| 1 | ###########################################################################
|
|---|
| 2 | #
|
|---|
| 3 | # PDFBoxConverter - helper plugin that does pdf document conversion with PDFBox
|
|---|
| 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) 2010 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 | package PDFBoxConverter;
|
|---|
| 27 |
|
|---|
| 28 | use BaseMediaConverter;
|
|---|
| 29 |
|
|---|
| 30 | use strict;
|
|---|
| 31 | no strict 'refs'; # allow filehandles to be variables and viceversa
|
|---|
| 32 | no strict 'subs'; # allow barewords (eg STDERR) as function arguments
|
|---|
| 33 |
|
|---|
| 34 | use gsprintf 'gsprintf';
|
|---|
| 35 |
|
|---|
| 36 | # these two variables mustn't be initialised here or they will get stuck
|
|---|
| 37 | # at those values.
|
|---|
| 38 | our $pdfbox_conversion_available;
|
|---|
| 39 | our $no_pdfbox_conversion_reason;
|
|---|
| 40 |
|
|---|
| 41 | BEGIN {
|
|---|
| 42 | @PDFBoxConverter::ISA = ('BaseMediaConverter');
|
|---|
| 43 |
|
|---|
| 44 | # Check that PDFBox is installed and available on the path
|
|---|
| 45 | $pdfbox_conversion_available = 1;
|
|---|
| 46 | $no_pdfbox_conversion_reason = "";
|
|---|
| 47 |
|
|---|
| 48 | if (!defined $ENV{'GEXT_PDFBOX'}) {
|
|---|
| 49 | $pdfbox_conversion_available = 0;
|
|---|
| 50 | $no_pdfbox_conversion_reason = "gextpdfboxnotinstalled";
|
|---|
| 51 | }
|
|---|
| 52 | else {
|
|---|
| 53 | my $gextpb_home = $ENV{'GEXT_PDFBOX'};
|
|---|
| 54 | my $pbajar = &util::filename_cat($gextpb_home,"lib","java","pdfbox-app.jar");
|
|---|
| 55 |
|
|---|
| 56 | if (!-e $pbajar) {
|
|---|
| 57 | &gsprintf(STDERR,"**** Failed to find $pbajar\n");
|
|---|
| 58 | $pdfbox_conversion_available = 0;
|
|---|
| 59 | $no_pdfbox_conversion_reason = "gextpdfboxjarnotinstalled";
|
|---|
| 60 | }
|
|---|
| 61 | else {
|
|---|
| 62 | # test to see if java is in path
|
|---|
| 63 | # Need to run java -version instead of just java, since the %ERRORLEVEL% returned
|
|---|
| 64 | # for `java` (which is checked below for failure of the command) is 0 for JDK 1.6*
|
|---|
| 65 | # while %ERRORLEVEL% is 1 for JDK 1.7*
|
|---|
| 66 | # If `java -version` is run however, %ERRORLEVEL% returned is 0 if java is
|
|---|
| 67 | # installed, regardless of whether the JDK version is 1.6* or 1.7*.
|
|---|
| 68 | my $java = &util::get_java_command();
|
|---|
| 69 |
|
|---|
| 70 | my $cmd = "$java -version";
|
|---|
| 71 | if ($ENV{'GSDLOS'} =~ /^windows/i) {
|
|---|
| 72 | $cmd .= " >nul 2>&1"; # java 2>&1 >null or java >null 2>&1 both work (%ERRORLEVEL% is 0)
|
|---|
| 73 | }
|
|---|
| 74 | else {
|
|---|
| 75 | # On Ubuntu, java >/dev/null 2>&1 works,
|
|---|
| 76 | # but java 2>&1 >/dev/null doesn't work: output goes to screen anyway
|
|---|
| 77 | $cmd .= " >/dev/null 2>&1"; # " >/dev/null 2>&1 &" - don't need & at end for Linux Centos anymore (Ubuntu was already fine without it)
|
|---|
| 78 | }
|
|---|
| 79 |
|
|---|
| 80 | my $status = system($cmd);
|
|---|
| 81 |
|
|---|
| 82 | if ($status != 0) {
|
|---|
| 83 |
|
|---|
| 84 | my $error_message = "**** Testing for java\n";
|
|---|
| 85 | $error_message .= "Failed to run: $cmd\n";
|
|---|
| 86 | $error_message .= "Error variable: |$!| and status: $status\n";
|
|---|
| 87 |
|
|---|
| 88 | &gsprintf(STDERR, "PDFBoxConverter: $error_message");
|
|---|
| 89 |
|
|---|
| 90 | $pdfbox_conversion_available = 0;
|
|---|
| 91 | $no_pdfbox_conversion_reason = "couldnotrunjava";
|
|---|
| 92 | }
|
|---|
| 93 | }
|
|---|
| 94 | }
|
|---|
| 95 |
|
|---|
| 96 | }
|
|---|
| 97 |
|
|---|
| 98 | my $arguments = [ ];
|
|---|
| 99 |
|
|---|
| 100 | my $options = { 'name' => "PDFBoxConverter",
|
|---|
| 101 | 'desc' => "{PDFBoxConverter.desc}",
|
|---|
| 102 | 'abstract' => "yes",
|
|---|
| 103 | 'inherits' => "yes",
|
|---|
| 104 | 'args' => $arguments };
|
|---|
| 105 |
|
|---|
| 106 | sub new {
|
|---|
| 107 | my ($class) = shift (@_);
|
|---|
| 108 | my ($pluginlist,$inputargs,$hashArgOptLists,$auxilary) = @_;
|
|---|
| 109 | push(@$pluginlist, $class);
|
|---|
| 110 |
|
|---|
| 111 | push(@{$hashArgOptLists->{"ArgList"}},@{$arguments});
|
|---|
| 112 | push(@{$hashArgOptLists->{"OptList"}},$options);
|
|---|
| 113 |
|
|---|
| 114 |
|
|---|
| 115 | my $self = new BaseMediaConverter($pluginlist, $inputargs,
|
|---|
| 116 | $hashArgOptLists, $auxilary);
|
|---|
| 117 |
|
|---|
| 118 | if ($self->{'info_only'}) {
|
|---|
| 119 | # don't worry about any options etc
|
|---|
| 120 | return bless $self, $class;
|
|---|
| 121 | }
|
|---|
| 122 | if ($pdfbox_conversion_available) {
|
|---|
| 123 | my $gextpb_home = $ENV{'GEXT_PDFBOX'};
|
|---|
| 124 | my $pbajar = &util::filename_cat($gextpb_home,"lib","java","pdfbox-app.jar");
|
|---|
| 125 | my $java = &util::get_java_command();
|
|---|
| 126 | my $launch_cmd = "$java -cp \"$pbajar\" -Dline.separator=\"<br />\" org.apache.pdfbox.ExtractText";
|
|---|
| 127 |
|
|---|
| 128 | $self->{'pdfbox_launch_cmd'} = $launch_cmd;
|
|---|
| 129 | $self->{'pdfbox_img_launch_cmd'} = "java -cp \"$pbajar\" org.apache.pdfbox.PDFToImage"; # cmd for converting pages to images (gif, jpg, png)
|
|---|
| 130 | }
|
|---|
| 131 | else {
|
|---|
| 132 | $self->{'no_pdfbox_conversion_reason'} = $no_pdfbox_conversion_reason;
|
|---|
| 133 |
|
|---|
| 134 | my $outhandle = $self->{'outhandle'};
|
|---|
| 135 | &gsprintf($outhandle, "PDFBoxConverter: {PDFBoxConverter.noconversionavailable} ({PDFBoxConverter.$no_pdfbox_conversion_reason})\n");
|
|---|
| 136 | }
|
|---|
| 137 |
|
|---|
| 138 | $self->{'pdfbox_conversion_available'} = $pdfbox_conversion_available;
|
|---|
| 139 |
|
|---|
| 140 | return bless $self, $class;
|
|---|
| 141 |
|
|---|
| 142 | }
|
|---|
| 143 |
|
|---|
| 144 | sub init {
|
|---|
| 145 | my $self = shift(@_);
|
|---|
| 146 | my ($verbosity, $outhandle, $failhandle) = @_;
|
|---|
| 147 |
|
|---|
| 148 | $self->{'pbtmp_file_paths'} = ();
|
|---|
| 149 | }
|
|---|
| 150 |
|
|---|
| 151 | sub deinit {
|
|---|
| 152 | my $self = shift(@_);
|
|---|
| 153 |
|
|---|
| 154 | $self->clean_up_temporary_files();
|
|---|
| 155 | }
|
|---|
| 156 |
|
|---|
| 157 |
|
|---|
| 158 | sub convert {
|
|---|
| 159 | my $self = shift(@_);
|
|---|
| 160 | my ($source_file_full_path, $target_file_type) = @_;
|
|---|
| 161 |
|
|---|
| 162 | return 0 unless $pdfbox_conversion_available;
|
|---|
| 163 | # check the filename
|
|---|
| 164 | return 0 if ( !-f $source_file_full_path);
|
|---|
| 165 |
|
|---|
| 166 | my $img_output_mode = 0;
|
|---|
| 167 |
|
|---|
| 168 | # the following line is necessary to avoid 'uninitialised variable' error
|
|---|
| 169 | # messages concerning the converted_to member variable when PDFPlugin's
|
|---|
| 170 | # use_sections option is checked.
|
|---|
| 171 | # PDFBox plugin now processes use_sections option, when working with v1.5.0
|
|---|
| 172 | # of the PDFBox jar file (which embeds each page in special <div> tags).
|
|---|
| 173 | if ($target_file_type eq "html") {
|
|---|
| 174 | $self->{'converted_to'} = "HTML";
|
|---|
| 175 | } elsif ($target_file_type eq "jpg" || $target_file_type eq "gif" || $target_file_type eq "png") {
|
|---|
| 176 | $self->{'converted_to'} = $target_file_type;
|
|---|
| 177 | $img_output_mode = 1;
|
|---|
| 178 | } else {
|
|---|
| 179 | $self->{'converted_to'} = "text";
|
|---|
| 180 | }
|
|---|
| 181 |
|
|---|
| 182 | my $outhandle = $self->{'outhandle'};
|
|---|
| 183 | my $verbosity = $self->{'verbosity'};
|
|---|
| 184 |
|
|---|
| 185 | my $source_file_no_path = &File::Basename::basename($source_file_full_path);
|
|---|
| 186 | # Determine the full name and path of the output file
|
|---|
| 187 | my $target_file_path;
|
|---|
| 188 | if ($self->{'enable_cache'}) {
|
|---|
| 189 | $self->init_cache_for_file($source_file_full_path);
|
|---|
| 190 | my $cache_dir = $self->{'cached_dir'};
|
|---|
| 191 | my $file_root = $self->{'cached_file_root'};
|
|---|
| 192 | #$file_root .= "_$convert_id" if ($convert_id ne "");
|
|---|
| 193 |
|
|---|
| 194 | # append the output filetype suffix only for non-image output formats, since for
|
|---|
| 195 | # images we can be outputting multiple image files per single PDF input file
|
|---|
| 196 | my $target_file = $img_output_mode ? "$file_root" : "$file_root.$target_file_type";
|
|---|
| 197 |
|
|---|
| 198 | $target_file_path = &util::filename_cat($cache_dir,$target_file);
|
|---|
| 199 | }
|
|---|
| 200 | else {
|
|---|
| 201 | # this is in gsdl/tmp. get a tmp filename in collection instead???
|
|---|
| 202 | $target_file_path = &util::get_tmp_filename($target_file_type);
|
|---|
| 203 |
|
|---|
| 204 | # for image files, remove the suffix, since we can have many output image files
|
|---|
| 205 | # per input PDF (one img for each page of the PDF, for example)
|
|---|
| 206 | if($img_output_mode) {
|
|---|
| 207 | $target_file_path =~ s/\.[^.]*$//g;
|
|---|
| 208 | if(!&util::dir_exists($target_file_path)) {
|
|---|
| 209 | mkdir($target_file_path);
|
|---|
| 210 | }
|
|---|
| 211 |
|
|---|
| 212 | # once the item file for the imgs has been created, need to adjust target_file_path
|
|---|
| 213 |
|
|---|
| 214 | # below, we'll store the dir just created to pbtmp_file_paths, so all imgs and the
|
|---|
| 215 | # item file generated in it can be deleted in one go on clean_up
|
|---|
| 216 | }
|
|---|
| 217 |
|
|---|
| 218 | push(@{$self->{'pbtmp_file_paths'}}, $target_file_path);
|
|---|
| 219 | }
|
|---|
| 220 |
|
|---|
| 221 | # Generate and run the convert command
|
|---|
| 222 | my $convert_cmd = "";
|
|---|
| 223 |
|
|---|
| 224 | # want the filename without extension, because any images
|
|---|
| 225 | # are to be generated with the same filename as the PDF
|
|---|
| 226 | my ($tailname, $dirname, $suffix) = &File::Basename::fileparse($source_file_full_path, "\\.[^\\.]+\$");
|
|---|
| 227 |
|
|---|
| 228 | if($img_output_mode) { # converting to images
|
|---|
| 229 | my $output_prefix = &util::filename_cat($target_file_path, $tailname);
|
|---|
| 230 |
|
|---|
| 231 | $convert_cmd = $self->{'pdfbox_img_launch_cmd'};
|
|---|
| 232 | $convert_cmd .= " -imageType $target_file_type";
|
|---|
| 233 | $convert_cmd .= " -outputPrefix \"$output_prefix\"";
|
|---|
| 234 | $convert_cmd .= " \"$source_file_full_path\"";
|
|---|
| 235 |
|
|---|
| 236 | } else { # html or text
|
|---|
| 237 | $convert_cmd = $self->{'pdfbox_launch_cmd'};
|
|---|
| 238 | $convert_cmd .= " -html" if ($target_file_type eq "html");
|
|---|
| 239 | $convert_cmd .= " \"$source_file_full_path\" \"$target_file_path\"";
|
|---|
| 240 | }
|
|---|
| 241 |
|
|---|
| 242 | if ($verbosity>2) {
|
|---|
| 243 | &gsprintf($outhandle,"Convert command: $convert_cmd\n");
|
|---|
| 244 | }
|
|---|
| 245 |
|
|---|
| 246 | my $print_info = { 'message_prefix' => "PDFBox Conversion",
|
|---|
| 247 | 'message' => "Converting $source_file_no_path to: $target_file_type" };
|
|---|
| 248 | # $print_info->{'cache_mode'} = $cache_mode if ($cache_mode ne "");
|
|---|
| 249 |
|
|---|
| 250 | my ($regenerated,$result,$had_error)
|
|---|
| 251 | = $self->autorun_general_cmd($convert_cmd,$source_file_full_path, $target_file_path,$print_info);
|
|---|
| 252 |
|
|---|
| 253 | if($img_output_mode) {
|
|---|
| 254 | # now the images have been generated, generate the "$target_file_path/tailname.item"
|
|---|
| 255 | # item file for them, which is also the target_file_path that needs to be returned
|
|---|
| 256 | $target_file_path = &util::create_itemfile($target_file_path, $tailname, $target_file_type);
|
|---|
| 257 | #print STDERR "**** item file: $target_file_path\n";
|
|---|
| 258 | }
|
|---|
| 259 |
|
|---|
| 260 | if ($had_error) {
|
|---|
| 261 | return (0, $result,$target_file_path);
|
|---|
| 262 | }
|
|---|
| 263 | return (1, $result,$target_file_path);
|
|---|
| 264 | }
|
|---|
| 265 |
|
|---|
| 266 | sub convert_without_result {
|
|---|
| 267 | my $self = shift(@_);
|
|---|
| 268 |
|
|---|
| 269 | my $source_file_path = shift(@_);
|
|---|
| 270 | my $target_file_type = shift(@_);
|
|---|
| 271 | my $convert_options = shift(@_) || "";
|
|---|
| 272 | my $convert_id = shift(@_) || "";
|
|---|
| 273 |
|
|---|
| 274 | return $self->convert($source_file_path,$target_file_type,
|
|---|
| 275 | $convert_options,$convert_id,"without_result");
|
|---|
| 276 | }
|
|---|
| 277 |
|
|---|
| 278 | sub clean_up_temporary_files {
|
|---|
| 279 | my $self = shift(@_);
|
|---|
| 280 |
|
|---|
| 281 | foreach my $pbtmp_file_path (@{$self->{'pbtmp_file_paths'}}) {
|
|---|
| 282 | if (-d $pbtmp_file_path) {
|
|---|
| 283 | #print STDERR "@@@@@@ cleanup called on $pbtmp_file_path\n";
|
|---|
| 284 | &util::rm_r($pbtmp_file_path);
|
|---|
| 285 | }
|
|---|
| 286 | elsif (-e $pbtmp_file_path) {
|
|---|
| 287 | &util::rm($pbtmp_file_path);
|
|---|
| 288 | }
|
|---|
| 289 | }
|
|---|
| 290 |
|
|---|
| 291 | $self->{'pbtmp_file_paths'} = ();
|
|---|
| 292 | }
|
|---|
| 293 |
|
|---|
| 294 |
|
|---|
| 295 | 1;
|
|---|
