1 package Koha::SearchEngine::Elasticsearch;
3 # Copyright 2015 Catalyst IT
5 # This file is part of Koha.
7 # Koha is free software; you can redistribute it and/or modify it under the
8 # terms of the GNU General Public License as published by the Free Software
9 # Foundation; either version 3 of the License, or (at your option) any later
12 # Koha is distributed in the hope that it will be useful, but WITHOUT ANY
13 # WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR
14 # A PARTICULAR PURPOSE. See the GNU General Public License for more details.
16 # You should have received a copy of the GNU General Public License along
17 # with Koha; if not, write to the Free Software Foundation, Inc.,
18 # 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
20 use base qw(Class::Accessor);
25 use Koha::Exceptions::Config;
26 use Koha::SearchFields;
27 use Koha::SearchMarcMaps;
33 use Search::Elasticsearch;
37 use List::Util qw( sum0 reduce );
40 use Encode qw(encode);
43 __PACKAGE__->mk_ro_accessors(qw( index ));
44 __PACKAGE__->mk_accessors(qw( sort_fields ));
46 # Constants to refer to the standard index names
47 Readonly our $BIBLIOS_INDEX => 'biblios';
48 Readonly our $AUTHORITIES_INDEX => 'authorities';
52 Koha::SearchEngine::Elasticsearch - Base module for things using elasticsearch
60 The name of the index to use, generally 'biblios' or 'authorities'.
70 my $self = $class->SUPER::new(@_);
71 # Check for a valid index
72 Koha::Exceptions::MissingParameter->throw('No index name provided') unless $self->index;
76 =head2 get_elasticsearch
78 my $elasticsearch_client = $self->get_elasticsearch();
80 Returns a C<Search::Elasticsearch> client. The client is cached on a C<Koha::SearchEngine::ElasticSearch>
81 instance level and will be reused if method is called multiple times.
85 sub get_elasticsearch {
87 unless (defined $self->{elasticsearch}) {
88 my $conf = $self->get_elasticsearch_params();
89 $self->{elasticsearch} = Search::Elasticsearch->new($conf);
91 return $self->{elasticsearch};
94 =head2 get_elasticsearch_params
96 my $params = $self->get_elasticsearch_params();
98 This provides a hashref that contains the parameters for connecting to the
99 ElasicSearch servers, in the form:
102 'nodes' => ['127.0.0.1:9200', 'anotherserver:9200'],
103 'index_name' => 'koha_instance_index',
106 This is configured by the following in the C<config> block in koha-conf.xml:
109 <server>127.0.0.1:9200</server>
110 <server>anotherserver:9200</server>
111 <index_name>koha_instance</index_name>
116 sub get_elasticsearch_params {
119 # Copy the hash so that we're not modifying the original
120 my $conf = C4::Context->config('elasticsearch');
121 die "No 'elasticsearch' block is defined in koha-conf.xml.\n" if ( !$conf );
122 my $es = { %{ $conf } };
124 # Helpfully, the multiple server lines end up in an array for us anyway
125 # if there are multiple ones, but not if there's only one.
126 my $server = $es->{server};
127 delete $es->{server};
128 if ( ref($server) eq 'ARRAY' ) {
130 # store it called 'nodes' (which is used by newer Search::Elasticsearch)
131 $es->{nodes} = $server;
134 $es->{nodes} = [$server];
137 die "No elasticsearch servers were specified in koha-conf.xml.\n";
139 die "No elasticsearch index_name was specified in koha-conf.xml.\n"
140 if ( !$es->{index_name} );
141 # Append the name of this particular index to our namespace
142 $es->{index_name} .= '_' . $self->index;
144 $es->{key_prefix} = 'es_';
145 $es->{client} //= '5_0::Direct';
146 $es->{cxn_pool} //= 'Sniff';
147 $es->{request_timeout} //= 60;
152 =head2 get_elasticsearch_settings
154 my $settings = $self->get_elasticsearch_settings();
156 This provides the settings provided to Elasticsearch when an index is created.
157 These can do things like define tokenization methods.
159 A hashref containing the settings is returned.
163 sub get_elasticsearch_settings {
166 # Use state to speed up repeated calls
167 state $settings = undef;
168 if (!defined $settings) {
169 my $config_file = C4::Context->config('elasticsearch_index_config');
170 $config_file ||= C4::Context->config('intranetdir') . '/admin/searchengine/elasticsearch/index_config.yaml';
171 $settings = LoadFile( $config_file );
177 =head2 get_elasticsearch_mappings
179 my $mappings = $self->get_elasticsearch_mappings();
181 This provides the mappings that get passed to Elasticsearch when an index is
186 sub get_elasticsearch_mappings {
189 # Use state to speed up repeated calls
193 if (!defined $all_mappings{$self->index}) {
194 $sort_fields{$self->index} = {};
196 data => scalar _get_elasticsearch_mapping('general', '')
198 my $marcflavour = lc C4::Context->preference('marcflavour');
199 $self->_foreach_mapping(
201 my ( $name, $type, $facet, $suggestible, $sort, $marc_type ) = @_;
202 return if $marc_type ne $marcflavour;
203 # TODO if this gets any sort of complexity to it, it should
204 # be broken out into its own function.
206 # TODO be aware of date formats, but this requires pre-parsing
207 # as ES will simply reject anything with an invalid date.
208 my $es_type = 'text';
209 if ($type eq 'boolean') {
210 $es_type = 'boolean';
211 } elsif ($type eq 'number' || $type eq 'sum') {
212 $es_type = 'integer';
213 } elsif ($type eq 'isbn' || $type eq 'stdno') {
217 $mappings->{data}{properties}{$name} = _get_elasticsearch_mapping('search', $es_type);
220 $mappings->{data}{properties}{ $name . '__facet' } = _get_elasticsearch_mapping('facet', $es_type);
223 $mappings->{data}{properties}{ $name . '__suggestion' } = _get_elasticsearch_mapping('suggestible', $es_type);
225 # Sort is a bit special as it can be true, false, undef.
226 # We care about "true" or "undef",
227 # "undef" means to do the default thing, which is make it sortable.
228 if (!defined $sort || $sort) {
229 $mappings->{data}{properties}{ $name . '__sort' } = _get_elasticsearch_mapping('sort', $es_type);
230 $sort_fields{$self->index}{$name} = 1;
234 $all_mappings{$self->index} = $mappings;
236 $self->sort_fields(\%{$sort_fields{$self->index}});
238 return $all_mappings{$self->index};
241 =head2 _get_elasticsearch_mapping
243 Get the Elasticsearch mappings for the given purpose and data type.
245 $mapping = _get_elasticsearch_mapping('search', 'text');
249 sub _get_elasticsearch_mapping {
251 my ( $purpose, $type ) = @_;
253 # Use state to speed up repeated calls
254 state $settings = undef;
255 if (!defined $settings) {
256 my $config_file = C4::Context->config('elasticsearch_field_config');
257 $config_file ||= C4::Context->config('intranetdir') . '/admin/searchengine/elasticsearch/field_config.yaml';
258 $settings = LoadFile( $config_file );
261 if (!defined $settings->{$purpose}) {
262 die "Field purpose $purpose not defined in field config";
265 return $settings->{$purpose};
267 if (defined $settings->{$purpose}{$type}) {
268 return $settings->{$purpose}{$type};
270 if (defined $settings->{$purpose}{'default'}) {
271 return $settings->{$purpose}{'default'};
276 sub reset_elasticsearch_mappings {
277 my ( $reset_fields ) = @_;
278 my $mappings_yaml = C4::Context->config('elasticsearch_index_mappings');
279 $mappings_yaml ||= C4::Context->config('intranetdir') . '/admin/searchengine/elasticsearch/mappings.yaml';
280 my $indexes = LoadFile( $mappings_yaml );
282 while ( my ( $index_name, $fields ) = each %$indexes ) {
283 while ( my ( $field_name, $data ) = each %$fields ) {
284 my $field_type = $data->{type};
285 my $field_label = $data->{label};
286 my $mappings = $data->{mappings};
287 my $facet_order = $data->{facet_order};
288 my $search_field = Koha::SearchFields->find_or_create({
290 label => $field_label,
296 $search_field->update(
298 facet_order => $facet_order
301 for my $mapping ( @$mappings ) {
302 my $marc_field = Koha::SearchMarcMaps->find_or_create({ index_name => $index_name, marc_type => $mapping->{marc_type}, marc_field => $mapping->{marc_field} });
303 $search_field->add_to_search_marc_maps($marc_field, { facet => $mapping->{facet} || 0, suggestible => $mapping->{suggestible} || 0, sort => $mapping->{sort} } );
309 # This overrides the accessor provided by Class::Accessor so that if
310 # sort_fields isn't set, then it'll generate it.
314 $self->_sort_fields_accessor(@_);
317 my $val = $self->_sort_fields_accessor();
320 # This will populate the accessor as a side effect
321 $self->get_elasticsearch_mappings();
322 return $self->_sort_fields_accessor();
325 =head2 _process_mappings($mappings, $data, $record_document, $altscript)
327 $self->_process_mappings($mappings, $marc_field_data, $record_document, 0)
329 Process all C<$mappings> targets operating on a specific MARC field C<$data>.
330 Since we group all mappings by MARC field targets C<$mappings> will contain
331 all targets for C<$data> and thus we need to fetch the MARC field only once.
332 C<$mappings> will be applied to C<$record_document> and new field values added.
333 The method has no return value.
339 Arrayref of mappings containing arrayrefs in the format
340 [C<$target>, C<$options>] where C<$target> is the name of the target field and
341 C<$options> is a hashref containing processing directives for this particular
346 The source data from a MARC record field.
348 =item C<$record_document>
350 Hashref representing the Elasticsearch document on which mappings should be
355 A boolean value indicating whether an alternate script presentation is being
362 sub _process_mappings {
363 my ($_self, $mappings, $data, $record_document, $altscript) = @_;
364 foreach my $mapping (@{$mappings}) {
365 my ($target, $options) = @{$mapping};
367 # Don't process sort fields for alternate scripts
368 my $sort = $target =~ /__sort$/;
369 if ($sort && $altscript) {
373 # Copy (scalar) data since can have multiple targets
374 # with differing options for (possibly) mutating data
375 # so need a different copy for each
377 $record_document->{$target} //= [];
378 if (defined $options->{substr}) {
379 my ($start, $length) = @{$options->{substr}};
380 $_data = length($data) > $start ? substr $data, $start, $length : '';
382 if (defined $options->{value_callbacks}) {
383 $_data = reduce { $b->($a) } ($_data, @{$options->{value_callbacks}});
385 if (defined $options->{property}) {
387 $options->{property} => $_data
390 push @{$record_document->{$target}}, $_data;
394 =head2 marc_records_to_documents($marc_records)
396 my @record_documents = $self->marc_records_to_documents($marc_records);
398 Using mappings stored in database convert C<$marc_records> to Elasticsearch documents.
400 Returns array of hash references, representing Elasticsearch documents,
401 acceptable as body payload in C<Search::Elasticsearch> requests.
405 =item C<$marc_documents>
407 Reference to array of C<MARC::Record> objects to be converted to Elasticsearch documents.
413 sub marc_records_to_documents {
414 my ($self, $records) = @_;
415 my $rules = $self->_get_marc_mapping_rules();
416 my $control_fields_rules = $rules->{control_fields};
417 my $data_fields_rules = $rules->{data_fields};
418 my $marcflavour = lc C4::Context->preference('marcflavour');
420 my @record_documents;
422 foreach my $record (@{$records}) {
423 my $record_document = {};
424 my $mappings = $rules->{leader};
426 $self->_process_mappings($mappings, $record->leader(), $record_document, 0);
428 foreach my $field ($record->fields()) {
429 if ($field->is_control_field()) {
430 my $mappings = $control_fields_rules->{$field->tag()};
432 $self->_process_mappings($mappings, $field->data(), $record_document, 0);
436 my $tag = $field->tag();
437 # Handle alternate scripts in MARC 21
439 if ($marcflavour eq 'marc21' && $tag eq '880') {
440 my $sub6 = $field->subfield('6');
441 if ($sub6 =~ /^(...)-\d+/) {
447 my $data_field_rules = $data_fields_rules->{$tag};
449 if ($data_field_rules) {
450 my $subfields_mappings = $data_field_rules->{subfields};
451 my $wildcard_mappings = $subfields_mappings->{'*'};
452 foreach my $subfield ($field->subfields()) {
453 my ($code, $data) = @{$subfield};
454 my $mappings = $subfields_mappings->{$code} // [];
455 if ($wildcard_mappings) {
456 $mappings = [@{$mappings}, @{$wildcard_mappings}];
459 $self->_process_mappings($mappings, $data, $record_document, $altscript);
463 my $subfields_join_mappings = $data_field_rules->{subfields_join};
464 if ($subfields_join_mappings) {
465 foreach my $subfields_group (keys %{$subfields_join_mappings}) {
466 # Map each subfield to values, remove empty values, join with space
471 map { join(' ', $field->subfield($_)) } split(//, $subfields_group)
475 $self->_process_mappings($subfields_join_mappings->{$subfields_group}, $data, $record_document, $altscript);
482 foreach my $field (keys %{$rules->{defaults}}) {
483 unless (defined $record_document->{$field}) {
484 $record_document->{$field} = $rules->{defaults}->{$field};
487 foreach my $field (@{$rules->{sum}}) {
488 if (defined $record_document->{$field}) {
489 # TODO: validate numeric? filter?
490 # TODO: Or should only accept fields without nested values?
491 # TODO: Quick and dirty, improve if needed
492 $record_document->{$field} = sum0(grep { !ref($_) && m/\d+(\.\d+)?/} @{$record_document->{$field}});
495 # Index all applicable ISBN forms (ISBN-10 and ISBN-13 with and without dashes)
496 foreach my $field (@{$rules->{isbn}}) {
497 if (defined $record_document->{$field}) {
499 foreach my $input_isbn (@{$record_document->{$field}}) {
500 my $isbn = Business::ISBN->new($input_isbn);
501 if (defined $isbn && $isbn->is_valid) {
502 my $isbn13 = $isbn->as_isbn13->as_string;
503 push @isbns, $isbn13;
505 push @isbns, $isbn13;
507 my $isbn10 = $isbn->as_isbn10;
509 $isbn10 = $isbn10->as_string;
510 push @isbns, $isbn10;
512 push @isbns, $isbn10;
515 push @isbns, $input_isbn;
518 $record_document->{$field} = \@isbns;
522 # Remove duplicate values and collapse sort fields
523 foreach my $field (keys %{$record_document}) {
524 if (ref($record_document->{$field}) eq 'ARRAY') {
525 @{$record_document->{$field}} = do {
527 grep { !$seen{ref($_) eq 'HASH' && defined $_->{input} ? $_->{input} : $_}++ } @{$record_document->{$field}};
529 if ($field =~ /__sort$/) {
530 # Make sure to keep the sort field length sensible. 255 was chosen as a nice round value.
531 $record_document->{$field} = [substr(join(' ', @{$record_document->{$field}}), 0, 255)];
536 # TODO: Perhaps should check if $records_document non empty, but really should never be the case
537 $record->encoding('UTF-8');
540 # Temporarily intercept all warn signals (MARC::Record carps when record length > 99999)
541 local $SIG{__WARN__} = sub {
542 push @warnings, $_[0];
544 $record_document->{'marc_data'} = encode_base64(encode('UTF-8', $record->as_usmarc()));
547 # Suppress warnings if record length exceeded
548 unless (substr($record->leader(), 0, 5) eq '99999') {
549 foreach my $warning (@warnings) {
553 $record_document->{'marc_data'} = $record->as_xml_record($marcflavour);
554 $record_document->{'marc_format'} = 'MARCXML';
557 $record_document->{'marc_format'} = 'base64ISO2709';
559 my $id = $record->subfield('999', 'c');
560 push @record_documents, [$id, $record_document];
562 return \@record_documents;
565 =head2 _field_mappings($facet, $suggestible, $sort, $target_name, $target_type, $range)
567 my @mappings = _field_mappings($facet, $suggestible, $sort, $target_name, $target_type, $range)
569 Get mappings, an internal data structure later used by
570 L<_process_mappings($mappings, $data, $record_document, $altscript)> to process MARC target
571 data for a MARC mapping.
573 The returned C<$mappings> is not to to be confused with mappings provided by
574 C<_foreach_mapping>, rather this sub accepts properties from a mapping as
575 provided by C<_foreach_mapping> and expands it to this internal data structure.
576 In the caller context (C<_get_marc_mapping_rules>) the returned C<@mappings>
577 is then applied to each MARC target (leader, control field data, subfield or
578 joined subfields) and integrated into the mapping rules data structure used in
579 C<marc_records_to_documents> to transform MARC records into Elasticsearch
586 Boolean indicating whether to create a facet field for this mapping.
588 =item C<$suggestible>
590 Boolean indicating whether to create a suggestion field for this mapping.
594 Boolean indicating whether to create a sort field for this mapping.
596 =item C<$target_name>
598 Elasticsearch document target field name.
600 =item C<$target_type>
602 Elasticsearch document target field type.
606 An optional range as a string in the format "<START>-<END>" or "<START>",
607 where "<START>" and "<END>" are integers specifying a range that will be used
608 for extracting a substring from MARC data as Elasticsearch field target value.
610 The first character position is "1", and the range is inclusive,
611 so "1-3" means the first three characters of MARC data.
613 If only "<START>" is provided only one character at position "<START>" will
620 sub _field_mappings {
621 my ($_self, $facet, $suggestible, $sort, $target_name, $target_type, $range) = @_;
622 my %mapping_defaults = ();
625 my $substr_args = undef;
627 # TODO: use value_callback instead?
628 my ($start, $end) = map(int, split /-/, $range, 2);
629 $substr_args = [$start];
630 push @{$substr_args}, (defined $end ? $end - $start + 1 : 1);
632 my $default_options = {};
634 $default_options->{substr} = $substr_args;
637 # TODO: Should probably have per type value callback/hook
638 # but hard code for now
639 if ($target_type eq 'boolean') {
640 $default_options->{value_callbacks} //= [];
641 push @{$default_options->{value_callbacks}}, sub {
643 # Trim whitespace at both ends
644 $value =~ s/^\s+|\s+$//g;
645 return $value ? 'true' : 'false';
649 my $mapping = [$target_name, $default_options];
650 push @mappings, $mapping;
653 push @suffixes, 'facet' if $facet;
654 push @suffixes, 'suggestion' if $suggestible;
655 push @suffixes, 'sort' if !defined $sort || $sort;
657 foreach my $suffix (@suffixes) {
658 my $mapping = ["${target_name}__$suffix"];
659 # TODO: Hack, fix later in less hideous manner
660 if ($suffix eq 'suggestion') {
661 push @{$mapping}, {%{$default_options}, property => 'input'};
664 push @{$mapping}, $default_options;
666 push @mappings, $mapping;
671 =head2 _get_marc_mapping_rules
673 my $mapping_rules = $self->_get_marc_mapping_rules()
675 Generates rules from mappings stored in database for MARC records to Elasticsearch JSON document conversion.
677 Since field retrieval is slow in C<MARC::Records> (all fields are itereted through for
678 each call to C<MARC::Record>->field) we create an optimized structure of mapping
679 rules keyed by MARC field tags holding all the mapping rules for that particular tag.
681 We can then iterate through all MARC fields for each record and apply all relevant
682 rules once per fields instead of retreiving fields multiple times for each mapping rule
683 which is terribly slow.
687 # TODO: This structure can be used for processing multiple MARC::Records so is currently
688 # rebuilt for each batch. Since it is cacheable it could also be stored in an in
689 # memory cache which it is currently not. The performance gain of caching
690 # would probably be marginal, but to do this could be a further improvement.
692 sub _get_marc_mapping_rules {
694 my $marcflavour = lc C4::Context->preference('marcflavour');
695 my $field_spec_regexp = qr/^([0-9]{3})([()0-9a-z]+)?(?:_\/(\d+(?:-\d+)?))?$/;
696 my $leader_regexp = qr/^leader(?:_\/(\d+(?:-\d+)?))?$/;
699 'control_fields' => {},
706 $self->_foreach_mapping(sub {
707 my ($name, $type, $facet, $suggestible, $sort, $marc_type, $marc_field) = @_;
708 return if $marc_type ne $marcflavour;
710 if ($type eq 'sum') {
711 push @{$rules->{sum}}, $name;
713 elsif ($type eq 'isbn') {
714 push @{$rules->{isbn}}, $name;
716 elsif ($type eq 'boolean') {
717 # boolean gets special handling, if value doesn't exist for a field,
719 $rules->{defaults}->{$name} = 'false';
722 if ($marc_field =~ $field_spec_regexp) {
727 # Parse and separate subfields form subfield groups
729 my $subfield_group = '';
732 foreach my $token (split //, $2) {
735 Koha::Exceptions::Elasticsearch::MARCFieldExprParseError->throw(
736 "Unmatched opening parenthesis for $marc_field"
743 elsif ($token eq ")") {
745 if ($subfield_group) {
746 push @subfield_groups, $subfield_group;
747 $subfield_group = '';
752 Koha::Exceptions::Elasticsearch::MARCFieldExprParseError->throw(
753 "Unmatched closing parenthesis for $marc_field"
757 elsif ($open_group) {
758 $subfield_group .= $token;
761 push @subfields, $token;
766 push @subfields, '*';
769 my $range = defined $3 ? $3 : undef;
770 my @mappings = $self->_field_mappings($facet, $suggestible, $sort, $name, $type, $range);
772 if ($field_tag < 10) {
773 $rules->{control_fields}->{$field_tag} //= [];
774 push @{$rules->{control_fields}->{$field_tag}}, @mappings;
777 $rules->{data_fields}->{$field_tag} //= {};
778 foreach my $subfield (@subfields) {
779 $rules->{data_fields}->{$field_tag}->{subfields}->{$subfield} //= [];
780 push @{$rules->{data_fields}->{$field_tag}->{subfields}->{$subfield}}, @mappings;
782 foreach my $subfield_group (@subfield_groups) {
783 $rules->{data_fields}->{$field_tag}->{subfields_join}->{$subfield_group} //= [];
784 push @{$rules->{data_fields}->{$field_tag}->{subfields_join}->{$subfield_group}}, @mappings;
788 elsif ($marc_field =~ $leader_regexp) {
789 my $range = defined $1 ? $1 : undef;
790 my @mappings = $self->_field_mappings($facet, $suggestible, $sort, $name, $type, $range);
791 push @{$rules->{leader}}, @mappings;
794 Koha::Exceptions::Elasticsearch::MARCFieldExprParseError->throw(
795 "Invalid MARC field expression: $marc_field"
802 =head2 _foreach_mapping
804 $self->_foreach_mapping(
806 my ( $name, $type, $facet, $suggestible, $sort, $marc_type,
809 return unless $marc_type eq 'marc21';
810 print "Data comes from: " . $marc_field . "\n";
814 This allows you to apply a function to each entry in the elasticsearch mappings
815 table, in order to build the mappings for whatever is needed.
817 In the provided function, the files are:
823 The field name for elasticsearch (corresponds to the 'mapping' column in the
828 The type for this value, e.g. 'string'.
832 True if this value should be facetised. This only really makes sense if the
833 field is understood by the facet processing code anyway.
837 True if this is a field that a) needs special sort handling, and b) if it
838 should be sorted on. False if a) but not b). Undef if not a). This allows,
839 for example, author to be sorted on but not everything marked with "author"
840 to be included in that sort.
844 A string that indicates the MARC type that this mapping is for, e.g. 'marc21',
845 'unimarc', 'normarc'.
849 A string that describes the MARC field that contains the data to extract.
850 These are of a form suited to Catmandu's MARC fixers.
856 sub _foreach_mapping {
857 my ( $self, $sub ) = @_;
859 # TODO use a caching framework here
860 my $search_fields = Koha::Database->schema->resultset('SearchField')->search(
862 'search_marc_map.index_name' => $self->index,
864 { join => { search_marc_to_fields => 'search_marc_map' },
866 'search_marc_to_fields.facet',
867 'search_marc_to_fields.suggestible',
868 'search_marc_to_fields.sort',
869 'search_marc_map.marc_type',
870 'search_marc_map.marc_field',
882 while ( my $search_field = $search_fields->next ) {
884 # Force lower case on indexed field names for case insensitive
885 # field name searches
886 lc($search_field->name),
888 $search_field->get_column('facet'),
889 $search_field->get_column('suggestible'),
890 $search_field->get_column('sort'),
891 $search_field->get_column('marc_type'),
892 $search_field->get_column('marc_field'),
899 die process_error($@);
901 This parses an Elasticsearch error message and produces a human-readable
902 result from it. This result is probably missing all the useful information
903 that you might want in diagnosing an issue, so the warning is also logged.
905 Note that currently the resulting message is not internationalised. This
906 will happen eventually by some method or other.
911 my ($self, $msg) = @_;
913 warn $msg; # simple logging
915 # This is super-primitive
916 return "Unable to understand your search query, please rephrase and try again.\n" if $msg =~ /ParseException/;
918 return "Unable to perform your search. Please try again.\n";
921 =head2 _read_configuration
923 my $conf = _read_configuration();
925 Reads the I<configuration file> and returns a hash structure with the
926 configuration information. It raises an exception if mandatory entries
929 The hashref structure has the following form:
932 'nodes' => ['127.0.0.1:9200', 'anotherserver:9200'],
933 'index_name' => 'koha_instance',
936 This is configured by the following in the C<config> block in koha-conf.xml:
939 <server>127.0.0.1:9200</server>
940 <server>anotherserver:9200</server>
941 <index_name>koha_instance</index_name>
946 sub _read_configuration {
950 my $conf = C4::Context->config('elasticsearch');
951 Koha::Exceptions::Config::MissingEntry->throw(
952 "Missing 'elasticsearch' block in config file")
953 unless defined $conf;
955 if ( $conf && $conf->{server} ) {
956 my $nodes = $conf->{server};
957 if ( ref($nodes) eq 'ARRAY' ) {
958 $configuration->{nodes} = $nodes;
961 $configuration->{nodes} = [$nodes];
965 Koha::Exceptions::Config::MissingEntry->throw(
966 "Missing 'server' entry in config file for elasticsearch");
969 if ( defined $conf->{index_name} ) {
970 $configuration->{index_name} = $conf->{index_name};
973 Koha::Exceptions::Config::MissingEntry->throw(
974 "Missing 'index_name' entry in config file for elasticsearch");
977 return $configuration;
980 =head2 get_facetable_fields
982 my @facetable_fields = Koha::SearchEngine::Elasticsearch->get_facetable_fields();
984 Returns the list of Koha::SearchFields marked to be faceted in the ES configuration
988 sub get_facetable_fields {
991 # These should correspond to the ES field names, as opposed to the CCL
992 # things that zebra uses.
993 my @search_field_names = qw( author itype location su-geo title-series subject ccode holdingbranch homebranch );
994 my @faceted_fields = Koha::SearchFields->search(
995 { name => { -in => \@search_field_names }, facet_order => { '!=' => undef } }, { order_by => ['facet_order'] }
997 my @not_faceted_fields = Koha::SearchFields->search(
998 { name => { -in => \@search_field_names }, facet_order => undef }, { order_by => ['facet_order'] }
1000 # This could certainly be improved
1001 return ( @faceted_fields, @not_faceted_fields );
1012 =item Chris Cormack C<< <chrisc@catalyst.net.nz> >>
1014 =item Robin Sheat C<< <robin@catalyst.net.nz> >>
1016 =item Jonathan Druart C<< <jonathan.druart@bugs.koha-community.org> >>