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;
129 my ($delete, $reset, $help, $man, $processes);
130 my ($index_biblios, $index_authorities);
131 my (@biblionumbers,@authids);
134 $|=1; # flushes output
137 'c|commit=i' => \$commit,
138 'd|delete' => \$delete,
139 'r|reset' => \$reset,
140 'a|authorities' => \$index_authorities,
141 'b|biblios' => \$index_biblios,
143 'bn|bnumber=i' => \@biblionumbers,
144 'ai|authid=i' => \@authids,
145 'p|processes=i' => \$processes,
146 'v|verbose+' => \$verbose,
151 # Default is to do both
152 unless ($index_authorities || $index_biblios) {
153 $index_authorities = $index_biblios = 1;
156 if ($processes && ( @biblionumbers || @authids) ) {
157 die "Argument p|processes cannot be combined with bn|bnumber or ai|authid";
160 pod2usage(1) if $help;
161 pod2usage( -exitstatus => 0, -verbose => 2 ) if $man;
166 Koha::SearchEngine::Elasticsearch->reset_elasticsearch_mappings;
170 _verify_index_state($Koha::SearchEngine::Elasticsearch::BIBLIOS_INDEX, $delete) if ($index_biblios);
171 _verify_index_state($Koha::SearchEngine::Elasticsearch::AUTHORITIES_INDEX, $delete) if ($index_authorities);
174 my $slice_count = ( $processes //= 1 );
175 my %iterator_options;
177 if ($slice_count > 1) {
178 # Fire up child processes for processing slices from 2 on. This main process will handle slice 1.
180 for (my $proc = 1; $proc < $slice_count; $proc++) {
182 die "Failed to fork a child process\n" unless defined $pid;
184 # Child process, give it a slice to process
185 $slice_index = $proc;
189 # Fudge the commit count a bit to spread out the Elasticsearch commits
190 $commit *= 1 + 0.10 * $slice_index;
191 _log(1, "Processing slice @{[$slice_index + 1]} of $slice_count\n");
192 $iterator_options{slice} = { index => $slice_index, count => $slice_count };
196 $iterator_options{desc} = 1;
200 if ($index_biblios) {
201 _log(1, "Indexing biblios\n");
202 if (@biblionumbers) {
204 my $r = shift @biblionumbers;
205 return () unless defined $r;
206 return ($r, Koha::BiblioUtils->get_from_biblionumber($r, item_data => 1 ));
209 my $records = Koha::BiblioUtils->get_all_biblios_iterator(%iterator_options);
214 _do_reindex($next, $Koha::SearchEngine::Elasticsearch::BIBLIOS_INDEX);
216 if ($index_authorities) {
217 _log(1, "Indexing authorities\n");
220 my $r = shift @authids;
221 return () unless defined $r;
222 my $a = Koha::MetadataRecord::Authority->get_from_authid($r);
226 my $records = Koha::MetadataRecord::Authority->get_all_authorities_iterator(%iterator_options);
231 _do_reindex($next, $Koha::SearchEngine::Elasticsearch::AUTHORITIES_INDEX);
234 if ($slice_index == 0) {
235 # Main process, wait for children
236 for (my $proc = 1; $proc < $processes; $proc++) {
241 =head1 INTERNAL METHODS
243 =head2 _verify_index_state
245 _verify_index_state($Koha::SearchEngine::Elasticsearch::BIBLIOS_INDEX, 1);
247 Checks the index state and recreates it if requested.
251 sub _verify_index_state {
252 my ( $index_name, $recreate ) = @_;
254 _log(1, "Checking state of $index_name index\n");
255 my $indexer = Koha::SearchEngine::Elasticsearch::Indexer->new( { index => $index_name } );
258 _log(1, "Dropping and recreating $index_name index\n");
259 $indexer->drop_index() if $indexer->index_exists();
260 $indexer->create_index();
262 elsif (!$indexer->index_exists) {
263 # Create index if does not exist
264 $indexer->create_index();
265 } elsif ($indexer->is_index_status_ok) {
266 # Update mapping unless index is some kind of problematic state
267 $indexer->update_mappings();
268 } elsif ($indexer->is_index_status_recreate_required) {
269 warn qq/Index "$index_name" has status "recreate required", suggesting it should be recreated/;
275 _do_reindex($callback, $Koha::SearchEngine::Elasticsearch::BIBLIOS_INDEX);
277 Does the actual reindexing. $callback is a function that always returns the next record.
278 For each index we iterate through the records, committing at specified count
283 my ( $next, $index_name ) = @_;
285 my $indexer = Koha::SearchEngine::Elasticsearch::Indexer->new( { index => $index_name } );
288 my $commit_count = $commit;
289 my ( @id_buffer, @commit_buffer );
290 while ( my $record = $next->() ) {
291 my $id = $record->id // $record->authid;
292 my $record = $record->record;
294 if ( $verbose == 1 ) {
295 _log( 1, "$count records processed\n" ) if ( $count % 1000 == 0);
300 push @id_buffer, $id;
301 push @commit_buffer, $record;
302 if ( !( --$commit_count ) ) {
303 _log( 1, "Committing $commit records...\n" );
304 my $response = $indexer->update_index( \@id_buffer, \@commit_buffer );
305 _handle_response($response);
306 $commit_count = $commit;
309 _log( 1, "Commit complete\n" );
313 # There are probably uncommitted records
314 _log( 1, "Committing final records...\n" );
315 my $response = $indexer->update_index( \@id_buffer, \@commit_buffer );
316 _handle_response($response);
317 _log( 1, "Total $count records indexed\n" );
324 Checks some basic stuff to ensure that it's sane before we start.
329 # Do we have an elasticsearch block defined?
330 my $conf = C4::Context->config('elasticsearch');
331 die "No 'elasticsearch' block is defined in koha-conf.xml.\n" if ( !$conf );
334 =head2 _handle_response
336 Parse the return from update_index and display errors depending on verbosity of the script
340 sub _handle_response {
342 if( $response->{errors} eq 'true' ){
343 _log( 1, "There were errors during indexing\n" );
345 foreach my $item (@{$response->{items}}){
346 next unless defined $item->{index}->{error};
347 print "Record #" . $item->{index}->{_id} . " " .
348 $item->{index}->{error}->{reason} . " (" . $item->{index}->{error}->{type} . ") : " .
349 $item->{index}->{error}->{caused_by}->{type} . " (" . $item->{index}->{error}->{caused_by}->{reason} . ")\n";
357 _log($level, "Message\n");
359 Output progress information.
361 Will output the message if verbosity level is set to $level or more. Will not
362 include a trailing newline automatically.
367 my ($level, $msg) = @_;
369 print "[$$] $msg" if ($verbose >= $level);