" },
|
---|
101 | # retain this for backward compatibility (w3mir option was replaced by
|
---|
102 | # file_is_url)
|
---|
103 | { 'name' => "w3mir",
|
---|
104 | # 'desc' => "{HTMLPlugin.w3mir}",
|
---|
105 | 'type' => "flag",
|
---|
106 | 'hiddengli' => "yes"},
|
---|
107 | { 'name' => "no_strip_metadata_html",
|
---|
108 | 'desc' => "{HTMLPlugin.no_strip_metadata_html}",
|
---|
109 | 'type' => "string",
|
---|
110 | 'deft' => "",
|
---|
111 | 'reqd' => "no"},
|
---|
112 | { 'name' => "sectionalise_using_h_tags",
|
---|
113 | 'desc' => "{HTMLPlugin.sectionalise_using_h_tags}",
|
---|
114 | 'type' => "flag" },
|
---|
115 | { 'name' => "use_realistic_book",
|
---|
116 | 'desc' => "{HTMLPlugin.tidy_html}",
|
---|
117 | 'type' => "flag"},
|
---|
118 | { 'name' => "old_style_HDL",
|
---|
119 | 'desc' => "{HTMLPlugin.old_style_HDL}",
|
---|
120 | 'type' => "flag"},
|
---|
121 | {'name' => "processing_tmp_files",
|
---|
122 | 'desc' => "{BasePlugin.processing_tmp_files}",
|
---|
123 | 'type' => "flag",
|
---|
124 | 'hiddengli' => "yes"}
|
---|
125 | ];
|
---|
126 |
|
---|
127 | my $options = { 'name' => "HTMLPlugin",
|
---|
128 | 'desc' => "{HTMLPlugin.desc}",
|
---|
129 | 'abstract' => "no",
|
---|
130 | 'inherits' => "yes",
|
---|
131 | 'args' => $arguments };
|
---|
132 |
|
---|
133 |
|
---|
134 | sub new {
|
---|
135 | my ($class) = shift (@_);
|
---|
136 | my ($pluginlist,$inputargs,$hashArgOptLists) = @_;
|
---|
137 | push(@$pluginlist, $class);
|
---|
138 |
|
---|
139 | push(@{$hashArgOptLists->{"ArgList"}},@{$arguments});
|
---|
140 | push(@{$hashArgOptLists->{"OptList"}},$options);
|
---|
141 |
|
---|
142 |
|
---|
143 | my $self = new ReadTextFile($pluginlist,$inputargs,$hashArgOptLists);
|
---|
144 |
|
---|
145 | if ($self->{'w3mir'}) {
|
---|
146 | $self->{'file_is_url'} = 1;
|
---|
147 | }
|
---|
148 | $self->{'aux_files'} = {};
|
---|
149 | $self->{'dir_num'} = 0;
|
---|
150 | $self->{'file_num'} = 0;
|
---|
151 |
|
---|
152 | return bless $self, $class;
|
---|
153 | }
|
---|
154 |
|
---|
155 | # may want to use (?i)\.(gif|jpe?g|jpe|png|css|js(?:@.*)?)$
|
---|
156 | # if have eg <script language="javascript" src="img/lib.js@123">
|
---|
157 | # blocking is now done by reading through the file and recording all the
|
---|
158 | # images and other files
|
---|
159 | sub get_default_block_exp {
|
---|
160 | my $self = shift (@_);
|
---|
161 |
|
---|
162 | #return q^(?i)\.(gif|jpe?g|jpe|jpg|png|css)$^;
|
---|
163 | return "";
|
---|
164 | }
|
---|
165 |
|
---|
166 | sub get_default_process_exp {
|
---|
167 | my $self = shift (@_);
|
---|
168 |
|
---|
169 | # the last option is an attempt to encode the concept of an html query ...
|
---|
170 | return q^(?i)(\.html?|\.shtml|\.shm|\.asp|\.php\d?|\.cgi|.+\?.+=.*)$^;
|
---|
171 | }
|
---|
172 |
|
---|
173 | sub store_block_files
|
---|
174 | {
|
---|
175 | my $self =shift (@_);
|
---|
176 | my ($filename_full_path, $block_hash) = @_;
|
---|
177 |
|
---|
178 | my $html_fname = $filename_full_path;
|
---|
179 | my @file_blocks;
|
---|
180 |
|
---|
181 | my ($language, $encoding) = $self->textcat_get_language_encoding ($filename_full_path);
|
---|
182 |
|
---|
183 | # read in file ($text will be in utf8)
|
---|
184 | my $raw_text = "";
|
---|
185 | $self->read_file_no_decoding ($filename_full_path, \$raw_text);
|
---|
186 |
|
---|
187 | my $textref = \$raw_text;
|
---|
188 | my $opencom = '(?:<!--|<!(?:—|—|--))';
|
---|
189 | my $closecom = '(?:-->|(?:—|—|--)>)';
|
---|
190 | $$textref =~ s/$opencom(.*?)$closecom//gs;
|
---|
191 |
|
---|
192 | my $attval = "\\\"[^\\\"]+\\\"|[^\\s>]+";
|
---|
193 | my @img_matches = ($$textref =~ m/<img[^>]*?src\s*=\s*($attval)[^>]*>/igs);
|
---|
194 | my @usemap_matches = ($$textref =~ m/<img[^>]*?usemap\s*=\s*($attval)[^>]*>/igs);
|
---|
195 | my @link_matches = ($$textref =~ m/<link[^>]*?href\s*=\s*($attval)[^>]*>/igs);
|
---|
196 | my @embed_matches = ($$textref =~ m/<embed[^>]*?src\s*=\s*($attval)[^>]*>/igs);
|
---|
197 | my @tabbg_matches = ($$textref =~ m/<(?:body|table|tr|td)[^>]*?background\s*=\s*($attval)[^>]*>/igs);
|
---|
198 | my @script_matches = ($$textref =~ m/<script[^>]*?src\s*=\s*($attval)[^>]*>/igs);
|
---|
199 |
|
---|
200 | if(!defined $self->{'utf8_to_original_filename'}) {
|
---|
201 | # maps from utf8 converted link name -> original filename referrred to by (possibly URL-encoded) src url
|
---|
202 | $self->{'utf8_to_original_filename'} = {};
|
---|
203 | }
|
---|
204 |
|
---|
205 | foreach my $link (@img_matches, @usemap_matches, @link_matches, @embed_matches, @tabbg_matches, @script_matches) {
|
---|
206 |
|
---|
207 | # remove quotes from link at start and end if necessary
|
---|
208 | if ($link=~/^\"/) {
|
---|
209 | $link=~s/^\"//;
|
---|
210 | $link=~s/\"$//;
|
---|
211 | }
|
---|
212 |
|
---|
213 | $link =~ s/\#.*$//s; # remove any anchor names, e.g. foo.html#name becomes foo.html
|
---|
214 | # some links may just be anchor names
|
---|
215 | next unless ($link =~ /\S+/);
|
---|
216 |
|
---|
217 | if ($link !~ m@^/@ && $link !~ m/^([A-Z]:?)\\/) {
|
---|
218 | # Turn relative file path into full path
|
---|
219 | my $dirname = &File::Basename::dirname($filename_full_path);
|
---|
220 | $link = &util::filename_cat($dirname, $link);
|
---|
221 | }
|
---|
222 | $link = $self->eval_dir_dots($link);
|
---|
223 |
|
---|
224 | # this is the actual filename on the filesystem (that the link refers to)
|
---|
225 | my $url_original_filename = $self->opt_url_decode($link);
|
---|
226 |
|
---|
227 | # Convert the url_original_filename into its utf8 version. Store the utf8 link along with the url_original_filename
|
---|
228 | my $utf8_link = "";
|
---|
229 | $self->decode_text($link,$encoding,$language,\$utf8_link);
|
---|
230 |
|
---|
231 | $self->{'utf8_to_original_filename'}->{$utf8_link} = $url_original_filename;
|
---|
232 | # print STDERR "**** utf8_encoded_link to original src filename:\n\t$utf8_link\n\t".$self->{'utf8_to_original_filename'}->{$utf8_link}."\n";
|
---|
233 |
|
---|
234 | if ($url_original_filename ne $utf8_link) {
|
---|
235 | my $outhandle = $self->{'outhandle'};
|
---|
236 |
|
---|
237 | print $outhandle "URL Encoding $url_original_filename\n";
|
---|
238 | print $outhandle " ->$utf8_link\n";
|
---|
239 | }
|
---|
240 |
|
---|
241 | $block_hash->{'file_blocks'}->{$url_original_filename} = 1;
|
---|
242 | }
|
---|
243 | }
|
---|
244 |
|
---|
245 | # Given a filename in any encoding, will URL decode it to get back the original filename
|
---|
246 | # in the original encoding. Because this method is intended to work out the *original*
|
---|
247 | # filename*, it does not URL decode any filename if a file by the name of the *URL-encoded*
|
---|
248 | # string already exists in the local folder.
|
---|
249 | # Return the original filename corresponding to the parameter URL-encoded filename, and
|
---|
250 | # a decoded flag that is set to true iff URL-decoding had to be applied.
|
---|
251 | sub opt_url_decode {
|
---|
252 | my $self = shift (@_);
|
---|
253 | my ($link) = @_;
|
---|
254 |
|
---|
255 | # Replace %XX's in URL with decoded value if required.
|
---|
256 | # Note that the filename may include the %XX in some situations
|
---|
257 | if ($link =~ m/\%[A-F0-9]{2}/i) {
|
---|
258 | if (!-e $link) {
|
---|
259 | $link = &unicode::url_decode($link);
|
---|
260 | }
|
---|
261 | }
|
---|
262 |
|
---|
263 | return $link;
|
---|
264 | }
|
---|
265 |
|
---|
266 | sub read_into_doc_obj
|
---|
267 | {
|
---|
268 | my $self = shift (@_);
|
---|
269 | my ($pluginfo, $base_dir, $file, $block_hash, $metadata, $processor, $maxdocs, $total_count, $gli) = @_;
|
---|
270 |
|
---|
271 | # get the input file
|
---|
272 | my $input_filename = $file;
|
---|
273 | my ($tailname, $dirname, $suffix) = &File::Basename::fileparse($input_filename, "\\.[^\\.]+\$");
|
---|
274 | $suffix = lc($suffix);
|
---|
275 |
|
---|
276 | if (($self->{'use_realistic_book'}) || ($self->{'old_style_HDL'}))
|
---|
277 | {
|
---|
278 | # because the document has to be sectionalized set the description tags
|
---|
279 | $self->{'description_tags'} = 1;
|
---|
280 |
|
---|
281 | # set the file to be tidied
|
---|
282 | $input_filename = &util::filename_cat($base_dir,$file) if $base_dir =~ m/\w/;
|
---|
283 |
|
---|
284 | # get the tidied file
|
---|
285 | #my $tidy_filename = $self->tmp_tidy_file($input_filename);
|
---|
286 | my $tidy_filename = $self->convert_tidy_or_oldHDL_file($input_filename);
|
---|
287 |
|
---|
288 | # derive tmp filename from input filename
|
---|
289 | my ($tailname, $dirname, $suffix) = &File::Basename::fileparse($tidy_filename, "\\.[^\\.]+\$");
|
---|
290 |
|
---|
291 | # set the new input file and base_dir to be from the tidied file
|
---|
292 | $file = "$tailname$suffix";
|
---|
293 | $base_dir = $dirname;
|
---|
294 | }
|
---|
295 |
|
---|
296 | # call the parent read_into_doc_obj
|
---|
297 | my ($process_status,$doc_obj) = $self->SUPER::read_into_doc_obj($pluginfo, $base_dir, $file, $block_hash, $metadata, $processor, $maxdocs, $total_count, $gli);
|
---|
298 |
|
---|
299 | return ($process_status,$doc_obj);
|
---|
300 | }
|
---|
301 |
|
---|
302 | # do plugin specific processing of doc_obj
|
---|
303 | sub process {
|
---|
304 | my $self = shift (@_);
|
---|
305 | my ($textref, $pluginfo, $base_dir, $file, $metadata, $doc_obj, $gli) = @_;
|
---|
306 | my $outhandle = $self->{'outhandle'};
|
---|
307 |
|
---|
308 | if ($ENV{'GSDLOS'} =~ m/^windows/i) {
|
---|
309 | # this makes life so much easier... perl can cope with unix-style '/'s.
|
---|
310 | $base_dir =~ s@(\\)+@/@g;
|
---|
311 | $file =~ s@(\\)+@/@g;
|
---|
312 | }
|
---|
313 |
|
---|
314 | # reset per-doc stuff...
|
---|
315 | $self->{'aux_files'} = {};
|
---|
316 | $self->{'dir_num'} = 0;
|
---|
317 | $self->{'file_num'} = 0;
|
---|
318 |
|
---|
319 | # process an HTML file where sections are divided by headings tags (H1, H2 ...)
|
---|
320 | # you can also include metadata in the format (X can be any number)
|
---|
321 | # <hX>Title<!--gsdl-metadata
|
---|
322 | # <Metadata name="name1">value1</Metadata>
|
---|
323 | # ...
|
---|
324 | # <Metadata name="nameN">valueN</Metadata>
|
---|
325 | #--></hX>
|
---|
326 | if ($self->{'sectionalise_using_h_tags'}) {
|
---|
327 | # description_tags should allways be activated because we convert headings to description tags
|
---|
328 | $self->{'description_tags'} = 1;
|
---|
329 |
|
---|
330 | my $arrSections = [];
|
---|
331 | $$textref =~ s/<h([0-9]+)[^>]*>(.*?)<\/h[0-9]+>/$self->process_heading($1, $2, $arrSections, $file)/isge;
|
---|
332 |
|
---|
333 | if (scalar(@$arrSections)) {
|
---|
334 | my $strMetadata = $self->update_section_data($arrSections, -1);
|
---|
335 | if (length($strMetadata)) {
|
---|
336 | $strMetadata = '<!--' . $strMetadata . "\n-->\n</body>";
|
---|
337 | $$textref =~ s/<\/body>/$strMetadata/ig;
|
---|
338 | }
|
---|
339 | }
|
---|
340 | }
|
---|
341 |
|
---|
342 | my $cursection = $doc_obj->get_top_section();
|
---|
343 |
|
---|
344 | $self->extract_metadata ($textref, $metadata, $doc_obj, $cursection)
|
---|
345 | unless $self->{'no_metadata'} || $self->{'description_tags'};
|
---|
346 |
|
---|
347 | # Store URL for page as metadata - this can be used for an
|
---|
348 | # altavista style search interface. The URL won't be valid
|
---|
349 | # unless the file structure contains the domain name (i.e.
|
---|
350 | # like when w3mir is used to download a website).
|
---|
351 |
|
---|
352 | # URL metadata (even invalid ones) are used to support internal
|
---|
353 | # links, so even if 'file_is_url' is off, still need to store info
|
---|
354 |
|
---|
355 | my ($tailname,$dirname,$suffix) = &File::Basename::fileparse($file, "\\.[^\\.]+\$");
|
---|
356 | my $utf8_file = $self->filename_to_utf8_metadata($file);
|
---|
357 | $utf8_file =~ s/&\#095;/_/g;
|
---|
358 | my $web_url = "http://";
|
---|
359 | if(defined $dirname) { # local directory
|
---|
360 | $dirname = $self->eval_dir_dots($dirname);
|
---|
361 | $dirname .= &util::get_dirsep() if $dirname ne ""; # if there's a directory, it should end on "/"
|
---|
362 | $web_url = $web_url.$dirname.$utf8_file;
|
---|
363 | } else {
|
---|
364 | $web_url = $web_url.$utf8_file;
|
---|
365 | }
|
---|
366 | $web_url =~ s/\\/\//g;
|
---|
367 | $doc_obj->add_utf8_metadata($cursection, "URL", $web_url);
|
---|
368 |
|
---|
369 | if ($self->{'file_is_url'}) {
|
---|
370 | $doc_obj->add_metadata($cursection, "weblink", "<a href=\"$web_url\">");
|
---|
371 | $doc_obj->add_metadata($cursection, "webicon", "_iconworld_");
|
---|
372 | $doc_obj->add_metadata($cursection, "/weblink", "</a>");
|
---|
373 | }
|
---|
374 |
|
---|
375 | if ($self->{'description_tags'}) {
|
---|
376 | # remove the html header - note that doing this here means any
|
---|
377 | # sections defined within the header will be lost (so all <Section>
|
---|
378 | # tags must appear within the body of the HTML)
|
---|
379 | my ($head_keep) = ($$textref =~ m/^(.*?)<body[^>]*>/is);
|
---|
380 |
|
---|
381 | $$textref =~ s/^.*?<body[^>]*>//is;
|
---|
382 | $$textref =~ s/(<\/body[^>]*>|<\/html[^>]*>)//isg;
|
---|
383 |
|
---|
384 | my $opencom = '(?:<!--|<!(?:—|—|--))';
|
---|
385 | my $closecom = '(?:-->|(?:—|—|--)>)';
|
---|
386 |
|
---|
387 | my $lt = '(?:<|<)';
|
---|
388 | my $gt = '(?:>|>)';
|
---|
389 | my $quot = '(?:"|"|”|“)';
|
---|
390 |
|
---|
391 | my $dont_strip = '';
|
---|
392 | if ($self->{'no_strip_metadata_html'}) {
|
---|
393 | ($dont_strip = $self->{'no_strip_metadata_html'}) =~ s{,}{|}g;
|
---|
394 | }
|
---|
395 |
|
---|
396 | my $found_something = 0; my $top = 1;
|
---|
397 | while ($$textref =~ s/^(.*?)$opencom(.*?)$closecom//s) {
|
---|
398 | my $text = $1;
|
---|
399 | my $comment = $2;
|
---|
400 | if (defined $text) {
|
---|
401 | # text before a comment - note that getting to here
|
---|
402 | # doesn't necessarily mean there are Section tags in
|
---|
403 | # the document
|
---|
404 | $self->process_section(\$text, $base_dir, $file, $doc_obj, $cursection);
|
---|
405 | }
|
---|
406 | while ($comment =~ s/$lt(.*?)$gt//s) {
|
---|
407 | my $tag = $1;
|
---|
408 | if ($tag eq "Section") {
|
---|
409 | $found_something = 1;
|
---|
410 | $cursection = $doc_obj->insert_section($doc_obj->get_end_child($cursection)) unless $top;
|
---|
411 | $top = 0;
|
---|
412 | } elsif ($tag eq "/Section") {
|
---|
413 | $found_something = 1;
|
---|
414 | $cursection = $doc_obj->get_parent_section ($cursection);
|
---|
415 | } elsif ($tag =~ m/^Metadata name=$quot(.*?)$quot/s) {
|
---|
416 | my $metaname = $1;
|
---|
417 | my $accumulate = $tag =~ m/mode=${quot}accumulate${quot}/ ? 1 : 0;
|
---|
418 | $comment =~ s/^(.*?)$lt\/Metadata$gt//s;
|
---|
419 | my $metavalue = $1;
|
---|
420 | $metavalue =~ s/^\s+//;
|
---|
421 | $metavalue =~ s/\s+$//;
|
---|
422 | # assume that no metadata value intentionally includes
|
---|
423 | # carriage returns or HTML tags (if they're there they
|
---|
424 | # were probably introduced when converting to HTML from
|
---|
425 | # some other format).
|
---|
426 | # actually some people want to have html tags in their
|
---|
427 | # metadata.
|
---|
428 | $metavalue =~ s/[\cJ\cM]/ /sg;
|
---|
429 | $metavalue =~ s/<[^>]+>//sg
|
---|
430 | unless $dont_strip && ($dont_strip eq 'all' || $metaname =~ m/^($dont_strip)$/);
|
---|
431 | $metavalue =~ s/\s+/ /sg;
|
---|
432 | if ($accumulate) {
|
---|
433 | $doc_obj->add_utf8_metadata($cursection, $metaname, $metavalue);
|
---|
434 | } else {
|
---|
435 | $doc_obj->set_utf8_metadata_element($cursection, $metaname, $metavalue);
|
---|
436 | }
|
---|
437 | } elsif ($tag eq "Description" || $tag eq "/Description") {
|
---|
438 | # do nothing with containing Description tags
|
---|
439 | } else {
|
---|
440 | # simple HTML tag (probably created by the conversion
|
---|
441 | # to HTML from some other format) - we'll ignore it and
|
---|
442 | # hope for the best ;-)
|
---|
443 | }
|
---|
444 | }
|
---|
445 | }
|
---|
446 | if ($cursection ne "") {
|
---|
447 | print $outhandle "HTMLPlugin: WARNING: $file contains unmatched <Section></Section> tags\n";
|
---|
448 | }
|
---|
449 |
|
---|
450 | $$textref =~ s/^.*?<body[^>]*>//is;
|
---|
451 | $$textref =~ s/(<\/body[^>]*>|<\/html[^>]*>)//isg;
|
---|
452 | if ($$textref =~ m/\S/) {
|
---|
453 | if (!$found_something) {
|
---|
454 | if ($self->{'verbosity'} > 2) {
|
---|
455 | print $outhandle "HTMLPlugin: WARNING: $file appears to contain no Section tags so\n";
|
---|
456 | print $outhandle " will be processed as a single section document\n";
|
---|
457 | }
|
---|
458 |
|
---|
459 | # go ahead and process single-section document
|
---|
460 | $self->process_section($textref, $base_dir, $file, $doc_obj, $cursection);
|
---|
461 |
|
---|
462 | # if document contains no Section tags we'll go ahead
|
---|
463 | # and extract metadata (this won't have been done
|
---|
464 | # above as the -description_tags option prevents it)
|
---|
465 | my $complete_text = $head_keep.$doc_obj->get_text($cursection);
|
---|
466 | $self->extract_metadata (\$complete_text, $metadata, $doc_obj, $cursection)
|
---|
467 | unless $self->{'no_metadata'};
|
---|
468 |
|
---|
469 | } else {
|
---|
470 | print $outhandle "HTMLPlugin: WARNING: $file contains the following text outside\n";
|
---|
471 | print $outhandle " of the final closing </Section> tag. This text will\n";
|
---|
472 | print $outhandle " be ignored.";
|
---|
473 |
|
---|
474 | my ($text);
|
---|
475 | if (length($$textref) > 30) {
|
---|
476 | $text = substr($$textref, 0, 30) . "...";
|
---|
477 | } else {
|
---|
478 | $text = $$textref;
|
---|
479 | }
|
---|
480 | $text =~ s/\n/ /isg;
|
---|
481 | print $outhandle " ($text)\n";
|
---|
482 | }
|
---|
483 | } elsif (!$found_something) {
|
---|
484 |
|
---|
485 | if ($self->{'verbosity'} > 2) {
|
---|
486 | # may get to here if document contained no valid Section
|
---|
487 | # tags but did contain some comments. The text will have
|
---|
488 | # been processed already but we should print the warning
|
---|
489 | # as above and extract metadata
|
---|
490 | print $outhandle "HTMLPlugin: WARNING: $file appears to contain no Section tags and\n";
|
---|
491 | print $outhandle " is blank or empty. Metadata will be assigned if present.\n";
|
---|
492 | }
|
---|
493 |
|
---|
494 | my $complete_text = $head_keep.$doc_obj->get_text($cursection);
|
---|
495 | $self->extract_metadata (\$complete_text, $metadata, $doc_obj, $cursection)
|
---|
496 | unless $self->{'no_metadata'};
|
---|
497 | }
|
---|
498 |
|
---|
499 | } else {
|
---|
500 |
|
---|
501 | # remove header and footer
|
---|
502 | if (!$self->{'keep_head'} || $self->{'description_tags'}) {
|
---|
503 | $$textref =~ s/^.*?<body[^>]*>//is;
|
---|
504 | $$textref =~ s/(<\/body[^>]*>|<\/html[^>]*>)//isg;
|
---|
505 | }
|
---|
506 |
|
---|
507 | # single section document
|
---|
508 | $self->process_section($textref, $base_dir, $file, $doc_obj, $cursection);
|
---|
509 | }
|
---|
510 | return 1;
|
---|
511 | }
|
---|
512 |
|
---|
513 |
|
---|
514 | sub process_heading
|
---|
515 | {
|
---|
516 | my ($self, $nHeadNo, $strHeadingText, $arrSections, $file) = @_;
|
---|
517 | $strHeadingText = '' if (!defined($strHeadingText));
|
---|
518 |
|
---|
519 | my $strMetadata = $self->update_section_data($arrSections, int($nHeadNo));
|
---|
520 |
|
---|
521 | my $strSecMetadata = '';
|
---|
522 | while ($strHeadingText =~ s/<!--gsdl-metadata(.*?)-->//is)
|
---|
523 | {
|
---|
524 | $strSecMetadata .= $1;
|
---|
525 | }
|
---|
526 |
|
---|
527 | $strHeadingText =~ s/^\s+//g;
|
---|
528 | $strHeadingText =~ s/\s+$//g;
|
---|
529 | $strSecMetadata =~ s/^\s+//g;
|
---|
530 | $strSecMetadata =~ s/\s+$//g;
|
---|
531 |
|
---|
532 | $strMetadata .= "\n<Section>\n\t<Description>\n\t\t<Metadata name=\"Title\">" . $strHeadingText . "</Metadata>\n";
|
---|
533 |
|
---|
534 | if (length($strSecMetadata)) {
|
---|
535 | $strMetadata .= "\t\t" . $strSecMetadata . "\n";
|
---|
536 | }
|
---|
537 |
|
---|
538 | $strMetadata .= "\t</Description>\n";
|
---|
539 |
|
---|
540 | return "<!--" . $strMetadata . "-->";
|
---|
541 | }
|
---|
542 |
|
---|
543 |
|
---|
544 | sub update_section_data
|
---|
545 | {
|
---|
546 | my ($self, $arrSections, $nCurTocNo) = @_;
|
---|
547 | my ($strBuffer, $nLast, $nSections) = ('', 0, scalar(@$arrSections));
|
---|
548 |
|
---|
549 | if ($nSections == 0) {
|
---|
550 | push @$arrSections, $nCurTocNo;
|
---|
551 | return $strBuffer;
|
---|
552 | }
|
---|
553 | $nLast = $arrSections->[$nSections - 1];
|
---|
554 | if ($nCurTocNo > $nLast) {
|
---|
555 | push @$arrSections, $nCurTocNo;
|
---|
556 | return $strBuffer;
|
---|
557 | }
|
---|
558 | for(my $i = $nSections - 1; $i >= 0; $i--) {
|
---|
559 | if ($nCurTocNo <= $arrSections->[$i]) {
|
---|
560 | $strBuffer .= "\n</Section>";
|
---|
561 | pop @$arrSections;
|
---|
562 | }
|
---|
563 | }
|
---|
564 | push @$arrSections, $nCurTocNo;
|
---|
565 | return $strBuffer;
|
---|
566 | }
|
---|
567 |
|
---|
568 |
|
---|
569 | # note that process_section may be called multiple times for a single
|
---|
570 | # section (relying on the fact that add_utf8_text appends the text to any
|
---|
571 | # that may exist already).
|
---|
572 | sub process_section {
|
---|
573 | my $self = shift (@_);
|
---|
574 | my ($textref, $base_dir, $file, $doc_obj, $cursection) = @_;
|
---|
575 | # trap links
|
---|
576 | if (!$self->{'nolinks'}) {
|
---|
577 | # usemap="./#index" not handled correctly => change to "#index"
|
---|
578 | ## $$textref =~ s/(<img[^>]*?usemap\s*=\s*[\"\']?)([^\"\'>\s]+)([\"\']?[^>]*>)/
|
---|
579 |
|
---|
580 | $$textref =~ s/(<img[^>]*?usemap\s*=\s*)((?:[\"][^\"]+[\"])|(?:[\'][^\']+[\'])|(?:[^\s\/>]+))([^>]*>)/
|
---|
581 | $self->replace_usemap_links($1, $2, $3)/isge;
|
---|
582 |
|
---|
583 | ## $$textref =~ s/(<(?:a|area|frame|link|script)\s+[^>]*?\s*(?:href|src)\s*=\s*[\"\']?)([^\"\'>\s]+)([\"\']?[^>]*>)/
|
---|
584 |
|
---|
585 | $$textref =~ s/(<(?:a|area|frame|link|script)\s+[^>]*?\s*(?:href|src)\s*=\s*)((?:[\"][^\"]+[\"])|(?:[\'][^\']+[\'])|(?:[^\s\/>]+))([^>]*>)/
|
---|
586 | $self->replace_href_links ($1, $2, $3, $base_dir, $file, $doc_obj, $cursection)/isge;
|
---|
587 | }
|
---|
588 |
|
---|
589 | # trap images
|
---|
590 |
|
---|
591 | # Previously, by default, HTMLPlugin would embed <img> tags inside anchor tags
|
---|
592 | # i.e. <a href="image><img src="image"></a> in order to overcome a problem that
|
---|
593 | # turned regular text succeeding images into links. That is, by embedding <imgs>
|
---|
594 | # inside <a href=""></a>, the text following images were no longer misbehaving.
|
---|
595 | # However, there would be many occasions whereby images were not meant to link
|
---|
596 | # to their source images but where the images would link to another web page.
|
---|
597 | # To allow this, the no_image_links option was introduced: it would prevent
|
---|
598 | # the behaviour of embedding images into links that referenced the source images.
|
---|
599 |
|
---|
600 | # Somewhere along the line, the problem of normal text turning into links when
|
---|
601 | # such text followed images which were not embedded in <a href=""></a> ceased
|
---|
602 | # to occur. This is why the following lines have been commented out (as well as
|
---|
603 | # two lines in replace_images). They appear to no longer apply.
|
---|
604 |
|
---|
605 | # If at any time, there is a need for having images embedded in <a> anchor tags,
|
---|
606 | # then it might be better to turn that into an HTMLPlugin option rather than make
|
---|
607 | # it the default behaviour. Also, eventually, no_image_links needs to become
|
---|
608 | # a deprecated option for HTMLPlugin as it has now become the default behaviour.
|
---|
609 |
|
---|
610 | #if(!$self->{'no_image_links'}){
|
---|
611 | $$textref =~ s/(<(?:img|embed|table|tr|td)[^>]*?(?:src|background)\s*=\s*)((?:[\"][^\"]+[\"])|(?:[\'][^\']+[\'])|(?:[^\s\/>]+))([^>]*>)/
|
---|
612 | $self->replace_images ($1, $2, $3, $base_dir, $file, $doc_obj, $cursection)/isge;
|
---|
613 | #}
|
---|
614 |
|
---|
615 | # add text to document object
|
---|
616 | # turn \ into \\ so that the rest of greenstone doesn't think there
|
---|
617 | # is an escape code following. (Macro parsing loses them...)
|
---|
618 | $$textref =~ s/\\/\\\\/go;
|
---|
619 |
|
---|
620 | $doc_obj->add_utf8_text($cursection, $$textref);
|
---|
621 | }
|
---|
622 |
|
---|
623 | sub replace_images {
|
---|
624 | my $self = shift (@_);
|
---|
625 | my ($front, $link, $back, $base_dir,
|
---|
626 | $file, $doc_obj, $section) = @_;
|
---|
627 |
|
---|
628 | # remove quotes from link at start and end if necessary
|
---|
629 | if ($link=~/^[\"\']/) {
|
---|
630 | $link=~s/^[\"\']//;
|
---|
631 | $link=~s/[\"\']$//;
|
---|
632 | $front.='"';
|
---|
633 | $back="\"$back";
|
---|
634 | }
|
---|
635 |
|
---|
636 | $link =~ s/\n/ /g;
|
---|
637 |
|
---|
638 | # Hack to overcome Windows wv 0.7.1 bug that causes embedded images to be broken
|
---|
639 | # If the Word file path has spaces in it, wv messes up and you end up with
|
---|
640 | # absolute paths for the images, and without the "file://" prefix
|
---|
641 | # So check for this special case and massage the data to be correct
|
---|
642 | if ($ENV{'GSDLOS'} =~ m/^windows/i && $self->{'plugin_type'} eq "WordPlug" && $link =~ m/^[A-Za-z]\:\\/) {
|
---|
643 | $link =~ s/^.*\\([^\\]+)$/$1/;
|
---|
644 | }
|
---|
645 |
|
---|
646 | my ($href, $hash_part, $rl) = $self->format_link ($link, $base_dir, $file);
|
---|
647 |
|
---|
648 | my $img_file = $self->add_file ($href, $rl, $hash_part, $base_dir, $doc_obj, $section);
|
---|
649 |
|
---|
650 | # print STDERR "**** link = $link\n**** href = $href\n**** img_file = $img_file, rl = $rl\n";
|
---|
651 |
|
---|
652 | my $anchor_name = $img_file;
|
---|
653 | #$anchor_name =~ s/^.*\///;
|
---|
654 | #$anchor_name = "<a name=\"$anchor_name\" ></a>";
|
---|
655 |
|
---|
656 | my $image_link = $front . $img_file .$back;
|
---|
657 | return $image_link;
|
---|
658 |
|
---|
659 | # The reasons for why the following two lines are no longer necessary can be
|
---|
660 | # found in subroutine process_section
|
---|
661 | #my $anchor_link = "<a href=\"$img_file\" >".$image_link."</a>";
|
---|
662 | #return $anchor_link;
|
---|
663 |
|
---|
664 | #return $front . $img_file . $back . $anchor_name;
|
---|
665 | }
|
---|
666 |
|
---|
667 | sub replace_href_links {
|
---|
668 | my $self = shift (@_);
|
---|
669 | my ($front, $link, $back, $base_dir, $file, $doc_obj, $section) = @_;
|
---|
670 |
|
---|
671 | # remove quotes from link at start and end if necessary
|
---|
672 | if ($link=~/^[\"\']/) {
|
---|
673 | $link=~s/^[\"\']//;
|
---|
674 | $link=~s/[\"\']$//;
|
---|
675 | $front.='"';
|
---|
676 | $back="\"$back";
|
---|
677 | }
|
---|
678 |
|
---|
679 | # attempt to sort out targets - frames are not handled
|
---|
680 | # well in this plugin and some cases will screw things
|
---|
681 | # up - e.g. the _parent target (so we'll just remove
|
---|
682 | # them all ;-)
|
---|
683 | $front =~ s/(target=\"?)_top(\"?)/$1_gsdltop_$2/is;
|
---|
684 | $back =~ s/(target=\"?)_top(\"?)/$1_gsdltop_$2/is;
|
---|
685 | $front =~ s/target=\"?_parent\"?//is;
|
---|
686 | $back =~ s/target=\"?_parent\"?//is;
|
---|
687 |
|
---|
688 | return $front . $link . $back if $link =~ m/^\#/s;
|
---|
689 | $link =~ s/\n/ /g;
|
---|
690 |
|
---|
691 | # Find file referred to by $link on file system
|
---|
692 | # This is more complicated than it sounds when char encodings
|
---|
693 | # is taken in to account
|
---|
694 | my ($href, $hash_part, $rl) = $self->format_link ($link, $base_dir, $file);
|
---|
695 |
|
---|
696 | # href may use '\'s where '/'s should be on Windows
|
---|
697 | $href =~ s/\\/\//g;
|
---|
698 | my ($filename) = $href =~ m/^(?:.*?):(?:\/\/)?(.*)/;
|
---|
699 |
|
---|
700 |
|
---|
701 | ##### leave all these links alone (they won't be picked up by intermediate
|
---|
702 | ##### pages). I think that's safest when dealing with frames, targets etc.
|
---|
703 | ##### (at least until I think of a better way to do it). Problems occur with
|
---|
704 | ##### mailto links from within small frames, the intermediate page is displayed
|
---|
705 | ##### within that frame and can't be seen. There is still potential for this to
|
---|
706 | ##### happen even with html pages - the solution seems to be to somehow tell
|
---|
707 | ##### the browser from the server side to display the page being sent (i.e.
|
---|
708 | ##### the intermediate page) in the top level window - I'm not sure if that's
|
---|
709 | ##### possible - the following line should probably be deleted if that can be done
|
---|
710 | return $front . $link . $back if $href =~ m/^(mailto|news|gopher|nntp|telnet|javascript):/is;
|
---|
711 |
|
---|
712 |
|
---|
713 | if (($rl == 0) || ($filename =~ m/$self->{'process_exp'}/) ||
|
---|
714 | ($href =~ m/\/$/) || ($href =~ m/^(mailto|news|gopher|nntp|telnet|javascript):/i)) {
|
---|
715 | &ghtml::urlsafe ($href);
|
---|
716 | return $front . "_httpextlink_&rl=" . $rl . "&href=" . $href . $hash_part . $back;
|
---|
717 | } else {
|
---|
718 | # link is to some other type of file (eg image) so we'll
|
---|
719 | # need to associate that file
|
---|
720 | return $front . $self->add_file ($href, $rl, $hash_part, $base_dir, $doc_obj, $section) . $back;
|
---|
721 | }
|
---|
722 | }
|
---|
723 |
|
---|
724 | sub add_file {
|
---|
725 | my $self = shift (@_);
|
---|
726 | my ($href, $rl, $hash_part, $base_dir, $doc_obj, $section) = @_;
|
---|
727 | my ($newname);
|
---|
728 |
|
---|
729 | my $filename = $href;
|
---|
730 | if ($base_dir eq "") {
|
---|
731 | # remove http:/ thereby leaving one slash at the start
|
---|
732 | $filename =~ s/^[^:]*:\///;
|
---|
733 | }
|
---|
734 | else {
|
---|
735 | # remove http://
|
---|
736 | $filename =~ s/^[^:]*:\/\///;
|
---|
737 | }
|
---|
738 |
|
---|
739 | $filename = &util::filename_cat($base_dir, $filename);
|
---|
740 | # Replace %XX's in URL with decoded value if required. Note that the filename may include the %XX in some
|
---|
741 | # situations. If the *original* file's name was in URL encoding, the following method will not decode it.
|
---|
742 | my $utf8_filename = $filename;
|
---|
743 | $filename = $self->opt_url_decode($utf8_filename);
|
---|
744 |
|
---|
745 | # some special processing if the intended filename was converted to utf8, but
|
---|
746 | # the actual file still needs to be renamed
|
---|
747 | if (!-e $filename) {
|
---|
748 | # try the original filename stored in map
|
---|
749 | my $original_filename = $self->{'utf8_to_original_filename'}->{$filename};
|
---|
750 | if (defined $original_filename && -e $original_filename) {
|
---|
751 | $filename = $original_filename;
|
---|
752 | }
|
---|
753 | }
|
---|
754 |
|
---|
755 | my ($ext) = $filename =~ m/(\.[^\.]*)$/;
|
---|
756 |
|
---|
757 | if ($rl == 0) {
|
---|
758 | if ((!defined $ext) || ($ext !~ m/$self->{'assoc_files'}/)) {
|
---|
759 | return "_httpextlink_&rl=0&el=prompt&href=" . $href . $hash_part;
|
---|
760 | }
|
---|
761 | else {
|
---|
762 | return "_httpextlink_&rl=0&el=direct&href=" . $href . $hash_part;
|
---|
763 | }
|
---|
764 | }
|
---|
765 |
|
---|
766 | if ((!defined $ext) || ($ext !~ m/$self->{'assoc_files'}/)) {
|
---|
767 | return "_httpextlink_&rl=" . $rl . "&href=" . $href . $hash_part;
|
---|
768 | }
|
---|
769 | # add the original image file as a source file
|
---|
770 | if (!$self->{'processing_tmp_files'} ) {
|
---|
771 | $doc_obj->associate_source_file($filename);
|
---|
772 | }
|
---|
773 | if ($self->{'rename_assoc_files'}) {
|
---|
774 | if (defined $self->{'aux_files'}->{$href}) {
|
---|
775 | $newname = $self->{'aux_files'}->{$href}->{'dir_num'} . "/" .
|
---|
776 | $self->{'aux_files'}->{$href}->{'file_num'} . $ext;
|
---|
777 | } else {
|
---|
778 | $newname = $self->{'dir_num'} . "/" . $self->{'file_num'} . $ext;
|
---|
779 | $self->{'aux_files'}->{$href} = {'dir_num' => $self->{'dir_num'}, 'file_num' => $self->{'file_num'}};
|
---|
780 | $self->inc_filecount ();
|
---|
781 | }
|
---|
782 | $doc_obj->associate_file($filename, $newname, undef, $section);
|
---|
783 | return "_httpdocimg_/$newname";
|
---|
784 | } else {
|
---|
785 | if(&unicode::is_url_encoded($utf8_filename)) {
|
---|
786 | # use the possibly-decoded filename instead to avoid double URL encoding
|
---|
787 | ($newname) = $filename =~ m/([^\/\\]*)$/;
|
---|
788 | } else {
|
---|
789 | ($newname) = $utf8_filename =~ m/([^\/\\]*)$/;
|
---|
790 | }
|
---|
791 |
|
---|
792 | # Make sure this name uses only ASCII characters.
|
---|
793 | # We use either base64 or URL encoding, as these preserve original encoding
|
---|
794 | $newname = &util::rename_file($newname, $self->{'file_rename_method'});
|
---|
795 |
|
---|
796 | $doc_obj->associate_file($filename, $newname, undef, $section);
|
---|
797 |
|
---|
798 | # Since the generated image will be URL-encoded to avoid file-system/browser mess-ups
|
---|
799 | # of filenames, URL-encode the additional percent signs of the URL-encoded filename
|
---|
800 | my $newname_url = $newname;
|
---|
801 | $newname_url = &unicode::filename_to_url($newname_url);
|
---|
802 | return "_httpdocimg_/$newname_url";
|
---|
803 | }
|
---|
804 | }
|
---|
805 |
|
---|
806 |
|
---|
807 | sub format_link {
|
---|
808 | my $self = shift (@_);
|
---|
809 | my ($link, $base_dir, $file) = @_;
|
---|
810 |
|
---|
811 | my ($before_hash, $hash_part) = $link =~ m/^([^\#]*)(\#?.*)$/;
|
---|
812 |
|
---|
813 | $hash_part = "" if !defined $hash_part;
|
---|
814 | if (!defined $before_hash || $before_hash !~ m/[\w\.\/]/) {
|
---|
815 | my $outhandle = $self->{'outhandle'};
|
---|
816 | print $outhandle "HTMLPlugin: ERROR - badly formatted tag ignored ($link)\n"
|
---|
817 | if $self->{'verbosity'};
|
---|
818 | return ($link, "", 0);
|
---|
819 | }
|
---|
820 |
|
---|
821 | if ($before_hash =~ s@^((?:http|https|ftp|file|mms)://)@@i) {
|
---|
822 | my $type = $1;
|
---|
823 |
|
---|
824 | if ($link =~ m/^(http|ftp):/i) {
|
---|
825 | # Turn url (using /) into file name (possibly using \ on windows)
|
---|
826 | my @http_dir_split = split('/', $before_hash);
|
---|
827 | $before_hash = &util::filename_cat(@http_dir_split);
|
---|
828 | }
|
---|
829 |
|
---|
830 | $before_hash = $self->eval_dir_dots($before_hash);
|
---|
831 |
|
---|
832 | my $linkfilename = &util::filename_cat ($base_dir, $before_hash);
|
---|
833 |
|
---|
834 | my $rl = 0;
|
---|
835 | $rl = 1 if (-e $linkfilename);
|
---|
836 |
|
---|
837 | # make sure there's a slash on the end if it's a directory
|
---|
838 | if ($before_hash !~ m/\/$/) {
|
---|
839 | $before_hash .= "/" if (-d $linkfilename);
|
---|
840 | }
|
---|
841 | return ($type . $before_hash, $hash_part, $rl);
|
---|
842 |
|
---|
843 | } elsif ($link !~ m/^(mailto|news|gopher|nntp|telnet|javascript):/i && $link !~ m/^\//) {
|
---|
844 |
|
---|
845 | if ($before_hash =~ s@^/@@ || $before_hash =~ m/\\/) {
|
---|
846 |
|
---|
847 | # the first directory will be the domain name if file_is_url
|
---|
848 | # to generate archives, otherwise we'll assume all files are
|
---|
849 | # from the same site and base_dir is the root
|
---|
850 |
|
---|
851 | if ($self->{'file_is_url'}) {
|
---|
852 | my @dirs = split /[\/\\]/, $file;
|
---|
853 | my $domname = shift (@dirs);
|
---|
854 | $before_hash = &util::filename_cat($domname, $before_hash);
|
---|
855 | $before_hash =~ s@\\@/@g; # for windows
|
---|
856 | }
|
---|
857 | else
|
---|
858 | {
|
---|
859 | # see if link shares directory with source document
|
---|
860 | # => turn into relative link if this is so!
|
---|
861 |
|
---|
862 | if ($ENV{'GSDLOS'} =~ m/^windows/i) {
|
---|
863 | # too difficult doing a pattern match with embedded '\'s...
|
---|
864 | my $win_before_hash=$before_hash;
|
---|
865 | $win_before_hash =~ s@(\\)+@/@g;
|
---|
866 | # $base_dir is already similarly "converted" on windows.
|
---|
867 | if ($win_before_hash =~ s@^$base_dir/@@o) {
|
---|
868 | # if this is true, we removed a prefix
|
---|
869 | $before_hash=$win_before_hash;
|
---|
870 | }
|
---|
871 | }
|
---|
872 | else {
|
---|
873 | # before_hash has lost leading slash by this point,
|
---|
874 | # -> add back in prior to substitution with $base_dir
|
---|
875 | $before_hash = "/$before_hash";
|
---|
876 |
|
---|
877 | $before_hash = &util::filename_cat("",$before_hash);
|
---|
878 | $before_hash =~ s@^$base_dir/@@;
|
---|
879 | }
|
---|
880 | }
|
---|
881 | } else {
|
---|
882 | # Turn relative file path into full path
|
---|
883 | my $dirname = &File::Basename::dirname($file);
|
---|
884 | $before_hash = &util::filename_cat($dirname, $before_hash);
|
---|
885 | $before_hash = $self->eval_dir_dots($before_hash);
|
---|
886 | }
|
---|
887 |
|
---|
888 | my $linkfilename = &util::filename_cat ($base_dir, $before_hash);
|
---|
889 | # make sure there's a slash on the end if it's a directory
|
---|
890 | if ($before_hash !~ m/\/$/) {
|
---|
891 | $before_hash .= "/" if (-d $linkfilename);
|
---|
892 | }
|
---|
893 | return ("http://" . $before_hash, $hash_part, 1);
|
---|
894 | } else {
|
---|
895 | # mailto, news, nntp, telnet, javascript or gopher link
|
---|
896 | return ($before_hash, "", 0);
|
---|
897 | }
|
---|
898 | }
|
---|
899 |
|
---|
900 | sub extract_first_NNNN_characters {
|
---|
901 | my $self = shift (@_);
|
---|
902 | my ($textref, $doc_obj, $thissection) = @_;
|
---|
903 |
|
---|
904 | foreach my $size (split /,/, $self->{'first'}) {
|
---|
905 | my $tmptext = $$textref;
|
---|
906 | # skip to the body
|
---|
907 | $tmptext =~ s/.*<body[^>]*>//i;
|
---|
908 | # remove javascript
|
---|
909 | $tmptext =~ s@<script.*?</script>@ @sig;
|
---|
910 | $tmptext =~ s/<[^>]*>/ /g;
|
---|
911 | $tmptext =~ s/ / /g;
|
---|
912 | $tmptext =~ s/^\s+//;
|
---|
913 | $tmptext =~ s/\s+$//;
|
---|
914 | $tmptext =~ s/\s+/ /gs;
|
---|
915 | $tmptext = &unicode::substr ($tmptext, 0, $size);
|
---|
916 | $tmptext =~ s/\s\S*$/…/; # adds an ellipse (...)
|
---|
917 | $doc_obj->add_utf8_metadata ($thissection, "First$size", $tmptext);
|
---|
918 | }
|
---|
919 | }
|
---|
920 |
|
---|
921 |
|
---|
922 | sub extract_metadata {
|
---|
923 | my $self = shift (@_);
|
---|
924 | my ($textref, $metadata, $doc_obj, $section) = @_;
|
---|
925 | my $outhandle = $self->{'outhandle'};
|
---|
926 | # if we don't want metadata, we may as well not be here ...
|
---|
927 | return if (!defined $self->{'metadata_fields'});
|
---|
928 |
|
---|
929 | my $separator = $self->{'metadata_field_separator'};
|
---|
930 | if ($separator eq "") {
|
---|
931 | undef $separator;
|
---|
932 | }
|
---|
933 |
|
---|
934 | # metadata fields to extract/save. 'key' is the (lowercase) name of the
|
---|
935 | # html meta, 'value' is the metadata name for greenstone to use
|
---|
936 | my %find_fields = ();
|
---|
937 |
|
---|
938 | my %creator_fields = (); # short-cut for lookups
|
---|
939 |
|
---|
940 |
|
---|
941 | foreach my $field (split /,/, $self->{'metadata_fields'}) {
|
---|
942 | $field =~ s/^\s+//; # remove leading whitespace
|
---|
943 | $field =~ s/\s+$//; # remove trailing whitespace
|
---|
944 |
|
---|
945 | # support tag<tagname>
|
---|
946 | if ($field =~ m/^(.*?)\s*<(.*?)>$/) {
|
---|
947 | # "$2" is the user's preferred gs metadata name
|
---|
948 | $find_fields{lc($1)}=$2; # lc = lowercase
|
---|
949 | } else { # no <tagname> for mapping
|
---|
950 | # "$field" is the user's preferred gs metadata name
|
---|
951 | $find_fields{lc($field)}=$field; # lc = lowercase
|
---|
952 | }
|
---|
953 | }
|
---|
954 |
|
---|
955 | if (defined $self->{'hunt_creator_metadata'} &&
|
---|
956 | $self->{'hunt_creator_metadata'} == 1 ) {
|
---|
957 | my @extra_fields =
|
---|
958 | (
|
---|
959 | 'author',
|
---|
960 | 'author.email',
|
---|
961 | 'creator',
|
---|
962 | 'dc.creator',
|
---|
963 | 'dc.creator.corporatename',
|
---|
964 | );
|
---|
965 |
|
---|
966 | # add the creator_metadata fields to search for
|
---|
967 | foreach my $field (@extra_fields) {
|
---|
968 | $creator_fields{$field}=0; # add to lookup hash
|
---|
969 | }
|
---|
970 | }
|
---|
971 |
|
---|
972 |
|
---|
973 | # find the header in the html file, which has the meta tags
|
---|
974 | $$textref =~ m@<head>(.*?)</head>@si;
|
---|
975 |
|
---|
976 | my $html_header=$1;
|
---|
977 |
|
---|
978 | # go through every <meta... tag defined in the html and see if it is
|
---|
979 | # one of the tags we want to match.
|
---|
980 |
|
---|
981 | # special case for title - we want to remember if its been found
|
---|
982 | my $found_title = 0;
|
---|
983 | # this assumes that ">" won't appear. (I don't think it's allowed to...)
|
---|
984 | $html_header =~ m/^/; # match the start of the string, for \G assertion
|
---|
985 |
|
---|
986 | while ($html_header =~ m/\G.*?<meta(.*?)>/sig) {
|
---|
987 | my $metatag=$1;
|
---|
988 | my ($tag, $value);
|
---|
989 |
|
---|
990 | # find the tag name
|
---|
991 | $metatag =~ m/(?:name|http-equiv)\s*=\s*([\"\'])?(.*?)\1/is;
|
---|
992 | $tag=$2;
|
---|
993 | # in case they're not using " or ', but they should...
|
---|
994 | if (! $tag) {
|
---|
995 | $metatag =~ m/(?:name|http-equiv)\s*=\s*([^\s\>]+)/is;
|
---|
996 | $tag=$1;
|
---|
997 | }
|
---|
998 |
|
---|
999 | if (!defined $tag) {
|
---|
1000 | print $outhandle "HTMLPlugin: can't find NAME in \"$metatag\"\n";
|
---|
1001 | next;
|
---|
1002 | }
|
---|
1003 |
|
---|
1004 | # don't need to assign this field if it was passed in from a previous
|
---|
1005 | # (recursive) plugin
|
---|
1006 | if (defined $metadata->{$tag}) {next}
|
---|
1007 |
|
---|
1008 | # find the tag content
|
---|
1009 | $metatag =~ m/content\s*=\s*([\"\'])?(.*?)\1/is;
|
---|
1010 | $value=$2;
|
---|
1011 |
|
---|
1012 | if (! $value) {
|
---|
1013 | $metatag =~ m/(?:name|http-equiv)\s*=\s*([^\s\>]+)/is;
|
---|
1014 | $value=$1;
|
---|
1015 | }
|
---|
1016 | if (!defined $value) {
|
---|
1017 | print $outhandle "HTMLPlugin: can't find VALUE in \"$metatag\"\n";
|
---|
1018 | next;
|
---|
1019 | }
|
---|
1020 |
|
---|
1021 | # clean up and add
|
---|
1022 | $value =~ s/\s+/ /gs;
|
---|
1023 | chomp($value); # remove trailing \n, if any
|
---|
1024 | if (exists $creator_fields{lc($tag)}) {
|
---|
1025 | # map this value onto greenstone's "Creator" metadata
|
---|
1026 | $tag='Creator';
|
---|
1027 | } elsif (!exists $find_fields{lc($tag)}) {
|
---|
1028 | next; # don't want this tag
|
---|
1029 | } else {
|
---|
1030 | # get the user's preferred capitalisation
|
---|
1031 | $tag = $find_fields{lc($tag)};
|
---|
1032 | }
|
---|
1033 | if (lc($tag) eq "title") {
|
---|
1034 | $found_title = 1;
|
---|
1035 | }
|
---|
1036 |
|
---|
1037 | if ($self->{'verbosity'} > 2) {
|
---|
1038 | print $outhandle " extracted \"$tag\" metadata \"$value\"\n";
|
---|
1039 | }
|
---|
1040 |
|
---|
1041 | # Do we still reply on the following? Surely there must
|
---|
1042 | # be a better way to go about this outside of the plugin?
|
---|
1043 | #
|
---|
1044 | #if ($tag =~ m/date.*/i){
|
---|
1045 | # $tag = lc($tag);
|
---|
1046 | #}
|
---|
1047 | if (defined $separator) {
|
---|
1048 | my @values = split($separator, $value);
|
---|
1049 | foreach my $v (@values) {
|
---|
1050 | $doc_obj->add_utf8_metadata($section, $tag, $v) if $v =~ /\S/;
|
---|
1051 | }
|
---|
1052 | }
|
---|
1053 | else {
|
---|
1054 | $doc_obj->add_utf8_metadata($section, $tag, $value);
|
---|
1055 | }
|
---|
1056 | }
|
---|
1057 |
|
---|
1058 | # TITLE: extract the document title
|
---|
1059 | if (exists $find_fields{'title'} && !$found_title) {
|
---|
1060 | # we want a title, and didn't find one in the meta tags
|
---|
1061 | # see if there's a <title> tag
|
---|
1062 | my $title;
|
---|
1063 | my $from = ""; # for debugging output only
|
---|
1064 | if ($html_header =~ m/<title[^>]*>([^<]+)<\/title[^>]*>/is) {
|
---|
1065 | $title = $1;
|
---|
1066 | $from = "<title> tags";
|
---|
1067 | }
|
---|
1068 |
|
---|
1069 | if (!defined $title) {
|
---|
1070 | $from = "first 100 chars";
|
---|
1071 | # if no title use first 100 or so characters
|
---|
1072 | $title = $$textref;
|
---|
1073 | $title =~ s/^\xFE\xFF//; # Remove unicode byte order mark
|
---|
1074 | $title =~ s/^.*?<body>//si;
|
---|
1075 | # ignore javascript!
|
---|
1076 | $title =~ s@<script.*?</script>@ @sig;
|
---|
1077 | $title =~ s/<\/([^>]+)><\1>//g; # (eg) </b><b> - no space
|
---|
1078 | $title =~ s/<[^>]*>/ /g; # remove all HTML tags
|
---|
1079 | $title = substr ($title, 0, 100);
|
---|
1080 | $title =~ s/\s\S*$/.../;
|
---|
1081 | }
|
---|
1082 | $title =~ s/<[^>]*>/ /g; # remove html tags
|
---|
1083 | $title =~ s/ / /g;
|
---|
1084 | $title =~ s/(?: |\xc2\xa0)/ /g; # utf-8 for nbsp...
|
---|
1085 | $title =~ s/\s+/ /gs; # collapse multiple spaces
|
---|
1086 | $title =~ s/^\s*//; # remove leading spaces
|
---|
1087 | $title =~ s/\s*$//; # remove trailing spaces
|
---|
1088 |
|
---|
1089 | $title =~ s/^$self->{'title_sub'}// if ($self->{'title_sub'});
|
---|
1090 | $title =~ s/^\s+//s; # in case title_sub introduced any...
|
---|
1091 | $doc_obj->add_utf8_metadata ($section, 'Title', $title);
|
---|
1092 | print $outhandle " extracted Title metadata \"$title\" from $from\n"
|
---|
1093 | if ($self->{'verbosity'} > 2);
|
---|
1094 | }
|
---|
1095 |
|
---|
1096 | # add FileFormat metadata
|
---|
1097 | $doc_obj->add_metadata($section,"FileFormat", "HTML");
|
---|
1098 |
|
---|
1099 | # Special, for metadata names such as tagH1 - extracts
|
---|
1100 | # the text between the first <H1> and </H1> tags into "H1" metadata.
|
---|
1101 |
|
---|
1102 | foreach my $field (keys %find_fields) {
|
---|
1103 | if ($field !~ m/^tag([a-z0-9]+)$/i) {next}
|
---|
1104 | my $tag = $1;
|
---|
1105 | if ($$textref =~ m@<$tag[^>]*>(.*?)</$tag[^>]*>@g) {
|
---|
1106 | my $content = $1;
|
---|
1107 | $content =~ s/ / /g;
|
---|
1108 | $content =~ s/<[^>]*>/ /g;
|
---|
1109 | $content =~ s/^\s+//;
|
---|
1110 | $content =~ s/\s+$//;
|
---|
1111 | $content =~ s/\s+/ /gs;
|
---|
1112 | if ($content) {
|
---|
1113 | $tag=$find_fields{"tag$tag"}; # get the user's capitalisation
|
---|
1114 | $tag =~ s/^tag//i;
|
---|
1115 | $doc_obj->add_utf8_metadata ($section, $tag, $content);
|
---|
1116 | print $outhandle " extracted \"$tag\" metadata \"$content\"\n"
|
---|
1117 | if ($self->{'verbosity'} > 2);
|
---|
1118 | }
|
---|
1119 | }
|
---|
1120 | }
|
---|
1121 | }
|
---|
1122 |
|
---|
1123 |
|
---|
1124 | # evaluate any "../" to next directory up
|
---|
1125 | # evaluate any "./" as here
|
---|
1126 | sub eval_dir_dots {
|
---|
1127 | my $self = shift (@_);
|
---|
1128 | my ($filename) = @_;
|
---|
1129 | my $dirsep_os = &util::get_os_dirsep();
|
---|
1130 | my @dirsep = split(/$dirsep_os/,$filename);
|
---|
1131 |
|
---|
1132 | my @eval_dirs = ();
|
---|
1133 | foreach my $d (@dirsep) {
|
---|
1134 | if ($d eq "..") {
|
---|
1135 | pop(@eval_dirs);
|
---|
1136 |
|
---|
1137 | } elsif ($d eq ".") {
|
---|
1138 | # do nothing!
|
---|
1139 |
|
---|
1140 | } else {
|
---|
1141 | push(@eval_dirs,$d);
|
---|
1142 | }
|
---|
1143 | }
|
---|
1144 |
|
---|
1145 | # Need to fiddle with number of elements in @eval_dirs if the
|
---|
1146 | # first one is the empty string. This is because of a
|
---|
1147 | # modification to util::filename_cat that supresses the addition
|
---|
1148 | # of a leading '/' character (or \ if windows) (intended to help
|
---|
1149 | # filename cat with relative paths) if the first entry in the
|
---|
1150 | # array is the empty string. Making the array start with *two*
|
---|
1151 | # empty strings is a way to defeat this "smart" option.
|
---|
1152 | #
|
---|
1153 | if (scalar(@eval_dirs) > 0) {
|
---|
1154 | if ($eval_dirs[0] eq ""){
|
---|
1155 | unshift(@eval_dirs,"");
|
---|
1156 | }
|
---|
1157 | }
|
---|
1158 |
|
---|
1159 | my $evaluated_filename = (scalar @eval_dirs > 0) ? &util::filename_cat(@eval_dirs) : "";
|
---|
1160 | return $evaluated_filename;
|
---|
1161 | }
|
---|
1162 |
|
---|
1163 | sub replace_usemap_links {
|
---|
1164 | my $self = shift (@_);
|
---|
1165 | my ($front, $link, $back) = @_;
|
---|
1166 |
|
---|
1167 | # remove quotes from link at start and end if necessary
|
---|
1168 | if ($link=~/^[\"\']/) {
|
---|
1169 | $link=~s/^[\"\']//;
|
---|
1170 | $link=~s/[\"\']$//;
|
---|
1171 | $front.='"';
|
---|
1172 | $back="\"$back";
|
---|
1173 | }
|
---|
1174 |
|
---|
1175 | $link =~ s/^\.\///;
|
---|
1176 | return $front . $link . $back;
|
---|
1177 | }
|
---|
1178 |
|
---|
1179 | sub inc_filecount {
|
---|
1180 | my $self = shift (@_);
|
---|
1181 |
|
---|
1182 | if ($self->{'file_num'} == 1000) {
|
---|
1183 | $self->{'dir_num'} ++;
|
---|
1184 | $self->{'file_num'} = 0;
|
---|
1185 | } else {
|
---|
1186 | $self->{'file_num'} ++;
|
---|
1187 | }
|
---|
1188 | }
|
---|
1189 |
|
---|
1190 |
|
---|
1191 | # Extend read_file so that strings like é are
|
---|
1192 | # converted to UTF8 internally.
|
---|
1193 | #
|
---|
1194 | # We don't convert < or > or & or " in case
|
---|
1195 | # they interfere with the GML files
|
---|
1196 |
|
---|
1197 | sub read_file {
|
---|
1198 | my $self = shift(@_);
|
---|
1199 | my ($filename, $encoding, $language, $textref) = @_;
|
---|
1200 |
|
---|
1201 | $self->SUPER::read_file($filename, $encoding, $language, $textref);
|
---|
1202 |
|
---|
1203 | # Convert entities to their UTF8 equivalents
|
---|
1204 | $$textref =~ s/&(lt|gt|amp|quot|nbsp);/&z$1;/go;
|
---|
1205 | $$textref =~ s/&([^;]+);/&ghtml::getcharequiv($1,1)/gseo;
|
---|
1206 | $$textref =~ s/&z(lt|gt|amp|quot|nbsp);/&$1;/go;
|
---|
1207 | }
|
---|
1208 |
|
---|
1209 | sub HB_read_html_file {
|
---|
1210 | my $self = shift (@_);
|
---|
1211 | my ($htmlfile, $text) = @_;
|
---|
1212 |
|
---|
1213 | # load in the file
|
---|
1214 | if (!open (FILE, $htmlfile)) {
|
---|
1215 | print STDERR "ERROR - could not open $htmlfile\n";
|
---|
1216 | return;
|
---|
1217 | }
|
---|
1218 |
|
---|
1219 | my $foundbody = 0;
|
---|
1220 | $self->HB_gettext (\$foundbody, $text, "FILE");
|
---|
1221 | close FILE;
|
---|
1222 |
|
---|
1223 | # just in case there was no <body> tag
|
---|
1224 | if (!$foundbody) {
|
---|
1225 | $foundbody = 1;
|
---|
1226 | open (FILE, $htmlfile) || return;
|
---|
1227 | $self->HB_gettext (\$foundbody, $text, "FILE");
|
---|
1228 | close FILE;
|
---|
1229 | }
|
---|
1230 | # text is in utf8
|
---|
1231 | }
|
---|
1232 |
|
---|
1233 | # converts the text to utf8, as ghtml does that for é etc.
|
---|
1234 | sub HB_gettext {
|
---|
1235 | my $self = shift (@_);
|
---|
1236 | my ($foundbody, $text, $handle) = @_;
|
---|
1237 |
|
---|
1238 | my $line = "";
|
---|
1239 | while (defined ($line = <$handle>)) {
|
---|
1240 | # look for body tag
|
---|
1241 | if (!$$foundbody) {
|
---|
1242 | if ($line =~ s/^.*<body[^>]*>//i) {
|
---|
1243 | $$foundbody = 1;
|
---|
1244 | } else {
|
---|
1245 | next;
|
---|
1246 | }
|
---|
1247 | }
|
---|
1248 |
|
---|
1249 | # check for symbol fonts
|
---|
1250 | if ($line =~ m/<font [^>]*?face\s*=\s*\"?(\w+)\"?/i) {
|
---|
1251 | my $font = $1;
|
---|
1252 | print STDERR "HBPlug::HB_gettext - warning removed font $font\n"
|
---|
1253 | if ($font !~ m/^arial$/i);
|
---|
1254 | }
|
---|
1255 |
|
---|
1256 | $$text .= $line;
|
---|
1257 | }
|
---|
1258 |
|
---|
1259 | if ($self->{'input_encoding'} eq "iso_8859_1") {
|
---|
1260 | # convert to utf-8
|
---|
1261 | $$text=&unicode::unicode2utf8(&unicode::convert2unicode("iso_8859_1", $text));
|
---|
1262 | }
|
---|
1263 | # convert any alphanumeric character entities to their utf-8
|
---|
1264 | # equivalent for indexing purposes
|
---|
1265 | #&ghtml::convertcharentities ($$text);
|
---|
1266 |
|
---|
1267 | $$text =~ s/\s+/ /g; # remove \n's
|
---|
1268 | }
|
---|
1269 |
|
---|
1270 | sub HB_clean_section {
|
---|
1271 | my $self = shift (@_);
|
---|
1272 | my ($section) = @_;
|
---|
1273 |
|
---|
1274 | # remove tags without a starting tag from the section
|
---|
1275 | my ($tag, $tagstart);
|
---|
1276 | while ($section =~ m/<\/([^>]{1,10})>/) {
|
---|
1277 | $tag = $1;
|
---|
1278 | $tagstart = index($section, "<$tag");
|
---|
1279 | last if (($tagstart >= 0) && ($tagstart < index($section, "<\/$tag")));
|
---|
1280 | $section =~ s/<\/$tag>//;
|
---|
1281 | }
|
---|
1282 |
|
---|
1283 | # remove extra paragraph tags
|
---|
1284 | while ($section =~ s/<p\b[^>]*>\s*<p\b/<p/ig) {}
|
---|
1285 |
|
---|
1286 | # remove extra stuff at the end of the section
|
---|
1287 | while ($section =~ s/(<u>|<i>|<b>|<p\b[^>]*>| |\s)$//i) {}
|
---|
1288 |
|
---|
1289 | # add a newline at the beginning of each paragraph
|
---|
1290 | $section =~ s/(.)\s*<p\b/$1\n\n<p/gi;
|
---|
1291 |
|
---|
1292 | # add a newline every 80 characters at a word boundary
|
---|
1293 | # Note: this regular expression puts a line feed before
|
---|
1294 | # the last word in each section, even when it is not
|
---|
1295 | # needed.
|
---|
1296 | $section =~ s/(.{1,80})\s/$1\n/g;
|
---|
1297 |
|
---|
1298 | # fix up the image links
|
---|
1299 | $section =~ s/<img[^>]*?src=\"?([^\">]+)\"?[^>]*>/
|
---|
1300 | <center><img src=\"$1\" \/><\/center><br\/>/ig;
|
---|
1301 | $section =~ s/<<I>>\s*([^\.]+\.(png|jpg|gif))/
|
---|
1302 | <center><img src=\"$1\" \/><\/center><br\/>/ig;
|
---|
1303 |
|
---|
1304 | return $section;
|
---|
1305 | }
|
---|
1306 |
|
---|
1307 | # Will convert the oldHDL format to the new HDL format (using the Section tag)
|
---|
1308 | sub convert_to_newHDLformat
|
---|
1309 | {
|
---|
1310 | my $self = shift (@_);
|
---|
1311 | my ($file,$cnfile) = @_;
|
---|
1312 | my $input_filename = $file;
|
---|
1313 | my $tmp_filename = $cnfile;
|
---|
1314 |
|
---|
1315 | # write HTML tmp file with new HDL format
|
---|
1316 | open (PROD, ">$tmp_filename") || die("Error Writing to File: $tmp_filename $!");
|
---|
1317 |
|
---|
1318 | # read in the file and do basic html cleaning (removing header etc)
|
---|
1319 | my $html = "";
|
---|
1320 | $self->HB_read_html_file ($input_filename, \$html);
|
---|
1321 |
|
---|
1322 | # process the file one section at a time
|
---|
1323 | my $curtoclevel = 1;
|
---|
1324 | my $firstsection = 1;
|
---|
1325 | my $toclevel = 0;
|
---|
1326 | while (length ($html) > 0) {
|
---|
1327 | if ($html =~ s/^.*?(?:<p\b[^>]*>)?((<b>|<i>|<u>|\s)*)<<TOC(\d+)>>\s*(.*?)<p\b/<p/i) {
|
---|
1328 | $toclevel = $3;
|
---|
1329 | my $title = $4;
|
---|
1330 | my $sectiontext = "";
|
---|
1331 | if ($html =~ s/^(.*?)((?:<p\b[^>]*>)?((<b>|<i>|<u>|\s)*)<<TOC\d+>>)/$2/i) {
|
---|
1332 | $sectiontext = $1;
|
---|
1333 | } else {
|
---|
1334 | $sectiontext = $html;
|
---|
1335 | $html = "";
|
---|
1336 | }
|
---|
1337 |
|
---|
1338 | # remove tags and extra spaces from the title
|
---|
1339 | $title =~ s/<\/?[^>]+>//g;
|
---|
1340 | $title =~ s/^\s+|\s+$//g;
|
---|
1341 |
|
---|
1342 | # close any sections below the current level and
|
---|
1343 | # create a new section (special case for the firstsection)
|
---|
1344 | print PROD "<!--\n";
|
---|
1345 | while (($curtoclevel > $toclevel) ||
|
---|
1346 | (!$firstsection && $curtoclevel == $toclevel)) {
|
---|
1347 | $curtoclevel--;
|
---|
1348 | print PROD "</Section>\n";
|
---|
1349 | }
|
---|
1350 | if ($curtoclevel+1 < $toclevel) {
|
---|
1351 | print STDERR "WARNING - jump in toc levels in $input_filename " .
|
---|
1352 | "from $curtoclevel to $toclevel\n";
|
---|
1353 | }
|
---|
1354 | while ($curtoclevel < $toclevel) {
|
---|
1355 | $curtoclevel++;
|
---|
1356 | }
|
---|
1357 |
|
---|
1358 | if ($curtoclevel == 1) {
|
---|
1359 | # add the header tag
|
---|
1360 | print PROD "-->\n";
|
---|
1361 | print PROD "<HTML>\n<HEAD>\n<TITLE>$title</TITLE>\n</HEAD>\n<BODY>\n";
|
---|
1362 | print PROD "<!--\n";
|
---|
1363 | }
|
---|
1364 |
|
---|
1365 | print PROD "<Section>\n\t<Description>\n\t\t<Metadata name=\"Title\">$title</Metadata>\n\t</Description>\n";
|
---|
1366 |
|
---|
1367 | print PROD "-->\n";
|
---|
1368 |
|
---|
1369 | # clean up the section html
|
---|
1370 | $sectiontext = $self->HB_clean_section($sectiontext);
|
---|
1371 |
|
---|
1372 | print PROD "$sectiontext\n";
|
---|
1373 |
|
---|
1374 | } else {
|
---|
1375 | print STDERR "WARNING - leftover text\n" , $self->shorten($html),
|
---|
1376 | "\nin $input_filename\n";
|
---|
1377 | last;
|
---|
1378 | }
|
---|
1379 | $firstsection = 0;
|
---|
1380 | }
|
---|
1381 |
|
---|
1382 | print PROD "<!--\n";
|
---|
1383 | while ($curtoclevel > 0) {
|
---|
1384 | $curtoclevel--;
|
---|
1385 | print PROD "</Section>\n";
|
---|
1386 | }
|
---|
1387 | print PROD "-->\n";
|
---|
1388 |
|
---|
1389 | close (PROD) || die("Error Closing File: $tmp_filename $!");
|
---|
1390 |
|
---|
1391 | return $tmp_filename;
|
---|
1392 | }
|
---|
1393 |
|
---|
1394 | sub shorten {
|
---|
1395 | my $self = shift (@_);
|
---|
1396 | my ($text) = @_;
|
---|
1397 |
|
---|
1398 | return "\"$text\"" if (length($text) < 100);
|
---|
1399 |
|
---|
1400 | return "\"" . substr ($text, 0, 50) . "\" ... \"" .
|
---|
1401 | substr ($text, length($text)-50) . "\"";
|
---|
1402 | }
|
---|
1403 |
|
---|
1404 | sub convert_tidy_or_oldHDL_file
|
---|
1405 | {
|
---|
1406 | my $self = shift (@_);
|
---|
1407 | my ($file) = @_;
|
---|
1408 | my $input_filename = $file;
|
---|
1409 |
|
---|
1410 | if (-d $input_filename)
|
---|
1411 | {
|
---|
1412 | return $input_filename;
|
---|
1413 | }
|
---|
1414 |
|
---|
1415 | # get the input filename
|
---|
1416 | my ($tailname, $dirname, $suffix) = &File::Basename::fileparse($input_filename, "\\.[^\\.]+\$");
|
---|
1417 | my $base_dirname = $dirname;
|
---|
1418 | $suffix = lc($suffix);
|
---|
1419 |
|
---|
1420 | # derive tmp filename from input filename
|
---|
1421 | # Remove any white space from filename -- no risk of name collision, and
|
---|
1422 | # makes later conversion by utils simpler. Leave spaces in path...
|
---|
1423 | # tidy up the filename with space, dot, hyphen between
|
---|
1424 | $tailname =~ s/\s+//g;
|
---|
1425 | $tailname =~ s/\.+//g;
|
---|
1426 | $tailname =~ s/\-+//g;
|
---|
1427 | # convert to utf-8 otherwise we have problems with the doc.xml file
|
---|
1428 | # later on
|
---|
1429 | &unicode::ensure_utf8(\$tailname);
|
---|
1430 |
|
---|
1431 | # softlink to collection tmp dir
|
---|
1432 | my $tmp_dirname = &util::filename_cat($ENV{'GSDLCOLLECTDIR'}, "tidytmp");
|
---|
1433 | &util::mk_dir($tmp_dirname) if (!-e $tmp_dirname);
|
---|
1434 |
|
---|
1435 | my $test_dirname = "";
|
---|
1436 | my $f_separator = &util::get_os_dirsep();
|
---|
1437 |
|
---|
1438 | if ($dirname =~ m/import$f_separator/)
|
---|
1439 | {
|
---|
1440 | $test_dirname = $'; #'
|
---|
1441 |
|
---|
1442 | #print STDERR "init $'\n";
|
---|
1443 |
|
---|
1444 | while ($test_dirname =~ m/[$f_separator]/)
|
---|
1445 | {
|
---|
1446 | my $folderdirname = $`;
|
---|
1447 | $tmp_dirname = &util::filename_cat($tmp_dirname,$folderdirname);
|
---|
1448 | &util::mk_dir($tmp_dirname) if (!-e $tmp_dirname);
|
---|
1449 | $test_dirname = $'; #'
|
---|
1450 | }
|
---|
1451 | }
|
---|
1452 |
|
---|
1453 | my $tmp_filename = &util::filename_cat($tmp_dirname, "$tailname$suffix");
|
---|
1454 |
|
---|
1455 | # tidy or convert the input file if it is a HTML-like file or it is accepted by the process_exp
|
---|
1456 | if (($suffix eq ".htm") || ($suffix eq ".html") || ($suffix eq ".shtml"))
|
---|
1457 | {
|
---|
1458 | #convert the input file to a new style HDL
|
---|
1459 | my $hdl_output_filename = $input_filename;
|
---|
1460 | if ($self->{'old_style_HDL'})
|
---|
1461 | {
|
---|
1462 | $hdl_output_filename = &util::filename_cat($tmp_dirname, "$tailname$suffix");
|
---|
1463 | $hdl_output_filename = $self->convert_to_newHDLformat($input_filename,$hdl_output_filename);
|
---|
1464 | }
|
---|
1465 |
|
---|
1466 | #just for checking copy all other file from the base dir to tmp dir if it is not exists
|
---|
1467 | opendir(DIR,$base_dirname) or die "Can't open base directory : $base_dirname!";
|
---|
1468 | my @files = grep {!/^\.+$/} readdir(DIR);
|
---|
1469 | close(DIR);
|
---|
1470 |
|
---|
1471 | foreach my $file (@files)
|
---|
1472 | {
|
---|
1473 | my $src_file = &util::filename_cat($base_dirname,$file);
|
---|
1474 | my $dest_file = &util::filename_cat($tmp_dirname,$file);
|
---|
1475 | if ((!-e $dest_file) && (!-d $src_file))
|
---|
1476 | {
|
---|
1477 | # just copy the original file back to the tmp directory
|
---|
1478 | copy($src_file,$dest_file) or die "Can't copy file $src_file to $dest_file $!";
|
---|
1479 | }
|
---|
1480 | }
|
---|
1481 |
|
---|
1482 | # tidy the input file
|
---|
1483 | my $tidy_output_filename = $hdl_output_filename;
|
---|
1484 | if ($self->{'use_realistic_book'})
|
---|
1485 | {
|
---|
1486 | $tidy_output_filename = &util::filename_cat($tmp_dirname, "$tailname$suffix");
|
---|
1487 | $tidy_output_filename = $self->tmp_tidy_file($hdl_output_filename,$tidy_output_filename);
|
---|
1488 | }
|
---|
1489 | $tmp_filename = $tidy_output_filename;
|
---|
1490 | }
|
---|
1491 | else
|
---|
1492 | {
|
---|
1493 | if (!-e $tmp_filename)
|
---|
1494 | {
|
---|
1495 | # just copy the original file back to the tmp directory
|
---|
1496 | copy($input_filename,$tmp_filename) or die "Can't copy file $input_filename to $tmp_filename $!";
|
---|
1497 | }
|
---|
1498 | }
|
---|
1499 |
|
---|
1500 | return $tmp_filename;
|
---|
1501 | }
|
---|
1502 |
|
---|
1503 |
|
---|
1504 | # Will make the html input file as a proper XML file with removed font tag and
|
---|
1505 | # image size added to the img tag.
|
---|
1506 | # The tidying process takes place in a collection specific 'tmp' directory so
|
---|
1507 | # that we don't accidentally damage the input.
|
---|
1508 | sub tmp_tidy_file
|
---|
1509 | {
|
---|
1510 | my $self = shift (@_);
|
---|
1511 | my ($file,$cnfile) = @_;
|
---|
1512 | my $input_filename = $file;
|
---|
1513 | my $tmp_filename = $cnfile;
|
---|
1514 |
|
---|
1515 | # get the input filename
|
---|
1516 | my ($tailname, $dirname, $suffix) = &File::Basename::fileparse($input_filename, "\\.[^\\.]+\$");
|
---|
1517 |
|
---|
1518 | require HTML::TokeParser::Simple;
|
---|
1519 |
|
---|
1520 | # create HTML parser to decode the input file
|
---|
1521 | my $parser = HTML::TokeParser::Simple->new($input_filename);
|
---|
1522 |
|
---|
1523 | # write HTML tmp file without the font tag and image size are added to the img tag
|
---|
1524 | open (PROD, ">$tmp_filename") || die("Error Writing to File: $tmp_filename $!");
|
---|
1525 | while (my $token = $parser->get_token())
|
---|
1526 | {
|
---|
1527 | # is it an img tag
|
---|
1528 | if ($token->is_start_tag('img'))
|
---|
1529 | {
|
---|
1530 | # get the attributes
|
---|
1531 | my $attr = $token->return_attr;
|
---|
1532 |
|
---|
1533 | # get the full path to the image
|
---|
1534 | my $img_file = &util::filename_cat($dirname,$attr->{src});
|
---|
1535 |
|
---|
1536 | # set the width and height attribute
|
---|
1537 | ($attr->{width}, $attr->{height}) = imgsize($img_file);
|
---|
1538 |
|
---|
1539 | # recreate the tag
|
---|
1540 | print PROD "<img";
|
---|
1541 | print PROD map { qq { $_="$attr->{$_}"} } keys %$attr;
|
---|
1542 | print PROD ">";
|
---|
1543 | }
|
---|
1544 | # is it a font tag
|
---|
1545 | else
|
---|
1546 | {
|
---|
1547 | if (($token->is_start_tag('font')) || ($token->is_end_tag('font')))
|
---|
1548 | {
|
---|
1549 | # remove font tag
|
---|
1550 | print PROD "";
|
---|
1551 | }
|
---|
1552 | else
|
---|
1553 | {
|
---|
1554 | # print without changes
|
---|
1555 | print PROD $token->as_is;
|
---|
1556 | }
|
---|
1557 | }
|
---|
1558 | }
|
---|
1559 | close (PROD) || die("Error Closing File: $tmp_filename $!");
|
---|
1560 |
|
---|
1561 | # run html-tidy on the tmp file to make it a proper XML file
|
---|
1562 | my $tidyfile = `tidy -utf8 -wrap 0 -asxml "$tmp_filename"`;
|
---|
1563 |
|
---|
1564 | # write result back to the tmp file
|
---|
1565 | open (PROD, ">$tmp_filename") || die("Error Writing to File: $tmp_filename $!");
|
---|
1566 | print PROD $tidyfile;
|
---|
1567 | close (PROD) || die("Error Closing File: $tmp_filename $!");
|
---|
1568 |
|
---|
1569 | # return the output filename
|
---|
1570 | return $tmp_filename;
|
---|
1571 | }
|
---|
1572 |
|
---|
1573 | 1;
|
---|