source: gsdl/trunk/perllib/plugins/WordPlugin.pm@ 15969

Last change on this file since 15969 was 15969, checked in by kjdon, 16 years ago

added in desc for metadata_fields arg

  • Property svn:keywords set to Author Date Id Revision
File size: 8.2 KB
Line 
1###########################################################################
2#
3# WordPlugin.pm -- plugin for importing Microsoft Word documents
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 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 WordPlugin;
26
27use ConvertBinaryFile;
28use strict;
29no strict 'refs'; # allow filehandles to be variables and viceversa
30
31sub BEGIN {
32 @WordPlugin::ISA = ('ConvertBinaryFile');
33}
34
35my $arguments =
36 [ { 'name' => "process_exp",
37 'desc' => "{BasePlugin.process_exp}",
38 'type' => "regexp",
39 'deft' => &get_default_process_exp(),
40 'reqd' => "no" },
41 { 'name' => "description_tags",
42 'desc' => "{HTMLPlugin.description_tags}",
43 'type' => "flag" }
44 ];
45
46my $options = { 'name' => "WordPlugin",
47 'desc' => "{WordPlugin.desc}",
48 'abstract' => "no",
49 'inherits' => "yes",
50 'srcreplaceable' => "yes", # Source docs in Word can be replaced with GS-generated html
51 'args' => $arguments };
52
53sub new {
54 my ($class) = shift (@_);
55 my ($pluginlist,$inputargs,$hashArgOptLists) = @_;
56 push(@$pluginlist, $class);
57
58 if ($ENV{'GSDLOS'} =~ m/^windows$/i) {
59 my $ws_arg = [ { 'name' => "windows_scripting",
60 'desc' => "{WordPlugin.windows_scripting}",
61 'type' => "flag",
62 'reqd' => "no" },
63 { 'name' => "metadata_fields",
64 'desc' => "{WordPlugin.metadata_fields}",
65 'type' => "string",
66 'deft' => "Title" },
67 { 'name' => "level1_header",
68 'desc' => "{StructuredHTMLPlugin.level1_header}",
69 'type' => "regexp",
70 'reqd' => "no",
71 'deft' => "" },
72 { 'name' => "level2_header",
73 'desc' => "{StructuredHTMLPlugin.level2_header}",
74 'type' => "regexp",
75 'reqd' => "no",
76 'deft' => "" },
77 { 'name' => "level3_header",
78 'desc' => "{StructuredHTMLPlugin.level3_header}",
79 'type' => "regexp",
80 'reqd' => "no",
81 'deft' => "" },
82 { 'name' => "title_header",
83 'desc' => "{StructuredHTMLPlugin.title_header}",
84 'type' => "regexp",
85 'reqd' => "no",
86 'deft' => "" },
87 { 'name' => "delete_toc",
88 'desc' => "{StructuredHTMLPlugin.delete_toc}",
89 'type' => "flag",
90 'reqd' => "no",
91 # set the mode so this doesn't show up unless
92 # all the following ones do
93 'modegli' => "3"},
94 { 'name' => "toc_header",
95 'desc' => "{StructuredHTMLPlugin.toc_header}",
96 'type' => "regexp",
97 'reqd' => "no",
98 'deft' => "" }
99 ];
100
101 push(@$arguments,@$ws_arg);
102 }
103
104 push(@{$hashArgOptLists->{"ArgList"}},@{$arguments});
105 push(@{$hashArgOptLists->{"OptList"}},$options);
106
107 my $self = new ConvertBinaryFile($pluginlist, $inputargs, $hashArgOptLists);
108
109 if ($self->{'info_only'}) {
110 # don't worry about any options etc
111 return bless $self, $class;
112 }
113
114 $self->{'filename_extension'} = "doc";
115 $self->{'file_type'} = "Word";
116
117 #this is passed through to gsConvert.pl by ConvertBinaryFile.pm
118 $self->{'convert_options'} = "-windows_scripting" if $self->{'windows_scripting'};
119
120 # we always save as utf-8
121# if ($self->{'input_encoding'} eq "auto") {
122# $self->{'input_encoding'} = "utf8";
123# }
124
125 my $secondary_plugin_options = $self->{'secondary_plugin_options'};
126 if (defined $self->{'windows_scripting'}) {
127 if (!defined $secondary_plugin_options->{'StructuredHTMLPlugin'}){
128 $secondary_plugin_options->{'StructuredHTMLPlugin'} = [];
129 my $structhtml_options = $secondary_plugin_options->{'StructuredHTMLPlugin'};
130
131 # Instruct HTMLPlugin (when eventually accessed through read_into_doc_obj)
132 # to extract these metadata fields from the HEAD META fields
133 push (@$structhtml_options,"-metadata_fields","Title,GENERATOR,date,author<Creator>");
134 push (@$structhtml_options , "-title_sub", '^(Page\s+\d+)?(\s*1\s+)?');
135 push (@$structhtml_options, "-description_tags") if $self->{'windows_scripting'};
136 push(@$structhtml_options,"-extract_language") if $self->{'extract_language'};
137 push (@$structhtml_options, "-delete_toc") if $self->{'delete_toc'};
138 push (@$structhtml_options, "-toc_header", $self->{'toc_header'}) if $self->{'toc_header'};
139 push (@$structhtml_options, "-title_header", $self->{'title_header'}) if $self->{'title_header'};
140 push (@$structhtml_options, "-level1_header", $self->{'level1_header'}) if $self->{'level1_header'};
141 push (@$structhtml_options, "-level2_header", $self->{'level2_header'})if $self->{'level2_header'};
142 push (@$structhtml_options, "-level3_header", $self->{'level3_header'}) if $self->{'level3_header'};
143 push (@$structhtml_options, "-metadata_fields", $self->{'metadata_fields'}) if $self->{'metadata_fields'};
144 }
145 }
146 if (!defined $secondary_plugin_options->{'HTMLPlugin'}) {
147 $secondary_plugin_options->{'HTMLPlugin'} = [];
148 }
149 if (!defined $secondary_plugin_options->{'TextPlugin'}) {
150 $secondary_plugin_options->{'TextPlugin'} = [];
151 }
152
153 my $html_options = $secondary_plugin_options->{'HTMLPlugin'};
154 my $text_options = $secondary_plugin_options->{'TextPlugin'};
155 my $structhtml_options = $secondary_plugin_options->{'StructuredHTMLPlugin'};
156 # wvWare will always produce html files encoded as utf-8, so make sure the secondary HTMLPlugin knows this
157 push(@$html_options,"-input_encoding", "utf8");
158 push(@$html_options,"-extract_language") if $self->{'extract_language'};
159 push(@$html_options, "-description_tags") if $self->{'description_tags'};
160
161 # Instruct HTMLPlugin (when eventually accessed through read_into_doc_obj)
162 # to extract these metadata fields from the HEAD META fields
163 push(@$html_options,"-metadata_fields","Title,GENERATOR,date,author<Creator>");
164 push(@$html_options , "-title_sub", '^(Page\s+\d+)?(\s*1\s+)?');
165
166 my $associate_tail_re = $self->{'associate_tail_re'};
167 if ((defined $associate_tail_re) && ($associate_tail_re ne "")) {
168 push(@$html_options, "-associate_tail_re", $associate_tail_re);
169 push(@$text_options, "-associate_tail_re", $associate_tail_re);
170 push(@$structhtml_options, "-associate_tail_re", $associate_tail_re);
171 }
172
173 $self = bless $self, $class;
174 $self->load_secondary_plugins($class,$secondary_plugin_options,$hashArgOptLists);
175
176 return bless $self;
177}
178
179sub get_default_process_exp {
180 my $self = shift (@_);
181
182 return q^(?i)\.(doc|dot)$^;
183}
184
185sub convert_post_process_old
186{
187 my $self = shift (@_);
188 my ($conv_filename) = @_;
189
190 my $outhandle=$self->{'outhandle'};
191
192 my ($language, $encoding) = $self->textcat_get_language_encoding ($conv_filename);
193
194 # read in file ($text will be in utf8)
195 my $text = "";
196 $self->read_file ($conv_filename, $encoding, $language, \$text);
197
198 # turn any high bytes that aren't valid utf-8 into utf-8.
199 #unicode::ensure_utf8(\$text);
200
201 # Write it out again!
202 #$self->utf8_write_file (\$text, $conv_filename);
203}
204
205# Modified to cache HTML files for efficieny reasons rather
206# than delete all. HTML is modified not to use IE's VML.
207# VML uses WML files, so these can be deleted.
208sub cleanup_tmp_area {
209 my ($self) = @_;
210 if (defined $self->{'files_dir'}) {
211 my $html_files_dir = $self->{'files_dir'};
212
213 if (opendir(DIN,$html_files_dir)) {
214 my @wmz_files = grep( /\.wmz$/, readdir(DIN));
215 foreach my $f (@wmz_files) {
216 my $full_f = &util::filename_cat($html_files_dir,$f);
217 &util::rm($full_f);
218 }
219 closedir(DIN);
220 }
221 else {
222 # if HTML file has no supporting images, then no _files dir made
223 # => do nothing
224 }
225 }
226}
227
228
2291;
230
Note: See TracBrowser for help on using the repository browser.