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