misc/search_tools/rebuild_elasticsearch.pl

   1 #!/usr/bin/perl
   2
   3 # This inserts records from a Koha database into elastic search
   4
   5 # Copyright 2014 Catalyst IT
   6 #
   7 # This file is part of Koha.
   8 #
   9 # Koha is free software; you can redistribute it and/or modify it
  10 # under the terms of the GNU General Public License as published by
  11 # the Free Software Foundation; either version 3 of the License, or
  12 # (at your option) any later version.
  13 #
  14 # Koha is distributed in the hope that it will be useful, but
  15 # WITHOUT ANY WARRANTY; without even the implied warranty of
  16 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  17 # GNU General Public License for more details.
  18 #
  19 # You should have received a copy of the GNU General Public License
  20 # along with Koha; if not, see <http://www.gnu.org/licenses>.
  21
  22 =head1 NAME
  23
  24 rebuild_elasticsearch.pl - inserts records from a Koha database into Elasticsearch
  25
  26 =head1 SYNOPSIS
  27
  28 B<rebuild_elasticsearch.pl>
  29 [B<-c|--commit>=C<count>]
  30 [B<-d|--delete>]
  31 [B<-r|--reset>]
  32 [B<-a|--authorities>]
  33 [B<-b|--biblios>]
  34 [B<-bn|--bnumber>]
  35 [B<-ai|--authid>]
  36 [B<-p|--processes>]
  37 [B<-v|--verbose>]
  38 [B<-h|--help>]
  39 [B<--man>]
  40
  41 =head1 DESCRIPTION
  42
  43 Inserts records from a Koha database into Elasticsearch.
  44
  45 =head1 OPTIONS
  46
  47 =over
  48
  49 =item B<-c|--commit>=C<count>
  50
  51 Specify how many records will be batched up before they're added to Elasticsearch.
  52 Higher should be faster, but will cause more RAM usage. Default is 5000.
  53
  54 =item B<-d|--delete>
  55
  56 Delete the index and recreate it before indexing.
  57
  58 =item B<-r|--reset>
  59
  60 Reload mappings from files (specified in koha-conf.xml) before indexing.
  61 Implies --delete.
  62
  63 =item B<-a|--authorities>
  64
  65 Index the authorities only. Combining this with B<-b> is the same as
  66 specifying neither and so both get indexed.
  67
  68 =item B<-b|--biblios>
  69
  70 Index the biblios only. Combining this with B<-a> is the same as
  71 specifying neither and so both get indexed.
  72
  73 =item B<-bn|--bnumber>
  74
  75 Only index the supplied biblionumber, mostly for testing purposes. May be
  76 repeated.
  77
  78 =item B<-ai|--authid>
  79
  80 Only index the supplied authority id, mostly for testing purposes. May be
  81 repeated.
  82
  83 =item B<-p|--processes>
  84
  85 Number of processes to use for indexing. This can be used to do more indexing
  86 work in parallel on multicore systems. By default, a single process is used.
  87
  88 =item B<-v|--verbose>
  89
  90 By default, this program only emits warnings and errors. This makes it talk
  91 more. Add more to make it even more wordy, in particular when debugging.
  92
  93 =item B<-h|--help>
  94
  95 Help!
  96
  97 =item B<--man>
  98
  99 Full documentation.
 100
 101 =back
 102
 103 =head1 IMPLEMENTATION
 104
 105 =cut
 106
 107 use autodie;
 108 use Getopt::Long;
 109 use Koha::Script;
 110 use C4::Context;
 111 use Koha::MetadataRecord::Authority;
 112 use Koha::BiblioUtils;
 113 use Koha::SearchEngine::Elasticsearch;
 114 use Koha::SearchEngine::Elasticsearch::Indexer;
 115 use MARC::Field;
 116 use MARC::Record;
 117 use Modern::Perl;
 118 use Pod::Usage;
 119
 120 my $verbose = 0;
 121 my $commit = 5000;
 122 my ($delete, $reset, $help, $man, $processes);
 123 my ($index_biblios, $index_authorities);
 124 my (@biblionumbers,@authids);
 125
 126 $|=1; # flushes output
 127
 128 GetOptions(
 129     'c|commit=i'    => \$commit,
 130     'd|delete'      => \$delete,
 131     'r|reset'       => \$reset,
 132     'a|authorities' => \$index_authorities,
 133     'b|biblios'     => \$index_biblios,
 134     'bn|bnumber=i' => \@biblionumbers,
 135     'ai|authid=i'  => \@authids,
 136     'p|processes=i' => \$processes,
 137     'v|verbose+'    => \$verbose,
 138     'h|help'        => \$help,
 139     'man'           => \$man,
 140 );
 141
 142 # Default is to do both
 143 unless ($index_authorities || $index_biblios) {
 144     $index_authorities = $index_biblios = 1;
 145 }
 146
 147 if ($processes && ( @biblionumbers || @authids) ) {
 148     die "Argument p|processes cannot be combined with bn|bnumber or ai|authid";
 149 }
 150
 151 pod2usage(1) if $help;
 152 pod2usage( -exitstatus => 0, -verbose => 2 ) if $man;
 153
 154 _sanity_check();
 155
 156 if ($reset){
 157     Koha::SearchEngine::Elasticsearch->reset_elasticsearch_mappings;
 158     $delete = 1;
 159 }
 160
 161 _verify_index_state($Koha::SearchEngine::Elasticsearch::BIBLIOS_INDEX, $delete) if ($index_biblios);
 162 _verify_index_state($Koha::SearchEngine::Elasticsearch::AUTHORITIES_INDEX, $delete) if ($index_authorities);
 163
 164 my $slice_index = 0;
 165 my $slice_count = ( $processes //= 1 );
 166 my %iterator_options;
 167
 168 if ($slice_count > 1) {
 169     # Fire up child processes for processing slices from 2 on. This main process will handle slice 1.
 170     $slice_index = 0;
 171     for (my $proc = 1; $proc < $slice_count; $proc++) {
 172         my $pid = fork();
 173         die "Failed to fork a child process\n" unless defined $pid;
 174         if ($pid == 0) {
 175             # Child process, give it a slice to process
 176             $slice_index = $proc;
 177             last;
 178         }
 179     }
 180     # Fudge the commit count a bit to spread out the Elasticsearch commits
 181     $commit *= 1 + 0.10 * $slice_index;
 182     _log(1, "Processing slice @{[$slice_index + 1]} of $slice_count\n");
 183     $iterator_options{slice} = { index => $slice_index, count => $slice_count };
 184 }
 185
 186 my $next;
 187 if ($index_biblios) {
 188     _log(1, "Indexing biblios\n");
 189     if (@biblionumbers) {
 190         $next = sub {
 191             my $r = shift @biblionumbers;
 192             return () unless defined $r;
 193             return ($r, Koha::BiblioUtils->get_from_biblionumber($r, item_data => 1 ));
 194         };
 195     } else {
 196         my $records = Koha::BiblioUtils->get_all_biblios_iterator(%iterator_options);
 197         $next = sub {
 198             $records->next();
 199         }
 200     }
 201     _do_reindex($next, $Koha::SearchEngine::Elasticsearch::BIBLIOS_INDEX);
 202 }
 203 if ($index_authorities) {
 204     _log(1, "Indexing authorities\n");
 205     if (@authids) {
 206         $next = sub {
 207             my $r = shift @authids;
 208             return () unless defined $r;
 209             my $a = Koha::MetadataRecord::Authority->get_from_authid($r);
 210             return ($r, $a);
 211         };
 212     } else {
 213         my $records = Koha::MetadataRecord::Authority->get_all_authorities_iterator(%iterator_options);
 214         $next = sub {
 215             $records->next();
 216         }
 217     }
 218     _do_reindex($next, $Koha::SearchEngine::Elasticsearch::AUTHORITIES_INDEX);
 219 }
 220
 221 if ($slice_index == 0) {
 222     # Main process, wait for children
 223     for (my $proc = 1; $proc < $processes; $proc++) {
 224         wait();
 225     }
 226 }
 227
 228 =head1 INTERNAL METHODS
 229
 230 =head2 _verify_index_state
 231
 232     _verify_index_state($Koha::SearchEngine::Elasticsearch::BIBLIOS_INDEX, 1);
 233
 234 Checks the index state and recreates it if requested.
 235
 236 =cut
 237
 238 sub _verify_index_state {
 239     my ( $index_name, $recreate ) = @_;
 240
 241     _log(1, "Checking state of $index_name index\n");
 242     my $indexer = Koha::SearchEngine::Elasticsearch::Indexer->new( { index => $index_name } );
 243
 244     if ($recreate) {
 245         _log(1, "Dropping and recreating $index_name index\n");
 246         $indexer->drop_index() if $indexer->index_exists();
 247         $indexer->create_index();
 248     }
 249     elsif (!$indexer->index_exists) {
 250         # Create index if does not exist
 251         $indexer->create_index();
 252     } elsif ($indexer->is_index_status_ok) {
 253         # Update mapping unless index is some kind of problematic state
 254         $indexer->update_mappings();
 255     } elsif ($indexer->is_index_status_recreate_required) {
 256         warn qq/Index "$index_name" has status "recreate required", suggesting it should be recreated/;
 257     }
 258 }
 259
 260 =head2 _do_reindex
 261
 262     _do_reindex($callback, $Koha::SearchEngine::Elasticsearch::BIBLIOS_INDEX);
 263
 264 Does the actual reindexing. $callback is a function that always returns the next record.
 265 For each index we iterate through the records, committing at specified count
 266
 267 =cut
 268
 269 sub _do_reindex {
 270     my ( $next, $index_name ) = @_;
 271
 272     my $indexer = Koha::SearchEngine::Elasticsearch::Indexer->new( { index => $index_name } );
 273
 274     my $count        = 0;
 275     my $commit_count = $commit;
 276     my ( @id_buffer, @commit_buffer );
 277     while ( my $record = $next->() ) {
 278         my $id     = $record->id // $record->authid;
 279         my $record = $record->record;
 280         $count++;
 281         if ( $verbose == 1 ) {
 282             _log( 1, "$count records processed\n" ) if ( $count % 1000 == 0);
 283         } else {
 284             _log( 2, "$id\n" );
 285         }
 286
 287         push @id_buffer,     $id;
 288         push @commit_buffer, $record;
 289         if ( !( --$commit_count ) ) {
 290             _log( 1, "Committing $commit records...\n" );
 291             my $response = $indexer->update_index( \@id_buffer, \@commit_buffer );
 292             _handle_response($response);
 293             $commit_count  = $commit;
 294             @id_buffer     = ();
 295             @commit_buffer = ();
 296             _log( 1, "Commit complete\n" );
 297         }
 298     }
 299
 300     # There are probably uncommitted records
 301     _log( 1, "Committing final records...\n" );
 302     my $response = $indexer->update_index( \@id_buffer, \@commit_buffer );
 303     _handle_response($response);
 304     _log( 1, "Total $count records indexed\n" );
 305 }
 306
 307 =head2 _sanity_check
 308
 309     _sanity_check();
 310
 311 Checks some basic stuff to ensure that it's sane before we start.
 312
 313 =cut
 314
 315 sub _sanity_check {
 316     # Do we have an elasticsearch block defined?
 317     my $conf = C4::Context->config('elasticsearch');
 318     die "No 'elasticsearch' block is defined in koha-conf.xml.\n" if ( !$conf );
 319 }
 320
 321 =head2 _handle_response
 322
 323 Parse the return from update_index and display errors depending on verbosity of the script
 324
 325 =cut
 326
 327 sub _handle_response {
 328     my ($response) = @_;
 329     if( $response->{errors} eq 'true' ){
 330         _log( 1, "There were errors during indexing\n" );
 331         if ( $verbose > 1 ){
 332             foreach my $item (@{$response->{items}}){
 333                 next unless defined $item->{index}->{error};
 334                 print "Record #" . $item->{index}->{_id} . " " .
 335                       $item->{index}->{error}->{reason} . " (" . $item->{index}->{error}->{type} . ") : " .
 336                       $item->{index}->{error}->{caused_by}->{type} . " (" . $item->{index}->{error}->{caused_by}->{reason} . ")\n";
 337             }
 338         }
 339     }
 340 }
 341
 342 =head2 _log
 343
 344     _log($level, "Message\n");
 345
 346 Output progress information.
 347
 348 Will output the message if verbosity level is set to $level or more. Will not
 349 include a trailing newline automatically.
 350
 351 =cut
 352
 353 sub _log {
 354     my ($level, $msg) = @_;
 355
 356     print "[$$] $msg" if ($verbose >= $level);
 357 }