1 package Koha::SearchEngine::Elasticsearch::Indexer;
3 # Copyright 2013 Catalyst IT
5 # This file is part of Koha.
7 # Koha is free software; you can redistribute it and/or modify it
8 # under the terms of the GNU General Public License as published by
9 # the Free Software Foundation; either version 3 of the License, or
10 # (at your option) any later version.
12 # Koha is distributed in the hope that it will be useful, but
13 # WITHOUT ANY WARRANTY; without even the implied warranty of
14 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
15 # GNU General Public License for more details.
17 # You should have received a copy of the GNU General Public License
18 # along with Koha; if not, see <http://www.gnu.org/licenses>.
20 use Carp qw( carp croak );
22 use Try::Tiny qw( catch try );
23 use List::Util qw( any );
24 use base qw(Koha::SearchEngine::Elasticsearch);
27 use Koha::Exceptions::Elasticsearch;
28 use Koha::SearchEngine::Zebra::Indexer;
29 use C4::AuthoritiesMarc qw//;
35 Koha::SearchEngine::Elasticsearch::Indexer - handles adding new records to the index
39 my $indexer = Koha::SearchEngine::Elasticsearch::Indexer->new(
40 { index => Koha::SearchEngine::BIBLIOS_INDEX } );
41 $indexer->drop_index();
42 $indexer->update_index(\@biblionumbers, \@records);
49 =item C<Koha::SearchEngine::Elasticsearch::Indexer::INDEX_STATUS_OK>
51 Represents an index state where index is created and in a working state.
53 =item C<Koha::SearchEngine::Elasticsearch::Indexer::INDEX_STATUS_REINDEX_REQUIRED>
55 Not currently used, but could be useful later, for example if can detect when new field or mapping added.
57 =item C<Koha::SearchEngine::Elasticsearch::Indexer::INDEX_STATUS_RECREATE_REQUIRED>
59 Representings an index state where index needs to be recreated and is not in a working state.
67 INDEX_STATUS_REINDEX_REQUIRED => 1,
68 INDEX_STATUS_RECREATE_REQUIRED => 2,
73 =head2 update_index($biblionums, $records)
76 $self->update_index($biblionums, $records);
78 die("Something went wrong trying to update index:" . $_[0]);
81 Converts C<MARC::Records> C<$records> to Elasticsearch documents and performs
82 an update request for these records on the Elasticsearch index.
88 Arrayref of biblio numbers for the C<$records>, the order must be the same as
89 and match up with C<$records>.
93 Arrayref of C<MARC::Record>s.
100 my ($self, $biblionums, $records) = @_;
102 my $documents = $self->marc_records_to_documents($records);
104 for (my $i = 0; $i < scalar @$biblionums; $i++) {
105 my $id = $biblionums->[$i];
106 my $document = $documents->[$i];
112 push @body, $document;
117 my $elasticsearch = $self->get_elasticsearch();
118 $response = $elasticsearch->bulk(
119 index => $self->index_name,
120 type => 'data', # is just hard coded in Indexer.pm?
123 if ($response->{errors}) {
124 carp "One or more ElasticSearch errors occurred when indexing documents";
127 Koha::Exceptions::Elasticsearch::BadResponse->throw(
129 details => $_->{text},
136 =head2 set_index_status_ok
138 Convenience method for setting index status to C<INDEX_STATUS_OK>.
142 sub set_index_status_ok {
144 $self->index_status(INDEX_STATUS_OK);
147 =head2 is_index_status_ok
149 Convenience method for checking if index status is C<INDEX_STATUS_OK>.
153 sub is_index_status_ok {
155 return $self->index_status == INDEX_STATUS_OK;
158 =head2 set_index_status_reindex_required
160 Convenience method for setting index status to C<INDEX_REINDEX_REQUIRED>.
164 sub set_index_status_reindex_required {
166 $self->index_status(INDEX_STATUS_REINDEX_REQUIRED);
169 =head2 is_index_status_reindex_required
171 Convenience method for checking if index status is C<INDEX_STATUS_REINDEX_REQUIRED>.
175 sub is_index_status_reindex_required {
177 return $self->index_status == INDEX_STATUS_REINDEX_REQUIRED;
180 =head2 set_index_status_recreate_required
182 Convenience method for setting index status to C<INDEX_STATUS_RECREATE_REQUIRED>.
186 sub set_index_status_recreate_required {
188 $self->index_status(INDEX_STATUS_RECREATE_REQUIRED);
191 =head2 is_index_status_recreate_required
193 Convenience method for checking if index status is C<INDEX_STATUS_RECREATE_REQUIRED>.
197 sub is_index_status_recreate_required {
199 return $self->index_status == INDEX_STATUS_RECREATE_REQUIRED;
202 =head2 index_status($status)
204 Will either set the current index status to C<$status> and return C<$status>,
205 or return the current index status if called with no arguments.
211 Optional argument. If passed will set current index status to C<$status> if C<$status> is
212 a valid status. See L</CONSTANTS>.
219 my ($self, $status) = @_;
220 my $key = 'ElasticsearchIndexStatus_' . $self->index;
222 if (defined $status) {
223 unless (any { $status == $_ } (
225 INDEX_STATUS_REINDEX_REQUIRED,
226 INDEX_STATUS_RECREATE_REQUIRED,
229 Koha::Exceptions::Exception->throw("Invalid index status: $status");
231 C4::Context->set_preference($key, $status);
235 return C4::Context->preference($key);
239 =head2 update_mappings
241 Generate Elasticsearch mappings from mappings stored in database and
242 perform a request to update Elasticsearch index mappings. Will throw an
243 error and set index status to C<INDEX_STATUS_RECREATE_REQUIRED> if update
248 sub update_mappings {
250 my $elasticsearch = $self->get_elasticsearch();
251 my $mappings = $self->get_elasticsearch_mappings();
253 foreach my $type (keys %{$mappings}) {
255 my $response = $elasticsearch->indices->put_mapping(
256 index => $self->index_name,
259 $type => $mappings->{$type}
263 $self->set_index_status_recreate_required();
264 my $reason = $_[0]->{vars}->{body}->{error}->{reason};
265 my $index_name = $self->index_name;
266 Koha::Exceptions::Exception->throw(
267 error => "Unable to update mappings for index \"$index_name\". Reason was: \"$reason\". Index needs to be recreated and reindexed",
271 $self->set_index_status_ok();
274 =head2 update_index_background($biblionums, $records)
276 This has exactly the same API as C<update_index> however it'll
277 return immediately. It'll start a background process that does the adding.
279 If it fails to add to Elasticsearch then it'll add to a queue that will cause
280 it to be updated by a regular index cron job in the future.
284 # TODO implement in the future - I don't know the best way of doing this yet.
285 # If fork: make sure process group is changed so apache doesn't wait for us.
287 sub update_index_background {
289 $self->update_index(@_);
294 This function takes an array of record numbers and fetches the records to send to update_index
297 If $records parameter is provided the records will be used as-is, this is only utilized for authorities
300 The other variables are used for parity with Zebra indexing calls. Currently the calls are passed through
306 my ( $self, $record_numbers, $op, $server, $records ) = @_;
307 $record_numbers = [$record_numbers] if ref $record_numbers ne 'ARRAY' && defined $record_numbers;
308 $records = [$records] if ref $records ne 'ARRAY' && defined $records;
309 if ( $op eq 'specialUpdate' ) {
310 my $index_record_numbers;
312 $index_record_numbers = $record_numbers;
314 foreach my $record_number ( @$record_numbers ){
315 my $record = _get_record( $record_number, $server );
317 push @$records, $record;
318 push @$index_record_numbers, $record_number;
322 $self->update_index_background( $index_record_numbers, $records ) if $index_record_numbers && $records;
324 elsif ( $op eq 'recordDelete' ) {
325 $self->delete_index_background( $record_numbers );
327 #FIXME Current behaviour is to index Zebra when using ES, at some point we should stop
328 Koha::SearchEngine::Zebra::Indexer::index_records( $self, $record_numbers, $op, $server, undef );
332 my ( $id, $server ) = @_;
333 return $server eq 'biblioserver'
334 ? C4::Biblio::GetMarcBiblio({ biblionumber => $id, embed_items => 1 })
335 : C4::AuthoritiesMarc::GetAuthority($id);
338 =head2 delete_index($biblionums)
340 C<$biblionums> is an arrayref of biblionumbers to delete from the index.
345 my ($self, $biblionums) = @_;
347 my $elasticsearch = $self->get_elasticsearch();
348 my @body = map { { delete => { _id => "$_" } } } @{$biblionums};
349 my $result = $elasticsearch->bulk(
350 index => $self->index_name,
354 if ($result->{errors}) {
355 croak "An Elasticsearch error occurred during bulk delete";
359 =head2 delete_index_background($biblionums)
361 Identical to L</delete_index($biblionums)>
365 # TODO: Should be made async
366 sub delete_index_background {
368 $self->delete_index(@_);
373 Drops the index from the Elasticsearch server.
379 if ($self->index_exists) {
380 my $elasticsearch = $self->get_elasticsearch();
381 $elasticsearch->indices->delete(index => $self->index_name);
382 $self->set_index_status_recreate_required();
388 Creates the index (including mappings) on the Elasticsearch server.
394 my $settings = $self->get_elasticsearch_settings();
395 my $elasticsearch = $self->get_elasticsearch();
396 $elasticsearch->indices->create(
397 index => $self->index_name,
399 settings => $settings
402 $self->update_mappings();
407 Checks if index has been created on the Elasticsearch server. Returns C<1> or the
408 empty string to indicate whether index exists or not.
414 my $elasticsearch = $self->get_elasticsearch();
415 return $elasticsearch->indices->exists(
416 index => $self->index_name,
428 =item Chris Cormack C<< <chrisc@catalyst.net.nz> >>
430 =item Robin Sheat C<< <robin@catalyst.net.nz> >>