| 1 | ###########################################################################
|
|---|
| 2 | #
|
|---|
| 3 | # OpenOfficeConverter - helper plugin that does office document conversion
|
|---|
| 4 | # using jodconverter combined with OpenOffice
|
|---|
| 5 | #
|
|---|
| 6 | # A component of the Greenstone digital library software
|
|---|
| 7 | # from the New Zealand Digital Library Project at the
|
|---|
| 8 | # University of Waikato, New Zealand.
|
|---|
| 9 | #
|
|---|
| 10 | # Copyright (C) 2010 New Zealand Digital Library Project
|
|---|
| 11 | #
|
|---|
| 12 | # This program is free software; you can redistribute it and/or modify
|
|---|
| 13 | # it under the terms of the GNU General Public License as published by
|
|---|
| 14 | # the Free Software Foundation; either version 2 of the License, or
|
|---|
| 15 | # (at your option) any later version.
|
|---|
| 16 | #
|
|---|
| 17 | # This program is distributed in the hope that it will be useful,
|
|---|
| 18 | # but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|---|
| 19 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|---|
| 20 | # GNU General Public License for more details.
|
|---|
| 21 | #
|
|---|
| 22 | # You should have received a copy of the GNU General Public License
|
|---|
| 23 | # along with this program; if not, write to the Free Software
|
|---|
| 24 | # Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
|
|---|
| 25 | #
|
|---|
| 26 | ###########################################################################
|
|---|
| 27 | package OpenOfficeConverter;
|
|---|
| 28 |
|
|---|
| 29 | use ConvertBinaryFile;
|
|---|
| 30 | use BaseMediaConverter;
|
|---|
| 31 |
|
|---|
| 32 | use strict;
|
|---|
| 33 | no strict 'refs'; # allow filehandles to be variables and viceversa
|
|---|
| 34 |
|
|---|
| 35 | use gsprintf 'gsprintf';
|
|---|
| 36 |
|
|---|
| 37 | # these two variables mustn't be initialised here or they will get stuck
|
|---|
| 38 | # at those values.
|
|---|
| 39 | our $openoffice_conversion_available;
|
|---|
| 40 | our $no_openoffice_conversion_reason;
|
|---|
| 41 |
|
|---|
| 42 | BEGIN {
|
|---|
| 43 | @OpenOfficeConverter::ISA = ('BaseMediaConverter');
|
|---|
| 44 |
|
|---|
| 45 | # Check that OpenOffice and jodconverter are installed and available on
|
|---|
| 46 | # the path
|
|---|
| 47 | $openoffice_conversion_available = 1;
|
|---|
| 48 | $no_openoffice_conversion_reason = "";
|
|---|
| 49 |
|
|---|
| 50 | if (! defined $ENV{'GEXT_OPENOFFICE'}) {
|
|---|
| 51 | $openoffice_conversion_available = 0;
|
|---|
| 52 | $no_openoffice_conversion_reason = "gextopenofficenotinstalled";
|
|---|
| 53 | }
|
|---|
| 54 | else {
|
|---|
| 55 | my $gextoo_home = $ENV{'GEXT_OPENOFFICE'};
|
|---|
| 56 | #my $jodjar = &FileUtils::filenameConcatenate($gextoo_home,"lib","java","jodconverter.jar");
|
|---|
| 57 | my $jodjar = &FileUtils::filenameConcatenate($gextoo_home,"jodconverter-2.2.2","lib","jodconverter-cli-2.2.2.jar");
|
|---|
| 58 |
|
|---|
| 59 | if (!-e $jodjar) {
|
|---|
| 60 | #print STDERR "Failed to find $jodjar\n";
|
|---|
| 61 | $openoffice_conversion_available = 0;
|
|---|
| 62 | $no_openoffice_conversion_reason = "gextjodconverternotinstalled";
|
|---|
| 63 | }
|
|---|
| 64 | else {
|
|---|
| 65 | # test to see if soffice is in path
|
|---|
| 66 |
|
|---|
| 67 | if(!defined $ENV{'SOFFICE_HOST'}) {
|
|---|
| 68 | $ENV{'SOFFICE_HOST'} = "localhost";
|
|---|
| 69 | }
|
|---|
| 70 | if(!defined $ENV{'SOFFICE_PORT'}) {
|
|---|
| 71 | $ENV{'SOFFICE_PORT'} = "8100";
|
|---|
| 72 | }
|
|---|
| 73 |
|
|---|
| 74 | #my $cmd = "soffice --headless 2>&1"; # for linux and mac
|
|---|
| 75 | my $cmd = "soffice \"--accept=socket,host=$ENV{'SOFFICE_HOST'},port=$ENV{'SOFFICE_PORT'};urp;StarOffice.ServiceManager\" --headless"; # basic shared command
|
|---|
| 76 | my $status = 0;
|
|---|
| 77 |
|
|---|
| 78 | # No more JODCONVERTER_PORT env var: the port that jodconverter uses is now the
|
|---|
| 79 | # same as SOFFICE_PORT, because the 2 are meant to communicate on the same port.
|
|---|
| 80 | # The default port that jodconverter expects OO (OpenOffice/libreoffice/soffice/staroffice) to listen on is 8100, so we try that.
|
|---|
| 81 | if ($ENV{'GSDLOS'} =~ m/^windows$/) {
|
|---|
| 82 |
|
|---|
| 83 | # important to have this set when the Greenstone server
|
|---|
| 84 | # and open office is on a remote windows machine
|
|---|
| 85 | if(!defined $ENV{'SOFFICE_HOME'}) {
|
|---|
| 86 | $ENV{'SOFFICE_HOME'} = &FileUtils::filenameConcatenate($ENV{'ProgramFiles'},"OpenOffice.org 3");
|
|---|
| 87 | }
|
|---|
| 88 | my $ooffice_dir_guess =
|
|---|
| 89 | &FileUtils::filenameConcatenate($ENV{'SOFFICE_HOME'},"program");
|
|---|
| 90 | if (-d $ooffice_dir_guess) {
|
|---|
| 91 | &util::envvar_append("PATH",$ooffice_dir_guess);
|
|---|
| 92 | }
|
|---|
| 93 |
|
|---|
| 94 | # for windows, when working on a remote system, want to be able to start OO if
|
|---|
| 95 | # not already running. We'll use the uno socket method to do so. Else client-gli
|
|---|
| 96 | # tends to hang, waiting for the prompt to return after OO has been started up
|
|---|
| 97 | # (which doesn't happen, so need to Ctrl-C GLI and run it again for it to work).
|
|---|
| 98 |
|
|---|
| 99 | # first need to silently check soffice exists else windows will display a popup
|
|---|
| 100 | $status = system("which soffice >nul 2>&1"); # which.exe should be in bin/windows
|
|---|
| 101 | if ($status == 0) {
|
|---|
| 102 | #$cmd = "start \"soffice process\" $cmd >nul 2>&1"; # order of >nul and 2>&1 matters, see below
|
|---|
| 103 |
|
|---|
| 104 | # On Windows, when perl/C/C++ uses system() to launch any standalone process in the background
|
|---|
| 105 | # with "start \"window title\" cmd", perl's child process (despite being mostly independent/detached
|
|---|
| 106 | # from perl) *appears* to share stdout/stderr (stdin) streams with the parent perl process. This is
|
|---|
| 107 | # noticeable with perl launching soffice or any program that neither writes to stderr/out nor sends
|
|---|
| 108 | # eof/eos to indicate parent perl's streams are closed (e.g. Notepad).
|
|---|
| 109 | # This is *not* a problem when perl scripts are run from command line. Command line programs run fine.
|
|---|
| 110 | # But when combined with SafeProcess used by GLI, such silent standalone programs like soffice
|
|---|
| 111 | # cause GLI to hang every time right until the silent program's terminated, because SafeProcess' join()
|
|---|
| 112 | # calls on the perl child process' stderr and stdout block at SafeProcess.InputStreamGobbler.readLine(),
|
|---|
| 113 | # because these stderr/out streams didn't receive eof/eos, implying the silent grandchild process
|
|---|
| 114 | # somehow kept them open.
|
|---|
| 115 | # This is true at Windows/C's system() command level: InputStreamGobblers using join() calls (as SafeProcess
|
|---|
| 116 | # does), combined with a c++ program that uses system() that launches something like Notepad, just all block
|
|---|
| 117 | # until Notepad is closed.
|
|---|
| 118 | # When perl calls a custom java BackgroundLauncher.java program to launch an external program with Runtime.exec(),
|
|---|
| 119 | # this hanging problem doesn't happen as java's exec() is not implemented with Windows C's system().
|
|---|
| 120 | # Similarly, a WScript file (vbs or js script) that uses WinScriptHost.Run method to run an external program
|
|---|
| 121 | # also doesn't exhibit this problem. So we use a custom vbs script to launch external programs like soffice in
|
|---|
| 122 | # the background from perl (to avoid having to set up Java or have the java program compiled up by the release-kits,
|
|---|
| 123 | # since perl building scripts are not just run from GLI but are also run directly from the cmd line).
|
|---|
| 124 |
|
|---|
| 125 | $cmd = "CScript //Nologo $ENV{'GSDLHOME'}\\bin\\windows\\background-launcher.vbs $cmd";
|
|---|
| 126 |
|
|---|
| 127 | #print STDERR "***** Tried to start-up OpenOffice with:\n$cmd\n";
|
|---|
| 128 | }
|
|---|
| 129 | # else cmd is still undefined
|
|---|
| 130 | }
|
|---|
| 131 | else {
|
|---|
| 132 | # Windows seems to launch OpenOffice as a service (i.e.
|
|---|
| 133 | # automatically puts it in the background).
|
|---|
| 134 | # For Unix putting it in the background needs to be done
|
|---|
| 135 | # explicitly by appending & to the end
|
|---|
| 136 |
|
|---|
| 137 | # IMPORTANT! When redirecting output from 2>&1 to a file or null, >/dev/null (or >out.txt) should come BEFORE 2>&1
|
|---|
| 138 | # i.e. $some_cmd >out.txt 2>&1 and likewise $some_cmd >/dev/null 2>&1
|
|---|
| 139 | # Doing it in the wrong order further causes problems with SafeProcess when GLI starts up and calls pluginfo on the pluginslist:
|
|---|
| 140 | # SafeProcess blocks forever on read() from stdout of the process running pluginfo, because there's never data including no eof/eos
|
|---|
| 141 | # on the process' stdout when running this command in the wrong order.
|
|---|
| 142 | $cmd .= " >/dev/null 2>&1 &";
|
|---|
| 143 | }
|
|---|
| 144 |
|
|---|
| 145 | #print STDERR "@@@@ running: $cmd\n";
|
|---|
| 146 |
|
|---|
| 147 | $status = system($cmd) if ($status == 0);
|
|---|
| 148 | if ($status != 0) {
|
|---|
| 149 | #print STDERR "Failed to run: $cmd\n";
|
|---|
| 150 | #print STDERR "$!\n";
|
|---|
| 151 | $openoffice_conversion_available = 0;
|
|---|
| 152 | $no_openoffice_conversion_reason = "openofficenotinstalled";
|
|---|
| 153 | }
|
|---|
| 154 | }
|
|---|
| 155 | }
|
|---|
| 156 | }
|
|---|
| 157 |
|
|---|
| 158 | my $arguments = [
|
|---|
| 159 | { 'name' => "openoffice_port",
|
|---|
| 160 | 'desc' => "{OpenOfficeConverter.openoffice_port}",
|
|---|
| 161 | 'type' => "int",
|
|---|
| 162 | 'deft' => "8100",
|
|---|
| 163 | 'range' => "81,",
|
|---|
| 164 | 'reqd' => "no" },
|
|---|
| 165 | ];
|
|---|
| 166 |
|
|---|
| 167 |
|
|---|
| 168 | my $options = { 'name' => "OpenOfficeConverter",
|
|---|
| 169 | 'desc' => "{OpenOfficeConverter.desc}",
|
|---|
| 170 | 'abstract' => "yes",
|
|---|
| 171 | 'inherits' => "yes",
|
|---|
| 172 | 'args' => $arguments };
|
|---|
| 173 |
|
|---|
| 174 | sub new {
|
|---|
| 175 | my ($class) = shift (@_);
|
|---|
| 176 | my ($pluginlist,$inputargs,$hashArgOptLists,$auxilary) = @_;
|
|---|
| 177 | push(@$pluginlist, $class);
|
|---|
| 178 |
|
|---|
| 179 | push(@{$hashArgOptLists->{"ArgList"}},@{$arguments});
|
|---|
| 180 | push(@{$hashArgOptLists->{"OptList"}},$options);
|
|---|
| 181 |
|
|---|
| 182 | my $self = new BaseMediaConverter($pluginlist, $inputargs,
|
|---|
| 183 | $hashArgOptLists, $auxilary);
|
|---|
| 184 |
|
|---|
| 185 | if ($self->{'info_only'}) {
|
|---|
| 186 | # don't worry about any options etc
|
|---|
| 187 | return bless $self, $class;
|
|---|
| 188 | }
|
|---|
| 189 | if (!$openoffice_conversion_available) {
|
|---|
| 190 | $self->{'no_openoffice_conversion_reason'} = $no_openoffice_conversion_reason;
|
|---|
| 191 |
|
|---|
| 192 | my $outhandle = $self->{'outhandle'};
|
|---|
| 193 | &gsprintf($outhandle, "OpenOfficeConverter: {OpenOfficeConverter.noconversionavailable} ({OpenOfficeConverter.$no_openoffice_conversion_reason})\n");
|
|---|
| 194 | }
|
|---|
| 195 |
|
|---|
| 196 | $self->{'openoffice_conversion_available'} = $openoffice_conversion_available;
|
|---|
| 197 |
|
|---|
| 198 | return bless $self, $class;
|
|---|
| 199 |
|
|---|
| 200 | }
|
|---|
| 201 |
|
|---|
| 202 | # launch_cmd variable seems to be entirely unused
|
|---|
| 203 | sub init {
|
|---|
| 204 | my $self = shift(@_);
|
|---|
| 205 | my ($verbosity, $outhandle, $failhandle) = @_;
|
|---|
| 206 |
|
|---|
| 207 | if ($openoffice_conversion_available) {
|
|---|
| 208 | my $oo_port = $self->{'openoffice_port'};
|
|---|
| 209 |
|
|---|
| 210 | my $launch_cmd = "soffice";
|
|---|
| 211 | $launch_cmd .= " \"--accept=socket,host=localhost,port=$oo_port;urp;StarOffice.ServiceManager\"";
|
|---|
| 212 | $launch_cmd .= " --headless";
|
|---|
| 213 | $self->{'openoffice_launch_cmd'} = $launch_cmd;
|
|---|
| 214 | }
|
|---|
| 215 |
|
|---|
| 216 | $self->{'ootmp_file_paths'} = ();
|
|---|
| 217 | }
|
|---|
| 218 |
|
|---|
| 219 | sub deinit {
|
|---|
| 220 | my $self = shift(@_);
|
|---|
| 221 |
|
|---|
| 222 | $self->clean_up_temporary_files();
|
|---|
| 223 | }
|
|---|
| 224 |
|
|---|
| 225 |
|
|---|
| 226 | sub convert {
|
|---|
| 227 | my $self = shift(@_);
|
|---|
| 228 | my $source_file_full_path = shift(@_);
|
|---|
| 229 | my $target_file_type = shift(@_);
|
|---|
| 230 | my $convert_options = shift(@_) || "";
|
|---|
| 231 | my $convert_id = shift(@_) || "";
|
|---|
| 232 | my $cache_mode = shift(@_) || "";
|
|---|
| 233 |
|
|---|
| 234 | return (0,undef,undef) unless $openoffice_conversion_available;
|
|---|
| 235 | # check the filename
|
|---|
| 236 | return (0,undef,undef) if ( !-f $source_file_full_path);
|
|---|
| 237 |
|
|---|
| 238 | my $outhandle = $self->{'outhandle'};
|
|---|
| 239 | my $verbosity = $self->{'verbosity'};
|
|---|
| 240 |
|
|---|
| 241 | my $source_file_no_path = &File::Basename::basename($source_file_full_path);
|
|---|
| 242 | # Determine the full name and path of the output file
|
|---|
| 243 | my $target_file_path;
|
|---|
| 244 | if ($self->{'enable_cache'}) {
|
|---|
| 245 | $self->init_cache_for_file($source_file_full_path);
|
|---|
| 246 | my $cache_dir = $self->{'cached_dir'};
|
|---|
| 247 | my $file_root = $self->{'cached_file_root'};
|
|---|
| 248 | $file_root .= "_$convert_id" if ($convert_id ne "");
|
|---|
| 249 | my $target_file = "$file_root.$target_file_type";
|
|---|
| 250 | $target_file_path = &FileUtils::filenameConcatenate($cache_dir,$target_file);
|
|---|
| 251 | }
|
|---|
| 252 | else {
|
|---|
| 253 | $target_file_path = &util::get_timestamped_tmp_filename_in_collection($source_file_full_path, $target_file_type);
|
|---|
| 254 | push(@{$self->{'ootmp_file_paths'}}, $target_file_path);
|
|---|
| 255 | }
|
|---|
| 256 |
|
|---|
| 257 | # Generate and run the convert command
|
|---|
| 258 |
|
|---|
| 259 | my $gextoo_home = $ENV{'GEXT_OPENOFFICE'};
|
|---|
| 260 | #my $jodjar = &FileUtils::filenameConcatenate($gextoo_home,"lib","java","jodconverter.jar");
|
|---|
| 261 | my $jodjar = &FileUtils::filenameConcatenate($gextoo_home,"jodconverter-2.2.2","lib","jodconverter-cli-2.2.2.jar");
|
|---|
| 262 |
|
|---|
| 263 | # debugging: print out all the env vars, when import's verbosity is high
|
|---|
| 264 | #foreach my $key (sort(keys %ENV)) {
|
|---|
| 265 | # print $outhandle "$key = $ENV{$key}\n";
|
|---|
| 266 | #}
|
|---|
| 267 |
|
|---|
| 268 | # *If* SOFFICE_HOME\program was not added to the PATH, *another* way is to insert:
|
|---|
| 269 | # -Doffice.home=C:\\PROGRA~1\\OPENOF~1.ORG
|
|---|
| 270 | # into the command to launch jodconverter so that, when using the remote GS server
|
|---|
| 271 | # on Windows, jodconverter can find the openoffice installation:
|
|---|
| 272 |
|
|---|
| 273 | my $office_short_path;
|
|---|
| 274 | if(defined $ENV{'SOFFICE_HOME'}) { # points to libreoffice or openoffice, since soffice launches both
|
|---|
| 275 | if ($ENV{'GSDLOS'} =~ m/^windows$/) {
|
|---|
| 276 | $office_short_path = Win32::GetShortPathName($ENV{'SOFFICE_HOME'});
|
|---|
| 277 | } else {
|
|---|
| 278 | $office_short_path = $ENV{'SOFFICE_HOME'};
|
|---|
| 279 | }
|
|---|
| 280 | }
|
|---|
| 281 | else { # SOFFICE_HOME not user-defined, try defaults
|
|---|
| 282 | if ($ENV{'GSDLOS'} =~ m/^windows$/) {
|
|---|
| 283 | $office_short_path = "C:\\PROGRA~1\\OPENOF~1.ORG";
|
|---|
| 284 | } else {
|
|---|
| 285 | if (-d "/usr/lib/libreoffice") { # try libreoffice first
|
|---|
| 286 | $office_short_path = "/usr/lib/libreoffice";
|
|---|
| 287 | } elsif (-d "/usr/lib/openoffice") { # perhaps it's more likely that openoffice exists (fallback)
|
|---|
| 288 | $office_short_path = "/usr/lib/openoffice";
|
|---|
| 289 | } elsif (-d "/usr/lib64/openoffice.org3") { # on 64 bit linux such as kjdon's CentOS
|
|---|
| 290 | $office_short_path = "/usr/lib64/openoffice.org3";
|
|---|
| 291 | } else { # set to default linux officeHome which jodconverter tries for
|
|---|
| 292 | $office_short_path = "/opt/openoffice.org3";
|
|---|
| 293 | }
|
|---|
| 294 | }
|
|---|
| 295 | }
|
|---|
| 296 |
|
|---|
| 297 | #if($self->{'openoffice_port'} != $ENV{'SOFFICE_PORT'}) { # if the user changed it
|
|---|
| 298 | #$ENV{'SOFFICE_PORT'} = $self->{'openoffice_port'};
|
|---|
| 299 | #}
|
|---|
| 300 |
|
|---|
| 301 | # jodjar always assumes that soffice (officeHome) is in "C:\Prog Files\OpenOffice org 3"
|
|---|
| 302 | # or "/opt/openoffice.org3" rather than checking for any user-defined SOFFICE_HOME var.
|
|---|
| 303 | # Therefore, we run jodjar with the -Doffice.home flag explicitly set, instead of without.
|
|---|
| 304 | # my $convert_cmd = "java -jar \"$jodjar\" --port $ENV{'SOFFICE_PORT'}";
|
|---|
| 305 | my $convert_cmd = "java -Doffice.home=\"".$office_short_path."\" -jar \"$jodjar\" --port $self->{'openoffice_port'}"; # --port $ENV{'SOFFICE_PORT'}
|
|---|
| 306 | $convert_cmd .= " \"$source_file_full_path\" \"$target_file_path\"";
|
|---|
| 307 |
|
|---|
| 308 | if ($verbosity>2) {
|
|---|
| 309 | print $outhandle "Convert command: $convert_cmd\n";
|
|---|
| 310 | }
|
|---|
| 311 |
|
|---|
| 312 | my $print_info = { 'message_prefix' => "OpenOffice Conversion",
|
|---|
| 313 | 'message' => "Converting $source_file_no_path to: $target_file_type" };
|
|---|
| 314 | $print_info->{'cache_mode'} = $cache_mode if ($cache_mode ne "");
|
|---|
| 315 |
|
|---|
| 316 | my ($regenerated,$result,$had_error)
|
|---|
| 317 | = $self->autorun_general_cmd($convert_cmd,$source_file_full_path, $target_file_path,$print_info);
|
|---|
| 318 | if ($had_error) {
|
|---|
| 319 | return (0, $result,$target_file_path);
|
|---|
| 320 | }
|
|---|
| 321 | return (1, $result,$target_file_path);
|
|---|
| 322 | }
|
|---|
| 323 |
|
|---|
| 324 |
|
|---|
| 325 | sub convert_without_result {
|
|---|
| 326 | my $self = shift(@_);
|
|---|
| 327 |
|
|---|
| 328 | my $source_file_path = shift(@_);
|
|---|
| 329 | my $target_file_type = shift(@_);
|
|---|
| 330 | my $convert_options = shift(@_) || "";
|
|---|
| 331 | my $convert_id = shift(@_) || "";
|
|---|
| 332 |
|
|---|
| 333 | return $self->convert($source_file_path,$target_file_type,
|
|---|
| 334 | $convert_options,$convert_id,"without_result");
|
|---|
| 335 | }
|
|---|
| 336 |
|
|---|
| 337 |
|
|---|
| 338 | sub tmp_area_convert_fileXX {
|
|---|
| 339 | my $self = shift (@_);
|
|---|
| 340 | my ($output_ext, $input_filename, $textref) = @_;
|
|---|
| 341 |
|
|---|
| 342 | my $outhandle = $self->{'outhandle'};
|
|---|
| 343 | my $convert_to = $self->{'convert_to'};
|
|---|
| 344 | my $failhandle = $self->{'failhandle'};
|
|---|
| 345 | my $convert_to_ext = $self->{'convert_to_ext'};
|
|---|
| 346 |
|
|---|
| 347 | # derive tmp filename from input filename
|
|---|
| 348 | my ($tailname, $dirname, $suffix)
|
|---|
| 349 | = &File::Basename::fileparse($input_filename, "\\.[^\\.]+\$");
|
|---|
| 350 | }
|
|---|
| 351 |
|
|---|
| 352 |
|
|---|
| 353 | sub clean_up_temporary_files {
|
|---|
| 354 | my $self = shift(@_);
|
|---|
| 355 |
|
|---|
| 356 | foreach my $ootmp_file_path (@{$self->{'ootmp_file_paths'}}) {
|
|---|
| 357 | if (-e $ootmp_file_path) {
|
|---|
| 358 | &FileUtils::removeFiles($ootmp_file_path);
|
|---|
| 359 | }
|
|---|
| 360 | }
|
|---|
| 361 |
|
|---|
| 362 | $self->{'ootmp_file_paths'} = ();
|
|---|
| 363 | }
|
|---|
| 364 |
|
|---|
| 365 |
|
|---|
| 366 |
|
|---|
| 367 | 1;
|
|---|
