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 | # check all the favourite haunts of openoffice in turn
|
---|
87 | $ENV{'SOFFICE_HOME'} = &util::get_first_existing_dir(
|
---|
88 | &FileUtils::filenameConcatenate($ENV{'ProgramFiles'},"OpenOffice.org 3"),
|
---|
89 | &FileUtils::filenameConcatenate($ENV{'ProgramFiles(X86)'},"OpenOffice.org 3"),
|
---|
90 | &FileUtils::filenameConcatenate($ENV{'ProgramFiles'},"LibreOffice"),
|
---|
91 | &FileUtils::filenameConcatenate($ENV{'ProgramFiles(X86)'},"LibreOffice")
|
---|
92 | );
|
---|
93 |
|
---|
94 | # These 2 env vars are diff and appear as expected when printed out in a cmd prompt on a 64 bit Windows. But on same machine
|
---|
95 | # in perl, at least when launched through (Java 32 bit) GLI, their values shown as the same: as "C:\Program Files (x86)"
|
---|
96 | #print STDERR "@@@@ ProgFiles: $ENV{'ProgramFiles'}\n";
|
---|
97 | #print STDERR "@@@@ ProgFiles x86: $ENV{'ProgramFiles(X86)'}\n";
|
---|
98 | }
|
---|
99 | if(defined $ENV{'SOFFICE_HOME'}) {
|
---|
100 | #print STDERR "@@@@ Found $ENV{'SOFFICE_HOME'}\n";
|
---|
101 | my $ooffice_dir_guess =
|
---|
102 | &FileUtils::filenameConcatenate($ENV{'SOFFICE_HOME'},"program");
|
---|
103 | if (-d $ooffice_dir_guess) {
|
---|
104 | &util::envvar_append("PATH",$ooffice_dir_guess);
|
---|
105 | }
|
---|
106 | }
|
---|
107 |
|
---|
108 | # for windows, when working on a remote system, want to be able to start OO if
|
---|
109 | # not already running. We'll use the uno socket method to do so. Else client-gli
|
---|
110 | # tends to hang, waiting for the prompt to return after OO has been started up
|
---|
111 | # (which doesn't happen, so need to Ctrl-C GLI and run it again for it to work).
|
---|
112 |
|
---|
113 | # first need to silently check soffice exists else windows will display a popup
|
---|
114 | $status = system("which soffice >nul 2>&1"); # which.exe should be in bin/windows
|
---|
115 | if ($status == 0) {
|
---|
116 | #$cmd = "start \"soffice process\" $cmd >nul 2>&1"; # order of >nul and 2>&1 matters, see below
|
---|
117 |
|
---|
118 | # On Windows, when perl/C/C++ uses system() to launch any standalone process in the background
|
---|
119 | # with "start \"window title\" cmd", perl's child process (despite being mostly independent/detached
|
---|
120 | # from perl) *appears* to share stdout/stderr (stdin) streams with the parent perl process. This is
|
---|
121 | # noticeable with perl launching soffice or any program that neither writes to stderr/out nor sends
|
---|
122 | # eof/eos to indicate parent perl's streams are closed (e.g. Notepad).
|
---|
123 | # This is *not* a problem when perl scripts are run from command line. Command line programs run fine.
|
---|
124 | # But when combined with SafeProcess used by GLI, such silent standalone programs like soffice
|
---|
125 | # cause GLI to hang every time right until the silent program's terminated, because SafeProcess' join()
|
---|
126 | # calls on the perl child process' stderr and stdout block at SafeProcess.InputStreamGobbler.readLine(),
|
---|
127 | # because these stderr/out streams didn't receive eof/eos, implying the silent grandchild process
|
---|
128 | # somehow kept them open.
|
---|
129 | # This is true at Windows/C's system() command level: InputStreamGobblers using join() calls (as SafeProcess
|
---|
130 | # does), combined with a c++ program that uses system() that launches something like Notepad, just all block
|
---|
131 | # until Notepad is closed.
|
---|
132 | # When perl calls a custom java BackgroundLauncher.java program to launch an external program with Runtime.exec(),
|
---|
133 | # this hanging problem doesn't happen as java's exec() is not implemented with Windows C's system().
|
---|
134 | # Similarly, a WScript file (vbs or js script) that uses WinScriptHost.Run method to run an external program
|
---|
135 | # also doesn't exhibit this problem. So we use a custom vbs script to launch external programs like soffice in
|
---|
136 | # the background from perl (to avoid having to set up Java or have the java program compiled up by the release-kits,
|
---|
137 | # since perl building scripts are not just run from GLI but are also run directly from the cmd line).
|
---|
138 |
|
---|
139 | $cmd = "CScript //Nologo $ENV{'GSDLHOME'}\\bin\\windows\\background-launcher.vbs $cmd";
|
---|
140 |
|
---|
141 | #print STDERR "***** Tried to start-up OpenOffice with:\n$cmd\n";
|
---|
142 | }
|
---|
143 | # else cmd is still undefined
|
---|
144 | }
|
---|
145 | else {
|
---|
146 | # Windows seems to launch OpenOffice as a service (i.e.
|
---|
147 | # automatically puts it in the background).
|
---|
148 | # For Unix putting it in the background needs to be done
|
---|
149 | # explicitly by appending & to the end
|
---|
150 |
|
---|
151 | # IMPORTANT! When redirecting output from 2>&1 to a file or null, >/dev/null (or >out.txt) should come BEFORE 2>&1
|
---|
152 | # i.e. $some_cmd >out.txt 2>&1 and likewise $some_cmd >/dev/null 2>&1
|
---|
153 | # Doing it in the wrong order further causes problems with SafeProcess when GLI starts up and calls pluginfo on the pluginslist:
|
---|
154 | # SafeProcess blocks forever on read() from stdout of the process running pluginfo, because there's never data including no eof/eos
|
---|
155 | # on the process' stdout when running this command in the wrong order.
|
---|
156 | $cmd .= " >/dev/null 2>&1 &";
|
---|
157 | }
|
---|
158 |
|
---|
159 | #print STDERR "@@@@ running: $cmd\n";
|
---|
160 |
|
---|
161 | $status = system($cmd) if ($status == 0);
|
---|
162 | if ($status != 0) {
|
---|
163 | #print STDERR "Failed to run: $cmd\n";
|
---|
164 | #print STDERR "$!\n";
|
---|
165 | $openoffice_conversion_available = 0;
|
---|
166 | $no_openoffice_conversion_reason = "openofficenotinstalled";
|
---|
167 | }
|
---|
168 | }
|
---|
169 | }
|
---|
170 | }
|
---|
171 |
|
---|
172 | my $arguments = [
|
---|
173 | { 'name' => "openoffice_port",
|
---|
174 | 'desc' => "{OpenOfficeConverter.openoffice_port}",
|
---|
175 | 'type' => "int",
|
---|
176 | 'deft' => "8100",
|
---|
177 | 'range' => "81,",
|
---|
178 | 'reqd' => "no" },
|
---|
179 | ];
|
---|
180 |
|
---|
181 |
|
---|
182 | my $options = { 'name' => "OpenOfficeConverter",
|
---|
183 | 'desc' => "{OpenOfficeConverter.desc}",
|
---|
184 | 'abstract' => "yes",
|
---|
185 | 'inherits' => "yes",
|
---|
186 | 'args' => $arguments };
|
---|
187 |
|
---|
188 | sub new {
|
---|
189 | my ($class) = shift (@_);
|
---|
190 | my ($pluginlist,$inputargs,$hashArgOptLists,$auxilary) = @_;
|
---|
191 | push(@$pluginlist, $class);
|
---|
192 |
|
---|
193 | push(@{$hashArgOptLists->{"ArgList"}},@{$arguments});
|
---|
194 | push(@{$hashArgOptLists->{"OptList"}},$options);
|
---|
195 |
|
---|
196 | my $self = new BaseMediaConverter($pluginlist, $inputargs,
|
---|
197 | $hashArgOptLists, $auxilary);
|
---|
198 |
|
---|
199 | if ($self->{'info_only'}) {
|
---|
200 | # don't worry about any options etc
|
---|
201 | return bless $self, $class;
|
---|
202 | }
|
---|
203 | if (!$openoffice_conversion_available) {
|
---|
204 | $self->{'no_openoffice_conversion_reason'} = $no_openoffice_conversion_reason;
|
---|
205 |
|
---|
206 | my $outhandle = $self->{'outhandle'};
|
---|
207 | &gsprintf($outhandle, "OpenOfficeConverter: {OpenOfficeConverter.noconversionavailable} ({OpenOfficeConverter.$no_openoffice_conversion_reason})\n");
|
---|
208 | }
|
---|
209 |
|
---|
210 | $self->{'openoffice_conversion_available'} = $openoffice_conversion_available;
|
---|
211 |
|
---|
212 | return bless $self, $class;
|
---|
213 |
|
---|
214 | }
|
---|
215 |
|
---|
216 | # launch_cmd variable seems to be entirely unused
|
---|
217 | sub init {
|
---|
218 | my $self = shift(@_);
|
---|
219 | my ($verbosity, $outhandle, $failhandle) = @_;
|
---|
220 |
|
---|
221 | if ($openoffice_conversion_available) {
|
---|
222 | my $oo_port = $self->{'openoffice_port'};
|
---|
223 |
|
---|
224 | my $launch_cmd = "soffice";
|
---|
225 | $launch_cmd .= " \"--accept=socket,host=localhost,port=$oo_port;urp;StarOffice.ServiceManager\"";
|
---|
226 | $launch_cmd .= " --headless";
|
---|
227 | $self->{'openoffice_launch_cmd'} = $launch_cmd;
|
---|
228 | }
|
---|
229 |
|
---|
230 | $self->{'ootmp_file_paths'} = ();
|
---|
231 | }
|
---|
232 |
|
---|
233 | sub deinit {
|
---|
234 | my $self = shift(@_);
|
---|
235 |
|
---|
236 | $self->clean_up_temporary_files();
|
---|
237 | }
|
---|
238 |
|
---|
239 |
|
---|
240 | sub convert {
|
---|
241 | my $self = shift(@_);
|
---|
242 | my $source_file_full_path = shift(@_);
|
---|
243 | my $target_file_type = shift(@_);
|
---|
244 | my $convert_options = shift(@_) || "";
|
---|
245 | my $convert_id = shift(@_) || "";
|
---|
246 | my $cache_mode = shift(@_) || "";
|
---|
247 |
|
---|
248 | return (0,undef,undef) unless $openoffice_conversion_available;
|
---|
249 | # check the filename
|
---|
250 | return (0,undef,undef) if ( !-f $source_file_full_path);
|
---|
251 |
|
---|
252 | my $outhandle = $self->{'outhandle'};
|
---|
253 | my $verbosity = $self->{'verbosity'};
|
---|
254 |
|
---|
255 | my $source_file_no_path = &File::Basename::basename($source_file_full_path);
|
---|
256 | # Determine the full name and path of the output file
|
---|
257 | my $target_file_path;
|
---|
258 | if ($self->{'enable_cache'}) {
|
---|
259 | $self->init_cache_for_file($source_file_full_path);
|
---|
260 | my $cache_dir = $self->{'cached_dir'};
|
---|
261 | my $file_root = $self->{'cached_file_root'};
|
---|
262 | $file_root .= "_$convert_id" if ($convert_id ne "");
|
---|
263 | my $target_file = "$file_root.$target_file_type";
|
---|
264 | $target_file_path = &FileUtils::filenameConcatenate($cache_dir,$target_file);
|
---|
265 | }
|
---|
266 | else {
|
---|
267 | $target_file_path = &util::get_timestamped_tmp_filename_in_collection($source_file_full_path, $target_file_type);
|
---|
268 | push(@{$self->{'ootmp_file_paths'}}, $target_file_path);
|
---|
269 | }
|
---|
270 |
|
---|
271 | # Generate and run the convert command
|
---|
272 |
|
---|
273 | my $gextoo_home = $ENV{'GEXT_OPENOFFICE'};
|
---|
274 | #my $jodjar = &FileUtils::filenameConcatenate($gextoo_home,"lib","java","jodconverter.jar");
|
---|
275 | my $jodjar = &FileUtils::filenameConcatenate($gextoo_home,"jodconverter-2.2.2","lib","jodconverter-cli-2.2.2.jar");
|
---|
276 |
|
---|
277 | # debugging: print out all the env vars, when import's verbosity is high
|
---|
278 | #foreach my $key (sort(keys %ENV)) {
|
---|
279 | # print $outhandle "$key = $ENV{$key}\n";
|
---|
280 | #}
|
---|
281 |
|
---|
282 | # *If* SOFFICE_HOME\program was not added to the PATH, *another* way is to insert:
|
---|
283 | # -Doffice.home=C:\\PROGRA~1\\OPENOF~1.ORG
|
---|
284 | # into the command to launch jodconverter so that, when using the remote GS server
|
---|
285 | # on Windows, jodconverter can find the openoffice installation:
|
---|
286 |
|
---|
287 | my $office_short_path;
|
---|
288 | if(defined $ENV{'SOFFICE_HOME'}) { # points to libreoffice or openoffice, since soffice launches both
|
---|
289 | if ($ENV{'GSDLOS'} =~ m/^windows$/) {
|
---|
290 | $office_short_path = Win32::GetShortPathName($ENV{'SOFFICE_HOME'});
|
---|
291 | } else {
|
---|
292 | $office_short_path = $ENV{'SOFFICE_HOME'};
|
---|
293 | }
|
---|
294 | }
|
---|
295 | else { # SOFFICE_HOME not user-defined, try defaults
|
---|
296 | if ($ENV{'GSDLOS'} =~ m/^windows$/) {
|
---|
297 | $office_short_path = "C:\\PROGRA~1\\OPENOF~1.ORG";
|
---|
298 | } else {
|
---|
299 | if (-d "/usr/lib/libreoffice") { # try libreoffice first
|
---|
300 | $office_short_path = "/usr/lib/libreoffice";
|
---|
301 | } elsif (-d "/usr/lib/openoffice") { # perhaps it's more likely that openoffice exists (fallback)
|
---|
302 | $office_short_path = "/usr/lib/openoffice";
|
---|
303 | } elsif (-d "/usr/lib64/openoffice.org3") { # on 64 bit linux such as kjdon's CentOS
|
---|
304 | $office_short_path = "/usr/lib64/openoffice.org3";
|
---|
305 | } else { # set to default linux officeHome which jodconverter tries for
|
---|
306 | $office_short_path = "/opt/openoffice.org3";
|
---|
307 | }
|
---|
308 | }
|
---|
309 | }
|
---|
310 |
|
---|
311 | #if($self->{'openoffice_port'} != $ENV{'SOFFICE_PORT'}) { # if the user changed it
|
---|
312 | #$ENV{'SOFFICE_PORT'} = $self->{'openoffice_port'};
|
---|
313 | #}
|
---|
314 |
|
---|
315 | # jodjar always assumes that soffice (officeHome) is in "C:\Prog Files\OpenOffice org 3"
|
---|
316 | # or "/opt/openoffice.org3" rather than checking for any user-defined SOFFICE_HOME var.
|
---|
317 | # Therefore, we run jodjar with the -Doffice.home flag explicitly set, instead of without.
|
---|
318 | # my $convert_cmd = "java -jar \"$jodjar\" --port $ENV{'SOFFICE_PORT'}";
|
---|
319 | my $convert_cmd = "java -Doffice.home=\"".$office_short_path."\" -jar \"$jodjar\" --port $self->{'openoffice_port'}"; # --port $ENV{'SOFFICE_PORT'}
|
---|
320 | $convert_cmd .= " \"$source_file_full_path\" \"$target_file_path\"";
|
---|
321 |
|
---|
322 | if ($verbosity>2) {
|
---|
323 | print $outhandle "Convert command: $convert_cmd\n";
|
---|
324 | }
|
---|
325 |
|
---|
326 | my $print_info = { 'message_prefix' => "OpenOffice Conversion",
|
---|
327 | 'message' => "Converting $source_file_no_path to: $target_file_type" };
|
---|
328 | $print_info->{'cache_mode'} = $cache_mode if ($cache_mode ne "");
|
---|
329 |
|
---|
330 | my ($regenerated,$result,$had_error)
|
---|
331 | = $self->autorun_general_cmd($convert_cmd,$source_file_full_path, $target_file_path,$print_info);
|
---|
332 | if ($had_error) {
|
---|
333 | return (0, $result,$target_file_path);
|
---|
334 | }
|
---|
335 | return (1, $result,$target_file_path);
|
---|
336 | }
|
---|
337 |
|
---|
338 |
|
---|
339 | sub convert_without_result {
|
---|
340 | my $self = shift(@_);
|
---|
341 |
|
---|
342 | my $source_file_path = shift(@_);
|
---|
343 | my $target_file_type = shift(@_);
|
---|
344 | my $convert_options = shift(@_) || "";
|
---|
345 | my $convert_id = shift(@_) || "";
|
---|
346 |
|
---|
347 | return $self->convert($source_file_path,$target_file_type,
|
---|
348 | $convert_options,$convert_id,"without_result");
|
---|
349 | }
|
---|
350 |
|
---|
351 |
|
---|
352 | sub tmp_area_convert_fileXX {
|
---|
353 | my $self = shift (@_);
|
---|
354 | my ($output_ext, $input_filename, $textref) = @_;
|
---|
355 |
|
---|
356 | my $outhandle = $self->{'outhandle'};
|
---|
357 | my $convert_to = $self->{'convert_to'};
|
---|
358 | my $failhandle = $self->{'failhandle'};
|
---|
359 | my $convert_to_ext = $self->{'convert_to_ext'};
|
---|
360 |
|
---|
361 | # derive tmp filename from input filename
|
---|
362 | my ($tailname, $dirname, $suffix)
|
---|
363 | = &File::Basename::fileparse($input_filename, "\\.[^\\.]+\$");
|
---|
364 | }
|
---|
365 |
|
---|
366 |
|
---|
367 | sub clean_up_temporary_files {
|
---|
368 | my $self = shift(@_);
|
---|
369 |
|
---|
370 | foreach my $ootmp_file_path (@{$self->{'ootmp_file_paths'}}) {
|
---|
371 | if (-e $ootmp_file_path) {
|
---|
372 | &FileUtils::removeFiles($ootmp_file_path);
|
---|
373 | }
|
---|
374 | }
|
---|
375 |
|
---|
376 | $self->{'ootmp_file_paths'} = ();
|
---|
377 | }
|
---|
378 |
|
---|
379 |
|
---|
380 |
|
---|
381 | 1;
|
---|