[2981] | 1 | ###########################################################################
|
---|
| 2 | #
|
---|
[17722] | 3 | # PowerPointPlugin.pm -- plugin for importing Microsoft PowerPoint files.
|
---|
[22709] | 4 | # (basic version supports versions 95 and 97)
|
---|
[22861] | 5 | # (through OpenOffice extension, supports all contemporary formats)
|
---|
[2981] | 6 | #
|
---|
| 7 | # A component of the Greenstone digital library software
|
---|
| 8 | # from the New Zealand Digital Library Project at the
|
---|
| 9 | # University of Waikato, New Zealand.
|
---|
| 10 | #
|
---|
| 11 | # Copyright (C) 2002 New Zealand Digital Library Project
|
---|
| 12 | #
|
---|
| 13 | # This program is free software; you can redistribute it and/or modify
|
---|
| 14 | # it under the terms of the GNU General Public License as published by
|
---|
| 15 | # the Free Software Foundation; either version 2 of the License, or
|
---|
| 16 | # (at your option) any later version.
|
---|
| 17 | #
|
---|
| 18 | # This program is distributed in the hope that it will be useful,
|
---|
| 19 | # but WITHOUT ANY WARRANTY; without even the implied warranty of
|
---|
| 20 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
---|
| 21 | # GNU General Public License for more details.
|
---|
| 22 | #
|
---|
| 23 | # You should have received a copy of the GNU General Public License
|
---|
| 24 | # along with this program; if not, write to the Free Software
|
---|
| 25 | # Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
|
---|
| 26 | #
|
---|
| 27 | ###########################################################################
|
---|
| 28 |
|
---|
[17722] | 29 | package PowerPointPlugin;
|
---|
[2981] | 30 |
|
---|
[10254] | 31 | use strict;
|
---|
| 32 | no strict 'refs'; # allow filehandles to be variables and viceversa
|
---|
[22515] | 33 | no strict 'subs';
|
---|
[22705] | 34 |
|
---|
[22515] | 35 | use gsprintf 'gsprintf';
|
---|
[2981] | 36 |
|
---|
[22861] | 37 | use AutoLoadConverters;
|
---|
| 38 | use ConvertBinaryFile;
|
---|
[2981] | 39 |
|
---|
[22861] | 40 | sub BEGIN {
|
---|
| 41 | @PowerPointPlugin::ISA = ('ConvertBinaryFile', 'AutoLoadConverters');
|
---|
| 42 | }
|
---|
[22705] | 43 |
|
---|
[22861] | 44 | my $openoffice_available = 0;
|
---|
[22705] | 45 |
|
---|
[22515] | 46 | my $windows_convert_to_list =
|
---|
[10466] | 47 | [ { 'name' => "auto",
|
---|
[15872] | 48 | 'desc' => "{ConvertBinaryFile.convert_to.auto}" },
|
---|
[10466] | 49 | { 'name' => "html",
|
---|
[15872] | 50 | 'desc' => "{ConvertBinaryFile.convert_to.html}" },
|
---|
[10466] | 51 | { 'name' => "text",
|
---|
[15872] | 52 | 'desc' => "{ConvertBinaryFile.convert_to.text}" },
|
---|
[10466] | 53 | { 'name' => "pagedimg_jpg",
|
---|
[15872] | 54 | 'desc' => "{ConvertBinaryFile.convert_to.pagedimg_jpg}" },
|
---|
[10466] | 55 | { 'name' => "pagedimg_gif",
|
---|
[15872] | 56 | 'desc' => "{ConvertBinaryFile.convert_to.pagedimg_gif}" },
|
---|
[10466] | 57 | { 'name' => "pagedimg_png",
|
---|
[15872] | 58 | 'desc' => "{ConvertBinaryFile.convert_to.pagedimg_png}" }
|
---|
[10466] | 59 | ];
|
---|
| 60 |
|
---|
[22861] | 61 | my $openoffice_convert_to_list =
|
---|
| 62 | [ { 'name' => "auto",
|
---|
| 63 | 'desc' => "{ConvertBinaryFile.convert_to.auto}" },
|
---|
| 64 | { 'name' => "html",
|
---|
| 65 | 'desc' => "{PowerPointPlugin.convert_to.oo_html}" },
|
---|
| 66 | { 'name' => "text",
|
---|
| 67 | 'desc' => "{ConvertBinaryFile.convert_to.text}" },
|
---|
| 68 | { 'name' => "pagedimg",
|
---|
| 69 | 'desc' => "{PowerPointPlugin.convert_to.pagedimg}" }
|
---|
| 70 | ];
|
---|
| 71 |
|
---|
[6408] | 72 | my $arguments =
|
---|
| 73 | [ { 'name' => "process_exp",
|
---|
[15872] | 74 | 'desc' => "{BasePlugin.process_exp}",
|
---|
[6408] | 75 | 'type' => "regexp",
|
---|
| 76 | 'reqd' => "no",
|
---|
[22861] | 77 | 'deft' => "&get_default_process_exp()", # delayed (see below)
|
---|
[22709] | 78 | }
|
---|
[6408] | 79 | ];
|
---|
| 80 |
|
---|
[22515] | 81 | my $opt_windows_args =
|
---|
| 82 | [ { 'name' => "convert_to",
|
---|
| 83 | 'desc' => "{ConvertBinaryFile.convert_to}",
|
---|
| 84 | 'type' => "enum",
|
---|
| 85 | 'reqd' => "yes",
|
---|
| 86 | 'list' => $windows_convert_to_list,
|
---|
| 87 | 'deft' => "html" },
|
---|
| 88 | { 'name' => "windows_scripting",
|
---|
| 89 | 'desc' => "{PowerPointPlugin.windows_scripting}",
|
---|
| 90 | 'type' => "flag",
|
---|
| 91 | 'reqd' => "no" }
|
---|
| 92 | ];
|
---|
| 93 |
|
---|
[22861] | 94 | my $opt_office_args =
|
---|
| 95 | [ { 'name' => "convert_to",
|
---|
| 96 | 'desc' => "{ConvertBinaryFile.convert_to}",
|
---|
| 97 | 'type' => "enum",
|
---|
| 98 | 'reqd' => "yes",
|
---|
| 99 | 'list' => $openoffice_convert_to_list,
|
---|
| 100 | 'deft' => "html" }
|
---|
| 101 | ];
|
---|
| 102 |
|
---|
[17722] | 103 | my $options = { 'name' => "PowerPointPlugin",
|
---|
[17744] | 104 | 'desc' => "{PowerPointPlugin.desc}",
|
---|
[6408] | 105 | 'abstract' => "no",
|
---|
[11679] | 106 | 'inherits' => "yes",
|
---|
[15114] | 107 | 'srcreplaceable' => "yes", # Source docs in PPT format can be replaced with GS-generated html
|
---|
[6408] | 108 | 'args' => $arguments };
|
---|
[4744] | 109 |
|
---|
[2981] | 110 | sub new {
|
---|
[10218] | 111 | my ($class) = shift (@_);
|
---|
| 112 | my ($pluginlist,$inputargs,$hashArgOptLists) = @_;
|
---|
| 113 | push(@$pluginlist, $class);
|
---|
[2981] | 114 |
|
---|
[22861] | 115 | # this bit needs to happen later after the arguments array has been
|
---|
| 116 | # finished - used for parsing the input args.
|
---|
| 117 | # push(@{$hashArgOptLists->{"ArgList"}},@{$arguments});
|
---|
| 118 | # this one needs to go in first, to get the print info in the right order
|
---|
| 119 | push(@{$hashArgOptLists->{"OptList"}},$options);
|
---|
| 120 |
|
---|
[10275] | 121 | if ($ENV{'GSDLOS'} =~ m/^windows$/i) {
|
---|
[22515] | 122 | push(@$arguments,@$opt_windows_args);
|
---|
[10275] | 123 | }
|
---|
[22515] | 124 |
|
---|
[22861] | 125 | my $auto_converter_self = new AutoLoadConverters($pluginlist,$inputargs,$hashArgOptLists,["OpenOfficeConverter"],1);
|
---|
[10218] | 126 |
|
---|
[22861] | 127 | if ($auto_converter_self->{'openoffice_available'}) {
|
---|
| 128 | push (@$arguments,@$opt_office_args);
|
---|
| 129 | $openoffice_available = 1;
|
---|
| 130 | }
|
---|
| 131 | # TODO need to do the case where they are both enabled!!! what will the convert to list be???
|
---|
[10427] | 132 |
|
---|
[22861] | 133 | # evaluate the default for process_exp - it needs to be delayed till here so we know if openoffice is available or not. But needs to be done before parsing the args.
|
---|
| 134 | foreach my $a (@$arguments) {
|
---|
[22709] | 135 | if ($a->{'name'} eq "process_exp") {
|
---|
| 136 | my $eval_expr = $a->{'deft'};
|
---|
| 137 | $a->{'deft'} = eval "$eval_expr";
|
---|
[22861] | 138 | last;
|
---|
[22709] | 139 | }
|
---|
| 140 | }
|
---|
| 141 |
|
---|
[22861] | 142 | push(@{$hashArgOptLists->{"ArgList"}},@{$arguments});
|
---|
| 143 |
|
---|
| 144 | my $cbf_self = new ConvertBinaryFile($pluginlist, $inputargs, $hashArgOptLists);
|
---|
| 145 | my $self = BasePlugin::merge_inheritance($auto_converter_self, $cbf_self);
|
---|
| 146 |
|
---|
[10580] | 147 | if ($self->{'info_only'}) {
|
---|
| 148 | # don't worry about any options etc
|
---|
| 149 | return bless $self, $class;
|
---|
| 150 | }
|
---|
| 151 |
|
---|
[22861] | 152 | $self = bless $self, $class;
|
---|
[15872] | 153 | $self->{'filename_extension'} = "ppt";
|
---|
| 154 | $self->{'file_type'} = "PPT";
|
---|
| 155 |
|
---|
[22597] | 156 | if ($self->{'convert_to'} eq "auto") {
|
---|
| 157 | if ($self->{'windows_scripting'}) {
|
---|
| 158 | $self->{'convert_to'} = "pagedimg_jpg";
|
---|
| 159 | }
|
---|
| 160 | else {
|
---|
| 161 | $self->{'convert_to'} = "html";
|
---|
| 162 | }
|
---|
| 163 | }
|
---|
| 164 |
|
---|
| 165 | my $outhandle = $self->{'outhandle'};
|
---|
| 166 |
|
---|
[22861] | 167 | # can't have windows_scripting and openoffice_conversion at the same time
|
---|
| 168 | if ($self->{'windows_scripting'} && $self->{'openoffice_conversion'}) {
|
---|
| 169 | print $outhandle "Warning: Cannot have -windows_scripting and -openoffice_conversion\n";
|
---|
[22515] | 170 | print $outhandle " on at the same time. Defaulting to -windows_scripting\n";
|
---|
[22861] | 171 | $self->{'openoffice_conversion'} = 0;
|
---|
[22515] | 172 | }
|
---|
| 173 |
|
---|
[15872] | 174 | #these are passed through to gsConvert.pl by ConvertBinaryFile.pm
|
---|
[10491] | 175 | $self->{'convert_options'} = "-windows_scripting" if $self->{'windows_scripting'};
|
---|
[22597] | 176 |
|
---|
| 177 | # set convert_to_plugin and convert_to_ext
|
---|
[22640] | 178 | $self->set_standard_convert_settings();
|
---|
[22597] | 179 |
|
---|
| 180 | my $secondary_plugin_name = $self->{'convert_to_plugin'};
|
---|
[10275] | 181 | my $secondary_plugin_options = $self->{'secondary_plugin_options'};
|
---|
[10218] | 182 |
|
---|
[22597] | 183 | if (!defined $secondary_plugin_options->{$secondary_plugin_name}) {
|
---|
| 184 | $secondary_plugin_options->{$secondary_plugin_name} = [];
|
---|
[10275] | 185 | }
|
---|
[22597] | 186 | my $specific_options = $secondary_plugin_options->{$secondary_plugin_name};
|
---|
| 187 |
|
---|
| 188 | push(@$specific_options, "-file_rename_method", "none");
|
---|
| 189 | push(@$specific_options, "-extract_language") if $self->{'extract_language'};
|
---|
| 190 |
|
---|
| 191 | if ($secondary_plugin_name eq "HTMLPlugin") {
|
---|
| 192 | push(@$specific_options, "-processing_tmp_files");
|
---|
| 193 | push(@$specific_options,"-metadata_fields","Title,GENERATOR,date,author<Creator>");
|
---|
[18406] | 194 | }
|
---|
[22597] | 195 | elsif ($secondary_plugin_name eq "PagedImagePlugin") {
|
---|
| 196 | push(@$specific_options, "-processing_tmp_files");
|
---|
[21958] | 197 | #is this true??
|
---|
[22597] | 198 | push(@$specific_options,"-input_encoding", "utf8");
|
---|
[22871] | 199 | if ($self->{'openoffice_conversion'}) {
|
---|
| 200 | push(@$specific_options, "-create_thumbnail", "false", "-create_screenview", "false");
|
---|
| 201 | }
|
---|
[15903] | 202 | }
|
---|
[2981] | 203 |
|
---|
[10275] | 204 | $self = bless $self, $class;
|
---|
| 205 |
|
---|
[10427] | 206 | $self->load_secondary_plugins($class,$secondary_plugin_options,$hashArgOptLists);
|
---|
[10275] | 207 | return $self;
|
---|
[2981] | 208 | }
|
---|
| 209 |
|
---|
| 210 | sub get_default_process_exp {
|
---|
| 211 | my $self = shift (@_);
|
---|
[22709] | 212 |
|
---|
[22861] | 213 | if ($openoffice_available) {
|
---|
[22709] | 214 | return q^(?i)\.(ppt|pptx|odp)$^;
|
---|
| 215 | }
|
---|
| 216 |
|
---|
[2981] | 217 | return q^(?i)\.ppt$^;
|
---|
| 218 | }
|
---|
[10275] | 219 |
|
---|
[22861] | 220 | sub init {
|
---|
| 221 | my $self = shift (@_);
|
---|
| 222 |
|
---|
| 223 | # ConvertBinaryFile init
|
---|
| 224 | $self->SUPER::init(@_);
|
---|
| 225 | $self->AutoLoadConverters::init();
|
---|
| 226 |
|
---|
| 227 | }
|
---|
| 228 |
|
---|
| 229 | sub begin {
|
---|
| 230 | my $self = shift (@_);
|
---|
| 231 |
|
---|
| 232 | $self->AutoLoadConverters::begin();
|
---|
| 233 | $self->SUPER::begin(@_);
|
---|
| 234 |
|
---|
| 235 | }
|
---|
| 236 |
|
---|
| 237 | sub deinit {
|
---|
| 238 | my $self = shift (@_);
|
---|
| 239 |
|
---|
| 240 | $self->AutoLoadConverters::deinit();
|
---|
| 241 | $self->SUPER::deinit(@_);
|
---|
| 242 |
|
---|
| 243 | }
|
---|
| 244 |
|
---|
| 245 | # override AutoLoadConverters version, as we need to do more stuff once its converted if we are converting to item file
|
---|
| 246 | sub tmp_area_convert_file {
|
---|
| 247 | my $self = shift (@_);
|
---|
| 248 | my ($output_ext, $input_filename, $textref) = @_;
|
---|
| 249 |
|
---|
| 250 | if ($self->{'openoffice_conversion'}) {
|
---|
| 251 | if ($self->{'convert_to'} eq "pagedimg") {
|
---|
| 252 | $output_ext = "html"; # first convert to html
|
---|
| 253 | }
|
---|
| 254 | my ($result, $result_str, $new_filename) = $self->OpenOfficeConverter::convert($input_filename, $output_ext);
|
---|
| 255 | if ($result == 0) {
|
---|
| 256 | my $outhandle=$self->{'outhandle'};
|
---|
| 257 | print $outhandle "OpenOfficeConverter Conversion error\n";
|
---|
| 258 | print $outhandle $result_str;
|
---|
| 259 | return "";
|
---|
| 260 |
|
---|
| 261 | }
|
---|
| 262 | #print STDERR "result = $result\n";
|
---|
| 263 | if ($self->{'convert_to'} eq "pagedimg") {
|
---|
[22871] | 264 | my $item_filename = $self->generate_item_file($new_filename);
|
---|
| 265 | return $item_filename;
|
---|
[22861] | 266 | }
|
---|
| 267 | return $new_filename;
|
---|
| 268 |
|
---|
| 269 | }
|
---|
| 270 | else {
|
---|
| 271 | return $self->ConvertBinaryFile::tmp_area_convert_file(@_);
|
---|
| 272 | }
|
---|
| 273 | # get tmp filename
|
---|
| 274 | }
|
---|
| 275 |
|
---|
| 276 | # override default read in some situations, as the conversion of ppt to html results in many files, and we want them all to be processed.
|
---|
| 277 | sub read_XX {
|
---|
| 278 | my $self = shift (@_);
|
---|
| 279 | my ($pluginfo, $base_dir, $file, $block_hash, $metadata, $processor, $maxdocs, $total_count, $gli) = @_;
|
---|
| 280 |
|
---|
| 281 | # can we process this file??
|
---|
| 282 | my ($filename_full_path, $filename_no_path) = &util::get_full_filenames($base_dir, $file);
|
---|
| 283 |
|
---|
| 284 | return undef unless $self->can_process_this_file($filename_full_path);
|
---|
| 285 |
|
---|
| 286 | my ($process_status,$doc_obj) = $self->read_into_doc_obj(@_);
|
---|
| 287 |
|
---|
| 288 | if ((defined $process_status) && ($process_status == 1)) {
|
---|
| 289 |
|
---|
| 290 | # process the document
|
---|
| 291 | $processor->process($doc_obj);
|
---|
| 292 |
|
---|
| 293 | $self->{'num_processed'} ++;
|
---|
| 294 | undef $doc_obj;
|
---|
| 295 | }
|
---|
| 296 | # delete any temp files that we may have created
|
---|
| 297 | $self->clean_up_after_doc_obj_processing();
|
---|
| 298 |
|
---|
| 299 |
|
---|
| 300 | # if process_status == 1, then the file has been processed.
|
---|
| 301 | return $process_status;
|
---|
| 302 |
|
---|
| 303 | }
|
---|
| 304 |
|
---|
[22871] | 305 | sub generate_item_file
|
---|
| 306 | {
|
---|
| 307 | my $self = shift(@_);
|
---|
| 308 | my ($input_filename) = @_;
|
---|
| 309 | my $outhandle = $self->{'outhandle'};
|
---|
| 310 | my ($tailname, $dirname, $suffix)
|
---|
| 311 | = &File::Basename::fileparse($input_filename, "\\.[^\\.]+\$");
|
---|
| 312 |
|
---|
| 313 | # find all the files in the directory
|
---|
| 314 | if (!opendir (DIR, $dirname)) {
|
---|
| 315 | print $outhandle "PowerPointPlugin: Couldn't read directory $dirname\n";
|
---|
| 316 | return $input_filename;
|
---|
| 317 | }
|
---|
| 318 |
|
---|
| 319 | my @dir = readdir (DIR);
|
---|
| 320 | closedir (DIR);
|
---|
| 321 |
|
---|
| 322 | # start the item file
|
---|
| 323 | my $itemfile_name = &util::filename_cat($dirname, "$tailname.item");
|
---|
| 324 |
|
---|
| 325 | # encoding specification????
|
---|
| 326 | if (!open (ITEMFILE, ">$itemfile_name")) {
|
---|
| 327 | print $outhandle "PowerPOintPlugin: Couldn't open $itemfile_name for writing\n";
|
---|
| 328 | }
|
---|
| 329 | print ITEMFILE "<GeneratedBy>PowerPointPlugin\n";
|
---|
| 330 | # print the first page
|
---|
| 331 | my @sorted_dir = sort alphanum_sort @dir;
|
---|
| 332 | for (my $i = 0; $i < scalar(@sorted_dir); $i++) {
|
---|
| 333 | my $file = $sorted_dir[$i];
|
---|
| 334 | if ($file =~ /^img(\d+)\.jpg$/) {
|
---|
| 335 | my $num = $1;
|
---|
| 336 | $self->tidy_up_html(&util::filename_cat($dirname, "text$num.html"));
|
---|
| 337 | print ITEMFILE "$num:img$num.jpg:text$num.html:\n";
|
---|
| 338 | }
|
---|
| 339 | }
|
---|
| 340 | close ITEMFILE;
|
---|
| 341 | return $itemfile_name;
|
---|
| 342 |
|
---|
| 343 |
|
---|
| 344 | }
|
---|
| 345 |
|
---|
| 346 | # want to sort img1, img2, ...img10, img11 etc.
|
---|
| 347 | sub alphanum_sort {
|
---|
| 348 |
|
---|
| 349 | my ($a_txt, $a_num) = $a =~ /^([^\d]*)(\d*)/;
|
---|
| 350 | my ($b_txt, $b_num) = $b =~ /^([^\d]*)(\d*)/;
|
---|
| 351 |
|
---|
| 352 | if ($a_txt ne $b_txt) { return ($a cmp $b) };
|
---|
| 353 | return ($a_num <=> $b_num);
|
---|
| 354 | }
|
---|
| 355 |
|
---|
| 356 | # Want to remove the line that links to first page, last page, next page, text etc.
|
---|
| 357 | sub tidy_up_html {
|
---|
| 358 |
|
---|
| 359 | my $self = shift(@_);
|
---|
| 360 | my ($filename) = @_;
|
---|
| 361 | return unless (-f $filename);
|
---|
| 362 | my $backup_filename = "$filename.bak";
|
---|
| 363 |
|
---|
| 364 | &File::Copy::copy($filename, $backup_filename);
|
---|
| 365 |
|
---|
| 366 | open (ORIGINAL, $backup_filename) || return;
|
---|
| 367 | open(HTMLFILE, ">$filename") || return;
|
---|
| 368 |
|
---|
| 369 | my $line ="";
|
---|
| 370 | while ($line = <ORIGINAL>) {
|
---|
| 371 | if ($line =~ /\<body\>/) {
|
---|
| 372 | print HTMLFILE $line;
|
---|
| 373 | $line = <ORIGINAL>;
|
---|
| 374 | next if $line =~ /\<center\>/;
|
---|
| 375 | }
|
---|
| 376 | next if $line =~ /First page/;
|
---|
| 377 | print HTMLFILE ($line);
|
---|
| 378 | }
|
---|
| 379 |
|
---|
| 380 | close HTMLFILE;
|
---|
| 381 | close ORIGINAL;
|
---|
| 382 | }
|
---|
[2981] | 383 | 1;
|
---|
[10275] | 384 |
|
---|