| 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 |
|
|---|
| 31 | require util;
|
|---|
| 32 | use gsprintf 'gsprintf';
|
|---|
| 33 |
|
|---|
| 34 | # global variables
|
|---|
| 35 | my $stats = {'num_processed' => 0,
|
|---|
| 36 | 'num_blocked' => 0,
|
|---|
| 37 | 'num_not_processed' => 0,
|
|---|
| 38 | 'num_not_recognised' => 0,
|
|---|
| 39 | 'num_archives' => 0
|
|---|
| 40 | };
|
|---|
| 41 |
|
|---|
| 42 | #globaloptions contains any options that should be passed to all plugins
|
|---|
| 43 | my ($verbosity, $outhandle, $failhandle, $globaloptions);
|
|---|
| 44 |
|
|---|
| 45 | sub load_plugins {
|
|---|
| 46 | my ($plugin_list) = shift @_;
|
|---|
| 47 | ($verbosity, $outhandle, $failhandle, $globaloptions) = @_; # globals
|
|---|
| 48 | my @plugin_objects = ();
|
|---|
| 49 |
|
|---|
| 50 | $verbosity = 2 unless defined $verbosity;
|
|---|
| 51 | $outhandle = 'STDERR' unless defined $outhandle;
|
|---|
| 52 | $failhandle = 'STDERR' unless defined $failhandle;
|
|---|
| 53 |
|
|---|
| 54 | map { $_ = "\"$_\""; } @$globaloptions;
|
|---|
| 55 | my $globals = join (",", @$globaloptions);
|
|---|
| 56 |
|
|---|
| 57 | foreach my $pluginoptions (@$plugin_list) {
|
|---|
| 58 | my $pluginname = shift @$pluginoptions;
|
|---|
| 59 | next unless defined $pluginname;
|
|---|
| 60 |
|
|---|
| 61 | # find the plugin
|
|---|
| 62 | my $colplugname = &util::filename_cat($ENV{'GSDLCOLLECTDIR'},"perllib/plugins",
|
|---|
| 63 | "${pluginname}.pm");
|
|---|
| 64 | my $mainplugname = &util::filename_cat($ENV{'GSDLHOME'},"perllib/plugins",
|
|---|
| 65 | "${pluginname}.pm");
|
|---|
| 66 | if (-e $colplugname) { require $colplugname; }
|
|---|
| 67 | elsif (-e $mainplugname) { require $mainplugname; }
|
|---|
| 68 | else {
|
|---|
| 69 | gsprintf($outhandle, "{plugin.could_not_find_plugin}\n",
|
|---|
| 70 | $pluginname);
|
|---|
| 71 | die "\n";
|
|---|
| 72 | }
|
|---|
| 73 |
|
|---|
| 74 | # create a plugin object
|
|---|
| 75 | my ($plugobj);
|
|---|
| 76 | map { $_ = "\"$_\""; } @$pluginoptions;
|
|---|
| 77 | my $options = join (",", @$pluginoptions);
|
|---|
| 78 | if ($globals) {
|
|---|
| 79 | if (@$pluginoptions) {
|
|---|
| 80 | $options .= ",";
|
|---|
| 81 | }
|
|---|
| 82 | $options .= "$globals";
|
|---|
| 83 | }
|
|---|
| 84 | $options =~ s/\$/\\\$/g;
|
|---|
| 85 |
|
|---|
| 86 | eval ("\$plugobj = new \$pluginname($options)");
|
|---|
| 87 | die "$@" if $@;
|
|---|
| 88 |
|
|---|
| 89 | # initialize plugin
|
|---|
| 90 | $plugobj->init($verbosity, $outhandle, $failhandle);
|
|---|
| 91 |
|
|---|
| 92 | # add this object to the list
|
|---|
| 93 | push (@plugin_objects, $plugobj);
|
|---|
| 94 | }
|
|---|
| 95 |
|
|---|
| 96 | return \@plugin_objects;
|
|---|
| 97 | }
|
|---|
| 98 |
|
|---|
| 99 |
|
|---|
| 100 | sub begin {
|
|---|
| 101 | my ($pluginfo, $base_dir, $processor, $maxdocs) = @_;
|
|---|
| 102 |
|
|---|
| 103 |
|
|---|
| 104 | map { $_->begin($pluginfo, $base_dir, $processor, $maxdocs); } @$pluginfo;
|
|---|
| 105 | }
|
|---|
| 106 |
|
|---|
| 107 | sub metadata_read {
|
|---|
| 108 | my ($pluginfo, $base_dir, $file, $metadata, $extrametakeys, $extrametadata, $processor, $maxdocs, $gli, $aux) = @_;
|
|---|
| 109 |
|
|---|
| 110 | $maxdocs = -1 unless defined $maxdocs && $maxdocs =~ /\d/;
|
|---|
| 111 | $gli = 0 unless defined $gli;
|
|---|
| 112 |
|
|---|
| 113 | my $rv = 0;
|
|---|
| 114 | my $glifile = $file;
|
|---|
| 115 |
|
|---|
| 116 | $glifile =~ s/^[\/\\]+//; # file sometimes starts with a / so get rid of it
|
|---|
| 117 |
|
|---|
| 118 | # Announce to GLI that we are handling a file
|
|---|
| 119 | print STDERR "<File n='$glifile'>\n" if $gli;
|
|---|
| 120 |
|
|---|
| 121 | # the .kill file is a handy (if not very elegant) way of aborting
|
|---|
| 122 | # an import.pl or buildcol.pl process
|
|---|
| 123 | if (-e &util::filename_cat ($ENV{'GSDLCOLLECTDIR'}, ".kill")) {
|
|---|
| 124 | gsprintf($outhandle, "{plugin.kill_file}\n");
|
|---|
| 125 | die "\n";
|
|---|
| 126 | }
|
|---|
| 127 |
|
|---|
| 128 | my $had_error = 0;
|
|---|
| 129 | # pass this file by each of the plugins in turn until one
|
|---|
| 130 | # is found which will process it
|
|---|
| 131 | # read must return:
|
|---|
| 132 | # undef - could not recognise
|
|---|
| 133 | # -1 - tried but error
|
|---|
| 134 | # 0 - blocked
|
|---|
| 135 | # anything else for successful processing
|
|---|
| 136 |
|
|---|
| 137 | foreach my $plugobj (@$pluginfo) {
|
|---|
| 138 |
|
|---|
| 139 | $rv = $plugobj->metadata_read($pluginfo, $base_dir, $file,
|
|---|
| 140 | $metadata, $extrametakeys, $extrametadata, $processor, $maxdocs, $gli, $aux);
|
|---|
| 141 |
|
|---|
| 142 | if (defined $rv) {
|
|---|
| 143 | if ($rv == -1) {
|
|---|
| 144 | # an error has occurred
|
|---|
| 145 | $had_error = 1;
|
|---|
| 146 | print STDERR "<ProcessingError n='$glifile'>\n" if $gli;
|
|---|
| 147 | } else {
|
|---|
| 148 | return $rv;
|
|---|
| 149 | }
|
|---|
| 150 | } # else undefined - was not recognised by the plugin
|
|---|
| 151 | }
|
|---|
| 152 |
|
|---|
| 153 | return 0;
|
|---|
| 154 | }
|
|---|
| 155 |
|
|---|
| 156 | sub read {
|
|---|
| 157 | my ($pluginfo, $base_dir, $file, $metadata, $processor, $maxdocs, $total_count, $gli, $aux) = @_;
|
|---|
| 158 |
|
|---|
| 159 | $maxdocs = -1 unless defined $maxdocs && $maxdocs =~ /\d/;
|
|---|
| 160 | $total_count = 0 unless defined $total_count && $total_count =~ /\d/;
|
|---|
| 161 | $gli = 0 unless defined $gli;
|
|---|
| 162 |
|
|---|
| 163 | my $rv = 0;
|
|---|
| 164 | my $glifile = $file;
|
|---|
| 165 |
|
|---|
| 166 | $glifile =~ s/^[\/\\]+//; # file sometimes starts with a / so get rid of it
|
|---|
| 167 |
|
|---|
| 168 | # Announce to GLI that we are handling a file
|
|---|
| 169 | print STDERR "<File n='$glifile'>\n" if $gli;
|
|---|
| 170 |
|
|---|
| 171 | # the .kill file is a handy (if not very elegant) way of aborting
|
|---|
| 172 | # an import.pl or buildcol.pl process
|
|---|
| 173 | if (-e &util::filename_cat ($ENV{'GSDLCOLLECTDIR'}, ".kill")) {
|
|---|
| 174 | gsprintf($outhandle, "{plugin.kill_file}\n");
|
|---|
| 175 | die "\n";
|
|---|
| 176 | }
|
|---|
| 177 |
|
|---|
| 178 | my $had_error = 0;
|
|---|
| 179 | # pass this file by each of the plugins in turn until one
|
|---|
| 180 | # is found which will process it
|
|---|
| 181 | # read must return:
|
|---|
| 182 | # undef - could not recognise
|
|---|
| 183 | # -1 - tried but error
|
|---|
| 184 | # 0 - blocked
|
|---|
| 185 | # anything else for successful processing
|
|---|
| 186 |
|
|---|
| 187 | foreach my $plugobj (@$pluginfo) {
|
|---|
| 188 |
|
|---|
| 189 | $rv = $plugobj->read($pluginfo, $base_dir, $file,
|
|---|
| 190 | $metadata, $processor, $maxdocs, $total_count, $gli, $aux);
|
|---|
| 191 |
|
|---|
| 192 | if (defined $rv) {
|
|---|
| 193 | if ($rv == -1) {
|
|---|
| 194 | # an error has occurred
|
|---|
| 195 | $had_error = 1;
|
|---|
| 196 | } else {
|
|---|
| 197 | return $rv;
|
|---|
| 198 | }
|
|---|
| 199 | } # else undefined - was not recognised by the plugin
|
|---|
| 200 | }
|
|---|
| 201 |
|
|---|
| 202 | if ($had_error) {
|
|---|
| 203 | # was recognised but couldn't be processed
|
|---|
| 204 | if ($verbosity >= 2) {
|
|---|
| 205 | gsprintf($outhandle, "{plugin.no_plugin_could_process}\n", $file);
|
|---|
| 206 | }
|
|---|
| 207 | # tell the GLI that it was not processed
|
|---|
| 208 | print STDERR "<NonProcessedFile n='$glifile'>\n" if $gli;
|
|---|
| 209 |
|
|---|
| 210 | $file =~ s/.*?([^\\\/]+)$/$1/;
|
|---|
| 211 | gsprintf($failhandle, "$file: {plugin.no_plugin_could_process_this_file}\n");
|
|---|
| 212 | $stats->{'num_not_processed'} ++;
|
|---|
| 213 | } else {
|
|---|
| 214 | # was not recognised
|
|---|
| 215 | if ($verbosity >= 2) {
|
|---|
| 216 | gsprintf($outhandle, "{plugin.no_plugin_could_recognise}\n",$file);
|
|---|
| 217 | }
|
|---|
| 218 | # tell the GLI that it was not processed
|
|---|
| 219 | print STDERR "<NonRecognisedFile n='$glifile'>\n" if $gli;
|
|---|
| 220 |
|
|---|
| 221 | $file =~ s/.*?([^\\\/]+)$/$1/;
|
|---|
| 222 | gsprintf($failhandle, "$file: {plugin.no_plugin_could_recognise_this_file}\n");
|
|---|
| 223 | $stats->{'num_not_recognised'} ++;
|
|---|
| 224 | }
|
|---|
| 225 | return 0;
|
|---|
| 226 | }
|
|---|
| 227 |
|
|---|
| 228 | # write out some general stats that the plugins have compiled - note that
|
|---|
| 229 | # the buildcol.pl process doesn't currently call this process so the stats
|
|---|
| 230 | # are only output after import.pl -
|
|---|
| 231 | sub write_stats {
|
|---|
| 232 | my ($pluginfo, $statshandle, $faillog, $gli) = @_;
|
|---|
| 233 |
|
|---|
| 234 | $gli = 0 unless defined $gli;
|
|---|
| 235 |
|
|---|
| 236 | foreach my $plugobj (@$pluginfo) {
|
|---|
| 237 | $plugobj->compile_stats($stats);
|
|---|
| 238 | }
|
|---|
| 239 |
|
|---|
| 240 | my $total = $stats->{'num_processed'} + $stats->{'num_blocked'} +
|
|---|
| 241 | $stats->{'num_not_processed'} + $stats->{'num_not_recognised'};
|
|---|
| 242 |
|
|---|
| 243 | print STDERR "<ImportComplete considered='$total' processed='$stats->{'num_processed'}' blocked='$stats->{'num_blocked'}' ignored='$stats->{'num_not_recognised'}' failed='$stats->{'num_not_processed'}'>\n" if $gli;
|
|---|
| 244 |
|
|---|
| 245 | if ($total == 1) {
|
|---|
| 246 | gsprintf($statshandle, "* {plugin.one_considered}\n");
|
|---|
| 247 | } else {
|
|---|
| 248 | gsprintf($statshandle, "* {plugin.n_considered}\n", $total);
|
|---|
| 249 | }
|
|---|
| 250 | if ($stats->{'num_archives'}) {
|
|---|
| 251 | if ($stats->{'num_archives'} == 1) {
|
|---|
| 252 | gsprintf($statshandle, " ({plugin.including_archive})\n");
|
|---|
| 253 | }
|
|---|
| 254 | else {
|
|---|
| 255 | gsprintf($statshandle, " ({plugin.including_archives})\n",
|
|---|
| 256 | $stats->{'num_archives'});
|
|---|
| 257 | }
|
|---|
| 258 | }
|
|---|
| 259 | if ($stats->{'num_processed'} == 1) {
|
|---|
| 260 | gsprintf($statshandle, "* {plugin.one_included}\n");
|
|---|
| 261 | } else {
|
|---|
| 262 | gsprintf($statshandle, "* {plugin.n_included}\n", $stats->{'num_processed'});
|
|---|
| 263 | }
|
|---|
| 264 | if ($stats->{'num_not_recognised'}) {
|
|---|
| 265 | if ($stats->{'num_not_recognised'} == 1) {
|
|---|
| 266 | gsprintf($statshandle, "* {plugin.one_unrecognised}\n");
|
|---|
| 267 | } else {
|
|---|
| 268 | gsprintf($statshandle, "* {plugin.n_unrecognised}\n",
|
|---|
| 269 | $stats->{'num_not_recognised'});
|
|---|
| 270 | }
|
|---|
| 271 |
|
|---|
| 272 | }
|
|---|
| 273 | if ($stats->{'num_not_processed'}) {
|
|---|
| 274 | if ($stats->{'num_not_processed'} == 1) {
|
|---|
| 275 | gsprintf($statshandle, "* {plugin.one_rejected}\n");
|
|---|
| 276 | } else {
|
|---|
| 277 | gsprintf($statshandle, "* {plugin.n_rejected}\n",
|
|---|
| 278 | $stats->{'num_not_processed'});
|
|---|
| 279 | }
|
|---|
| 280 | }
|
|---|
| 281 | if ($stats->{'num_not_processed'} || $stats->{'num_not_recognised'}) {
|
|---|
| 282 | gsprintf($statshandle, " {plugin.see_faillog}\n", $faillog);
|
|---|
| 283 | }
|
|---|
| 284 | }
|
|---|
| 285 |
|
|---|
| 286 | sub end {
|
|---|
| 287 | my ($pluginfo, $processor) = @_;
|
|---|
| 288 | map { $_->end($processor); } @$pluginfo;
|
|---|
| 289 | }
|
|---|
| 290 |
|
|---|
| 291 | 1;
|
|---|
