root/main/trunk/greenstone2/perllib/plugins/PDFPlugin.pm @ 21800

Revision 21800, 12.8 KB (checked in by kjdon, 10 years ago)

added a new option, metadata_field_separator, which specifies what to split on for multi-valued metadata

  • Property svn:keywords set to Author Date Id Revision
Line 
1###########################################################################
2#
3# PDFPlugin.pm -- reasonably with-it pdf plugin
4# A component of the Greenstone digital library software
5# from the New Zealand Digital Library Project at the
6# University of Waikato, New Zealand.
7#
8# Copyright (C) 1999-2001 New Zealand Digital Library Project
9#
10# This program is free software; you can redistribute it and/or modify
11# it under the terms of the GNU General Public License as published by
12# the Free Software Foundation; either version 2 of the License, or
13# (at your option) any later version.
14#
15# This program is distributed in the hope that it will be useful,
16# but WITHOUT ANY WARRANTY; without even the implied warranty of
17# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
18# GNU General Public License for more details.
19#
20# You should have received a copy of the GNU General Public License
21# along with this program; if not, write to the Free Software
22# Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
23#
24###########################################################################
25package PDFPlugin;
26
27use ConvertBinaryFile;
28use ReadTextFile;
29use unicode;
30use strict;
31no strict 'refs'; # so we can use a var for filehandles (eg STDERR)
32
33sub BEGIN {
34    @PDFPlugin::ISA = ('ConvertBinaryFile', 'ReadTextFile');
35}
36
37my $convert_to_list =
38    [ { 'name' => "auto",
39    'desc' => "{ConvertBinaryFile.convert_to.auto}" },
40      { 'name' => "html",
41    'desc' => "{ConvertBinaryFile.convert_to.html}" },
42      { 'name' => "text",
43    'desc' => "{ConvertBinaryFile.convert_to.text}" },
44      { 'name' => "pagedimg_jpg",
45    'desc' => "{ConvertBinaryFile.convert_to.pagedimg_jpg}"},
46      { 'name' => "pagedimg_gif",
47    'desc' => "{ConvertBinaryFile.convert_to.pagedimg_gif}"},
48      { 'name' => "pagedimg_png",
49    'desc' => "{ConvertBinaryFile.convert_to.pagedimg_png}"},
50      ];
51
52
53my $arguments =
54    [
55     { 'name' => "convert_to",
56       'desc' => "{ConvertBinaryFile.convert_to}",
57       'type' => "enum",
58       'reqd' => "yes",
59       'list' => $convert_to_list,
60       'deft' => "html" },   
61     { 'name' => "process_exp",
62       'desc' => "{BasePlugin.process_exp}",
63       'type' => "regexp",
64       'deft' => &get_default_process_exp(),
65       'reqd' => "no" },
66     { 'name' => "block_exp",
67       'desc' => "{BasePlugin.block_exp}",
68       'type' => "regexp",
69       'deft' => &get_default_block_exp() },
70     { 'name' => "metadata_fields",
71       'desc' => "{HTMLPlugin.metadata_fields}",
72       'type' => "string",
73       'deft' => "" },
74      { 'name' => "metadata_field_separator",
75    'desc' => "{HTMLPlugin.metadata_field_separator}",
76    'type' => "string",
77    'deft' => "" },
78     { 'name' => "noimages",
79       'desc' => "{PDFPlugin.noimages}",
80       'type' => "flag" },
81     { 'name' => "allowimagesonly",
82       'desc' => "{PDFPlugin.allowimagesonly}",
83       'type' => "flag" },
84     { 'name' => "complex",
85       'desc' => "{PDFPlugin.complex}",
86       'type' => "flag" },
87     { 'name' => "nohidden",
88       'desc' => "{PDFPlugin.nohidden}",
89       'type' => "flag" },
90     { 'name' => "zoom",
91       'desc' => "{PDFPlugin.zoom}",
92       'deft' => "2",
93       'range' => "1,3", # actually the range is 0.5-3
94       'type' => "int" },
95     { 'name' => "use_sections",
96       'desc' => "{PDFPlugin.use_sections}",
97       'type' => "flag" },
98     { 'name' => "description_tags",
99       'desc' => "{HTMLPlugin.description_tags}",
100       'type' => "flag" }
101     ];
102
103my $options = { 'name'     => "PDFPlugin",
104        'desc'     => "{PDFPlugin.desc}",
105        'abstract' => "no",
106        'inherits' => "yes",
107        'srcreplaceable' => "yes", # Source docs in PDF can be replaced with GS-generated html     
108        'args'     => $arguments };
109
110sub new {
111    my ($class) = shift (@_);
112    my ($pluginlist,$inputargs,$hashArgOptLists) = @_;
113    push(@$pluginlist, $class);
114
115    push(@$inputargs,"-title_sub");
116    push(@$inputargs,'^(Page\s+\d+)?(\s*1\s+)?');
117
118    push(@{$hashArgOptLists->{"ArgList"}},@{$arguments});
119    push(@{$hashArgOptLists->{"OptList"}},$options);
120
121    my @arg_array = @$inputargs;
122    my $self = new ConvertBinaryFile($pluginlist, $inputargs, $hashArgOptLists);
123   
124    if ($self->{'info_only'}) {
125    # don't worry about any options etc
126    return bless $self, $class;
127    }
128
129    $self->{'filename_extension'} = "pdf";
130    $self->{'file_type'} = "PDF";
131
132    # these are passed through to gsConvert.pl by ConvertBinaryFile.pm
133    my $zoom = $self->{"zoom"};
134    $self->{'convert_options'} = "-pdf_zoom $zoom";
135    $self->{'convert_options'} .= " -pdf_complex" if $self->{"complex"};
136    $self->{'convert_options'} .= " -pdf_nohidden" if $self->{"nohidden"};
137    $self->{'convert_options'} .= " -pdf_ignore_images" if $self->{"noimages"};
138    $self->{'convert_options'} .= " -pdf_allow_images_only" if $self->{"allowimagesonly"};
139
140    my $secondary_plugin_options = $self->{'secondary_plugin_options'};
141
142    if (!defined $secondary_plugin_options->{'HTMLPlugin'}) {
143    $secondary_plugin_options->{'HTMLPlugin'} = [];
144    }
145    if (!defined $secondary_plugin_options->{'TextPlugin'}) {
146    $secondary_plugin_options->{'TextPlugin'} = [];
147    }
148    if (defined $self->{'convert_to'} && $self->{'convert_to'} =~ m/(pagedimage|pagedimg).*/i) {
149    if (!defined $secondary_plugin_options->{'PagedImagePlugin'}){
150        $secondary_plugin_options->{'PagedImagePlugin'} = [];
151        my $pagedimg_options = $secondary_plugin_options->{'PagedImagePlugin'};
152        push(@$pagedimg_options, "-title_sub", '^(Page\s+\d+)?(\s*1\s+)?');
153        push(@$pagedimg_options, "-screenviewsize", "1000");
154        push(@$pagedimg_options, "-enable_cache");
155    }
156    }
157    my $html_options = $secondary_plugin_options->{'HTMLPlugin'};
158    my $text_options = $secondary_plugin_options->{'TextPlugin'};
159    my $pagedimg_options = $secondary_plugin_options->{'PagedImagePlugin'};
160
161#    if ($self->{'input_encoding'} eq "auto") {
162#   $self->{'input_encoding'} = "utf8";
163#    }
164
165    # if pdftohtml is always producing utf8, then htmlplug always needs this option
166    push(@$html_options,"-input_encoding", "utf8");
167    push(@$html_options,"-extract_language") if $self->{'extract_language'};
168
169    push(@$html_options, "-processing_tmp_files");
170    push(@$pagedimg_options, "-processing_tmp_files");
171
172    # Instruct HTMLPlug (when eventually accessed through read_into_doc_obj)
173    # to extract these metadata fields from the HEAD META fields
174    my $required_metadata;
175    if (defined $self->{'metadata_fields'} && $self->{'metadata_fields'} =~ /\S/) {
176    push(@$html_options,"-metadata_fields",$self->{'metadata_fields'});
177    } else {
178    push(@$html_options,"-metadata_fields","Title,GENERATOR,date,author<Creator>");
179    }
180    if (defined $self->{'metadata_field_separator'} && $self->{'metadata_field_separator'} =~ /\S/) {
181    push(@$html_options,"-metadata_field_separator",$self->{'metadata_field_separator'});
182    }
183       
184    if ($self->{'use_sections'} || $self->{'description_tags'}) {
185    $self->{'description_tags'} = 1;
186    push(@$html_options,"-description_tags");
187    }
188
189    # following title_sub removes "Page 1" added by pdftohtml, and a leading
190    # "1", which is often the page number at the top of the page. Bad Luck
191    # if your document title actually starts with "1 " - is there a better way?
192    push(@$html_options , "-title_sub", '^(Page\s+\d+)?(\s*1\s+)?');
193    push(@$text_options , "-title_sub", '^(Page\s+\d+)?(\s*1\s+)?');
194
195    my $associate_tail_re = $self->{'associate_tail_re'};
196    if ((defined $associate_tail_re) && ($associate_tail_re ne "")) {
197    push(@$html_options, "-associate_tail_re", $associate_tail_re);
198    push(@$text_options, "-associate_tail_re", $associate_tail_re);
199    push(@$pagedimg_options, "-associate_tail_re", $associate_tail_re) if defined $pagedimg_options;
200    }
201
202    push(@$html_options, "-file_rename_method", "none");
203    push(@$text_options, "-file_rename_method", "none");
204    push(@$pagedimg_options, "-file_rename_method", "none") if defined $pagedimg_options;
205
206    $self = bless $self, $class;
207    $self->load_secondary_plugins($class,$secondary_plugin_options,$hashArgOptLists);
208    return $self;
209}
210
211sub get_default_process_exp {
212    my $self = shift (@_);
213
214    return q^(?i)\.pdf$^;
215}
216
217# so we don't inherit HTMLPlug's block exp...
218sub get_default_block_exp {
219    return "";
220}
221   
222sub convert_post_process
223{
224    my $self = shift (@_);
225    my ($conv_filename) = @_;
226
227    my $outhandle=$self->{'outhandle'};
228
229    #$self->{'input_encoding'} = "utf8"; # The output is always in utf8 (is it?? it is for html, but what about other types?)
230    #my ($language, $encoding) = $self->textcat_get_language_encoding ($conv_filename);
231
232    # read in file ($text will be in utf8)
233    my $text = "";
234    # encoding will be utf8 for html files - what about other types? will we do this step for them anyway?
235    $self->read_file ($conv_filename, "utf8", "", \$text);
236
237    # Calculate number of pages based on <a ...> tags (we have a <a name=1> etc
238    # for each page).  Metadata based on this calculation not set until process()
239    #
240    # Note: this is done even if we are not breaking to document into pages as it might
241    # be useful to give an indication of document length in browser through setting
242    # num_pages as metadata.
243    my @pages = ($text =~ /\<[Aa] name=\"?\w+\"?>/ig);
244    my $num_pages = scalar(@pages);
245    $self->{'num_pages'} = $num_pages;
246
247    if ($self->{'use_sections'}
248    && $self->{'converted_to'} eq "HTML") {
249
250    print $outhandle "PDFPlugin: Calculating sections...\n";
251
252    # we have "<a name=1></a>" etc for each page
253    # it may be <A name=
254    my @sections = split('<[Aa] name=', $text);
255
256    my $top_section = "";
257
258    if (scalar (@sections) == 1) { #only one section - no split!
259        print $outhandle "PDFPlugin: warning - no sections found\n";
260    } else {
261        $top_section .= shift @sections; # keep HTML header etc as top_section
262    }
263
264    # handle first section specially for title? Or all use first 100...
265   
266    my $title = $sections[0];
267    $title =~ s/^\"?\w+\"?>//; # specific for pdftohtml...
268    $title =~ s/<\/([^>]+)><\1>//g; # (eg) </b><b> - no space
269    $title =~ s/<[^>]*>/ /g;
270    $title =~ s/(?:&nbsp;|\xc2\xa0)/ /g; # utf-8 for nbsp...
271    $title =~ s/^\s+//s;
272    $title =~ s/\s+$//;
273    $title =~ s/\s+/ /gs;
274    $title =~ s/^$self->{'title_sub'}// if ($self->{'title_sub'});
275    $title =~ s/^\s+//s; # in case title_sub introduced any...
276    $title = substr ($title, 0, 100);
277    $title =~ s/\s\S*$/.../;
278
279
280    if (scalar (@sections) == 1) { # no sections found
281        $top_section .= $sections[0];
282        @sections=();
283    } else {
284        $top_section .= "<!--<Section>\n<Metadata name=\"Title\">$title</Metadata>\n-->\n <!--</Section>-->\n";
285    }
286
287    # add metadata per section...
288    foreach my $section (@sections) {
289        # section names are not always just digits, may be like "outline"
290        $section =~ s@^\"?(\w+)\"?></a>@@; # leftover from split expression...
291
292        $title = $1; # Greenstone does magic if sections are titled digits
293        if (! defined($title) ) {
294        print STDERR "no title: $section\n";
295        $title = " "; # get rid of the undefined warning in next line
296        }
297        my $newsection = "<!-- from PDFPlugin -->\n<!-- <Section>\n";
298        $newsection .= "<Metadata name=\"Title\">" . $title
299        . "</Metadata>\n--><p>\n";
300        $newsection .= $section;
301        $newsection .= "<!--</Section>-->\n";
302        $section = $newsection;
303    }
304
305    $text=join('', ($top_section, @sections));
306    }
307
308    # turn any high bytes that aren't valid utf-8 into utf-8.
309    unicode::ensure_utf8(\$text);
310
311    # Write it out again!
312    $self->utf8_write_file (\$text, $conv_filename);
313}
314
315
316# do plugin specific processing of doc_obj for HTML type
317sub process {
318    my $self = shift (@_);
319    my ($pluginfo, $base_dir, $file, $metadata, $doc_obj, $gli) = @_;
320
321    my $result = $self->process_type($base_dir,$file,$doc_obj);
322
323    # fix up the extracted date metadata to be in Greenstone date format,
324    # and fix the capitalisation of 'date'
325    my $cursection = $doc_obj->get_top_section();
326    foreach my $datemeta (@{$doc_obj->get_metadata($cursection, "date")}) {
327    $doc_obj->delete_metadata($cursection, "date", $datemeta);
328
329    # We're just interested in the date bit, not the time
330    # some pdf creators (eg "Acrobat 5.0 Scan Plug-in for Windows")
331    # set a /CreationDate, and set /ModDate to 000000000. pdftohtml
332    # extracts the ModDate, so it is 0...
333    $datemeta =~ /(\d+)-(\d+)-(\d+)/;
334    my ($year, $month, $day) = ($1,$2,$3);
335    if (defined($year) && defined($month) && defined($day)) {
336        if ($year == 0) {next}
337        if ($year < 100) {$year += 1900} # just to be safe
338        if ($month =~ /^\d$/) {$month="0$month"} # single digit
339        if ($day =~ /^\d$/) {$day="0$day"} # single digit
340        my $date="$year$month$day";
341        $doc_obj->add_utf8_metadata($cursection, "Date", $date);
342    }
343    }
344
345    $doc_obj->add_utf8_metadata($cursection, "NumPages", $self->{'num_pages'});
346   
347    if ($self->{'use_sections'} && $self->{'converted_to'} eq "HTML") {
348    # we explicitly make it a paged document, cos greenstone won't get it
349    # right if any section has an empty title, or one with letters in it
350    $doc_obj->set_utf8_metadata_element ($cursection, "gsdlthistype", "Paged");
351    }
352
353    return $result;
354}
355
3561;
Note: See TracBrowser for help on using the browser.