3 # This inserts records from a Koha database into elastic search
5 # Copyright 2014 Catalyst IT
7 # This file is part of Koha.
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.
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.
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>.
24 rebuild_elasticsearch.pl - inserts records from a Koha database into Elasticsearch
28 B<rebuild_elasticsearch.pl>
29 [B<-c|--commit>=C<count>]
44 Inserts records from a Koha database into Elasticsearch.
50 =item B<-c|--commit>=C<count>
52 Specify how many records will be batched up before they're added to Elasticsearch.
53 Higher should be faster, but will cause more RAM usage. Default is 5000.
57 Delete the index and recreate it before indexing.
61 Reload mappings from files (specified in koha-conf.xml) before indexing.
64 =item B<-a|--authorities>
66 Index the authorities only. Combining this with B<-b> is the same as
67 specifying neither and so both get indexed.
71 Index the biblios only. Combining this with B<-a> is the same as
72 specifying neither and so both get indexed.
76 Index the records in descending id order. Intended to index newer record before older records.
77 Default is to index in ascending order.
78 Does not work with --bnumber or --authid
80 =item B<-bn|--bnumber>
82 Only index the supplied biblionumber, mostly for testing purposes. May be
87 Only index the supplied authority id, mostly for testing purposes. May be
90 =item B<-p|--processes>
92 Number of processes to use for indexing. This can be used to do more indexing
93 work in parallel on multicore systems. By default, a single process is used.
97 By default, this program only emits warnings and errors. This makes it talk
98 more. Add more to make it even more wordy, in particular when debugging.
110 =head1 IMPLEMENTATION
118 use Koha::MetadataRecord::Authority;
119 use Koha::BiblioUtils;
120 use Koha::SearchEngine::Elasticsearch;
121 use Koha::SearchEngine::Elasticsearch::Indexer;
130 my ($delete, $reset, $help, $man, $processes);
131 my ($index_biblios, $index_authorities);
132 my (@biblionumbers,@authids);
135 $|=1; # flushes output
138 'c|commit=i' => \$commit,
139 'd|delete' => \$delete,
140 'r|reset' => \$reset,
141 'a|authorities' => \$index_authorities,
142 'b|biblios' => \$index_biblios,
144 'bn|bnumber=i' => \@biblionumbers,
145 'ai|authid=i' => \@authids,
146 'p|processes=i' => \$processes,
147 'v|verbose+' => \$verbose,
152 # Default is to do both
153 unless ($index_authorities || $index_biblios) {
154 $index_authorities = $index_biblios = 1;
157 if ($processes && ( @biblionumbers || @authids) ) {
158 die "Argument p|processes cannot be combined with bn|bnumber or ai|authid";
161 pod2usage(1) if $help;
162 pod2usage( -exitstatus => 0, -verbose => 2 ) if $man;
167 Koha::SearchEngine::Elasticsearch->reset_elasticsearch_mappings;
171 _verify_index_state($Koha::SearchEngine::Elasticsearch::BIBLIOS_INDEX, $delete) if ($index_biblios);
172 _verify_index_state($Koha::SearchEngine::Elasticsearch::AUTHORITIES_INDEX, $delete) if ($index_authorities);
175 my $slice_count = ( $processes //= 1 );
176 my %iterator_options;
178 if ($slice_count > 1) {
179 # Fire up child processes for processing slices from 2 on. This main process will handle slice 1.
181 for (my $proc = 1; $proc < $slice_count; $proc++) {
183 die "Failed to fork a child process\n" unless defined $pid;
185 # Child process, give it a slice to process
186 $slice_index = $proc;
190 # Fudge the commit count a bit to spread out the Elasticsearch commits
191 $commit *= 1 + 0.10 * $slice_index;
192 $commit = int( $commit );
193 _log(1, "Processing slice @{[$slice_index + 1]} of $slice_count\n");
194 $iterator_options{slice} = { index => $slice_index, count => $slice_count };
198 $iterator_options{desc} = 1;
202 if ($index_biblios) {
203 _log(1, "Indexing biblios\n");
204 if (@biblionumbers) {
206 my $r = shift @biblionumbers;
207 return () unless defined $r;
208 return ($r, Koha::BiblioUtils->get_from_biblionumber($r, item_data => 1 ));
211 my $records = Koha::BiblioUtils->get_all_biblios_iterator(%iterator_options);
216 _do_reindex($next, $Koha::SearchEngine::Elasticsearch::BIBLIOS_INDEX);
218 if ($index_authorities) {
219 _log(1, "Indexing authorities\n");
222 my $r = shift @authids;
223 return () unless defined $r;
224 my $a = Koha::MetadataRecord::Authority->get_from_authid($r);
228 my $records = Koha::MetadataRecord::Authority->get_all_authorities_iterator(%iterator_options);
233 _do_reindex($next, $Koha::SearchEngine::Elasticsearch::AUTHORITIES_INDEX);
236 if ($slice_index == 0) {
237 # Main process, wait for children
238 for (my $proc = 1; $proc < $processes; $proc++) {
243 =head1 INTERNAL METHODS
245 =head2 _verify_index_state
247 _verify_index_state($Koha::SearchEngine::Elasticsearch::BIBLIOS_INDEX, 1);
249 Checks the index state and recreates it if requested.
253 sub _verify_index_state {
254 my ( $index_name, $recreate ) = @_;
256 _log(1, "Checking state of $index_name index\n");
257 my $indexer = Koha::SearchEngine::Elasticsearch::Indexer->new( { index => $index_name } );
260 _log(1, "Dropping and recreating $index_name index\n");
261 $indexer->drop_index() if $indexer->index_exists();
262 $indexer->create_index();
264 elsif (!$indexer->index_exists) {
265 # Create index if does not exist
266 $indexer->create_index();
267 } elsif ($indexer->is_index_status_ok) {
268 # Update mapping unless index is some kind of problematic state
269 $indexer->update_mappings();
270 } elsif ($indexer->is_index_status_recreate_required) {
271 warn qq/Index "$index_name" has status "recreate required", suggesting it should be recreated/;
277 _do_reindex($callback, $Koha::SearchEngine::Elasticsearch::BIBLIOS_INDEX);
279 Does the actual reindexing. $callback is a function that always returns the next record.
280 For each index we iterate through the records, committing at specified count
285 my ( $next, $index_name ) = @_;
287 my $indexer = Koha::SearchEngine::Elasticsearch::Indexer->new( { index => $index_name } );
290 my $commit_count = $commit;
291 my ( @id_buffer, @commit_buffer );
292 while ( my $record = $next->() ) {
293 my $id = $record->id // $record->authid;
294 my $record = $record->record;
296 if ( $verbose == 1 ) {
297 _log( 1, "$count records processed\n" ) if ( $count % 1000 == 0);
302 push @id_buffer, $id;
303 push @commit_buffer, $record;
304 if ( !( --$commit_count ) ) {
305 _log( 1, "Committing $commit records...\n" );
308 $response = $indexer->update_index( \@id_buffer, \@commit_buffer );
309 _handle_response($response);
310 _log( 1, "Commit complete\n" );
312 if( ref $_ eq 'Koha::Exceptions::Elasticsearch::BadResponse'){
314 _log(2,$_->{details});
317 $commit_count = $commit;
323 # There are probably uncommitted records
324 _log( 1, "Committing final records...\n" );
325 my $response = $indexer->update_index( \@id_buffer, \@commit_buffer );
326 _handle_response($response);
327 _log( 1, "Total $count records indexed\n" );
334 Checks some basic stuff to ensure that it's sane before we start.
339 # Do we have an elasticsearch block defined?
340 my $conf = C4::Context->config('elasticsearch');
341 die "No 'elasticsearch' block is defined in koha-conf.xml.\n" if ( !$conf );
344 =head2 _handle_response
346 Parse the return from update_index and display errors depending on verbosity of the script
350 sub _handle_response {
352 if( $response->{errors} eq 'true' ){
353 _log( 1, "There were errors during indexing\n" );
355 foreach my $item (@{$response->{items}}){
356 next unless defined $item->{index}->{error};
357 print "Record #" . $item->{index}->{_id} . " " .
358 $item->{index}->{error}->{reason} . " (" . $item->{index}->{error}->{type} . ") : " .
359 $item->{index}->{error}->{caused_by}->{type} . " (" . $item->{index}->{error}->{caused_by}->{reason} . ")\n";
367 _log($level, "Message\n");
369 Output progress information.
371 Will output the message if verbosity level is set to $level or more. Will not
372 include a trailing newline automatically.
377 my ($level, $msg) = @_;
379 print "[$$] $msg" if ($verbose >= $level);