=head1 NAME
-Koha::SearchEngine::ElasticSearch::Search - search functions for Elasticsearch
+Koha::SearchEngine::Elasticsearch::Search - search functions for Elasticsearch
=head1 SYNOPSIS
- my $searcher = Koha::SearchEngine::ElasticSearch::Search->new();
- my $builder = Koha::SearchEngine::Elasticsearch::QueryBuilder->new();
+ my $searcher =
+ Koha::SearchEngine::Elasticsearch::Search->new( { index => $index } );
+ my $builder = Koha::SearchEngine::Elasticsearch::QueryBuilder->new(
+ { index => $index } );
my $query = $builder->build_query('perl');
my $results = $searcher->search($query);
print "There were " . $results->total . " results.\n";
=cut
-use base qw(Koha::ElasticSearch);
-use Koha::ItemTypes;
+use Modern::Perl;
+use base qw(Koha::SearchEngine::Elasticsearch);
+use C4::Context;
+use Koha::ItemTypes;
+use Koha::AuthorisedValues;
+use Koha::SearchEngine::QueryBuilder;
+use Koha::SearchEngine::Search;
+use MARC::Record;
use Catmandu::Store::ElasticSearch;
use Data::Dumper; #TODO remove
}
$self->store(
Catmandu::Store::ElasticSearch->new(
- %$params, trace_calls => 1,
+ %$params,
)
) unless $self->store;
- my $results = $self->store->bag->search( %$query, %paging );
+ my $results = eval {
+ $self->store->bag->search( %$query, %paging );
+ };
+ if ($@) {
+ die $self->process_error($@);
+ }
return $results;
}
my $count = $searcher->count($query);
This mimics a search request, but just gets the result count instead. That's
-faster than pulling all the data in, ususally.
+faster than pulling all the data in, usually.
=cut
Catmandu::Store::ElasticSearch->new( %$params, trace_calls => 0, ) )
unless $self->store;
- my $searcher = $self->store->bag->searcher(query => $query);
- my $count = $searcher->count();
+ my $search = $self->store->bag->search( %$query);
+ my $count = $search->total() || 0;
return $count;
}
$servers, $results_per_page, $offset, $expanded_facet,
$branches, $query_type, $scan
) = @_;
-
my %options;
+ if ( !defined $offset or $offset < 0 ) {
+ $offset = 0;
+ }
$options{offset} = $offset;
+ $options{expanded_facet} = $expanded_facet;
my $results = $self->search($query, undef, $results_per_page, %options);
# Convert each result into a MARC::Record
# right place in the array, according to $offset
$results->each(sub {
# The results come in an array for some reason
- my $marc_json = @_[0]->{record};
+ my $marc_json = $_[0]->{record};
my $marc = $self->json2marc($marc_json);
$records[$index++] = $marc;
});
my %result;
$result{biblioserver}{hits} = $results->total;
$result{biblioserver}{RECORDS} = \@records;
- return (undef, \%result, $self->_convert_facets($results->{facets}));
+ return (undef, \%result, $self->_convert_facets($results->{aggregations}, $expanded_facet));
}
=head2 search_auth_compat
$res->each(
sub {
my %result;
- my $record = @_[0];
+ my $record = $_[0];
my $marc_json = $record->{record};
# I wonder if these should be real values defined in the mapping
# rather than hard-coded conversions.
- # Our results often come through as nested arrays, to fix this
- # requires changes in catmandu.
- my $authid = $record->{ 'Local-Number' }[0][0];
+ # Handle legacy nested arrays indexed with splitting enabled.
+ my $authid = $record->{ 'Local-number' }[0];
+ $authid = @$authid[0] if (ref $authid eq 'ARRAY');
$result{authid} = $authid;
# TODO put all this info into the record at index time so we
# with the record. It's not documented why this is the case, so
# it's not reproduced here yet.
my $authtype = $rs->single;
- my $auth_tag_to_report = $authtype->auth_tag_to_report;
+ my $auth_tag_to_report = $authtype ? $authtype->auth_tag_to_report : "";
my $marc = $self->json2marc($marc_json);
my $mainentry = $marc->field($auth_tag_to_report);
my $reported_tag;
}
}
# Turn the resultset into a hash
- my %authtype_cols;
- foreach my $col ($authtype->result_source->columns) {
- $authtype_cols{$col} = $authtype->get_column($col);
- }
- $result{authtype} = $authtype->authtypetext;
+ $result{authtype} = $authtype ? $authtype->authtypetext : $authtypecode;
$result{reported_tag} = $reported_tag;
# Reimplementing BuildSummary is out of scope because it'll be hard
my $query = {
query => {
- filtered => {
- query => { match_all => {} },
+ bool => {
+# query => { match_all => {} },
filter => { term => { an => $authid } }
}
}
$bib_searcher->count($query);
}
+=head2 simple_search_compat
+
+ my ( $error, $marcresults, $total_hits ) =
+ $searcher->simple_search( $query, $offset, $max_results, %options );
+
+This is a simpler interface to the searching, intended to be similar enough to
+L<C4::Search::SimpleSearch>.
+
+Arguments:
+
+=over 4
+
+=item C<$query>
+
+A thing to search for. It could be a simple string, or something constructed
+with the appropriate QueryBuilder module.
+
+=item C<$offset>
+
+How many results to skip from the start of the results.
+
+=item C<$max_results>
+
+The max number of results to return. The default is 100 (because unlimited
+is a pretty terrible thing to do.)
+
+=item C<%options>
+
+These options are unused by Elasticsearch
+
+=back
+
+Returns:
+
+=over 4
+
+=item C<$error>
+
+if something went wrong, this'll contain some kind of error
+message.
+
+=item C<$marcresults>
+
+an arrayref of MARC::Records (note that this is different from the
+L<C4::Search> version which will return plain XML, but too bad.)
+
+=item C<$total_hits>
+
+the total number of results that this search could have returned.
+
+=back
+
+=cut
+sub simple_search_compat {
+ my ($self, $query, $offset, $max_results) = @_;
+
+ return ('No query entered', undef, undef) unless $query;
+
+ my %options;
+ $offset = 0 if not defined $offset or $offset < 0;
+ $options{offset} = $offset;
+ $max_results //= 100;
+
+ unless (ref $query) {
+ # We'll push it through the query builder to sanitise everything.
+ my $qb = Koha::SearchEngine::QueryBuilder->new({index => $self->index});
+ (undef,$query) = $qb->build_query_compat(undef, [$query]);
+ }
+ my $results = $self->search($query, undef, $max_results, %options);
+ my @records;
+ $results->each(sub {
+ # The results come in an array for some reason
+ my $marc_json = $_[0]->{record};
+ my $marc = $self->json2marc($marc_json);
+ push @records, $marc;
+ });
+ return (undef, \@records, $results->total);
+}
+
+=head2 extract_biblionumber
+
+ my $biblionumber = $searcher->extract_biblionumber( $searchresult );
+
+$searchresult comes from simple_search_compat.
+
+Returns the biblionumber from the search result record.
+
+=cut
+
+sub extract_biblionumber {
+ my ( $self, $searchresultrecord ) = @_;
+ return Koha::SearchEngine::Search::extract_biblionumber( $searchresultrecord );
+}
=head2 json2marc
# fields are like:
# [ '245', '1', '2', 'a' => 'Title', 'b' => 'Subtitle' ]
+ # or
+ # [ '001', undef, undef, '_', 'a value' ]
# conveniently, this is the form that MARC::Field->new() likes
- foreach $field (@$marcjson) {
- next if @$field < 5; # Shouldn't be possible, but...
+ foreach my $field (@$marcjson) {
+ next if @$field < 5;
if ( $field->[0] eq 'LDR' ) {
$marc->leader( $field->[4] );
}
else {
- my $marc_field = MARC::Field->new(@$field);
+ my $tag = $field->[0];
+ my $marc_field;
+ if ( MARC::Field->is_controlfield_tag( $field->[0] ) ) {
+ $marc_field = MARC::Field->new($field->[0], $field->[4]);
+ } else {
+ $marc_field = MARC::Field->new(@$field);
+ }
$marc->append_fields($marc_field);
}
}
return $marc;
}
+sub max_result_window {
+ my ($self) = @_;
+
+ $self->store(
+ Catmandu::Store::ElasticSearch->new(%{ $self->get_elasticsearch_params })
+ ) unless $self->store;
+
+ my $index_name = $self->store->index_name;
+ my $settings = $self->store->es->indices->get_settings(
+ index => $index_name,
+ params => { include_defaults => 1, flat_settings => 1 },
+ );
+
+ my $max_result_window = $settings->{$index_name}->{settings}->{'index.max_result_window'};
+ $max_result_window //= $settings->{$index_name}->{defaults}->{'index.max_result_window'};
+
+ return $max_result_window;
+}
+
=head2 _convert_facets
- my $koha_facets = _convert_facets($es_facets);
+ my $koha_facets = _convert_facets($es_facets, $expanded_facet);
Converts elasticsearch facets types to the form that Koha expects.
It expects the ES facet name to match the Koha type, for example C<itype>,
C<au>, C<su-to>, etc.
+C<$expanded_facet> is the facet that we want to show FacetMaxCount entries for, rather
+than just 5 like normal.
+
=cut
sub _convert_facets {
- my ( $self, $es ) = @_;
+ my ( $self, $es, $exp_facet ) = @_;
- return undef if !$es;
+ return if !$es;
# These should correspond to the ES field names, as opposed to the CCL
# things that zebra uses.
+ # TODO let the library define the order using the interface.
my %type_to_label = (
- author => 'Authors',
- location => 'Location',
- itype => 'ItemTypes',
- se => 'Series',
- subject => 'Topics',
- 'su-geo' => 'Places',
+ author => { order => 1, label => 'Authors', },
+ itype => { order => 2, label => 'ItemTypes', },
+ location => { order => 3, label => 'Location', },
+ 'su-geo' => { order => 4, label => 'Places', },
+ se => { order => 5, label => 'Series', },
+ subject => { order => 6, label => 'Topics', },
+ ccode => { order => 7, label => 'CollectionCodes',},
+ holdingbranch => { order => 8, label => 'HoldingLibrary' },
+ homebranch => { order => 9, label => 'HomeLibrary' }
);
# We also have some special cases, e.g. itypes that need to show the
# value rather than the code.
- my $itypes = Koha::ItemTypes->new();
- my %special = ( itype => sub { $itypes->get_description_for_code(@_) }, );
- my @res;
- while ( ( $type, $data ) = each %$es ) {
+ my @itypes = Koha::ItemTypes->search;
+ my @libraries = Koha::Libraries->search;
+ my $library_names = { map { $_->branchcode => $_->branchname } @libraries };
+ my @locations = Koha::AuthorisedValues->search( { category => 'LOC' } );
+ my $opac = C4::Context->interface eq 'opac' ;
+ my %special = (
+ itype => { map { $_->itemtype => $_->description } @itypes },
+ location => { map { $_->authorised_value => ( $opac ? ( $_->lib_opac || $_->lib ) : $_->lib ) } @locations },
+ holdingbranch => $library_names,
+ homebranch => $library_names
+ );
+ my @facets;
+ $exp_facet //= '';
+ while ( my ( $type, $data ) = each %$es ) {
next if !exists( $type_to_label{$type} );
+
+ # We restrict to the most popular $limit !results
+ my $limit = ( $type eq $exp_facet ) ? C4::Context->preference('FacetMaxCount') : 5;
my $facet = {
- type_id => $type . '_id',
- expand => $type,
- expandable => 1, # TODO figure how that's supposed to work
- "type_label_$type_to_label{$type}" => 1,
+ type_id => $type . '_id',
+ expand => $type,
+ expandable => ( $type ne $exp_facet )
+ && ( @{ $data->{buckets} } > $limit ),
+ "type_label_$type_to_label{$type}{label}" => 1,
type_link_value => $type,
+ order => $type_to_label{$type}{order},
};
- foreach my $term ( @{ $data->{terms} } ) {
- my $t = $term->{term};
- my $c = $term->{count};
+ $limit = @{ $data->{buckets} } if ( $limit > @{ $data->{buckets} } );
+ foreach my $term ( @{ $data->{buckets} }[ 0 .. $limit - 1 ] ) {
+ my $t = $term->{key};
+ my $c = $term->{doc_count};
+ my $label;
if ( exists( $special{$type} ) ) {
- $label = $special{$type}->($t);
+ $label = $special{$type}->{$t} // $t;
}
else {
$label = $t;
facet_count => $c,
facet_link_value => $t,
facet_title_value => $t . " ($c)",
- facet_label_value => $label, # TODO either truncate this,
+ facet_label_value => $label, # TODO either truncate this,
# or make the template do it like it should anyway
type_link_value => $type,
};
}
- push @res, $facet if exists $facet->{facets};
+ push @facets, $facet if exists $facet->{facets};
}
- return \@res;
+
+ @facets = sort { $a->{order} cmp $b->{order} } @facets;
+ return \@facets;
}