root/main/trunk/greenstone2/perllib/plugins/PDFPlugin.pm @ 22597

Revision 22597, 12.3 KB (checked in by kjdon, 9 years ago)

code tidy up. rearranged how convertbinaryfile plugins set up their secondary plugins - now only set up the options for the one they are using. all subclass specific code moved out of convertbinaryfile.new into the appropriate plugin file.

  • 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 $self = new ConvertBinaryFile($pluginlist, $inputargs, $hashArgOptLists);
122   
123    if ($self->{'info_only'}) {
124    # don't worry about any options etc
125    return bless $self, $class;
126    }
127
128    $self->{'filename_extension'} = "pdf";
129    $self->{'file_type'} = "PDF";
130
131    # these are passed through to gsConvert.pl by ConvertBinaryFile.pm
132    my $zoom = $self->{"zoom"};
133    $self->{'convert_options'} = "-pdf_zoom $zoom";
134    $self->{'convert_options'} .= " -pdf_complex" if $self->{"complex"};
135    $self->{'convert_options'} .= " -pdf_nohidden" if $self->{"nohidden"};
136    $self->{'convert_options'} .= " -pdf_ignore_images" if $self->{"noimages"};
137    $self->{'convert_options'} .= " -pdf_allow_images_only" if $self->{"allowimagesonly"};
138
139    # check convert_to
140    if ($self->{'convert_to'} eq "text" && $ENV{'GSDLOS'} =~ /^windows$/i) {
141    print STDERR "Windows does not support pdf to text. PDFs will be converted to HTML instead\n";
142    $self->{'convert_to'} = "html";
143    }
144    elsif ($self->{'convert_to'} eq "auto") {
145    # choose html ?? is this the best option
146    $self->{'convert_to'} = "html";
147    }
148    # set convert_to_plugin and convert_to_ext
149    $self->ConvertBinaryFile::set_standard_convert_settings();
150
151    my $secondary_plugin_name = $self->{'convert_to_plugin'};
152    my $secondary_plugin_options = $self->{'secondary_plugin_options'};
153
154    if (!defined $secondary_plugin_options->{$secondary_plugin_name}) {
155    $secondary_plugin_options->{$secondary_plugin_name} = [];
156    }
157    my $specific_options = $secondary_plugin_options->{$secondary_plugin_name};
158
159    # following title_sub removes "Page 1" added by pdftohtml, and a leading
160    # "1", which is often the page number at the top of the page. Bad Luck
161    # if your document title actually starts with "1 " - is there a better way?
162    push(@$specific_options , "-title_sub", '^(Page\s+\d+)?(\s*1\s+)?');
163    my $associate_tail_re = $self->{'associate_tail_re'};
164    if ((defined $associate_tail_re) && ($associate_tail_re ne "")) {
165    push(@$specific_options, "-associate_tail_re", $associate_tail_re);
166    }
167    push(@$specific_options, "-file_rename_method", "none");
168   
169    if ($secondary_plugin_name eq "HTMLPlugin") {
170    # pdftohtml always produces utf8
171    push(@$specific_options, "-input_encoding", "utf8");
172    push(@$specific_options, "-extract_language") if $self->{'extract_language'};
173    push(@$specific_options, "-processing_tmp_files");
174    # Instruct HTMLPlug (when eventually accessed through read_into_doc_obj)
175    # to extract these metadata fields from the HEAD META fields
176    if (defined $self->{'metadata_fields'} && $self->{'metadata_fields'} =~ /\S/) {
177        push(@$specific_options,"-metadata_fields",$self->{'metadata_fields'});
178    } else {
179        push(@$specific_options,"-metadata_fields","Title,GENERATOR,date,author<Creator>");
180    }
181    if (defined $self->{'metadata_field_separator'} && $self->{'metadata_field_separator'} =~ /\S/) {
182        push(@$specific_options,"-metadata_field_separator",$self->{'metadata_field_separator'});
183    }
184    if ($self->{'use_sections'} || $self->{'description_tags'}) {
185        $self->{'description_tags'} = 1;
186        push(@$specific_options, "-description_tags");
187    }
188    }
189    elsif ($secondary_plugin_name eq "PagedImagePlugin") {
190    push(@$specific_options, "-screenviewsize", "1000");
191    push(@$specific_options, "-enable_cache");
192    push(@$specific_options, "-processing_tmp_files");
193    }
194
195    $self = bless $self, $class;
196    $self->load_secondary_plugins($class,$secondary_plugin_options,$hashArgOptLists);
197    return $self;
198}
199
200sub get_default_process_exp {
201    my $self = shift (@_);
202
203    return q^(?i)\.pdf$^;
204}
205
206# so we don't inherit HTMLPlug's block exp...
207sub get_default_block_exp {
208    return "";
209}
210   
211sub convert_post_process
212{
213    my $self = shift (@_);
214    my ($conv_filename) = @_;
215
216    my $outhandle=$self->{'outhandle'};
217
218    #$self->{'input_encoding'} = "utf8"; # The output is always in utf8 (is it?? it is for html, but what about other types?)
219    #my ($language, $encoding) = $self->textcat_get_language_encoding ($conv_filename);
220
221    # read in file ($text will be in utf8)
222    my $text = "";
223    # encoding will be utf8 for html files - what about other types? will we do this step for them anyway?
224    $self->read_file ($conv_filename, "utf8", "", \$text);
225
226    # Calculate number of pages based on <a ...> tags (we have a <a name=1> etc
227    # for each page).  Metadata based on this calculation not set until process()
228    #
229    # Note: this is done even if we are not breaking to document into pages as it might
230    # be useful to give an indication of document length in browser through setting
231    # num_pages as metadata.
232    my @pages = ($text =~ /\<[Aa] name=\"?\w+\"?>/ig);
233    my $num_pages = scalar(@pages);
234    $self->{'num_pages'} = $num_pages;
235
236    if ($self->{'use_sections'}
237    && $self->{'converted_to'} eq "HTML") {
238
239    print $outhandle "PDFPlugin: Calculating sections...\n";
240
241    # we have "<a name=1></a>" etc for each page
242    # it may be <A name=
243    my @sections = split('<[Aa] name=', $text);
244
245    my $top_section = "";
246
247    if (scalar (@sections) == 1) { #only one section - no split!
248        print $outhandle "PDFPlugin: warning - no sections found\n";
249    } else {
250        $top_section .= shift @sections; # keep HTML header etc as top_section
251    }
252
253    # handle first section specially for title? Or all use first 100...
254   
255    my $title = $sections[0];
256    $title =~ s/^\"?\w+\"?>//; # specific for pdftohtml...
257    $title =~ s/<\/([^>]+)><\1>//g; # (eg) </b><b> - no space
258    $title =~ s/<[^>]*>/ /g;
259    $title =~ s/(?:&nbsp;|\xc2\xa0)/ /g; # utf-8 for nbsp...
260    $title =~ s/^\s+//s;
261    $title =~ s/\s+$//;
262    $title =~ s/\s+/ /gs;
263    $title =~ s/^$self->{'title_sub'}// if ($self->{'title_sub'});
264    $title =~ s/^\s+//s; # in case title_sub introduced any...
265    $title = substr ($title, 0, 100);
266    $title =~ s/\s\S*$/.../;
267
268
269    if (scalar (@sections) == 1) { # no sections found
270        $top_section .= $sections[0];
271        @sections=();
272    } else {
273        $top_section .= "<!--<Section>\n<Metadata name=\"Title\">$title</Metadata>\n-->\n <!--</Section>-->\n";
274    }
275
276    # add metadata per section...
277    foreach my $section (@sections) {
278        # section names are not always just digits, may be like "outline"
279        $section =~ s@^\"?(\w+)\"?></a>@@; # leftover from split expression...
280
281        $title = $1; # Greenstone does magic if sections are titled digits
282        if (! defined($title) ) {
283        print STDERR "no title: $section\n";
284        $title = " "; # get rid of the undefined warning in next line
285        }
286        my $newsection = "<!-- from PDFPlugin -->\n<!-- <Section>\n";
287        $newsection .= "<Metadata name=\"Title\">" . $title
288        . "</Metadata>\n--><p>\n";
289        $newsection .= $section;
290        $newsection .= "<!--</Section>-->\n";
291        $section = $newsection;
292    }
293
294    $text=join('', ($top_section, @sections));
295    }
296
297    # turn any high bytes that aren't valid utf-8 into utf-8.
298    unicode::ensure_utf8(\$text);
299
300    # Write it out again!
301    $self->utf8_write_file (\$text, $conv_filename);
302}
303
304
305# do plugin specific processing of doc_obj for HTML type
306sub process {
307    my $self = shift (@_);
308    my ($pluginfo, $base_dir, $file, $metadata, $doc_obj, $gli) = @_;
309
310    my $result = $self->process_type($base_dir,$file,$doc_obj);
311
312    # fix up the extracted date metadata to be in Greenstone date format,
313    # and fix the capitalisation of 'date'
314    my $cursection = $doc_obj->get_top_section();
315    foreach my $datemeta (@{$doc_obj->get_metadata($cursection, "date")}) {
316    $doc_obj->delete_metadata($cursection, "date", $datemeta);
317
318    # We're just interested in the date bit, not the time
319    # some pdf creators (eg "Acrobat 5.0 Scan Plug-in for Windows")
320    # set a /CreationDate, and set /ModDate to 000000000. pdftohtml
321    # extracts the ModDate, so it is 0...
322    $datemeta =~ /(\d+)-(\d+)-(\d+)/;
323    my ($year, $month, $day) = ($1,$2,$3);
324    if (defined($year) && defined($month) && defined($day)) {
325        if ($year == 0) {next}
326        if ($year < 100) {$year += 1900} # just to be safe
327        if ($month =~ /^\d$/) {$month="0$month"} # single digit
328        if ($day =~ /^\d$/) {$day="0$day"} # single digit
329        my $date="$year$month$day";
330        $doc_obj->add_utf8_metadata($cursection, "Date", $date);
331    }
332    }
333
334    $doc_obj->add_utf8_metadata($cursection, "NumPages", $self->{'num_pages'});
335   
336    if ($self->{'use_sections'} && $self->{'converted_to'} eq "HTML") {
337    # we explicitly make it a paged document, cos greenstone won't get it
338    # right if any section has an empty title, or one with letters in it
339    $doc_obj->set_utf8_metadata_element ($cursection, "gsdlthistype", "Paged");
340    }
341
342    return $result;
343}
344
3451;
Note: See TracBrowser for help on using the browser.