3 # This file is part of Koha.
5 # Koha is free software; you can redistribute it and/or modify it under the
6 # terms of the GNU General Public License as published by the Free Software
7 # Foundation; either version 2 of the License, or (at your option) any later
10 # Koha is distributed in the hope that it will be useful, but WITHOUT ANY
11 # WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR
12 # A PARTICULAR PURPOSE. See the GNU General Public License for more details.
14 # You should have received a copy of the GNU General Public License along with
15 # Koha; if not, write to the Free Software Foundation, Inc., 59 Temple Place,
16 # Suite 330, Boston, MA 02111-1307 USA
19 # use warnings; # FIXME
22 use C4::Biblio; # GetMarcFromKohaField
23 use C4::Koha; # getFacets
25 use C4::Search::PazPar2;
27 use C4::Dates qw(format_date);
30 use vars qw($VERSION @ISA @EXPORT @EXPORT_OK %EXPORT_TAGS $DEBUG);
32 # set the version for version checking
35 $DEBUG = ($ENV{DEBUG}) ? 1 : 0;
40 C4::Search - Functions for searching the Koha catalog.
44 See opac/opac-search.pl or catalogue/search.pl for example of usage
48 This module provides searching functions for Koha's bibliographic databases
65 # make all your functions, whether exported or not;
67 =head2 findseealso($dbh,$fields);
69 C<$dbh> is a link to the DB handler.
72 my $dbh =C4::Context->dbh;
74 C<$fields> is a reference to the fields array
76 This function modifies the @$fields array and adds related fields to search on.
78 FIXME: this function is probably deprecated in Koha 3
83 my ( $dbh, $fields ) = @_;
84 my $tagslib = GetMarcStructure(1);
85 for ( my $i = 0 ; $i <= $#{$fields} ; $i++ ) {
86 my ($tag) = substr( @$fields[$i], 1, 3 );
87 my ($subfield) = substr( @$fields[$i], 4, 1 );
88 @$fields[$i] .= ',' . $tagslib->{$tag}->{$subfield}->{seealso}
89 if ( $tagslib->{$tag}->{$subfield}->{seealso} );
95 ($biblionumber,$biblionumber,$title) = FindDuplicate($record);
97 This function attempts to find duplicate records using a hard-coded, fairly simplistic algorithm
103 my $dbh = C4::Context->dbh;
104 my $result = TransformMarcToKoha( $dbh, $record, '' );
109 my ( $biblionumber, $title );
111 # search duplicate on ISBN, easy and fast..
112 # ... normalize first
113 if ( $result->{isbn} ) {
114 $result->{isbn} =~ s/\(.*$//;
115 $result->{isbn} =~ s/\s+$//;
116 $query = "isbn=$result->{isbn}";
119 $result->{title} =~ s /\\//g;
120 $result->{title} =~ s /\"//g;
121 $result->{title} =~ s /\(//g;
122 $result->{title} =~ s /\)//g;
124 # FIXME: instead of removing operators, could just do
125 # quotes around the value
126 $result->{title} =~ s/(and|or|not)//g;
127 $query = "ti,ext=$result->{title}";
128 $query .= " and itemtype=$result->{itemtype}"
129 if ( $result->{itemtype} );
130 if ( $result->{author} ) {
131 $result->{author} =~ s /\\//g;
132 $result->{author} =~ s /\"//g;
133 $result->{author} =~ s /\(//g;
134 $result->{author} =~ s /\)//g;
136 # remove valid operators
137 $result->{author} =~ s/(and|or|not)//g;
138 $query .= " and au,ext=$result->{author}";
142 # FIXME: add error handling
143 my ( $error, $searchresults ) = SimpleSearch($query); # FIXME :: hardcoded !
145 foreach my $possible_duplicate_record (@$searchresults) {
147 MARC::Record->new_from_usmarc($possible_duplicate_record);
148 my $result = TransformMarcToKoha( $dbh, $marcrecord, '' );
150 # FIXME :: why 2 $biblionumber ?
152 push @results, $result->{'biblionumber'};
153 push @results, $result->{'title'};
161 ( $error, $results, $total_hits ) = SimpleSearch( $query, $offset, $max_results, [@servers] );
163 This function provides a simple search API on the bibliographic catalog
169 * $query can be a simple keyword or a complete CCL query
170 * @servers is optional. Defaults to biblioserver as found in koha-conf.xml
171 * $offset - If present, represents the number of records at the beggining to omit. Defaults to 0
172 * $max_results - if present, determines the maximum number of records to fetch. undef is All. defaults to undef.
177 * $error is a empty unless an error is detected
178 * \@results is an array of records.
179 * $total_hits is the number of hits that would have been returned with no limit
181 =item C<usage in the script:>
185 my ( $error, $marcresults, $total_hits ) = SimpleSearch($query);
187 if (defined $error) {
188 $template->param(query_error => $error);
189 warn "error: ".$error;
190 output_html_with_http_headers $input, $cookie, $template->output;
194 my $hits = scalar @$marcresults;
197 for my $i (0..$hits) {
199 my $marcrecord = MARC::File::USMARC::decode($marcresults->[$i]);
200 my $biblio = TransformMarcToKoha(C4::Context->dbh,$marcrecord,'');
202 #build the hash for the template.
203 $resultsloop{highlight} = ($i % 2)?(1):(0);
204 $resultsloop{title} = $biblio->{'title'};
205 $resultsloop{subtitle} = $biblio->{'subtitle'};
206 $resultsloop{biblionumber} = $biblio->{'biblionumber'};
207 $resultsloop{author} = $biblio->{'author'};
208 $resultsloop{publishercode} = $biblio->{'publishercode'};
209 $resultsloop{publicationyear} = $biblio->{'publicationyear'};
211 push @results, \%resultsloop;
214 $template->param(result=>\@results);
219 my ( $query, $offset, $max_results, $servers ) = @_;
221 if ( C4::Context->preference('NoZebra') ) {
222 my $result = NZorder( NZanalyse($query) )->{'biblioserver'};
225 && $result->{hits} > 0 ? $result->{'RECORDS'} : [] );
226 return ( undef, $search_result, scalar($search_result) );
229 # FIXME hardcoded value. See catalog/search.pl & opac-search.pl too.
230 my @servers = defined ( $servers ) ? @$servers : ( "biblioserver" );
236 return ( "No query entered", undef, undef ) unless $query;
238 # Initialize & Search Zebra
239 for ( my $i = 0 ; $i < @servers ; $i++ ) {
241 $zconns[$i] = C4::Context->Zconn( $servers[$i], 1 );
242 $zoom_queries[$i] = new ZOOM::Query::CCL2RPN( $query, $zconns[$i]);
243 $tmpresults[$i] = $zconns[$i]->search( $zoom_queries[$i] );
247 $zconns[$i]->errmsg() . " ("
248 . $zconns[$i]->errcode() . ") "
249 . $zconns[$i]->addinfo() . " "
250 . $zconns[$i]->diagset();
252 return ( $error, undef, undef ) if $zconns[$i]->errcode();
256 # caught a ZOOM::Exception
260 . $@->addinfo() . " "
263 return ( $error, undef, undef );
266 while ( ( my $i = ZOOM::event( \@zconns ) ) != 0 ) {
267 my $event = $zconns[ $i - 1 ]->last_event();
268 if ( $event == ZOOM::Event::ZEND ) {
270 my $first_record = defined( $offset ) ? $offset+1 : 1;
271 my $hits = $tmpresults[ $i - 1 ]->size();
272 $total_hits += $hits;
273 my $last_record = $hits;
274 if ( defined $max_results && $offset + $max_results < $hits ) {
275 $last_record = $offset + $max_results;
278 for my $j ( $first_record..$last_record ) {
279 my $record = $tmpresults[ $i - 1 ]->record( $j-1 )->raw(); # 0 indexed
280 push @results, $record;
285 foreach my $result (@tmpresults) {
288 foreach my $zoom_query (@zoom_queries) {
289 $zoom_query->destroy();
292 return ( undef, \@results, $total_hits );
298 ( undef, $results_hashref, \@facets_loop ) = getRecords (
300 $koha_query, $simple_query, $sort_by_ref, $servers_ref,
301 $results_per_page, $offset, $expanded_facet, $branches,
305 The all singing, all dancing, multi-server, asynchronous, scanning,
306 searching, record nabbing, facet-building
308 See verbse embedded documentation.
314 $koha_query, $simple_query, $sort_by_ref, $servers_ref,
315 $results_per_page, $offset, $expanded_facet, $branches,
319 my @servers = @$servers_ref;
320 my @sort_by = @$sort_by_ref;
322 # Initialize variables for the ZOOM connection and results object
326 my $results_hashref = ();
328 # Initialize variables for the faceted results objects
329 my $facets_counter = ();
330 my $facets_info = ();
331 my $facets = getFacets();
334 ; # stores the ref to array of hashes for template facets loop
336 ### LOOP THROUGH THE SERVERS
337 for ( my $i = 0 ; $i < @servers ; $i++ ) {
338 $zconns[$i] = C4::Context->Zconn( $servers[$i], 1 );
340 # perform the search, create the results objects
341 # if this is a local search, use the $koha-query, if it's a federated one, use the federated-query
342 my $query_to_use = ($servers[$i] =~ /biblioserver/) ? $koha_query : $simple_query;
344 #$query_to_use = $simple_query if $scan;
345 warn $simple_query if ( $scan and $DEBUG );
347 # Check if we've got a query_type defined, if so, use it
351 if ( $query_type =~ /^ccl/ ) {
353 s/\:/\=/g; # change : to = last minute (FIXME)
356 new ZOOM::Query::CCL2RPN( $query_to_use, $zconns[$i] )
359 elsif ( $query_type =~ /^cql/ ) {
362 new ZOOM::Query::CQL( $query_to_use, $zconns[$i] ) );
364 elsif ( $query_type =~ /^pqf/ ) {
367 new ZOOM::Query::PQF( $query_to_use, $zconns[$i] ) );
374 new ZOOM::Query::CCL2RPN( $query_to_use, $zconns[$i] )
380 new ZOOM::Query::CCL2RPN( $query_to_use, $zconns[$i] )
386 warn "WARNING: query problem with $query_to_use " . $@;
389 # Concatenate the sort_by limits and pass them to the results object
390 # Note: sort will override rank
392 foreach my $sort (@sort_by) {
393 if ( $sort eq "author_az" ) {
394 $sort_by .= "1=1003 <i ";
396 elsif ( $sort eq "author_za" ) {
397 $sort_by .= "1=1003 >i ";
399 elsif ( $sort eq "popularity_asc" ) {
400 $sort_by .= "1=9003 <i ";
402 elsif ( $sort eq "popularity_dsc" ) {
403 $sort_by .= "1=9003 >i ";
405 elsif ( $sort eq "call_number_asc" ) {
406 $sort_by .= "1=20 <i ";
408 elsif ( $sort eq "call_number_dsc" ) {
409 $sort_by .= "1=20 >i ";
411 elsif ( $sort eq "pubdate_asc" ) {
412 $sort_by .= "1=31 <i ";
414 elsif ( $sort eq "pubdate_dsc" ) {
415 $sort_by .= "1=31 >i ";
417 elsif ( $sort eq "acqdate_asc" ) {
418 $sort_by .= "1=32 <i ";
420 elsif ( $sort eq "acqdate_dsc" ) {
421 $sort_by .= "1=32 >i ";
423 elsif ( $sort eq "title_az" ) {
424 $sort_by .= "1=4 <i ";
426 elsif ( $sort eq "title_za" ) {
427 $sort_by .= "1=4 >i ";
430 warn "Ignoring unrecognized sort '$sort' requested" if $sort_by;
434 if ( $results[$i]->sort( "yaz", $sort_by ) < 0 ) {
435 warn "WARNING sort $sort_by failed";
438 } # finished looping through servers
440 # The big moment: asynchronously retrieve results from all servers
441 while ( ( my $i = ZOOM::event( \@zconns ) ) != 0 ) {
442 my $ev = $zconns[ $i - 1 ]->last_event();
443 if ( $ev == ZOOM::Event::ZEND ) {
444 next unless $results[ $i - 1 ];
445 my $size = $results[ $i - 1 ]->size();
449 # loop through the results
450 $results_hash->{'hits'} = $size;
452 if ( $offset + $results_per_page <= $size ) {
453 $times = $offset + $results_per_page;
458 for ( my $j = $offset ; $j < $times ; $j++ ) {
463 ## Check if it's an index scan
465 my ( $term, $occ ) = $results[ $i - 1 ]->term($j);
467 # here we create a minimal MARC record and hand it off to the
468 # template just like a normal result ... perhaps not ideal, but
470 my $tmprecord = MARC::Record->new();
471 $tmprecord->encoding('UTF-8');
475 # the minimal record in author/title (depending on MARC flavour)
476 if (C4::Context->preference("marcflavour") eq "UNIMARC") {
477 $tmptitle = MARC::Field->new('200',' ',' ', a => $term, f => $occ);
478 $tmprecord->append_fields($tmptitle);
480 $tmptitle = MARC::Field->new('245',' ',' ', a => $term,);
481 $tmpauthor = MARC::Field->new('100',' ',' ', a => $occ,);
482 $tmprecord->append_fields($tmptitle);
483 $tmprecord->append_fields($tmpauthor);
485 $results_hash->{'RECORDS'}[$j] = $tmprecord->as_usmarc();
490 $record = $results[ $i - 1 ]->record($j)->raw();
492 # warn "RECORD $j:".$record;
493 $results_hash->{'RECORDS'}[$j] = $record;
495 # Fill the facets while we're looping, but only for the biblioserver
496 $facet_record = MARC::Record->new_from_usmarc($record)
497 if $servers[ $i - 1 ] =~ /biblioserver/;
499 #warn $servers[$i-1]."\n".$record; #.$facet_record->title();
501 for ( my $k = 0 ; $k <= @$facets ; $k++ ) {
502 ($facets->[$k]) or next;
503 my @fields = map {$facet_record->field($_)} @{$facets->[$k]->{'tags'}} ;
504 for my $field (@fields) {
505 my @subfields = $field->subfields();
506 for my $subfield (@subfields) {
507 my ( $code, $data ) = @$subfield;
508 ($code eq $facets->[$k]->{'subfield'}) or next;
509 $facets_counter->{ $facets->[$k]->{'link_value'} }->{$data}++;
512 $facets_info->{ $facets->[$k]->{'link_value'} }->{'label_value'} =
513 $facets->[$k]->{'label_value'};
514 $facets_info->{ $facets->[$k]->{'link_value'} }->{'expanded'} =
515 $facets->[$k]->{'expanded'};
520 $results_hashref->{ $servers[ $i - 1 ] } = $results_hash;
523 # warn "connection ", $i-1, ": $size hits";
524 # warn $results[$i-1]->record(0)->render() if $size > 0;
527 if ( $servers[ $i - 1 ] =~ /biblioserver/ ) {
529 sort { $facets_counter->{$b} <=> $facets_counter->{$a} }
530 keys %$facets_counter )
533 my $number_of_facets;
534 my @this_facets_array;
537 $facets_counter->{$link_value}
538 ->{$b} <=> $facets_counter->{$link_value}->{$a}
539 } keys %{ $facets_counter->{$link_value} }
543 if ( ( $number_of_facets < 6 )
544 || ( $expanded_facet eq $link_value )
545 || ( $facets_info->{$link_value}->{'expanded'} ) )
548 # Sanitize the link value ), ( will cause errors with CCL,
549 my $facet_link_value = $one_facet;
550 $facet_link_value =~ s/(\(|\))/ /g;
552 # fix the length that will display in the label,
553 my $facet_label_value = $one_facet;
555 substr( $one_facet, 0, 20 ) . "..."
556 unless length($facet_label_value) <= 20;
558 # if it's a branch, label by the name, not the code,
559 if ( $link_value =~ /branch/ ) {
561 $branches->{$one_facet}->{'branchname'};
564 # but we're down with the whole label being in the link's title.
565 my $facet_title_value = $one_facet;
567 push @this_facets_array,
571 $facets_counter->{$link_value}
573 facet_label_value => $facet_label_value,
574 facet_title_value => $facet_title_value,
575 facet_link_value => $facet_link_value,
576 type_link_value => $link_value,
582 # handle expanded option
583 unless ( $facets_info->{$link_value}->{'expanded'} ) {
585 if ( ( $number_of_facets > 6 )
586 && ( $expanded_facet ne $link_value ) );
591 type_link_value => $link_value,
592 type_id => $link_value . "_id",
593 "type_label_" . $facets_info->{$link_value}->{'label_value'} => 1,
594 facets => \@this_facets_array,
595 expandable => $expandable,
596 expand => $link_value,
598 ) unless ( ($facets_info->{$link_value}->{'label_value'} =~ /Libraries/) and (C4::Context->preference('singleBranchMode')) );
603 return ( undef, $results_hashref, \@facets_loop );
608 $koha_query, $simple_query, $sort_by_ref, $servers_ref,
609 $results_per_page, $offset, $expanded_facet, $branches,
613 my $paz = C4::Search::PazPar2->new(C4::Context->config('pazpar2url'));
615 $paz->search($simple_query);
619 my $results_hashref = {};
620 my $stats = XMLin($paz->stat);
621 my $results = XMLin($paz->show($offset, $results_per_page, 'work-title:1'), forcearray => 1);
623 # for a grouped search result, the number of hits
624 # is the number of groups returned; 'bib_hits' will have
625 # the total number of bibs.
626 $results_hashref->{'biblioserver'}->{'hits'} = $results->{'merged'}->[0];
627 $results_hashref->{'biblioserver'}->{'bib_hits'} = $stats->{'hits'};
629 HIT: foreach my $hit (@{ $results->{'hit'} }) {
630 my $recid = $hit->{recid}->[0];
632 my $work_title = $hit->{'md-work-title'}->[0];
634 if (exists $hit->{'md-work-author'}) {
635 $work_author = $hit->{'md-work-author'}->[0];
637 my $group_label = (defined $work_author) ? "$work_title / $work_author" : $work_title;
639 my $result_group = {};
640 $result_group->{'group_label'} = $group_label;
641 $result_group->{'group_merge_key'} = $recid;
644 if (exists $hit->{count}) {
645 $count = $hit->{count}->[0];
647 $result_group->{'group_count'} = $count;
649 for (my $i = 0; $i < $count; $i++) {
650 # FIXME -- may need to worry about diacritics here
651 my $rec = $paz->record($recid, $i);
652 push @{ $result_group->{'RECORDS'} }, $rec;
655 push @{ $results_hashref->{'biblioserver'}->{'GROUPS'} }, $result_group;
658 # pass through facets
659 my $termlist_xml = $paz->termlist('author,subject');
660 my $terms = XMLin($termlist_xml, forcearray => 1);
661 my @facets_loop = ();
662 #die Dumper($results);
663 # foreach my $list (sort keys %{ $terms->{'list'} }) {
665 # foreach my $facet (sort @{ $terms->{'list'}->{$list}->{'term'} } ) {
667 # facet_label_value => $facet->{'name'}->[0],
670 # push @facets_loop, ( {
671 # type_label => $list,
672 # facets => \@facets,
676 return ( undef, $results_hashref, \@facets_loop );
680 sub _remove_stopwords {
681 my ( $operand, $index ) = @_;
682 my @stopwords_removed;
684 # phrase and exact-qualified indexes shouldn't have stopwords removed
685 if ( $index !~ m/phr|ext/ ) {
687 # remove stopwords from operand : parse all stopwords & remove them (case insensitive)
688 # we use IsAlpha unicode definition, to deal correctly with diacritics.
689 # otherwise, a French word like "leçon" woudl be split into "le" "çon", "le"
690 # is a stopword, we'd get "çon" and wouldn't find anything...
691 foreach ( keys %{ C4::Context->stopwords } ) {
692 next if ( $_ =~ /(and|or|not)/ ); # don't remove operators
694 /(\P{IsAlpha}$_\P{IsAlpha}|^$_\P{IsAlpha}|\P{IsAlpha}$_$|^$_$)/ )
696 $operand =~ s/\P{IsAlpha}$_\P{IsAlpha}/ /gi;
697 $operand =~ s/^$_\P{IsAlpha}/ /gi;
698 $operand =~ s/\P{IsAlpha}$_$/ /gi;
699 $operand =~ s/$1//gi;
700 push @stopwords_removed, $_;
704 return ( $operand, \@stopwords_removed );
708 sub _detect_truncation {
709 my ( $operand, $index ) = @_;
710 my ( @nontruncated, @righttruncated, @lefttruncated, @rightlefttruncated,
713 my @wordlist = split( /\s/, $operand );
714 foreach my $word (@wordlist) {
715 if ( $word =~ s/^\*([^\*]+)\*$/$1/ ) {
716 push @rightlefttruncated, $word;
718 elsif ( $word =~ s/^\*([^\*]+)$/$1/ ) {
719 push @lefttruncated, $word;
721 elsif ( $word =~ s/^([^\*]+)\*$/$1/ ) {
722 push @righttruncated, $word;
724 elsif ( index( $word, "*" ) < 0 ) {
725 push @nontruncated, $word;
728 push @regexpr, $word;
732 \@nontruncated, \@righttruncated, \@lefttruncated,
733 \@rightlefttruncated, \@regexpr
738 sub _build_stemmed_operand {
742 # If operand contains a digit, it is almost certainly an identifier, and should
743 # not be stemmed. This is particularly relevant for ISBNs and ISSNs, which
744 # can contain the letter "X" - for example, _build_stemmend_operand would reduce
745 # "014100018X" to "x ", which for a MARC21 database would bring up irrelevant
746 # results (e.g., "23 x 29 cm." from the 300$c). Bug 2098.
747 return $operand if $operand =~ /\d/;
749 # FIXME: the locale should be set based on the user's language and/or search choice
750 my $stemmer = Lingua::Stem->new( -locale => 'EN-US' );
752 # FIXME: these should be stored in the db so the librarian can modify the behavior
753 $stemmer->add_exceptions(
760 my @words = split( / /, $operand );
761 my $stems = $stemmer->stem(@words);
762 for my $stem (@$stems) {
763 $stemmed_operand .= "$stem";
764 $stemmed_operand .= "?"
765 unless ( $stem =~ /(and$|or$|not$)/ ) || ( length($stem) < 3 );
766 $stemmed_operand .= " ";
768 warn "STEMMED OPERAND: $stemmed_operand" if $DEBUG;
769 return $stemmed_operand;
773 sub _build_weighted_query {
775 # FIELD WEIGHTING - This is largely experimental stuff. What I'm committing works
776 # pretty well but could work much better if we had a smarter query parser
777 my ( $operand, $stemmed_operand, $index ) = @_;
778 my $stemming = C4::Context->preference("QueryStemming") || 0;
779 my $weight_fields = C4::Context->preference("QueryWeightFields") || 0;
780 my $fuzzy_enabled = C4::Context->preference("QueryFuzzy") || 0;
782 my $weighted_query .= "(rk=("; # Specifies that we're applying rank
784 # Keyword, or, no index specified
785 if ( ( $index eq 'kw' ) || ( !$index ) ) {
787 "Title-cover,ext,r1=\"$operand\""; # exact title-cover
788 $weighted_query .= " or ti,ext,r2=\"$operand\""; # exact title
789 $weighted_query .= " or ti,phr,r3=\"$operand\""; # phrase title
790 #$weighted_query .= " or any,ext,r4=$operand"; # exact any
791 #$weighted_query .=" or kw,wrdl,r5=\"$operand\""; # word list any
792 $weighted_query .= " or wrdl,fuzzy,r8=\"$operand\""
793 if $fuzzy_enabled; # add fuzzy, word list
794 $weighted_query .= " or wrdl,right-Truncation,r9=\"$stemmed_operand\""
795 if ( $stemming and $stemmed_operand )
796 ; # add stemming, right truncation
797 $weighted_query .= " or wrdl,r9=\"$operand\"";
799 # embedded sorting: 0 a-z; 1 z-a
800 # $weighted_query .= ") or (sort1,aut=1";
803 # Barcode searches should skip this process
804 elsif ( $index eq 'bc' ) {
805 $weighted_query .= "bc=\"$operand\"";
808 # Authority-number searches should skip this process
809 elsif ( $index eq 'an' ) {
810 $weighted_query .= "an=\"$operand\"";
813 # If the index already has more than one qualifier, wrap the operand
814 # in quotes and pass it back (assumption is that the user knows what they
815 # are doing and won't appreciate us mucking up their query
816 elsif ( $index =~ ',' ) {
817 $weighted_query .= " $index=\"$operand\"";
820 #TODO: build better cases based on specific search indexes
822 $weighted_query .= " $index,ext,r1=\"$operand\""; # exact index
823 #$weighted_query .= " or (title-sort-az=0 or $index,startswithnt,st-word,r3=$operand #)";
824 $weighted_query .= " or $index,phr,r3=\"$operand\""; # phrase index
826 " or $index,rt,wrdl,r3=\"$operand\""; # word list index
829 $weighted_query .= "))"; # close rank specification
830 return $weighted_query;
836 $simple_query, $query_cgi,
838 $limit_cgi, $limit_desc,
839 $stopwords_removed, $query_type ) = getRecords ( $operators, $operands, $indexes, $limits, $sort_by, $scan);
841 Build queries and limits in CCL, CGI, Human,
842 handle truncation, stemming, field weighting, stopwords, fuzziness, etc.
844 See verbose embedded documentation.
850 my ( $operators, $operands, $indexes, $limits, $sort_by, $scan ) = @_;
852 warn "---------\nEnter buildQuery\n---------" if $DEBUG;
855 my @operators = @$operators if $operators;
856 my @indexes = @$indexes if $indexes;
857 my @operands = @$operands if $operands;
858 my @limits = @$limits if $limits;
859 my @sort_by = @$sort_by if $sort_by;
861 my $stemming = C4::Context->preference("QueryStemming") || 0;
862 my $auto_truncation = C4::Context->preference("QueryAutoTruncate") || 0;
863 my $weight_fields = C4::Context->preference("QueryWeightFields") || 0;
864 my $fuzzy_enabled = C4::Context->preference("QueryFuzzy") || 0;
865 my $remove_stopwords = C4::Context->preference("QueryRemoveStopwords") || 0;
867 # no stemming/weight/fuzzy in NoZebra
868 if ( C4::Context->preference("NoZebra") ) {
874 my $query = $operands[0];
875 my $simple_query = $operands[0];
877 # initialize the variables we're passing back
886 my $stopwords_removed; # flag to determine if stopwords have been removed
888 # for handling ccl, cql, pqf queries in diagnostic mode, skip the rest of the steps
890 if ( $query =~ /^ccl=/ ) {
891 return ( undef, $', $', $', $', '', '', '', '', 'ccl' );
893 if ( $query =~ /^cql=/ ) {
894 return ( undef, $', $', $', $', '', '', '', '', 'cql' );
896 if ( $query =~ /^pqf=/ ) {
897 return ( undef, $', $', $', $', '', '', '', '', 'pqf' );
900 # pass nested queries directly
901 # FIXME: need better handling of some of these variables in this case
902 if ( $query =~ /(\(|\))/ ) {
904 undef, $query, $simple_query, $query_cgi,
905 $query, $limit, $limit_cgi, $limit_desc,
906 $stopwords_removed, 'ccl'
910 # Form-based queries are non-nested and fixed depth, so we can easily modify the incoming
911 # query operands and indexes and add stemming, truncation, field weighting, etc.
912 # Once we do so, we'll end up with a value in $query, just like if we had an
913 # incoming $query from the user
916 ; # clear it out so we can populate properly with field-weighted, stemmed, etc. query
918 ; # a flag used to keep track if there was a previous query
919 # if there was, we can apply the current operator
921 for ( my $i = 0 ; $i <= @operands ; $i++ ) {
923 # COMBINE OPERANDS, INDEXES AND OPERATORS
924 if ( $operands[$i] ) {
926 # A flag to determine whether or not to add the index to the query
929 # If the user is sophisticated enough to specify an index, turn off field weighting, stemming, and stopword handling
930 if ( $operands[$i] =~ /(:|=)/ || $scan ) {
933 $remove_stopwords = 0;
935 my $operand = $operands[$i];
936 my $index = $indexes[$i];
938 # Add index-specific attributes
939 # Date of Publication
940 if ( $index eq 'yr' ) {
941 $index .= ",st-numeric";
943 $stemming = $auto_truncation = $weight_fields = $fuzzy_enabled = $remove_stopwords = 0;
946 # Date of Acquisition
947 elsif ( $index eq 'acqdate' ) {
948 $index .= ",st-date-normalized";
950 $stemming = $auto_truncation = $weight_fields = $fuzzy_enabled = $remove_stopwords = 0;
952 # ISBN,ISSN,Standard Number, don't need special treatment
953 elsif ( $index eq 'nb' || $index eq 'ns' ) {
956 $stemming, $auto_truncation,
957 $weight_fields, $fuzzy_enabled,
959 ) = ( 0, 0, 0, 0, 0 );
962 # Set default structure attribute (word list)
964 unless ( $indexes_set || !$index || $index =~ /(st-|phr|ext|wrdl)/ ) {
965 $struct_attr = ",wrdl";
968 # Some helpful index variants
969 my $index_plus = $index . $struct_attr . ":" if $index;
970 my $index_plus_comma = $index . $struct_attr . "," if $index;
971 if ($auto_truncation){
972 # FIXME Auto Truncation is only valid for LTR languages
974 # use C4::Languages qw(regex_lang_subtags get_bidi);
975 # $lang = $query->cookie('KohaOpacLanguage') if (defined $query && $query->cookie('KohaOpacLanguage'));
976 # my $current_lang = regex_lang_subtags($lang);
978 # $bidi = get_bidi($current_lang->{script}) if $current_lang->{script};
979 $index_plus_comma .= "rtrn:";
983 if ($remove_stopwords) {
984 ( $operand, $stopwords_removed ) =
985 _remove_stopwords( $operand, $index );
986 warn "OPERAND w/out STOPWORDS: >$operand<" if $DEBUG;
987 warn "REMOVED STOPWORDS: @$stopwords_removed"
988 if ( $stopwords_removed && $DEBUG );
992 my $truncated_operand;
993 my( $nontruncated, $righttruncated, $lefttruncated,
994 $rightlefttruncated, $regexpr
995 ) = _detect_truncation( $operand, $index );
997 "TRUNCATION: NON:>@$nontruncated< RIGHT:>@$righttruncated< LEFT:>@$lefttruncated< RIGHTLEFT:>@$rightlefttruncated< REGEX:>@$regexpr<"
1002 scalar(@$righttruncated) + scalar(@$lefttruncated) +
1003 scalar(@$rightlefttruncated) > 0 )
1006 # Don't field weight or add the index to the query, we do it here
1008 undef $weight_fields;
1009 my $previous_truncation_operand;
1010 if ( scalar(@$nontruncated) > 0 ) {
1011 $truncated_operand .= "$index_plus @$nontruncated ";
1012 $previous_truncation_operand = 1;
1014 if ( scalar(@$righttruncated) > 0 ) {
1015 $truncated_operand .= "and "
1016 if $previous_truncation_operand;
1017 $truncated_operand .=
1018 "$index_plus_comma" . "rtrn:@$righttruncated ";
1019 $previous_truncation_operand = 1;
1021 if ( scalar(@$lefttruncated) > 0 ) {
1022 $truncated_operand .= "and "
1023 if $previous_truncation_operand;
1024 $truncated_operand .=
1025 "$index_plus_comma" . "ltrn:@$lefttruncated ";
1026 $previous_truncation_operand = 1;
1028 if ( scalar(@$rightlefttruncated) > 0 ) {
1029 $truncated_operand .= "and "
1030 if $previous_truncation_operand;
1031 $truncated_operand .=
1032 "$index_plus_comma" . "rltrn:@$rightlefttruncated ";
1033 $previous_truncation_operand = 1;
1036 $operand = $truncated_operand if $truncated_operand;
1037 warn "TRUNCATED OPERAND: >$truncated_operand<" if $DEBUG;
1040 my $stemmed_operand;
1041 $stemmed_operand = _build_stemmed_operand($operand)
1043 warn "STEMMED OPERAND: >$stemmed_operand<" if $DEBUG;
1045 # Handle Field Weighting
1046 my $weighted_operand;
1048 _build_weighted_query( $operand, $stemmed_operand, $index )
1050 warn "FIELD WEIGHTED OPERAND: >$weighted_operand<" if $DEBUG;
1051 $operand = $weighted_operand if $weight_fields;
1052 $indexes_set = 1 if $weight_fields;
1054 # If there's a previous operand, we need to add an operator
1055 if ($previous_operand) {
1057 # User-specified operator
1058 if ( $operators[ $i - 1 ] ) {
1059 $query .= " $operators[$i-1] ";
1060 $query .= " $index_plus " unless $indexes_set;
1061 $query .= " $operand";
1062 $query_cgi .= "&op=$operators[$i-1]";
1063 $query_cgi .= "&idx=$index" if $index;
1064 $query_cgi .= "&q=$operands[$i]" if $operands[$i];
1066 " $operators[$i-1] $index_plus $operands[$i]";
1069 # Default operator is and
1072 $query .= "$index_plus " unless $indexes_set;
1073 $query .= "$operand";
1074 $query_cgi .= "&op=and&idx=$index" if $index;
1075 $query_cgi .= "&q=$operands[$i]" if $operands[$i];
1076 $query_desc .= " and $index_plus $operands[$i]";
1080 # There isn't a pervious operand, don't need an operator
1083 # Field-weighted queries already have indexes set
1084 $query .= " $index_plus " unless $indexes_set;
1086 $query_desc .= " $index_plus $operands[$i]";
1087 $query_cgi .= "&idx=$index" if $index;
1088 $query_cgi .= "&q=$operands[$i]" if $operands[$i];
1089 $previous_operand = 1;
1094 warn "QUERY BEFORE LIMITS: >$query<" if $DEBUG;
1097 my $group_OR_limits;
1098 my $availability_limit;
1099 foreach my $this_limit (@limits) {
1100 if ( $this_limit =~ /available/ ) {
1102 # 'available' is defined as (items.onloan is NULL) and (items.itemlost = 0)
1104 # all records not indexed in the onloan register (zebra) and all records with a value of lost equal to 0
1105 $availability_limit .=
1106 "( ( allrecords,AlwaysMatches='' not onloan,AlwaysMatches='') and (lost,st-numeric=0) )"; #or ( allrecords,AlwaysMatches='' not lost,AlwaysMatches='')) )";
1107 $limit_cgi .= "&limit=available";
1111 # group_OR_limits, prefixed by mc-
1112 # OR every member of the group
1113 elsif ( $this_limit =~ /mc/ ) {
1114 $group_OR_limits .= " or " if $group_OR_limits;
1115 $limit_desc .= " or " if $group_OR_limits;
1116 $group_OR_limits .= "$this_limit";
1117 $limit_cgi .= "&limit=$this_limit";
1118 $limit_desc .= " $this_limit";
1121 # Regular old limits
1124 $limit .= " and " if $limit || $query;
1125 $limit .= "$this_limit";
1126 $limit_cgi .= "&limit=$this_limit";
1127 $limit_desc .= " $this_limit";
1131 if ($group_OR_limits) {
1132 $limit .= " and " if ( $query || $limit );
1133 $limit .= "($group_OR_limits)";
1135 if ($availability_limit) {
1136 $limit .= " and " if ( $query || $limit );
1137 $limit .= "($availability_limit)";
1140 # Normalize the query and limit strings
1143 for ( $query, $query_desc, $limit, $limit_desc ) {
1144 $_ =~ s/ / /g; # remove extra spaces
1145 $_ =~ s/^ //g; # remove any beginning spaces
1146 $_ =~ s/ $//g; # remove any ending spaces
1147 $_ =~ s/==/=/g; # remove double == from query
1149 $query_cgi =~ s/^&//; # remove unnecessary & from beginning of the query cgi
1151 for ($query_cgi,$simple_query) {
1154 # append the limit to the query
1155 $query .= " " . $limit;
1159 warn "QUERY:" . $query;
1160 warn "QUERY CGI:" . $query_cgi;
1161 warn "QUERY DESC:" . $query_desc;
1162 warn "LIMIT:" . $limit;
1163 warn "LIMIT CGI:" . $limit_cgi;
1164 warn "LIMIT DESC:" . $limit_desc;
1165 warn "---------\nLeave buildQuery\n---------";
1168 undef, $query, $simple_query, $query_cgi,
1169 $query_desc, $limit, $limit_cgi, $limit_desc,
1170 $stopwords_removed, $query_type
1174 =head2 searchResults
1176 Format results in a form suitable for passing to the template
1180 # IMO this subroutine is pretty messy still -- it's responsible for
1181 # building the HTML output for the template
1183 my ( $searchdesc, $hits, $results_per_page, $offset, $scan, @marcresults ) = @_;
1184 my $dbh = C4::Context->dbh;
1188 # add search-term highlighting via <span>s on the search terms
1189 my $span_terms_hashref;
1190 for my $span_term ( split( / /, $searchdesc ) ) {
1191 $span_term =~ s/(.*=|\)|\(|\+|\.|\*)//g;
1192 $span_terms_hashref->{$span_term}++;
1195 #Build branchnames hash
1197 #get branch information.....
1200 $dbh->prepare("SELECT branchcode,branchname FROM branches")
1201 ; # FIXME : use C4::Koha::GetBranches
1203 while ( my $bdata = $bsth->fetchrow_hashref ) {
1204 $branches{ $bdata->{'branchcode'} } = $bdata->{'branchname'};
1206 # FIXME - We build an authorised values hash here, using the default framework
1207 # though it is possible to have different authvals for different fws.
1209 my $shelflocations =GetKohaAuthorisedValues('items.location','');
1211 # get notforloan authorised value list (see $shelflocations FIXME)
1212 my $notforloan_authorised_value = GetAuthValCode('items.notforloan','');
1214 #Build itemtype hash
1215 #find itemtype & itemtype image
1219 "SELECT itemtype,description,imageurl,summary,notforloan FROM itemtypes"
1222 while ( my $bdata = $bsth->fetchrow_hashref ) {
1223 foreach (qw(description imageurl summary notforloan)) {
1224 $itemtypes{ $bdata->{'itemtype'} }->{$_} = $bdata->{$_};
1228 #search item field code
1231 "SELECT tagfield FROM marc_subfield_structure WHERE kohafield LIKE 'items.itemnumber'"
1234 my ($itemtag) = $sth->fetchrow;
1236 ## find column names of items related to MARC
1237 my $sth2 = $dbh->prepare("SHOW COLUMNS FROM items");
1239 my %subfieldstosearch;
1240 while ( ( my $column ) = $sth2->fetchrow ) {
1241 my ( $tagfield, $tagsubfield ) =
1242 &GetMarcFromKohaField( "items." . $column, "" );
1243 $subfieldstosearch{$column} = $tagsubfield;
1246 # handle which records to actually retrieve
1248 if ( $hits && $offset + $results_per_page <= $hits ) {
1249 $times = $offset + $results_per_page;
1252 $times = $hits; # FIXME: if $hits is undefined, why do we want to equal it?
1255 # loop through all of the records we've retrieved
1256 for ( my $i = $offset ; $i <= $times - 1 ; $i++ ) {
1257 my $marcrecord = MARC::File::USMARC::decode( $marcresults[$i] );
1258 my $oldbiblio = TransformMarcToKoha( $dbh, $marcrecord, '' );
1259 $oldbiblio->{subtitle} = C4::Biblio::get_koha_field_from_marc('bibliosubtitle', 'subtitle', $marcrecord, '');
1260 $oldbiblio->{result_number} = $i + 1;
1262 # add imageurl to itemtype if there is one
1263 $oldbiblio->{imageurl} = getitemtypeimagelocation( 'opac', $itemtypes{ $oldbiblio->{itemtype} }->{imageurl} );
1265 $oldbiblio->{'authorised_value_images'} = C4::Items::get_authorised_value_images( C4::Biblio::get_biblio_authorised_values( $oldbiblio->{'biblionumber'}, $marcrecord ) );
1266 (my $aisbn) = $oldbiblio->{isbn} =~ /([\d-]*[X]*)/;
1268 $oldbiblio->{amazonisbn} = $aisbn;
1269 $oldbiblio->{description} = $itemtypes{ $oldbiblio->{itemtype} }->{description};
1270 # Build summary if there is one (the summary is defined in the itemtypes table)
1271 # FIXME: is this used anywhere, I think it can be commented out? -- JF
1272 if ( $itemtypes{ $oldbiblio->{itemtype} }->{summary} ) {
1273 my $summary = $itemtypes{ $oldbiblio->{itemtype} }->{summary};
1274 my @fields = $marcrecord->fields();
1275 foreach my $field (@fields) {
1276 my $tag = $field->tag();
1277 my $tagvalue = $field->as_string();
1278 if (! utf8::is_utf8($tagvalue)) {
1279 utf8::decode($tagvalue);
1283 s/\[(.?.?.?.?)$tag\*(.*?)]/$1$tagvalue$2\[$1$tag$2]/g;
1284 unless ( $tag < 10 ) {
1285 my @subf = $field->subfields;
1286 for my $i ( 0 .. $#subf ) {
1287 my $subfieldcode = $subf[$i][0];
1288 my $subfieldvalue = $subf[$i][1];
1289 if (! utf8::is_utf8($subfieldvalue)) {
1290 utf8::decode($subfieldvalue);
1292 my $tagsubf = $tag . $subfieldcode;
1294 s/\[(.?.?.?.?)$tagsubf(.*?)]/$1$subfieldvalue$2\[$1$tagsubf$2]/g;
1299 $summary =~ s/\[(.*?)]//g;
1300 $summary =~ s/\n/<br\/>/g;
1301 $oldbiblio->{summary} = $summary;
1304 # save an author with no <span> tag, for the <a href=search.pl?q=<!--tmpl_var name="author"-->> link
1305 $oldbiblio->{'author_nospan'} = $oldbiblio->{'author'};
1306 $oldbiblio->{'title_nospan'} = $oldbiblio->{'title'};
1307 $oldbiblio->{'subtitle_nospan'} = $oldbiblio->{'subtitle'};
1308 # Add search-term highlighting to the whole record where they match using <span>s
1309 if (C4::Context->preference("OpacHighlightedWords")){
1310 my $searchhighlightblob;
1311 for my $highlight_field ( $marcrecord->fields ) {
1313 # FIXME: need to skip title, subtitle, author, etc., as they are handled below
1314 next if $highlight_field->tag() =~ /(^00)/; # skip fixed fields
1315 for my $subfield ($highlight_field->subfields()) {
1317 next if $subfield->[0] eq '9';
1318 my $field = $subfield->[1];
1319 for my $term ( keys %$span_terms_hashref ) {
1320 if ( ( $field =~ /$term/i ) && (( length($term) > 3 ) || ($field =~ / $term /i)) ) {
1321 $field =~ s/$term/<span class=\"term\">$&<\/span>/gi;
1325 $searchhighlightblob .= $field . " ... " if $match;
1329 $searchhighlightblob = ' ... '.$searchhighlightblob if $searchhighlightblob;
1330 $oldbiblio->{'searchhighlightblob'} = $searchhighlightblob;
1333 # Add search-term highlighting to the title, subtitle, etc. fields
1334 for my $term ( keys %$span_terms_hashref ) {
1335 my $old_term = $term;
1336 if ( length($term) > 3 ) {
1337 $term =~ s/(.*=|\)|\(|\+|\.|\?|\[|\]|\\|\*)//g;
1338 foreach(qw(title subtitle author publishercode place pages notes size)) {
1339 $oldbiblio->{$_} =~ s/$term/<span class=\"term\">$&<\/span>/gi;
1344 ($i % 2) and $oldbiblio->{'toggle'} = 1;
1346 # Pull out the items fields
1347 my @fields = $marcrecord->field($itemtag);
1349 # Setting item statuses for display
1350 my @available_items_loop;
1351 my @onloan_items_loop;
1352 my @other_items_loop;
1354 my $available_items;
1358 my $ordered_count = 0;
1359 my $available_count = 0;
1360 my $onloan_count = 0;
1361 my $longoverdue_count = 0;
1362 my $other_count = 0;
1363 my $wthdrawn_count = 0;
1364 my $itemlost_count = 0;
1365 my $itembinding_count = 0;
1366 my $itemdamaged_count = 0;
1367 my $item_in_transit_count = 0;
1368 my $can_place_holds = 0;
1369 my $items_count = scalar(@fields);
1371 ( C4::Context->preference('maxItemsinSearchResults') )
1372 ? C4::Context->preference('maxItemsinSearchResults') - 1
1375 # loop through every item
1376 foreach my $field (@fields) {
1379 # populate the items hash
1380 foreach my $code ( keys %subfieldstosearch ) {
1381 $item->{$code} = $field->subfield( $subfieldstosearch{$code} );
1383 my $hbranch = C4::Context->preference('HomeOrHoldingBranch') eq 'homebranch' ? 'homebranch' : 'holdingbranch';
1384 my $otherbranch = C4::Context->preference('HomeOrHoldingBranch') eq 'homebranch' ? 'holdingbranch' : 'homebranch';
1385 # set item's branch name, use HomeOrHoldingBranch syspref first, fall back to the other one
1386 if ($item->{$hbranch}) {
1387 $item->{'branchname'} = $branches{$item->{$hbranch}};
1389 elsif ($item->{$otherbranch}) { # Last resort
1390 $item->{'branchname'} = $branches{$item->{$otherbranch}};
1393 my $prefix = $item->{$hbranch} . '--' . $item->{location} . $item->{itype} . $item->{itemcallnumber};
1394 # For each grouping of items (onloan, available, unavailable), we build a key to store relevant info about that item
1395 if ( $item->{onloan} ) {
1397 my $key = $prefix . $item->{due_date};
1398 $onloan_items->{$key}->{due_date} = format_date($item->{onloan});
1399 $onloan_items->{$key}->{count}++ if $item->{$hbranch};
1400 $onloan_items->{$key}->{branchname} = $item->{branchname};
1401 $onloan_items->{$key}->{location} = $shelflocations->{ $item->{location} };
1402 $onloan_items->{$key}->{itemcallnumber} = $item->{itemcallnumber};
1403 $onloan_items->{$key}->{imageurl} = getitemtypeimagelocation( 'opac', $itemtypes{ $item->{itype} }->{imageurl} );
1404 $onloan_items->{$key}->{barcode} = $item->{barcode};
1405 # if something's checked out and lost, mark it as 'long overdue'
1406 if ( $item->{itemlost} ) {
1407 $onloan_items->{$prefix}->{longoverdue}++;
1408 $longoverdue_count++;
1409 } else { # can place holds as long as item isn't lost
1410 $can_place_holds = 1;
1414 # items not on loan, but still unavailable ( lost, withdrawn, damaged )
1418 if ( $item->{notforloan} == -1 ) {
1422 # is item in transit?
1423 my $transfertwhen = '';
1424 my ($transfertfrom, $transfertto);
1426 unless ($item->{wthdrawn}
1427 || $item->{itemlost}
1429 || $item->{notforloan}
1430 || $items_count > 20) {
1432 # A couple heuristics to limit how many times
1433 # we query the database for item transfer information, sacrificing
1434 # accuracy in some cases for speed;
1436 # 1. don't query if item has one of the other statuses
1437 # 2. don't check transit status if the bib has
1438 # more than 20 items
1440 # FIXME: to avoid having the query the database like this, and to make
1441 # the in transit status count as unavailable for search limiting,
1442 # should map transit status to record indexed in Zebra.
1444 ($transfertwhen, $transfertfrom, $transfertto) = C4::Circulation::GetTransfers($item->{itemnumber});
1447 # item is withdrawn, lost or damaged
1448 if ( $item->{wthdrawn}
1449 || $item->{itemlost}
1451 || $item->{notforloan}
1452 || ($transfertwhen ne ''))
1454 $wthdrawn_count++ if $item->{wthdrawn};
1455 $itemlost_count++ if $item->{itemlost};
1456 $itemdamaged_count++ if $item->{damaged};
1457 $item_in_transit_count++ if $transfertwhen ne '';
1458 $item->{status} = $item->{wthdrawn} . "-" . $item->{itemlost} . "-" . $item->{damaged} . "-" . $item->{notforloan};
1461 my $key = $prefix . $item->{status};
1462 foreach (qw(wthdrawn itemlost damaged branchname itemcallnumber)) {
1463 $other_items->{$key}->{$_} = $item->{$_};
1465 $other_items->{$key}->{intransit} = ($transfertwhen ne '') ? 1 : 0;
1466 $other_items->{$key}->{notforloan} = GetAuthorisedValueDesc('','',$item->{notforloan},'','',$notforloan_authorised_value) if $notforloan_authorised_value;
1467 $other_items->{$key}->{count}++ if $item->{$hbranch};
1468 $other_items->{$key}->{location} = $shelflocations->{ $item->{location} };
1469 $other_items->{$key}->{imageurl} = getitemtypeimagelocation( 'opac', $itemtypes{ $item->{itype} }->{imageurl} );
1470 $other_items->{$key}->{barcode} = $item->{barcode};
1474 $can_place_holds = 1;
1476 $available_items->{$prefix}->{count}++ if $item->{$hbranch};
1477 foreach (qw(branchname itemcallnumber barcode)) {
1478 $available_items->{$prefix}->{$_} = $item->{$_};
1480 $available_items->{$prefix}->{location} = $shelflocations->{ $item->{location} };
1481 $available_items->{$prefix}->{imageurl} = getitemtypeimagelocation( 'opac', $itemtypes{ $item->{itype} }->{imageurl} );
1484 } # notforloan, item level and biblioitem level
1485 my ( $availableitemscount, $onloanitemscount, $otheritemscount );
1487 ( C4::Context->preference('maxItemsinSearchResults') )
1488 ? C4::Context->preference('maxItemsinSearchResults') - 1
1490 for my $key ( sort keys %$onloan_items ) {
1491 (++$onloanitemscount > $maxitems) and last;
1492 push @onloan_items_loop, $onloan_items->{$key};
1494 for my $key ( sort keys %$other_items ) {
1495 (++$otheritemscount > $maxitems) and last;
1496 push @other_items_loop, $other_items->{$key};
1498 for my $key ( sort keys %$available_items ) {
1499 (++$availableitemscount > $maxitems) and last;
1500 push @available_items_loop, $available_items->{$key}
1503 # XSLT processing of some stuff
1504 if (C4::Context->preference("XSLTResultsDisplay") && !$scan) {
1505 my $newxmlrecord = XSLTParse4Display($oldbiblio->{biblionumber}, $marcrecord, C4::Context->config('opachtdocs')."/prog/en/xslt/MARC21slim2OPACResults.xsl");
1506 $oldbiblio->{XSLTResultsRecord} = $newxmlrecord;
1509 # last check for norequest : if itemtype is notforloan, it can't be reserved either, whatever the items
1510 $can_place_holds = 0
1511 if $itemtypes{ $oldbiblio->{itemtype} }->{notforloan};
1512 $oldbiblio->{norequests} = 1 unless $can_place_holds;
1513 $oldbiblio->{itemsplural} = 1 if $items_count > 1;
1514 $oldbiblio->{items_count} = $items_count;
1515 $oldbiblio->{available_items_loop} = \@available_items_loop;
1516 $oldbiblio->{onloan_items_loop} = \@onloan_items_loop;
1517 $oldbiblio->{other_items_loop} = \@other_items_loop;
1518 $oldbiblio->{availablecount} = $available_count;
1519 $oldbiblio->{availableplural} = 1 if $available_count > 1;
1520 $oldbiblio->{onloancount} = $onloan_count;
1521 $oldbiblio->{onloanplural} = 1 if $onloan_count > 1;
1522 $oldbiblio->{othercount} = $other_count;
1523 $oldbiblio->{otherplural} = 1 if $other_count > 1;
1524 $oldbiblio->{wthdrawncount} = $wthdrawn_count;
1525 $oldbiblio->{itemlostcount} = $itemlost_count;
1526 $oldbiblio->{damagedcount} = $itemdamaged_count;
1527 $oldbiblio->{intransitcount} = $item_in_transit_count;
1528 $oldbiblio->{orderedcount} = $ordered_count;
1529 $oldbiblio->{isbn} =~
1530 s/-//g; # deleting - in isbn to enable amazon content
1531 push( @newresults, $oldbiblio );
1536 #----------------------------------------------------------------------
1538 # Non-Zebra GetRecords#
1539 #----------------------------------------------------------------------
1543 NZgetRecords has the same API as zera getRecords, even if some parameters are not managed
1549 $query, $simple_query, $sort_by_ref, $servers_ref,
1550 $results_per_page, $offset, $expanded_facet, $branches,
1553 warn "query =$query" if $DEBUG;
1554 my $result = NZanalyse($query);
1555 warn "results =$result" if $DEBUG;
1557 NZorder( $result, @$sort_by_ref[0], $results_per_page, $offset ),
1563 NZanalyse : get a CQL string as parameter, and returns a list of biblionumber;title,biblionumber;title,...
1564 the list is built from an inverted index in the nozebra SQL table
1565 note that title is here only for convenience : the sorting will be very fast when requested on title
1566 if the sorting is requested on something else, we will have to reread all results, and that may be longer.
1571 my ( $string, $server ) = @_;
1572 # warn "---------" if $DEBUG;
1573 warn " NZanalyse" if $DEBUG;
1574 # warn "---------" if $DEBUG;
1576 # $server contains biblioserver or authorities, depending on what we search on.
1577 #warn "querying : $string on $server";
1578 $server = 'biblioserver' unless $server;
1580 # if we have a ", replace the content to discard temporarily any and/or/not inside
1582 if ( $string =~ /"/ ) {
1583 $string =~ s/"(.*?)"/__X__/;
1585 warn "commacontent : $commacontent" if $DEBUG;
1588 # split the query string in 3 parts : X AND Y means : $left="X", $operand="AND" and $right="Y"
1589 # then, call again NZanalyse with $left and $right
1590 # (recursive until we find a leaf (=> something without and/or/not)
1591 # delete repeated operator... Would then go in infinite loop
1592 while ( $string =~ s/( and| or| not| AND| OR| NOT)\1/$1/g ) {
1595 #process parenthesis before.
1596 if ( $string =~ /^\s*\((.*)\)(( and | or | not | AND | OR | NOT )(.*))?/ ) {
1599 my $operator = lc($3); # FIXME: and/or/not are operators, not operands
1601 "dealing w/parenthesis before recursive sub call. left :$left operator:$operator right:$right"
1603 my $leftresult = NZanalyse( $left, $server );
1605 my $rightresult = NZanalyse( $right, $server );
1607 # OK, we have the results for right and left part of the query
1608 # depending of operand, intersect, union or exclude both lists
1609 # to get a result list
1610 if ( $operator eq ' and ' ) {
1611 return NZoperatorAND($leftresult,$rightresult);
1613 elsif ( $operator eq ' or ' ) {
1615 # just merge the 2 strings
1616 return $leftresult . $rightresult;
1618 elsif ( $operator eq ' not ' ) {
1619 return NZoperatorNOT($leftresult,$rightresult);
1623 # this error is impossible, because of the regexp that isolate the operand, but just in case...
1627 warn "string :" . $string if $DEBUG;
1631 if ($string =~ /(.*?)( and | or | not | AND | OR | NOT )(.*)/) {
1634 $operator = lc($2); # FIXME: and/or/not are operators, not operands
1636 warn "no parenthesis. left : $left operator: $operator right: $right"
1639 # it's not a leaf, we have a and/or/not
1642 # reintroduce comma content if needed
1643 $right =~ s/__X__/"$commacontent"/ if $commacontent;
1644 $left =~ s/__X__/"$commacontent"/ if $commacontent;
1645 warn "node : $left / $operator / $right\n" if $DEBUG;
1646 my $leftresult = NZanalyse( $left, $server );
1647 my $rightresult = NZanalyse( $right, $server );
1648 warn " leftresult : $leftresult" if $DEBUG;
1649 warn " rightresult : $rightresult" if $DEBUG;
1650 # OK, we have the results for right and left part of the query
1651 # depending of operand, intersect, union or exclude both lists
1652 # to get a result list
1653 if ( $operator eq ' and ' ) {
1655 return NZoperatorAND($leftresult,$rightresult);
1657 elsif ( $operator eq ' or ' ) {
1659 # just merge the 2 strings
1660 return $leftresult . $rightresult;
1662 elsif ( $operator eq ' not ' ) {
1663 return NZoperatorNOT($leftresult,$rightresult);
1667 # this error is impossible, because of the regexp that isolate the operand, but just in case...
1668 die "error : operand unknown : $operator for $string";
1671 # it's a leaf, do the real SQL query and return the result
1674 $string =~ s/__X__/"$commacontent"/ if $commacontent;
1675 $string =~ s/-|\.|\?|,|;|!|'|\(|\)|\[|\]|{|}|"|&|\+|\*|\// /g;
1676 #remove trailing blank at the beginning
1678 warn "leaf:$string" if $DEBUG;
1680 # parse the string in in operator/operand/value again
1684 if ($string =~ /(.*)(>=|<=)(.*)/) {
1691 # warn "handling leaf... left:$left operator:$operator right:$right"
1693 unless ($operator) {
1694 if ($string =~ /(.*)(>|<|=)(.*)/) {
1699 "handling unless (operator)... left:$left operator:$operator right:$right"
1707 # strip adv, zebra keywords, currently not handled in nozebra: wrdl, ext, phr...
1710 # automatic replace for short operators
1711 $left = 'title' if $left =~ '^ti$';
1712 $left = 'author' if $left =~ '^au$';
1713 $left = 'publisher' if $left =~ '^pb$';
1714 $left = 'subject' if $left =~ '^su$';
1715 $left = 'koha-Auth-Number' if $left =~ '^an$';
1716 $left = 'keyword' if $left =~ '^kw$';
1717 warn "handling leaf... left:$left operator:$operator right:$right" if $DEBUG;
1718 if ( $operator && $left ne 'keyword' ) {
1720 #do a specific search
1721 my $dbh = C4::Context->dbh;
1722 $operator = 'LIKE' if $operator eq '=' and $right =~ /%/;
1725 "SELECT biblionumbers,value FROM nozebra WHERE server=? AND indexname=? AND value $operator ?"
1727 warn "$left / $operator / $right\n" if $DEBUG;
1729 # split each word, query the DB and build the biblionumbers result
1730 #sanitizing leftpart
1731 $left =~ s/^\s+|\s+$//;
1732 foreach ( split / /, $right ) {
1734 $_ =~ s/^\s+|\s+$//;
1736 warn "EXECUTE : $server, $left, $_" if $DEBUG;
1737 $sth->execute( $server, $left, $_ )
1738 or warn "execute failed: $!";
1739 while ( my ( $line, $value ) = $sth->fetchrow ) {
1741 # if we are dealing with a numeric value, use only numeric results (in case of >=, <=, > or <)
1742 # otherwise, fill the result
1743 $biblionumbers .= $line
1744 unless ( $right =~ /^\d+$/ && $value =~ /\D/ );
1745 warn "result : $value "
1746 . ( $right =~ /\d/ ) . "=="
1747 . ( $value =~ /\D/?$line:"" ) if $DEBUG; #= $line";
1750 # do a AND with existing list if there is one, otherwise, use the biblionumbers list as 1st result list
1752 warn "NZAND" if $DEBUG;
1753 $results = NZoperatorAND($biblionumbers,$results);
1756 $results = $biblionumbers;
1762 #do a complete search (all indexes), if index='kw' do complete search too.
1763 my $dbh = C4::Context->dbh;
1766 "SELECT biblionumbers FROM nozebra WHERE server=? AND value LIKE ?"
1769 # split each word, query the DB and build the biblionumbers result
1770 foreach ( split / /, $string ) {
1771 next if C4::Context->stopwords->{ uc($_) }; # skip if stopword
1772 warn "search on all indexes on $_" if $DEBUG;
1775 $sth->execute( $server, $_ );
1776 while ( my $line = $sth->fetchrow ) {
1777 $biblionumbers .= $line;
1780 # do a AND with existing list if there is one, otherwise, use the biblionumbers list as 1st result list
1782 $results = NZoperatorAND($biblionumbers,$results);
1785 warn "NEW RES for $_ = $biblionumbers" if $DEBUG;
1786 $results = $biblionumbers;
1790 warn "return : $results for LEAF : $string" if $DEBUG;
1793 warn "---------\nLeave NZanalyse\n---------" if $DEBUG;
1797 my ($rightresult, $leftresult)=@_;
1799 my @leftresult = split /;/, $leftresult;
1800 warn " @leftresult / $rightresult \n" if $DEBUG;
1802 # my @rightresult = split /;/,$leftresult;
1805 # parse the left results, and if the biblionumber exist in the right result, save it in finalresult
1806 # the result is stored twice, to have the same weight for AND than OR.
1807 # example : TWO : 61,61,64,121 (two is twice in the biblio #61) / TOWER : 61,64,130
1808 # result : 61,61,61,61,64,64 for two AND tower : 61 has more weight than 64
1809 foreach (@leftresult) {
1812 ( $value, $countvalue ) = ( $1, $2 ) if ($value=~/(.*)-(\d+)$/);
1813 if ( $rightresult =~ /\Q$value\E-(\d+);/ ) {
1814 $countvalue = ( $1 > $countvalue ? $countvalue : $1 );
1816 "$value-$countvalue;$value-$countvalue;";
1819 warn "NZAND DONE : $finalresult \n" if $DEBUG;
1820 return $finalresult;
1824 my ($rightresult, $leftresult)=@_;
1825 return $rightresult.$leftresult;
1829 my ($leftresult, $rightresult)=@_;
1831 my @leftresult = split /;/, $leftresult;
1833 # my @rightresult = split /;/,$leftresult;
1835 foreach (@leftresult) {
1837 $value=$1 if $value=~m/(.*)-\d+$/;
1838 unless ($rightresult =~ "$value-") {
1839 $finalresult .= "$_;";
1842 return $finalresult;
1847 $finalresult = NZorder($biblionumbers, $ordering,$results_per_page,$offset);
1854 my ( $biblionumbers, $ordering, $results_per_page, $offset ) = @_;
1855 warn "biblionumbers = $biblionumbers and ordering = $ordering\n" if $DEBUG;
1857 # order title asc by default
1858 # $ordering = '1=36 <i' unless $ordering;
1859 $results_per_page = 20 unless $results_per_page;
1860 $offset = 0 unless $offset;
1861 my $dbh = C4::Context->dbh;
1864 # order by POPULARITY
1866 if ( $ordering =~ /popularity/ ) {
1870 # popularity is not in MARC record, it's builded from a specific query
1872 $dbh->prepare("select sum(issues) from items where biblionumber=?");
1873 foreach ( split /;/, $biblionumbers ) {
1874 my ( $biblionumber, $title ) = split /,/, $_;
1875 $result{$biblionumber} = GetMarcBiblio($biblionumber);
1876 $sth->execute($biblionumber);
1877 my $popularity = $sth->fetchrow || 0;
1879 # hint : the key is popularity.title because we can have
1880 # many results with the same popularity. In this cas, sub-ordering is done by title
1881 # we also have biblionumber to avoid bug for 2 biblios with the same title & popularity
1882 # (un-frequent, I agree, but we won't forget anything that way ;-)
1883 $popularity{ sprintf( "%10d", $popularity ) . $title
1884 . $biblionumber } = $biblionumber;
1887 # sort the hash and return the same structure as GetRecords (Zebra querying)
1890 if ( $ordering eq 'popularity_dsc' ) { # sort popularity DESC
1891 foreach my $key ( sort { $b cmp $a } ( keys %popularity ) ) {
1892 $result_hash->{'RECORDS'}[ $numbers++ ] =
1893 $result{ $popularity{$key} }->as_usmarc();
1896 else { # sort popularity ASC
1897 foreach my $key ( sort ( keys %popularity ) ) {
1898 $result_hash->{'RECORDS'}[ $numbers++ ] =
1899 $result{ $popularity{$key} }->as_usmarc();
1902 my $finalresult = ();
1903 $result_hash->{'hits'} = $numbers;
1904 $finalresult->{'biblioserver'} = $result_hash;
1905 return $finalresult;
1911 elsif ( $ordering =~ /author/ ) {
1913 foreach ( split /;/, $biblionumbers ) {
1914 my ( $biblionumber, $title ) = split /,/, $_;
1915 my $record = GetMarcBiblio($biblionumber);
1917 if ( C4::Context->preference('marcflavour') eq 'UNIMARC' ) {
1918 $author = $record->subfield( '200', 'f' );
1919 $author = $record->subfield( '700', 'a' ) unless $author;
1922 $author = $record->subfield( '100', 'a' );
1925 # hint : the result is sorted by title.biblionumber because we can have X biblios with the same title
1926 # and we don't want to get only 1 result for each of them !!!
1927 $result{ $author . $biblionumber } = $record;
1930 # sort the hash and return the same structure as GetRecords (Zebra querying)
1933 if ( $ordering eq 'author_za' ) { # sort by author desc
1934 foreach my $key ( sort { $b cmp $a } ( keys %result ) ) {
1935 $result_hash->{'RECORDS'}[ $numbers++ ] =
1936 $result{$key}->as_usmarc();
1939 else { # sort by author ASC
1940 foreach my $key ( sort ( keys %result ) ) {
1941 $result_hash->{'RECORDS'}[ $numbers++ ] =
1942 $result{$key}->as_usmarc();
1945 my $finalresult = ();
1946 $result_hash->{'hits'} = $numbers;
1947 $finalresult->{'biblioserver'} = $result_hash;
1948 return $finalresult;
1951 # ORDER BY callnumber
1954 elsif ( $ordering =~ /callnumber/ ) {
1956 foreach ( split /;/, $biblionumbers ) {
1957 my ( $biblionumber, $title ) = split /,/, $_;
1958 my $record = GetMarcBiblio($biblionumber);
1960 my ( $callnumber_tag, $callnumber_subfield ) =
1961 GetMarcFromKohaField( 'items.itemcallnumber','' );
1962 ( $callnumber_tag, $callnumber_subfield ) =
1963 GetMarcFromKohaField('biblioitems.callnumber','')
1964 unless $callnumber_tag;
1965 if ( C4::Context->preference('marcflavour') eq 'UNIMARC' ) {
1966 $callnumber = $record->subfield( '200', 'f' );
1969 $callnumber = $record->subfield( '100', 'a' );
1972 # hint : the result is sorted by title.biblionumber because we can have X biblios with the same title
1973 # and we don't want to get only 1 result for each of them !!!
1974 $result{ $callnumber . $biblionumber } = $record;
1977 # sort the hash and return the same structure as GetRecords (Zebra querying)
1980 if ( $ordering eq 'call_number_dsc' ) { # sort by title desc
1981 foreach my $key ( sort { $b cmp $a } ( keys %result ) ) {
1982 $result_hash->{'RECORDS'}[ $numbers++ ] =
1983 $result{$key}->as_usmarc();
1986 else { # sort by title ASC
1987 foreach my $key ( sort { $a cmp $b } ( keys %result ) ) {
1988 $result_hash->{'RECORDS'}[ $numbers++ ] =
1989 $result{$key}->as_usmarc();
1992 my $finalresult = ();
1993 $result_hash->{'hits'} = $numbers;
1994 $finalresult->{'biblioserver'} = $result_hash;
1995 return $finalresult;
1997 elsif ( $ordering =~ /pubdate/ ) { #pub year
1999 foreach ( split /;/, $biblionumbers ) {
2000 my ( $biblionumber, $title ) = split /,/, $_;
2001 my $record = GetMarcBiblio($biblionumber);
2002 my ( $publicationyear_tag, $publicationyear_subfield ) =
2003 GetMarcFromKohaField( 'biblioitems.publicationyear', '' );
2004 my $publicationyear =
2005 $record->subfield( $publicationyear_tag,
2006 $publicationyear_subfield );
2008 # hint : the result is sorted by title.biblionumber because we can have X biblios with the same title
2009 # and we don't want to get only 1 result for each of them !!!
2010 $result{ $publicationyear . $biblionumber } = $record;
2013 # sort the hash and return the same structure as GetRecords (Zebra querying)
2016 if ( $ordering eq 'pubdate_dsc' ) { # sort by pubyear desc
2017 foreach my $key ( sort { $b cmp $a } ( keys %result ) ) {
2018 $result_hash->{'RECORDS'}[ $numbers++ ] =
2019 $result{$key}->as_usmarc();
2022 else { # sort by pub year ASC
2023 foreach my $key ( sort ( keys %result ) ) {
2024 $result_hash->{'RECORDS'}[ $numbers++ ] =
2025 $result{$key}->as_usmarc();
2028 my $finalresult = ();
2029 $result_hash->{'hits'} = $numbers;
2030 $finalresult->{'biblioserver'} = $result_hash;
2031 return $finalresult;
2037 elsif ( $ordering =~ /title/ ) {
2039 # the title is in the biblionumbers string, so we just need to build a hash, sort it and return
2041 foreach ( split /;/, $biblionumbers ) {
2042 my ( $biblionumber, $title ) = split /,/, $_;
2044 # hint : the result is sorted by title.biblionumber because we can have X biblios with the same title
2045 # and we don't want to get only 1 result for each of them !!!
2046 # hint & speed improvement : we can order without reading the record
2047 # so order, and read records only for the requested page !
2048 $result{ $title . $biblionumber } = $biblionumber;
2051 # sort the hash and return the same structure as GetRecords (Zebra querying)
2054 if ( $ordering eq 'title_az' ) { # sort by title desc
2055 foreach my $key ( sort ( keys %result ) ) {
2056 $result_hash->{'RECORDS'}[ $numbers++ ] = $result{$key};
2059 else { # sort by title ASC
2060 foreach my $key ( sort { $b cmp $a } ( keys %result ) ) {
2061 $result_hash->{'RECORDS'}[ $numbers++ ] = $result{$key};
2065 # limit the $results_per_page to result size if it's more
2066 $results_per_page = $numbers - 1 if $numbers < $results_per_page;
2068 # for the requested page, replace biblionumber by the complete record
2069 # speed improvement : avoid reading too much things
2071 my $counter = $offset ;
2072 $counter <= $offset + $results_per_page ;
2076 $result_hash->{'RECORDS'}[$counter] =
2077 GetMarcBiblio( $result_hash->{'RECORDS'}[$counter] )->as_usmarc;
2079 my $finalresult = ();
2080 $result_hash->{'hits'} = $numbers;
2081 $finalresult->{'biblioserver'} = $result_hash;
2082 return $finalresult;
2089 # we need 2 hashes to order by ranking : the 1st one to count the ranking, the 2nd to order by ranking
2092 foreach ( split /;/, $biblionumbers ) {
2093 my ( $biblionumber, $title ) = split /,/, $_;
2094 $title =~ /(.*)-(\d)/;
2099 # note that we + the ranking because ranking is calculated on weight of EACH term requested.
2100 # if we ask for "two towers", and "two" has weight 2 in biblio N, and "towers" has weight 4 in biblio N
2101 # biblio N has ranking = 6
2102 $count_ranking{$biblionumber} += $ranking;
2105 # build the result by "inverting" the count_ranking hash
2106 # hing : as usual, we don't order by ranking only, to avoid having only 1 result for each rank. We build an hash on concat(ranking,biblionumber) instead
2108 foreach ( keys %count_ranking ) {
2109 $result{ sprintf( "%10d", $count_ranking{$_} ) . '-' . $_ } = $_;
2112 # sort the hash and return the same structure as GetRecords (Zebra querying)
2115 foreach my $key ( sort { $b cmp $a } ( keys %result ) ) {
2116 $result_hash->{'RECORDS'}[ $numbers++ ] = $result{$key};
2119 # limit the $results_per_page to result size if it's more
2120 $results_per_page = $numbers - 1 if $numbers < $results_per_page;
2122 # for the requested page, replace biblionumber by the complete record
2123 # speed improvement : avoid reading too much things
2125 my $counter = $offset ;
2126 $counter <= $offset + $results_per_page ;
2130 $result_hash->{'RECORDS'}[$counter] =
2131 GetMarcBiblio( $result_hash->{'RECORDS'}[$counter] )->as_usmarc
2132 if $result_hash->{'RECORDS'}[$counter];
2134 my $finalresult = ();
2135 $result_hash->{'hits'} = $numbers;
2136 $finalresult->{'biblioserver'} = $result_hash;
2137 return $finalresult;
2141 END { } # module clean-up code here (global destructor)
2148 Koha Developement team <info@koha.org>