1 | ###########################################################################
|
---|
2 | #
|
---|
3 | # plugin.pm -- functions to handle using plugins
|
---|
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 | ###########################################################################
|
---|
25 |
|
---|
26 | package plugin;
|
---|
27 |
|
---|
28 | use inexport;
|
---|
29 |
|
---|
30 | use strict; # to pick up typos and undeclared variables...
|
---|
31 | no strict 'refs'; # ...but allow filehandles to be variables and vice versa
|
---|
32 | no strict 'subs';
|
---|
33 |
|
---|
34 | require util;
|
---|
35 | use gsprintf 'gsprintf';
|
---|
36 |
|
---|
37 | # mapping from old plugin names to new ones for backwards compatibility
|
---|
38 | # can remove at sometime in future when we no longer want to support old xxPlug names in the config file
|
---|
39 | my $plugin_name_map = {
|
---|
40 | 'GAPlug' => 'GreenstoneXMLPlugin',
|
---|
41 | 'ArcPlug' => 'ArchivesInfPlugin',
|
---|
42 | 'RecPlug' => 'DirectoryPlugin',
|
---|
43 | 'TEXTPlug' => 'TextPlugin',
|
---|
44 | 'XMLPlug' => 'ReadXMLFile',
|
---|
45 | 'EMAILPlug' => 'EmailPlugin',
|
---|
46 | 'SRCPlug' => 'SourceCodePlugin',
|
---|
47 | 'NULPlug' => 'NulPlugin',
|
---|
48 | 'W3ImgPlug' => 'HTMLImagePlugin',
|
---|
49 | 'PagedImgPlug' => 'PagedImagePlugin',
|
---|
50 | 'METSPlug' => 'GreenstoneMETSPlugin',
|
---|
51 | 'PPTPlug' => 'PowerPointPlugin',
|
---|
52 | 'PSPlug' => 'PostScriptPlugin',
|
---|
53 | 'DBPlug' => 'DatabasePlugin'
|
---|
54 | };
|
---|
55 |
|
---|
56 | # global variables
|
---|
57 | my $stats = {'num_processed' => 0,
|
---|
58 | 'num_blocked' => 0,
|
---|
59 | 'num_not_processed' => 0,
|
---|
60 | 'num_not_recognised' => 0,
|
---|
61 | 'num_archives' => 0
|
---|
62 | };
|
---|
63 |
|
---|
64 | #globaloptions contains any options that should be passed to all plugins
|
---|
65 | my ($verbosity, $outhandle, $failhandle, $globaloptions);
|
---|
66 |
|
---|
67 | sub get_valid_pluginname {
|
---|
68 | my ($pluginname) = @_;
|
---|
69 | my $valid_name = $pluginname;
|
---|
70 | if (defined $plugin_name_map->{$pluginname}) {
|
---|
71 | $valid_name = $plugin_name_map->{$pluginname};
|
---|
72 | } elsif ($pluginname =~ /Plug$/) {
|
---|
73 | $valid_name =~ s/Plug/Plugin/;
|
---|
74 |
|
---|
75 | }
|
---|
76 | return $valid_name;
|
---|
77 | }
|
---|
78 |
|
---|
79 | sub load_plugin_require
|
---|
80 | {
|
---|
81 | my ($pluginname) = @_;
|
---|
82 |
|
---|
83 | my @check_list = ();
|
---|
84 |
|
---|
85 | # pp_plugname shorthand for 'perllib' 'plugin' '$pluginname.pm'
|
---|
86 | my $pp_plugname
|
---|
87 | = &util::filename_cat('perllib', 'plugins', "${pluginname}.pm");
|
---|
88 | my $collectdir = $ENV{'GSDLCOLLECTDIR'};
|
---|
89 |
|
---|
90 | # find the plugin
|
---|
91 | if (defined($ENV{'GSDLCOLLECTION'}))
|
---|
92 | {
|
---|
93 | my $customplugname
|
---|
94 | = &util::filename_cat($collectdir, "custom",$ENV{'GSDLCOLLECTION'},
|
---|
95 | $pp_plugname);
|
---|
96 | push(@check_list,$customplugname);
|
---|
97 | }
|
---|
98 |
|
---|
99 | my $colplugname = &util::filename_cat($collectdir, $pp_plugname);
|
---|
100 | push(@check_list,$colplugname);
|
---|
101 |
|
---|
102 | if (defined $ENV{'GSDLEXTS'}) {
|
---|
103 |
|
---|
104 | my $ext_prefix = &util::filename_cat($ENV{'GSDLHOME'}, "ext");
|
---|
105 |
|
---|
106 | my @extensions = split(/:/,$ENV{'GSDLEXTS'});
|
---|
107 | foreach my $e (@extensions) {
|
---|
108 | my $extplugname = &util::filename_cat($ext_prefix, $e, $pp_plugname);
|
---|
109 | push(@check_list,$extplugname);
|
---|
110 |
|
---|
111 | }
|
---|
112 | }
|
---|
113 | if (defined $ENV{'GSDL3EXTS'}) {
|
---|
114 |
|
---|
115 | my $ext_prefix = &util::filename_cat($ENV{'GSDL3SRCHOME'}, "ext");
|
---|
116 |
|
---|
117 | my @extensions = split(/:/,$ENV{'GSDL3EXTS'});
|
---|
118 | foreach my $e (@extensions) {
|
---|
119 | my $extplugname = &util::filename_cat($ext_prefix, $e, $pp_plugname);
|
---|
120 | push(@check_list,$extplugname);
|
---|
121 |
|
---|
122 | }
|
---|
123 | }
|
---|
124 |
|
---|
125 |
|
---|
126 | my $mainplugname = &util::filename_cat($ENV{'GSDLHOME'}, $pp_plugname);
|
---|
127 | push(@check_list,$mainplugname);
|
---|
128 |
|
---|
129 | my $success=0;
|
---|
130 | foreach my $plugname (@check_list) {
|
---|
131 | if (-e $plugname) {
|
---|
132 | require $plugname;
|
---|
133 | $success=1;
|
---|
134 | last;
|
---|
135 | }
|
---|
136 | }
|
---|
137 |
|
---|
138 | if (!$success) {
|
---|
139 | &gsprintf(STDERR, "{plugin.could_not_find_plugin}\n",
|
---|
140 | $pluginname);
|
---|
141 | die "\n";
|
---|
142 | }
|
---|
143 | }
|
---|
144 |
|
---|
145 | sub load_plugin_for_info {
|
---|
146 | my ($pluginname) = shift @_;
|
---|
147 | $pluginname = &get_valid_pluginname($pluginname);
|
---|
148 | load_plugin_require($pluginname);
|
---|
149 |
|
---|
150 | # create a plugin object
|
---|
151 | my ($plugobj);
|
---|
152 | my $options = "-gsdlinfo";
|
---|
153 |
|
---|
154 | eval ("\$plugobj = new \$pluginname([],[$options])");
|
---|
155 | die "$@" if $@;
|
---|
156 |
|
---|
157 | return $plugobj;
|
---|
158 | }
|
---|
159 |
|
---|
160 | sub load_plugins {
|
---|
161 | my ($plugin_list) = shift @_;
|
---|
162 | my $incremental_mode;
|
---|
163 | ($verbosity, $outhandle, $failhandle, $globaloptions, $incremental_mode) = @_; # globals
|
---|
164 | my @plugin_objects = ();
|
---|
165 | $verbosity = 2 unless defined $verbosity;
|
---|
166 | $outhandle = 'STDERR' unless defined $outhandle;
|
---|
167 | $failhandle = 'STDERR' unless defined $failhandle;
|
---|
168 |
|
---|
169 | my $colplugindir = &util::filename_cat($ENV{'GSDLCOLLECTDIR'},"perllib/plugins");
|
---|
170 | unshift (@INC, $colplugindir);
|
---|
171 |
|
---|
172 | map { $_ = "\"$_\""; } @$globaloptions;
|
---|
173 | my $globals = join (",", @$globaloptions);
|
---|
174 |
|
---|
175 | foreach my $pluginoptions (@$plugin_list) {
|
---|
176 | my $pluginname = shift @$pluginoptions;
|
---|
177 | next unless defined $pluginname;
|
---|
178 | $pluginname = &get_valid_pluginname($pluginname);
|
---|
179 | load_plugin_require($pluginname);
|
---|
180 |
|
---|
181 | # create a plugin object
|
---|
182 | my ($plugobj);
|
---|
183 | map { $_ = "\"$_\""; } @$pluginoptions;
|
---|
184 | my $options = join (",", @$pluginoptions);
|
---|
185 | if ($globals) {
|
---|
186 | if (@$pluginoptions) {
|
---|
187 | $options .= ",";
|
---|
188 | }
|
---|
189 | $options .= "$globals";
|
---|
190 | }
|
---|
191 | # need to escape backslash before putting in to the eval
|
---|
192 | $options =~ s/\\/\\\\/g;
|
---|
193 | $options =~ s/\$/\\\$/g;
|
---|
194 | eval ("\$plugobj = new \$pluginname([],[$options])");
|
---|
195 | die "$@" if $@;
|
---|
196 |
|
---|
197 | # initialize plugin
|
---|
198 | $plugobj->init($verbosity, $outhandle, $failhandle);
|
---|
199 |
|
---|
200 | $plugobj->set_incremental($incremental_mode);
|
---|
201 |
|
---|
202 | # add this object to the list
|
---|
203 | push (@plugin_objects, $plugobj);
|
---|
204 | }
|
---|
205 |
|
---|
206 | return \@plugin_objects;
|
---|
207 | }
|
---|
208 |
|
---|
209 |
|
---|
210 | sub begin {
|
---|
211 | my ($pluginfo, $base_dir, $processor, $maxdocs, $gli) = @_;
|
---|
212 |
|
---|
213 | map { $_->{'gli'} = $gli; } @$pluginfo;
|
---|
214 | map { $_->begin($pluginfo, $base_dir, $processor, $maxdocs); } @$pluginfo;
|
---|
215 | }
|
---|
216 |
|
---|
217 | sub remove_all {
|
---|
218 | my ($pluginfo, $base_dir, $processor, $maxdocs) = @_;
|
---|
219 |
|
---|
220 | map { $_->remove_all($pluginfo, $base_dir, $processor, $maxdocs); } @$pluginfo;
|
---|
221 | }
|
---|
222 |
|
---|
223 | sub remove_some {
|
---|
224 | my ($pluginfo, $archivedir, $deleted_files) = @_;
|
---|
225 | print STDERR "in remove some\n";
|
---|
226 | return if (scalar(@$deleted_files)==0);
|
---|
227 | my $arcinfo_src_filename = &inexport::src_db_file($archivedir);
|
---|
228 |
|
---|
229 | foreach my $file (@$deleted_files) {
|
---|
230 | # use 'archiveinf-src' GDBM file to look up all the OIDs
|
---|
231 | # that this file is used in (note in most cases, it's just one OID)
|
---|
232 |
|
---|
233 | my $src_rec = GDBMUtils::gdbmRecordToHash($arcinfo_src_filename,$file);
|
---|
234 | my $oids = $src_rec->{'oid'};
|
---|
235 | map {$_->remove_one($file, $oids); } @$pluginfo;
|
---|
236 | }
|
---|
237 |
|
---|
238 | }
|
---|
239 | sub file_block_read {
|
---|
240 | my ($pluginfo, $base_dir, $file, $block_hash, $metadata, $gli) = @_;
|
---|
241 |
|
---|
242 |
|
---|
243 | $gli = 0 unless defined $gli;
|
---|
244 |
|
---|
245 | my $rv = 0;
|
---|
246 | my $glifile = $file;
|
---|
247 |
|
---|
248 | $glifile =~ s/^[\/\\]+//; # file sometimes starts with a / so get rid of it
|
---|
249 |
|
---|
250 | # Announce to GLI that we are handling a file
|
---|
251 | print STDERR "<File n='$glifile'>\n" if $gli;
|
---|
252 |
|
---|
253 | # the .kill file is a handy (if not very elegant) way of aborting
|
---|
254 | # an import.pl or buildcol.pl process
|
---|
255 | if (-e &util::filename_cat ($ENV{'GSDLCOLLECTDIR'}, ".kill")) {
|
---|
256 | gsprintf($outhandle, "{plugin.kill_file}\n");
|
---|
257 | die "\n";
|
---|
258 | }
|
---|
259 |
|
---|
260 | foreach my $plugobj (@$pluginfo) {
|
---|
261 |
|
---|
262 | $rv = $plugobj->file_block_read($pluginfo, $base_dir, $file, $block_hash, $metadata, $gli);
|
---|
263 | #last if (defined $rv && $rv==1); # stop this file once we have found something to 'process' it
|
---|
264 | }
|
---|
265 |
|
---|
266 | }
|
---|
267 |
|
---|
268 |
|
---|
269 | sub metadata_read {
|
---|
270 | my ($pluginfo, $base_dir, $file, $block_hash,
|
---|
271 | $extrametakeys, $extrametadata, $extrametafile,
|
---|
272 | $processor, $maxdocs, $gli, $aux) = @_;
|
---|
273 |
|
---|
274 | $maxdocs = -1 unless defined $maxdocs && $maxdocs =~ /\d/;
|
---|
275 | $gli = 0 unless defined $gli;
|
---|
276 |
|
---|
277 | my $rv = 0;
|
---|
278 | my $glifile = $file;
|
---|
279 |
|
---|
280 | $glifile =~ s/^[\/\\]+//; # file sometimes starts with a / so get rid of it
|
---|
281 |
|
---|
282 | # Announce to GLI that we are handling a file
|
---|
283 | print STDERR "<File n='$glifile'>\n" if $gli;
|
---|
284 |
|
---|
285 | # the .kill file is a handy (if not very elegant) way of aborting
|
---|
286 | # an import.pl or buildcol.pl process
|
---|
287 | if (-e &util::filename_cat ($ENV{'GSDLCOLLECTDIR'}, ".kill")) {
|
---|
288 | gsprintf($outhandle, "{plugin.kill_file}\n");
|
---|
289 | die "\n";
|
---|
290 | }
|
---|
291 |
|
---|
292 | my $had_error = 0;
|
---|
293 | # pass this file by each of the plugins in turn until one
|
---|
294 | # is found which will process it
|
---|
295 | # read must return:
|
---|
296 | # undef - could not recognise
|
---|
297 | # -1 - tried but error
|
---|
298 | # 0 - blocked
|
---|
299 | # anything else for successful processing
|
---|
300 |
|
---|
301 | foreach my $plugobj (@$pluginfo) {
|
---|
302 |
|
---|
303 | $rv = $plugobj->metadata_read($pluginfo, $base_dir, $file, $block_hash,
|
---|
304 | $extrametakeys, $extrametadata, $extrametafile,
|
---|
305 | $processor, $maxdocs, $gli, $aux);
|
---|
306 |
|
---|
307 | if (defined $rv) {
|
---|
308 | if ($rv == -1) {
|
---|
309 | # an error has occurred
|
---|
310 | $had_error = 1;
|
---|
311 | print STDERR "<ProcessingError n='$glifile'>\n" if $gli;
|
---|
312 | } else {
|
---|
313 | return $rv;
|
---|
314 | }
|
---|
315 | } # else undefined - was not recognised by the plugin
|
---|
316 | }
|
---|
317 |
|
---|
318 | return 0;
|
---|
319 | }
|
---|
320 |
|
---|
321 | sub read {
|
---|
322 | my ($pluginfo, $base_dir, $file, $block_hash, $metadata, $processor, $maxdocs, $total_count, $gli, $aux) = @_;
|
---|
323 |
|
---|
324 | $maxdocs = -1 unless defined $maxdocs && $maxdocs =~ /\d/;
|
---|
325 | $total_count = 0 unless defined $total_count && $total_count =~ /\d/;
|
---|
326 | $gli = 0 unless defined $gli;
|
---|
327 |
|
---|
328 | my $rv = 0;
|
---|
329 | my $glifile = $file;
|
---|
330 |
|
---|
331 | $glifile =~ s/^[\/\\]+//; # file sometimes starts with a / so get rid of it
|
---|
332 |
|
---|
333 | # Announce to GLI that we are handling a file
|
---|
334 | print STDERR "<File n='$glifile'>\n" if $gli;
|
---|
335 |
|
---|
336 | # the .kill file is a handy (if not very elegant) way of aborting
|
---|
337 | # an import.pl or buildcol.pl process
|
---|
338 | if (-e &util::filename_cat ($ENV{'GSDLCOLLECTDIR'}, ".kill")) {
|
---|
339 | gsprintf($outhandle, "{plugin.kill_file}\n");
|
---|
340 | die "\n";
|
---|
341 | }
|
---|
342 |
|
---|
343 | my $had_error = 0;
|
---|
344 | # pass this file by each of the plugins in turn until one
|
---|
345 | # is found which will process it
|
---|
346 | # read must return:
|
---|
347 | # undef - could not recognise
|
---|
348 | # -1 - tried but error
|
---|
349 | # 0 - blocked
|
---|
350 | # anything else for successful processing
|
---|
351 |
|
---|
352 | foreach my $plugobj (@$pluginfo) {
|
---|
353 |
|
---|
354 | $rv = $plugobj->read($pluginfo, $base_dir, $file,
|
---|
355 | $block_hash, $metadata, $processor, $maxdocs,
|
---|
356 | $total_count, $gli, $aux);
|
---|
357 |
|
---|
358 | if (defined $rv) {
|
---|
359 | if ($rv == -1) {
|
---|
360 | # an error has occurred
|
---|
361 | $had_error = 1;
|
---|
362 | } else {
|
---|
363 | return $rv;
|
---|
364 | }
|
---|
365 | } # else undefined - was not recognised by the plugin
|
---|
366 | }
|
---|
367 |
|
---|
368 | if ($had_error) {
|
---|
369 | # was recognised but couldn't be processed
|
---|
370 | if ($verbosity >= 2) {
|
---|
371 | gsprintf($outhandle, "{plugin.no_plugin_could_process}\n", $file);
|
---|
372 | }
|
---|
373 | # tell the GLI that it was not processed
|
---|
374 | print STDERR "<NonProcessedFile n='$glifile'>\n" if $gli;
|
---|
375 |
|
---|
376 | gsprintf($failhandle, "$file: {plugin.no_plugin_could_process_this_file}\n");
|
---|
377 | $stats->{'num_not_processed'} ++;
|
---|
378 | } else {
|
---|
379 | # was not recognised
|
---|
380 | if ($verbosity >= 2) {
|
---|
381 | gsprintf($outhandle, "{plugin.no_plugin_could_recognise}\n",$file);
|
---|
382 | }
|
---|
383 | # tell the GLI that it was not processed
|
---|
384 | print STDERR "<NonRecognisedFile n='$glifile'>\n" if $gli;
|
---|
385 |
|
---|
386 | gsprintf($failhandle, "$file: {plugin.no_plugin_could_recognise_this_file}\n");
|
---|
387 | $stats->{'num_not_recognised'} ++;
|
---|
388 | }
|
---|
389 | return 0;
|
---|
390 | }
|
---|
391 |
|
---|
392 | # write out some general stats that the plugins have compiled - note that
|
---|
393 | # the buildcol.pl process doesn't currently call this process so the stats
|
---|
394 | # are only output after import.pl -
|
---|
395 | sub write_stats {
|
---|
396 | my ($pluginfo, $statshandle, $faillog, $gli) = @_;
|
---|
397 |
|
---|
398 | $gli = 0 unless defined $gli;
|
---|
399 |
|
---|
400 | foreach my $plugobj (@$pluginfo) {
|
---|
401 | $plugobj->compile_stats($stats);
|
---|
402 | }
|
---|
403 |
|
---|
404 | my $total = $stats->{'num_processed'} + $stats->{'num_blocked'} +
|
---|
405 | $stats->{'num_not_processed'} + $stats->{'num_not_recognised'};
|
---|
406 |
|
---|
407 | print STDERR "<ImportComplete considered='$total' processed='$stats->{'num_processed'}' blocked='$stats->{'num_blocked'}' ignored='$stats->{'num_not_recognised'}' failed='$stats->{'num_not_processed'}'>\n" if $gli;
|
---|
408 |
|
---|
409 | if ($total == 1) {
|
---|
410 | gsprintf($statshandle, "* {plugin.one_considered}\n");
|
---|
411 | } else {
|
---|
412 | gsprintf($statshandle, "* {plugin.n_considered}\n", $total);
|
---|
413 | }
|
---|
414 | if ($stats->{'num_archives'}) {
|
---|
415 | if ($stats->{'num_archives'} == 1) {
|
---|
416 | gsprintf($statshandle, " ({plugin.including_archive})\n");
|
---|
417 | }
|
---|
418 | else {
|
---|
419 | gsprintf($statshandle, " ({plugin.including_archives})\n",
|
---|
420 | $stats->{'num_archives'});
|
---|
421 | }
|
---|
422 | }
|
---|
423 | if ($stats->{'num_processed'} == 1) {
|
---|
424 | gsprintf($statshandle, "* {plugin.one_included}\n");
|
---|
425 | } else {
|
---|
426 | gsprintf($statshandle, "* {plugin.n_included}\n", $stats->{'num_processed'});
|
---|
427 | }
|
---|
428 | if ($stats->{'num_not_recognised'}) {
|
---|
429 | if ($stats->{'num_not_recognised'} == 1) {
|
---|
430 | gsprintf($statshandle, "* {plugin.one_unrecognised}\n");
|
---|
431 | } else {
|
---|
432 | gsprintf($statshandle, "* {plugin.n_unrecognised}\n",
|
---|
433 | $stats->{'num_not_recognised'});
|
---|
434 | }
|
---|
435 |
|
---|
436 | }
|
---|
437 | if ($stats->{'num_not_processed'}) {
|
---|
438 | if ($stats->{'num_not_processed'} == 1) {
|
---|
439 | gsprintf($statshandle, "* {plugin.one_rejected}\n");
|
---|
440 | } else {
|
---|
441 | gsprintf($statshandle, "* {plugin.n_rejected}\n",
|
---|
442 | $stats->{'num_not_processed'});
|
---|
443 | }
|
---|
444 | }
|
---|
445 | if ($stats->{'num_not_processed'} || $stats->{'num_not_recognised'}) {
|
---|
446 | gsprintf($statshandle, " {plugin.see_faillog}\n", $faillog);
|
---|
447 | }
|
---|
448 | }
|
---|
449 |
|
---|
450 | sub end {
|
---|
451 | my ($pluginfo, $processor) = @_;
|
---|
452 | map { $_->end($processor); } @$pluginfo;
|
---|
453 | }
|
---|
454 |
|
---|
455 | sub deinit {
|
---|
456 | my ($pluginfo, $processor) = @_;
|
---|
457 |
|
---|
458 |
|
---|
459 | map { $_->deinit($processor); } @$pluginfo;
|
---|
460 | }
|
---|
461 |
|
---|
462 | 1;
|
---|