source: main/trunk/greenstone2/perllib/plugin.pm@ 22037

Last change on this file since 22037 was 22037, checked in by davidb, 14 years ago

Manifest file processing upgraded to support OIDs. The code then uses the archiveinf-doc database to look up which files it uses.
  • Property svn:keywords set to Author Date Id Revision
File size: 13.7 KB
Line 
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
26package plugin;
27
28
29use strict; # to pick up typos and undeclared variables...
30no strict 'refs'; # ...but allow filehandles to be variables and vice versa
31no strict 'subs';
32
33require util;
34use gsprintf 'gsprintf';
35
36# mapping from old plugin names to new ones for backwards compatibility
37# can remove at sometime in future when we no longer want to support old xxPlug names in the config file
38my $plugin_name_map = {
39 'GAPlug' => 'GreenstoneXMLPlugin',
40 'ArcPlug' => 'ArchivesInfPlugin',
41 'RecPlug' => 'DirectoryPlugin',
42 'TEXTPlug' => 'TextPlugin',
43 'XMLPlug' => 'ReadXMLFile',
44 'EMAILPlug' => 'EmailPlugin',
45 'SRCPlug' => 'SourceCodePlugin',
46 'NULPlug' => 'NulPlugin',
47 'W3ImgPlug' => 'HTMLImagePlugin',
48 'PagedImgPlug' => 'PagedImagePlugin',
49 'METSPlug' => 'GreenstoneMETSPlugin',
50 'PPTPlug' => 'PowerPointPlugin',
51 'PSPlug' => 'PostScriptPlugin',
52 'DBPlug' => 'DatabasePlugin'
53 };
54
55# global variables
56my $stats = {'num_processed' => 0,
57 'num_blocked' => 0,
58 'num_not_processed' => 0,
59 'num_not_recognised' => 0,
60 'num_archives' => 0
61 };
62
63#globaloptions contains any options that should be passed to all plugins
64my ($verbosity, $outhandle, $failhandle, $globaloptions);
65
66sub get_valid_pluginname {
67 my ($pluginname) = @_;
68 my $valid_name = $pluginname;
69 if (defined $plugin_name_map->{$pluginname}) {
70 $valid_name = $plugin_name_map->{$pluginname};
71 } elsif ($pluginname =~ /Plug$/) {
72 $valid_name =~ s/Plug/Plugin/;
73
74 }
75 return $valid_name;
76}
77
78sub load_plugin_require
79{
80 my ($pluginname) = @_;
81
82 my @check_list = ();
83
84 # pp_plugname shorthand for 'perllib' 'plugin' '$pluginname.pm'
85 my $pp_plugname
86 = &util::filename_cat('perllib', 'plugins', "${pluginname}.pm");
87 my $collectdir = $ENV{'GSDLCOLLECTDIR'};
88
89 # find the plugin
90 if (defined($ENV{'GSDLCOLLECTION'}))
91 {
92 my $customplugname
93 = &util::filename_cat($collectdir, "custom",$ENV{'GSDLCOLLECTION'},
94 $pp_plugname);
95 push(@check_list,$customplugname);
96 }
97
98 my $colplugname = &util::filename_cat($collectdir, $pp_plugname);
99 push(@check_list,$colplugname);
100
101 if (defined $ENV{'GSDLEXTS'}) {
102
103 my $ext_prefix = &util::filename_cat($ENV{'GSDLHOME'}, "ext");
104
105 my @extensions = split(/:/,$ENV{'GSDLEXTS'});
106 foreach my $e (@extensions) {
107 my $extplugname = &util::filename_cat($ext_prefix, $e, $pp_plugname);
108 push(@check_list,$extplugname);
109
110 }
111 }
112 if (defined $ENV{'GSDL3EXTS'}) {
113
114 my $ext_prefix = &util::filename_cat($ENV{'GSDL3SRCHOME'}, "ext");
115
116 my @extensions = split(/:/,$ENV{'GSDL3EXTS'});
117 foreach my $e (@extensions) {
118 my $extplugname = &util::filename_cat($ext_prefix, $e, $pp_plugname);
119 push(@check_list,$extplugname);
120
121 }
122 }
123
124
125 my $mainplugname = &util::filename_cat($ENV{'GSDLHOME'}, $pp_plugname);
126 push(@check_list,$mainplugname);
127
128 my $success=0;
129 foreach my $plugname (@check_list) {
130 if (-e $plugname) {
131 require $plugname;
132 $success=1;
133 last;
134 }
135 }
136
137 if (!$success) {
138 &gsprintf(STDERR, "{plugin.could_not_find_plugin}\n",
139 $pluginname);
140 die "\n";
141 }
142}
143
144sub load_plugin_for_info {
145 my ($pluginname) = shift @_;
146 $pluginname = &get_valid_pluginname($pluginname);
147 load_plugin_require($pluginname);
148
149 # create a plugin object
150 my ($plugobj);
151 my $options = "-gsdlinfo";
152
153 eval ("\$plugobj = new \$pluginname([],[$options])");
154 die "$@" if $@;
155
156 return $plugobj;
157}
158
159sub load_plugins {
160 my ($plugin_list) = shift @_;
161 my $incremental_mode;
162 ($verbosity, $outhandle, $failhandle, $globaloptions, $incremental_mode) = @_; # globals
163 my @plugin_objects = ();
164 $verbosity = 2 unless defined $verbosity;
165 $outhandle = 'STDERR' unless defined $outhandle;
166 $failhandle = 'STDERR' unless defined $failhandle;
167
168 my $colplugindir = &util::filename_cat($ENV{'GSDLCOLLECTDIR'},"perllib/plugins");
169 unshift (@INC, $colplugindir);
170
171 map { $_ = "\"$_\""; } @$globaloptions;
172 my $globals = join (",", @$globaloptions);
173
174 foreach my $pluginoptions (@$plugin_list) {
175 my $pluginname = shift @$pluginoptions;
176 next unless defined $pluginname;
177 $pluginname = &get_valid_pluginname($pluginname);
178 load_plugin_require($pluginname);
179
180 # create a plugin object
181 my ($plugobj);
182 map { $_ = "\"$_\""; } @$pluginoptions;
183 my $options = join (",", @$pluginoptions);
184 if ($globals) {
185 if (@$pluginoptions) {
186 $options .= ",";
187 }
188 $options .= "$globals";
189 }
190 # need to escape backslash before putting in to the eval
191 $options =~ s/\\/\\\\/g;
192 $options =~ s/\$/\\\$/g;
193 eval ("\$plugobj = new \$pluginname([],[$options])");
194 die "$@" if $@;
195
196 # initialize plugin
197 $plugobj->init($verbosity, $outhandle, $failhandle);
198
199 $plugobj->set_incremental($incremental_mode);
200
201 # add this object to the list
202 push (@plugin_objects, $plugobj);
203 }
204
205 return \@plugin_objects;
206}
207
208
209sub begin {
210 my ($pluginfo, $base_dir, $processor, $maxdocs, $gli) = @_;
211
212 map { $_->{'gli'} = $gli; } @$pluginfo;
213 map { $_->begin($pluginfo, $base_dir, $processor, $maxdocs); } @$pluginfo;
214}
215
216 sub remove_all {
217 my ($pluginfo, $base_dir, $processor, $maxdocs) = @_;
218
219 map { $_->remove_all($pluginfo, $base_dir, $processor, $maxdocs); } @$pluginfo;
220}
221
222sub remove_some {
223 my ($pluginfo, $infodbtype, $archivedir, $deleted_files) = @_;
224 return if (scalar(@$deleted_files)==0);
225 my $arcinfo_src_filename = &dbutil::get_infodb_file_path($infodbtype, "archiveinf-src", $archivedir);
226
227 foreach my $file (@$deleted_files) {
228 # use 'archiveinf-src' info database to look up all the OIDs
229 # that this file is used in (note in most cases, it's just one OID)
230
231 my $src_rec_string = &dbutil::read_infodb_entry($infodbtype, $arcinfo_src_filename, $file);
232 my $src_rec = &dbutil::convert_infodb_string_to_hash($src_rec_string);
233 my $oids = $src_rec->{'oid'};
234 my $rv;
235 foreach my $plugobj (@$pluginfo) {
236
237 $rv = $plugobj->remove_one($file, $oids, $archivedir);
238 if (defined $rv && $rv != -1) {
239 return $rv;
240 } # else undefined (was not recognised by the plugin) or there was an error, try the next one
241 }
242 return 0;
243 }
244
245}
246sub file_block_read {
247 my ($pluginfo, $base_dir, $file, $block_hash, $metadata, $gli) = @_;
248
249
250 $gli = 0 unless defined $gli;
251
252 my $rv = 0;
253 my $glifile = $file;
254
255 $glifile =~ s/^[\/\\]+//; # file sometimes starts with a / so get rid of it
256
257 # Announce to GLI that we are handling a file
258 print STDERR "<File n='$glifile'>\n" if $gli;
259
260 # the .kill file is a handy (if not very elegant) way of aborting
261 # an import.pl or buildcol.pl process
262 if (-e &util::filename_cat ($ENV{'GSDLCOLLECTDIR'}, ".kill")) {
263 gsprintf($outhandle, "{plugin.kill_file}\n");
264 die "\n";
265 }
266
267 foreach my $plugobj (@$pluginfo) {
268
269 $rv = $plugobj->file_block_read($pluginfo, $base_dir, $file, $block_hash, $metadata, $gli);
270 #last if (defined $rv && $rv==1); # stop this file once we have found something to 'process' it
271 }
272
273}
274
275
276sub metadata_read {
277 my ($pluginfo, $base_dir, $file, $block_hash,
278 $extrametakeys, $extrametadata, $extrametafile,
279 $processor, $maxdocs, $gli, $aux) = @_;
280
281 $maxdocs = -1 unless defined $maxdocs && $maxdocs =~ /\d/;
282 $gli = 0 unless defined $gli;
283
284 my $rv = 0;
285 my $glifile = $file;
286
287 $glifile =~ s/^[\/\\]+//; # file sometimes starts with a / so get rid of it
288
289 # Announce to GLI that we are handling a file
290 print STDERR "<File n='$glifile'>\n" if $gli;
291
292 # the .kill file is a handy (if not very elegant) way of aborting
293 # an import.pl or buildcol.pl process
294 if (-e &util::filename_cat ($ENV{'GSDLCOLLECTDIR'}, ".kill")) {
295 gsprintf($outhandle, "{plugin.kill_file}\n");
296 die "\n";
297 }
298
299 my $had_error = 0;
300 # pass this file by each of the plugins in turn until one
301 # is found which will process it
302 # read must return:
303 # undef - could not recognise
304 # -1 - tried but error
305 # 0 - blocked
306 # anything else for successful processing
307
308 foreach my $plugobj (@$pluginfo) {
309
310 $rv = $plugobj->metadata_read($pluginfo, $base_dir, $file, $block_hash,
311 $extrametakeys, $extrametadata, $extrametafile,
312 $processor, $maxdocs, $gli, $aux);
313
314 if (defined $rv) {
315 if ($rv == -1) {
316 # an error has occurred
317 $had_error = 1;
318 print STDERR "<ProcessingError n='$glifile'>\n" if $gli;
319 } else {
320 return $rv;
321 }
322 } # else undefined - was not recognised by the plugin
323 }
324
325 return 0;
326}
327
328sub read {
329 my ($pluginfo, $base_dir, $file, $block_hash, $metadata, $processor, $maxdocs, $total_count, $gli, $aux) = @_;
330
331 $maxdocs = -1 unless defined $maxdocs && $maxdocs =~ /\d/;
332 $total_count = 0 unless defined $total_count && $total_count =~ /\d/;
333 $gli = 0 unless defined $gli;
334
335 my $rv = 0;
336 my $glifile = $file;
337
338 $glifile =~ s/^[\/\\]+//; # file sometimes starts with a / so get rid of it
339
340 # Announce to GLI that we are handling a file
341 print STDERR "<File n='$glifile'>\n" if $gli;
342
343 # the .kill file is a handy (if not very elegant) way of aborting
344 # an import.pl or buildcol.pl process
345 if (-e &util::filename_cat ($ENV{'GSDLCOLLECTDIR'}, ".kill")) {
346 gsprintf($outhandle, "{plugin.kill_file}\n");
347 die "\n";
348 }
349
350 my $had_error = 0;
351 # pass this file by each of the plugins in turn until one
352 # is found which will process it
353 # read must return:
354 # undef - could not recognise
355 # -1 - tried but error
356 # 0 - blocked
357 # anything else for successful processing
358
359 foreach my $plugobj (@$pluginfo) {
360
361 $rv = $plugobj->read($pluginfo, $base_dir, $file,
362 $block_hash, $metadata, $processor, $maxdocs,
363 $total_count, $gli, $aux);
364
365 if (defined $rv) {
366 if ($rv == -1) {
367 # an error has occurred
368 $had_error = 1;
369 } else {
370 return $rv;
371 }
372 } # else undefined - was not recognised by the plugin
373 }
374
375 if ($had_error) {
376 # was recognised but couldn't be processed
377 if ($verbosity >= 2) {
378 gsprintf($outhandle, "{plugin.no_plugin_could_process}\n", $file);
379 }
380 # tell the GLI that it was not processed
381 print STDERR "<NonProcessedFile n='$glifile'>\n" if $gli;
382
383 gsprintf($failhandle, "$file: {plugin.no_plugin_could_process_this_file}\n");
384 $stats->{'num_not_processed'} ++;
385 } else {
386 # was not recognised
387 if ($verbosity >= 2) {
388 gsprintf($outhandle, "{plugin.no_plugin_could_recognise}\n",$file);
389 }
390 # tell the GLI that it was not processed
391 print STDERR "<NonRecognisedFile n='$glifile'>\n" if $gli;
392
393 gsprintf($failhandle, "$file: {plugin.no_plugin_could_recognise_this_file}\n");
394 $stats->{'num_not_recognised'} ++;
395 }
396 return 0;
397}
398
399# write out some general stats that the plugins have compiled - note that
400# the buildcol.pl process doesn't currently call this process so the stats
401# are only output after import.pl -
402sub write_stats {
403 my ($pluginfo, $statshandle, $faillog, $gli) = @_;
404
405 $gli = 0 unless defined $gli;
406
407 foreach my $plugobj (@$pluginfo) {
408 $plugobj->compile_stats($stats);
409 }
410
411 my $total = $stats->{'num_processed'} + $stats->{'num_blocked'} +
412 $stats->{'num_not_processed'} + $stats->{'num_not_recognised'};
413
414 print STDERR "<ImportComplete considered='$total' processed='$stats->{'num_processed'}' blocked='$stats->{'num_blocked'}' ignored='$stats->{'num_not_recognised'}' failed='$stats->{'num_not_processed'}'>\n" if $gli;
415
416 if ($total == 1) {
417 gsprintf($statshandle, "* {plugin.one_considered}\n");
418 } else {
419 gsprintf($statshandle, "* {plugin.n_considered}\n", $total);
420 }
421 if ($stats->{'num_archives'}) {
422 if ($stats->{'num_archives'} == 1) {
423 gsprintf($statshandle, " ({plugin.including_archive})\n");
424 }
425 else {
426 gsprintf($statshandle, " ({plugin.including_archives})\n",
427 $stats->{'num_archives'});
428 }
429 }
430 if ($stats->{'num_processed'} == 1) {
431 gsprintf($statshandle, "* {plugin.one_included}\n");
432 } else {
433 gsprintf($statshandle, "* {plugin.n_included}\n", $stats->{'num_processed'});
434 }
435 if ($stats->{'num_not_recognised'}) {
436 if ($stats->{'num_not_recognised'} == 1) {
437 gsprintf($statshandle, "* {plugin.one_unrecognised}\n");
438 } else {
439 gsprintf($statshandle, "* {plugin.n_unrecognised}\n",
440 $stats->{'num_not_recognised'});
441 }
442
443 }
444 if ($stats->{'num_not_processed'}) {
445 if ($stats->{'num_not_processed'} == 1) {
446 gsprintf($statshandle, "* {plugin.one_rejected}\n");
447 } else {
448 gsprintf($statshandle, "* {plugin.n_rejected}\n",
449 $stats->{'num_not_processed'});
450 }
451 }
452 if ($stats->{'num_not_processed'} || $stats->{'num_not_recognised'}) {
453 gsprintf($statshandle, " {plugin.see_faillog}\n", $faillog);
454 }
455}
456
457sub end {
458 my ($pluginfo, $processor) = @_;
459 map { $_->end($processor); } @$pluginfo;
460}
461
462sub deinit {
463 my ($pluginfo, $processor) = @_;
464
465
466 map { $_->deinit($processor); } @$pluginfo;
467}
468
4691;
Note: See TracBrowser for help on using the repository browser.