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);
31 use vars qw($VERSION @ISA @EXPORT @EXPORT_OK %EXPORT_TAGS $DEBUG);
33 # set the version for version checking
36 $DEBUG = ($ENV{DEBUG}) ? 1 : 0;
41 C4::Search - Functions for searching the Koha catalog.
45 See opac/opac-search.pl or catalogue/search.pl for example of usage
49 This module provides searching functions for Koha's bibliographic databases
65 # make all your functions, whether exported or not;
69 ($biblionumber,$biblionumber,$title) = FindDuplicate($record);
71 This function attempts to find duplicate records using a hard-coded, fairly simplistic algorithm
77 my $dbh = C4::Context->dbh;
78 my $result = TransformMarcToKoha( $dbh, $record, '' );
83 my ( $biblionumber, $title );
85 # search duplicate on ISBN, easy and fast..
87 if ( $result->{isbn} ) {
88 $result->{isbn} =~ s/\(.*$//;
89 $result->{isbn} =~ s/\s+$//;
90 $query = "isbn=$result->{isbn}";
93 $result->{title} =~ s /\\//g;
94 $result->{title} =~ s /\"//g;
95 $result->{title} =~ s /\(//g;
96 $result->{title} =~ s /\)//g;
98 # FIXME: instead of removing operators, could just do
99 # quotes around the value
100 $result->{title} =~ s/(and|or|not)//g;
101 $query = "ti,ext=$result->{title}";
102 $query .= " and itemtype=$result->{itemtype}"
103 if ( $result->{itemtype} );
104 if ( $result->{author} ) {
105 $result->{author} =~ s /\\//g;
106 $result->{author} =~ s /\"//g;
107 $result->{author} =~ s /\(//g;
108 $result->{author} =~ s /\)//g;
110 # remove valid operators
111 $result->{author} =~ s/(and|or|not)//g;
112 $query .= " and au,ext=$result->{author}";
116 # FIXME: add error handling
117 my ( $error, $searchresults ) = SimpleSearch($query); # FIXME :: hardcoded !
119 foreach my $possible_duplicate_record (@$searchresults) {
121 MARC::Record->new_from_usmarc($possible_duplicate_record);
122 my $result = TransformMarcToKoha( $dbh, $marcrecord, '' );
124 # FIXME :: why 2 $biblionumber ?
126 push @results, $result->{'biblionumber'};
127 push @results, $result->{'title'};
135 ( $error, $results, $total_hits ) = SimpleSearch( $query, $offset, $max_results, [@servers] );
137 This function provides a simple search API on the bibliographic catalog
143 * $query can be a simple keyword or a complete CCL query
144 * @servers is optional. Defaults to biblioserver as found in koha-conf.xml
145 * $offset - If present, represents the number of records at the beggining to omit. Defaults to 0
146 * $max_results - if present, determines the maximum number of records to fetch. undef is All. defaults to undef.
151 * $error is a empty unless an error is detected
152 * \@results is an array of records.
153 * $total_hits is the number of hits that would have been returned with no limit
155 =item C<usage in the script:>
159 my ( $error, $marcresults, $total_hits ) = SimpleSearch($query);
161 if (defined $error) {
162 $template->param(query_error => $error);
163 warn "error: ".$error;
164 output_html_with_http_headers $input, $cookie, $template->output;
168 my $hits = scalar @$marcresults;
171 for my $i (0..$hits) {
173 my $marcrecord = MARC::File::USMARC::decode($marcresults->[$i]);
174 my $biblio = TransformMarcToKoha(C4::Context->dbh,$marcrecord,'');
176 #build the hash for the template.
177 $resultsloop{highlight} = ($i % 2)?(1):(0);
178 $resultsloop{title} = $biblio->{'title'};
179 $resultsloop{subtitle} = $biblio->{'subtitle'};
180 $resultsloop{biblionumber} = $biblio->{'biblionumber'};
181 $resultsloop{author} = $biblio->{'author'};
182 $resultsloop{publishercode} = $biblio->{'publishercode'};
183 $resultsloop{publicationyear} = $biblio->{'publicationyear'};
185 push @results, \%resultsloop;
188 $template->param(result=>\@results);
193 my ( $query, $offset, $max_results, $servers ) = @_;
195 if ( C4::Context->preference('NoZebra') ) {
196 my $result = NZorder( NZanalyse($query) )->{'biblioserver'};
199 && $result->{hits} > 0 ? $result->{'RECORDS'} : [] );
200 return ( undef, $search_result, scalar($result->{hits}) );
203 # FIXME hardcoded value. See catalog/search.pl & opac-search.pl too.
204 my @servers = defined ( $servers ) ? @$servers : ( "biblioserver" );
210 return ( "No query entered", undef, undef ) unless $query;
212 # Initialize & Search Zebra
213 for ( my $i = 0 ; $i < @servers ; $i++ ) {
215 $zconns[$i] = C4::Context->Zconn( $servers[$i], 1 );
216 $zoom_queries[$i] = new ZOOM::Query::CCL2RPN( $query, $zconns[$i]);
217 $tmpresults[$i] = $zconns[$i]->search( $zoom_queries[$i] );
221 $zconns[$i]->errmsg() . " ("
222 . $zconns[$i]->errcode() . ") "
223 . $zconns[$i]->addinfo() . " "
224 . $zconns[$i]->diagset();
226 return ( $error, undef, undef ) if $zconns[$i]->errcode();
230 # caught a ZOOM::Exception
234 . $@->addinfo() . " "
237 return ( $error, undef, undef );
240 while ( ( my $i = ZOOM::event( \@zconns ) ) != 0 ) {
241 my $event = $zconns[ $i - 1 ]->last_event();
242 if ( $event == ZOOM::Event::ZEND ) {
244 my $first_record = defined( $offset ) ? $offset+1 : 1;
245 my $hits = $tmpresults[ $i - 1 ]->size();
246 $total_hits += $hits;
247 my $last_record = $hits;
248 if ( defined $max_results && $offset + $max_results < $hits ) {
249 $last_record = $offset + $max_results;
252 for my $j ( $first_record..$last_record ) {
253 my $record = $tmpresults[ $i - 1 ]->record( $j-1 )->raw(); # 0 indexed
254 push @results, $record;
259 foreach my $result (@tmpresults) {
262 foreach my $zoom_query (@zoom_queries) {
263 $zoom_query->destroy();
266 return ( undef, \@results, $total_hits );
272 ( undef, $results_hashref, \@facets_loop ) = getRecords (
274 $koha_query, $simple_query, $sort_by_ref, $servers_ref,
275 $results_per_page, $offset, $expanded_facet, $branches,
279 The all singing, all dancing, multi-server, asynchronous, scanning,
280 searching, record nabbing, facet-building
282 See verbse embedded documentation.
288 $koha_query, $simple_query, $sort_by_ref, $servers_ref,
289 $results_per_page, $offset, $expanded_facet, $branches,
293 my @servers = @$servers_ref;
294 my @sort_by = @$sort_by_ref;
296 # Initialize variables for the ZOOM connection and results object
300 my $results_hashref = ();
302 # Initialize variables for the faceted results objects
303 my $facets_counter = ();
304 my $facets_info = ();
305 my $facets = getFacets();
308 ; # stores the ref to array of hashes for template facets loop
310 ### LOOP THROUGH THE SERVERS
311 for ( my $i = 0 ; $i < @servers ; $i++ ) {
312 $zconns[$i] = C4::Context->Zconn( $servers[$i], 1 );
314 # perform the search, create the results objects
315 # if this is a local search, use the $koha-query, if it's a federated one, use the federated-query
316 my $query_to_use = ($servers[$i] =~ /biblioserver/) ? $koha_query : $simple_query;
318 #$query_to_use = $simple_query if $scan;
319 warn $simple_query if ( $scan and $DEBUG );
321 # Check if we've got a query_type defined, if so, use it
324 if ($query_type =~ /^ccl/) {
325 $query_to_use =~ s/\:/\=/g; # change : to = last minute (FIXME)
326 $results[$i] = $zconns[$i]->search(new ZOOM::Query::CCL2RPN($query_to_use, $zconns[$i]));
327 } elsif ($query_type =~ /^cql/) {
328 $results[$i] = $zconns[$i]->search(new ZOOM::Query::CQL($query_to_use, $zconns[$i]));
329 } elsif ($query_type =~ /^pqf/) {
330 $results[$i] = $zconns[$i]->search(new ZOOM::Query::PQF($query_to_use, $zconns[$i]));
332 warn "Unknown query_type '$query_type'. Results undetermined.";
335 $results[$i] = $zconns[$i]->scan( new ZOOM::Query::CCL2RPN($query_to_use, $zconns[$i]));
337 $results[$i] = $zconns[$i]->search(new ZOOM::Query::CCL2RPN($query_to_use, $zconns[$i]));
341 warn "WARNING: query problem with $query_to_use " . $@;
344 # Concatenate the sort_by limits and pass them to the results object
345 # Note: sort will override rank
347 foreach my $sort (@sort_by) {
348 if ( $sort eq "author_az" ) {
349 $sort_by .= "1=1003 <i ";
351 elsif ( $sort eq "author_za" ) {
352 $sort_by .= "1=1003 >i ";
354 elsif ( $sort eq "popularity_asc" ) {
355 $sort_by .= "1=9003 <i ";
357 elsif ( $sort eq "popularity_dsc" ) {
358 $sort_by .= "1=9003 >i ";
360 elsif ( $sort eq "call_number_asc" ) {
361 $sort_by .= "1=20 <i ";
363 elsif ( $sort eq "call_number_dsc" ) {
364 $sort_by .= "1=20 >i ";
366 elsif ( $sort eq "pubdate_asc" ) {
367 $sort_by .= "1=31 <i ";
369 elsif ( $sort eq "pubdate_dsc" ) {
370 $sort_by .= "1=31 >i ";
372 elsif ( $sort eq "acqdate_asc" ) {
373 $sort_by .= "1=32 <i ";
375 elsif ( $sort eq "acqdate_dsc" ) {
376 $sort_by .= "1=32 >i ";
378 elsif ( $sort eq "title_az" ) {
379 $sort_by .= "1=4 <i ";
381 elsif ( $sort eq "title_za" ) {
382 $sort_by .= "1=4 >i ";
385 warn "Ignoring unrecognized sort '$sort' requested" if $sort_by;
389 if ( $results[$i]->sort( "yaz", $sort_by ) < 0 ) {
390 warn "WARNING sort $sort_by failed";
393 } # finished looping through servers
395 # The big moment: asynchronously retrieve results from all servers
396 while ( ( my $i = ZOOM::event( \@zconns ) ) != 0 ) {
397 my $ev = $zconns[ $i - 1 ]->last_event();
398 if ( $ev == ZOOM::Event::ZEND ) {
399 next unless $results[ $i - 1 ];
400 my $size = $results[ $i - 1 ]->size();
404 # loop through the results
405 $results_hash->{'hits'} = $size;
407 if ( $offset + $results_per_page <= $size ) {
408 $times = $offset + $results_per_page;
413 for ( my $j = $offset ; $j < $times ; $j++ ) {
418 ## Check if it's an index scan
420 my ( $term, $occ ) = $results[ $i - 1 ]->term($j);
422 # here we create a minimal MARC record and hand it off to the
423 # template just like a normal result ... perhaps not ideal, but
425 my $tmprecord = MARC::Record->new();
426 $tmprecord->encoding('UTF-8');
430 # the minimal record in author/title (depending on MARC flavour)
431 if (C4::Context->preference("marcflavour") eq "UNIMARC") {
432 $tmptitle = MARC::Field->new('200',' ',' ', a => $term, f => $occ);
433 $tmprecord->append_fields($tmptitle);
435 $tmptitle = MARC::Field->new('245',' ',' ', a => $term,);
436 $tmpauthor = MARC::Field->new('100',' ',' ', a => $occ,);
437 $tmprecord->append_fields($tmptitle);
438 $tmprecord->append_fields($tmpauthor);
440 $results_hash->{'RECORDS'}[$j] = $tmprecord->as_usmarc();
445 $record = $results[ $i - 1 ]->record($j)->raw();
447 # warn "RECORD $j:".$record;
448 $results_hash->{'RECORDS'}[$j] = $record;
450 # Fill the facets while we're looping, but only for the biblioserver
451 $facet_record = MARC::Record->new_from_usmarc($record)
452 if $servers[ $i - 1 ] =~ /biblioserver/;
454 #warn $servers[$i-1]."\n".$record; #.$facet_record->title();
456 for ( my $k = 0 ; $k <= @$facets ; $k++ ) {
457 ($facets->[$k]) or next;
458 my @fields = map {$facet_record->field($_)} @{$facets->[$k]->{'tags'}} ;
459 for my $field (@fields) {
460 my @subfields = $field->subfields();
461 for my $subfield (@subfields) {
462 my ( $code, $data ) = @$subfield;
463 ($code eq $facets->[$k]->{'subfield'}) or next;
464 $facets_counter->{ $facets->[$k]->{'link_value'} }->{$data}++;
467 $facets_info->{ $facets->[$k]->{'link_value'} }->{'label_value'} =
468 $facets->[$k]->{'label_value'};
469 $facets_info->{ $facets->[$k]->{'link_value'} }->{'expanded'} =
470 $facets->[$k]->{'expanded'};
475 $results_hashref->{ $servers[ $i - 1 ] } = $results_hash;
478 # warn "connection ", $i-1, ": $size hits";
479 # warn $results[$i-1]->record(0)->render() if $size > 0;
482 if ( $servers[ $i - 1 ] =~ /biblioserver/ ) {
484 sort { $facets_counter->{$b} <=> $facets_counter->{$a} }
485 keys %$facets_counter )
488 my $number_of_facets;
489 my @this_facets_array;
492 $facets_counter->{$link_value}->{$b}
493 <=> $facets_counter->{$link_value}->{$a}
494 } keys %{ $facets_counter->{$link_value} }
498 if ( ( $number_of_facets < 6 )
499 || ( $expanded_facet eq $link_value )
500 || ( $facets_info->{$link_value}->{'expanded'} ) )
503 # Sanitize the link value ), ( will cause errors with CCL,
504 my $facet_link_value = $one_facet;
505 $facet_link_value =~ s/(\(|\))/ /g;
507 # fix the length that will display in the label,
508 my $facet_label_value = $one_facet;
510 substr( $one_facet, 0, 20 ) . "..."
511 unless length($facet_label_value) <= 20;
513 # if it's a branch, label by the name, not the code,
514 if ( $link_value =~ /branch/ ) {
516 $branches->{$one_facet}->{'branchname'};
519 # but we're down with the whole label being in the link's title.
520 push @this_facets_array, {
521 facet_count => $facets_counter->{$link_value}->{$one_facet},
522 facet_label_value => $facet_label_value,
523 facet_title_value => $one_facet,
524 facet_link_value => $facet_link_value,
525 type_link_value => $link_value,
530 # handle expanded option
531 unless ( $facets_info->{$link_value}->{'expanded'} ) {
533 if ( ( $number_of_facets > 6 )
534 && ( $expanded_facet ne $link_value ) );
537 type_link_value => $link_value,
538 type_id => $link_value . "_id",
539 "type_label_" . $facets_info->{$link_value}->{'label_value'} => 1,
540 facets => \@this_facets_array,
541 expandable => $expandable,
542 expand => $link_value,
543 } unless ( ($facets_info->{$link_value}->{'label_value'} =~ /Libraries/) and (C4::Context->preference('singleBranchMode')) );
548 return ( undef, $results_hashref, \@facets_loop );
553 $koha_query, $simple_query, $sort_by_ref, $servers_ref,
554 $results_per_page, $offset, $expanded_facet, $branches,
558 my $paz = C4::Search::PazPar2->new(C4::Context->config('pazpar2url'));
560 $paz->search($simple_query);
564 my $results_hashref = {};
565 my $stats = XMLin($paz->stat);
566 my $results = XMLin($paz->show($offset, $results_per_page, 'work-title:1'), forcearray => 1);
568 # for a grouped search result, the number of hits
569 # is the number of groups returned; 'bib_hits' will have
570 # the total number of bibs.
571 $results_hashref->{'biblioserver'}->{'hits'} = $results->{'merged'}->[0];
572 $results_hashref->{'biblioserver'}->{'bib_hits'} = $stats->{'hits'};
574 HIT: foreach my $hit (@{ $results->{'hit'} }) {
575 my $recid = $hit->{recid}->[0];
577 my $work_title = $hit->{'md-work-title'}->[0];
579 if (exists $hit->{'md-work-author'}) {
580 $work_author = $hit->{'md-work-author'}->[0];
582 my $group_label = (defined $work_author) ? "$work_title / $work_author" : $work_title;
584 my $result_group = {};
585 $result_group->{'group_label'} = $group_label;
586 $result_group->{'group_merge_key'} = $recid;
589 if (exists $hit->{count}) {
590 $count = $hit->{count}->[0];
592 $result_group->{'group_count'} = $count;
594 for (my $i = 0; $i < $count; $i++) {
595 # FIXME -- may need to worry about diacritics here
596 my $rec = $paz->record($recid, $i);
597 push @{ $result_group->{'RECORDS'} }, $rec;
600 push @{ $results_hashref->{'biblioserver'}->{'GROUPS'} }, $result_group;
603 # pass through facets
604 my $termlist_xml = $paz->termlist('author,subject');
605 my $terms = XMLin($termlist_xml, forcearray => 1);
606 my @facets_loop = ();
607 #die Dumper($results);
608 # foreach my $list (sort keys %{ $terms->{'list'} }) {
610 # foreach my $facet (sort @{ $terms->{'list'}->{$list}->{'term'} } ) {
612 # facet_label_value => $facet->{'name'}->[0],
615 # push @facets_loop, ( {
616 # type_label => $list,
617 # facets => \@facets,
621 return ( undef, $results_hashref, \@facets_loop );
625 sub _remove_stopwords {
626 my ( $operand, $index ) = @_;
627 my @stopwords_removed;
629 # phrase and exact-qualified indexes shouldn't have stopwords removed
630 if ( $index !~ m/phr|ext/ ) {
632 # remove stopwords from operand : parse all stopwords & remove them (case insensitive)
633 # we use IsAlpha unicode definition, to deal correctly with diacritics.
634 # otherwise, a French word like "leçon" woudl be split into "le" "çon", "le"
635 # is a stopword, we'd get "çon" and wouldn't find anything...
636 foreach ( keys %{ C4::Context->stopwords } ) {
637 next if ( $_ =~ /(and|or|not)/ ); # don't remove operators
638 if ( my ($matched) = ($operand =~
639 /(\P{IsAlnum}\Q$_\E\P{IsAlnum}|^\Q$_\E\P{IsAlnum}|\P{IsAlnum}\Q$_\E$|^\Q$_\E$)/gi) )
641 $operand =~ s/\Q$matched\E/ /gi;
642 push @stopwords_removed, $_;
646 return ( $operand, \@stopwords_removed );
650 sub _detect_truncation {
651 my ( $operand, $index ) = @_;
652 my ( @nontruncated, @righttruncated, @lefttruncated, @rightlefttruncated,
655 my @wordlist = split( /\s/, $operand );
656 foreach my $word (@wordlist) {
657 if ( $word =~ s/^\*([^\*]+)\*$/$1/ ) {
658 push @rightlefttruncated, $word;
660 elsif ( $word =~ s/^\*([^\*]+)$/$1/ ) {
661 push @lefttruncated, $word;
663 elsif ( $word =~ s/^([^\*]+)\*$/$1/ ) {
664 push @righttruncated, $word;
666 elsif ( index( $word, "*" ) < 0 ) {
667 push @nontruncated, $word;
670 push @regexpr, $word;
674 \@nontruncated, \@righttruncated, \@lefttruncated,
675 \@rightlefttruncated, \@regexpr
680 sub _build_stemmed_operand {
684 # If operand contains a digit, it is almost certainly an identifier, and should
685 # not be stemmed. This is particularly relevant for ISBNs and ISSNs, which
686 # can contain the letter "X" - for example, _build_stemmend_operand would reduce
687 # "014100018X" to "x ", which for a MARC21 database would bring up irrelevant
688 # results (e.g., "23 x 29 cm." from the 300$c). Bug 2098.
689 return $operand if $operand =~ /\d/;
691 # FIXME: the locale should be set based on the user's language and/or search choice
692 my $stemmer = Lingua::Stem->new( -locale => 'EN-US' );
694 # FIXME: these should be stored in the db so the librarian can modify the behavior
695 $stemmer->add_exceptions(
702 my @words = split( / /, $operand );
703 my $stems = $stemmer->stem(@words);
704 for my $stem (@$stems) {
705 $stemmed_operand .= "$stem";
706 $stemmed_operand .= "?"
707 unless ( $stem =~ /(and$|or$|not$)/ ) || ( length($stem) < 3 );
708 $stemmed_operand .= " ";
710 warn "STEMMED OPERAND: $stemmed_operand" if $DEBUG;
711 return $stemmed_operand;
715 sub _build_weighted_query {
717 # FIELD WEIGHTING - This is largely experimental stuff. What I'm committing works
718 # pretty well but could work much better if we had a smarter query parser
719 my ( $operand, $stemmed_operand, $index ) = @_;
720 my $stemming = C4::Context->preference("QueryStemming") || 0;
721 my $weight_fields = C4::Context->preference("QueryWeightFields") || 0;
722 my $fuzzy_enabled = C4::Context->preference("QueryFuzzy") || 0;
724 my $weighted_query .= "(rk=("; # Specifies that we're applying rank
726 # Keyword, or, no index specified
727 if ( ( $index eq 'kw' ) || ( !$index ) ) {
729 "Title-cover,ext,r1=\"$operand\""; # exact title-cover
730 $weighted_query .= " or ti,ext,r2=\"$operand\""; # exact title
731 $weighted_query .= " or ti,phr,r3=\"$operand\""; # phrase title
732 #$weighted_query .= " or any,ext,r4=$operand"; # exact any
733 #$weighted_query .=" or kw,wrdl,r5=\"$operand\""; # word list any
734 $weighted_query .= " or wrdl,fuzzy,r8=\"$operand\""
735 if $fuzzy_enabled; # add fuzzy, word list
736 $weighted_query .= " or wrdl,right-Truncation,r9=\"$stemmed_operand\""
737 if ( $stemming and $stemmed_operand )
738 ; # add stemming, right truncation
739 $weighted_query .= " or wrdl,r9=\"$operand\"";
741 # embedded sorting: 0 a-z; 1 z-a
742 # $weighted_query .= ") or (sort1,aut=1";
745 # Barcode searches should skip this process
746 elsif ( $index eq 'bc' ) {
747 $weighted_query .= "bc=\"$operand\"";
750 # Authority-number searches should skip this process
751 elsif ( $index eq 'an' ) {
752 $weighted_query .= "an=\"$operand\"";
755 # If the index already has more than one qualifier, wrap the operand
756 # in quotes and pass it back (assumption is that the user knows what they
757 # are doing and won't appreciate us mucking up their query
758 elsif ( $index =~ ',' ) {
759 $weighted_query .= " $index=\"$operand\"";
762 #TODO: build better cases based on specific search indexes
764 $weighted_query .= " $index,ext,r1=\"$operand\""; # exact index
765 #$weighted_query .= " or (title-sort-az=0 or $index,startswithnt,st-word,r3=$operand #)";
766 $weighted_query .= " or $index,phr,r3=\"$operand\""; # phrase index
768 " or $index,rt,wrdl,r3=\"$operand\""; # word list index
771 $weighted_query .= "))"; # close rank specification
772 return $weighted_query;
778 $simple_query, $query_cgi,
780 $limit_cgi, $limit_desc,
781 $stopwords_removed, $query_type ) = getRecords ( $operators, $operands, $indexes, $limits, $sort_by, $scan);
783 Build queries and limits in CCL, CGI, Human,
784 handle truncation, stemming, field weighting, stopwords, fuzziness, etc.
786 See verbose embedded documentation.
792 my ( $operators, $operands, $indexes, $limits, $sort_by, $scan ) = @_;
794 warn "---------\nEnter buildQuery\n---------" if $DEBUG;
797 my @operators = $operators ? @$operators : ();
798 my @indexes = $indexes ? @$indexes : ();
799 my @operands = $operands ? @$operands : ();
800 my @limits = $limits ? @$limits : ();
801 my @sort_by = $sort_by ? @$sort_by : ();
803 my $stemming = C4::Context->preference("QueryStemming") || 0;
804 my $auto_truncation = C4::Context->preference("QueryAutoTruncate") || 0;
805 my $weight_fields = C4::Context->preference("QueryWeightFields") || 0;
806 my $fuzzy_enabled = C4::Context->preference("QueryFuzzy") || 0;
807 my $remove_stopwords = C4::Context->preference("QueryRemoveStopwords") || 0;
809 # no stemming/weight/fuzzy in NoZebra
810 if ( C4::Context->preference("NoZebra") ) {
816 my $query = $operands[0];
817 my $simple_query = $operands[0];
819 # initialize the variables we're passing back
828 my $stopwords_removed; # flag to determine if stopwords have been removed
830 # for handling ccl, cql, pqf queries in diagnostic mode, skip the rest of the steps
832 if ( $query =~ /^ccl=/ ) {
833 return ( undef, $', $', "q=ccl=$'", $', '', '', '', '', 'ccl' );
835 if ( $query =~ /^cql=/ ) {
836 return ( undef, $', $', "q=cql=$'", $', '', '', '', '', 'cql' );
838 if ( $query =~ /^pqf=/ ) {
839 return ( undef, $', $', "q=pqf=$'", $', '', '', '', '', 'pqf' );
842 # pass nested queries directly
843 # FIXME: need better handling of some of these variables in this case
844 if ( $query =~ /(\(|\))/ ) {
846 undef, $query, $simple_query, $query_cgi,
847 $query, $limit, $limit_cgi, $limit_desc,
848 $stopwords_removed, 'ccl'
852 # Form-based queries are non-nested and fixed depth, so we can easily modify the incoming
853 # query operands and indexes and add stemming, truncation, field weighting, etc.
854 # Once we do so, we'll end up with a value in $query, just like if we had an
855 # incoming $query from the user
858 ; # clear it out so we can populate properly with field-weighted, stemmed, etc. query
860 ; # a flag used to keep track if there was a previous query
861 # if there was, we can apply the current operator
863 for ( my $i = 0 ; $i <= @operands ; $i++ ) {
865 # COMBINE OPERANDS, INDEXES AND OPERATORS
866 if ( $operands[$i] ) {
868 # A flag to determine whether or not to add the index to the query
871 # If the user is sophisticated enough to specify an index, turn off field weighting, stemming, and stopword handling
872 if ( $operands[$i] =~ /(:|=)/ || $scan ) {
875 $remove_stopwords = 0;
877 my $operand = $operands[$i];
878 my $index = $indexes[$i];
880 # Add index-specific attributes
881 # Date of Publication
882 if ( $index eq 'yr' ) {
883 $index .= ",st-numeric";
885 $stemming = $auto_truncation = $weight_fields = $fuzzy_enabled = $remove_stopwords = 0;
888 # Date of Acquisition
889 elsif ( $index eq 'acqdate' ) {
890 $index .= ",st-date-normalized";
892 $stemming = $auto_truncation = $weight_fields = $fuzzy_enabled = $remove_stopwords = 0;
894 # ISBN,ISSN,Standard Number, don't need special treatment
895 elsif ( $index eq 'nb' || $index eq 'ns' ) {
898 $stemming, $auto_truncation,
899 $weight_fields, $fuzzy_enabled,
901 ) = ( 0, 0, 0, 0, 0 );
904 # Set default structure attribute (word list)
906 unless ( $indexes_set || !$index || $index =~ /(st-|phr|ext|wrdl)/ ) {
907 $struct_attr = ",wrdl";
910 # Some helpful index variants
911 my $index_plus = $index . $struct_attr . ":" if $index;
912 my $index_plus_comma = $index . $struct_attr . "," if $index;
915 if ($remove_stopwords) {
916 ( $operand, $stopwords_removed ) =
917 _remove_stopwords( $operand, $index );
918 warn "OPERAND w/out STOPWORDS: >$operand<" if $DEBUG;
919 warn "REMOVED STOPWORDS: @$stopwords_removed"
920 if ( $stopwords_removed && $DEBUG );
923 if ($auto_truncation){
924 #FIXME only valid with LTR scripts
925 $operand=join(" ",map{
927 }split (/\s+/,$operand));
928 warn $operand if $DEBUG;
932 my $truncated_operand;
933 my( $nontruncated, $righttruncated, $lefttruncated,
934 $rightlefttruncated, $regexpr
935 ) = _detect_truncation( $operand, $index );
937 "TRUNCATION: NON:>@$nontruncated< RIGHT:>@$righttruncated< LEFT:>@$lefttruncated< RIGHTLEFT:>@$rightlefttruncated< REGEX:>@$regexpr<"
942 scalar(@$righttruncated) + scalar(@$lefttruncated) +
943 scalar(@$rightlefttruncated) > 0 )
946 # Don't field weight or add the index to the query, we do it here
948 undef $weight_fields;
949 my $previous_truncation_operand;
950 if (scalar @$nontruncated) {
951 $truncated_operand .= "$index_plus @$nontruncated ";
952 $previous_truncation_operand = 1;
954 if (scalar @$righttruncated) {
955 $truncated_operand .= "and " if $previous_truncation_operand;
956 $truncated_operand .= $index_plus_comma . "rtrn:@$righttruncated ";
957 $previous_truncation_operand = 1;
959 if (scalar @$lefttruncated) {
960 $truncated_operand .= "and " if $previous_truncation_operand;
961 $truncated_operand .= $index_plus_comma . "ltrn:@$lefttruncated ";
962 $previous_truncation_operand = 1;
964 if (scalar @$rightlefttruncated) {
965 $truncated_operand .= "and " if $previous_truncation_operand;
966 $truncated_operand .= $index_plus_comma . "rltrn:@$rightlefttruncated ";
967 $previous_truncation_operand = 1;
970 $operand = $truncated_operand if $truncated_operand;
971 warn "TRUNCATED OPERAND: >$truncated_operand<" if $DEBUG;
975 $stemmed_operand = _build_stemmed_operand($operand) if $stemming;
977 warn "STEMMED OPERAND: >$stemmed_operand<" if $DEBUG;
979 # Handle Field Weighting
980 my $weighted_operand;
981 if ($weight_fields) {
982 $weighted_operand = _build_weighted_query( $operand, $stemmed_operand, $index );
983 $operand = $weighted_operand;
987 warn "FIELD WEIGHTED OPERAND: >$weighted_operand<" if $DEBUG;
989 # If there's a previous operand, we need to add an operator
990 if ($previous_operand) {
992 # User-specified operator
993 if ( $operators[ $i - 1 ] ) {
994 $query .= " $operators[$i-1] ";
995 $query .= " $index_plus " unless $indexes_set;
996 $query .= " $operand";
997 $query_cgi .= "&op=$operators[$i-1]";
998 $query_cgi .= "&idx=$index" if $index;
999 $query_cgi .= "&q=$operands[$i]" if $operands[$i];
1001 " $operators[$i-1] $index_plus $operands[$i]";
1004 # Default operator is and
1007 $query .= "$index_plus " unless $indexes_set;
1008 $query .= "$operand";
1009 $query_cgi .= "&op=and&idx=$index" if $index;
1010 $query_cgi .= "&q=$operands[$i]" if $operands[$i];
1011 $query_desc .= " and $index_plus $operands[$i]";
1015 # There isn't a pervious operand, don't need an operator
1018 # Field-weighted queries already have indexes set
1019 $query .= " $index_plus " unless $indexes_set;
1021 $query_desc .= " $index_plus $operands[$i]";
1022 $query_cgi .= "&idx=$index" if $index;
1023 $query_cgi .= "&q=$operands[$i]" if $operands[$i];
1024 $previous_operand = 1;
1029 warn "QUERY BEFORE LIMITS: >$query<" if $DEBUG;
1032 my $group_OR_limits;
1033 my $availability_limit;
1034 foreach my $this_limit (@limits) {
1035 if ( $this_limit =~ /available/ ) {
1037 # 'available' is defined as (items.onloan is NULL) and (items.itemlost = 0)
1039 # all records not indexed in the onloan register (zebra) and all records with a value of lost equal to 0
1040 $availability_limit .=
1041 "( ( allrecords,AlwaysMatches='' not onloan,AlwaysMatches='') and (lost,st-numeric=0) )"; #or ( allrecords,AlwaysMatches='' not lost,AlwaysMatches='')) )";
1042 $limit_cgi .= "&limit=available";
1046 # group_OR_limits, prefixed by mc-
1047 # OR every member of the group
1048 elsif ( $this_limit =~ /mc/ ) {
1049 $group_OR_limits .= " or " if $group_OR_limits;
1050 $limit_desc .= " or " if $group_OR_limits;
1051 $group_OR_limits .= "$this_limit";
1052 $limit_cgi .= "&limit=$this_limit";
1053 $limit_desc .= " $this_limit";
1056 # Regular old limits
1058 $limit .= " and " if $limit || $query;
1059 $limit .= "$this_limit";
1060 $limit_cgi .= "&limit=$this_limit";
1061 if ($this_limit =~ /^branch:(.+)/) {
1062 my $branchcode = $1;
1063 my $branchname = GetBranchName($branchcode);
1064 if (defined $branchname) {
1065 $limit_desc .= " branch:$branchname";
1067 $limit_desc .= " $this_limit";
1070 $limit_desc .= " $this_limit";
1074 if ($group_OR_limits) {
1075 $limit .= " and " if ( $query || $limit );
1076 $limit .= "($group_OR_limits)";
1078 if ($availability_limit) {
1079 $limit .= " and " if ( $query || $limit );
1080 $limit .= "($availability_limit)";
1083 # Normalize the query and limit strings
1086 for ( $query, $query_desc, $limit, $limit_desc ) {
1087 s/ / /g; # remove extra spaces
1088 s/^ //g; # remove any beginning spaces
1089 s/ $//g; # remove any ending spaces
1090 s/==/=/g; # remove double == from query
1092 $query_cgi =~ s/^&//; # remove unnecessary & from beginning of the query cgi
1094 for ($query_cgi,$simple_query) {
1097 # append the limit to the query
1098 $query .= " " . $limit;
1102 warn "QUERY:" . $query;
1103 warn "QUERY CGI:" . $query_cgi;
1104 warn "QUERY DESC:" . $query_desc;
1105 warn "LIMIT:" . $limit;
1106 warn "LIMIT CGI:" . $limit_cgi;
1107 warn "LIMIT DESC:" . $limit_desc;
1108 warn "---------\nLeave buildQuery\n---------";
1111 undef, $query, $simple_query, $query_cgi,
1112 $query_desc, $limit, $limit_cgi, $limit_desc,
1113 $stopwords_removed, $query_type
1117 =head2 searchResults
1119 Format results in a form suitable for passing to the template
1123 # IMO this subroutine is pretty messy still -- it's responsible for
1124 # building the HTML output for the template
1126 my ( $searchdesc, $hits, $results_per_page, $offset, $scan, @marcresults ) = @_;
1127 my $dbh = C4::Context->dbh;
1131 # add search-term highlighting via <span>s on the search terms
1132 my $span_terms_hashref;
1133 for my $span_term ( split( / /, $searchdesc ) ) {
1134 $span_term =~ s/(.*=|\)|\(|\+|\.|\*)//g;
1135 $span_terms_hashref->{$span_term}++;
1138 #Build branchnames hash
1140 #get branch information.....
1143 $dbh->prepare("SELECT branchcode,branchname FROM branches")
1144 ; # FIXME : use C4::Koha::GetBranches
1146 while ( my $bdata = $bsth->fetchrow_hashref ) {
1147 $branches{ $bdata->{'branchcode'} } = $bdata->{'branchname'};
1149 # FIXME - We build an authorised values hash here, using the default framework
1150 # though it is possible to have different authvals for different fws.
1152 my $shelflocations =GetKohaAuthorisedValues('items.location','');
1154 # get notforloan authorised value list (see $shelflocations FIXME)
1155 my $notforloan_authorised_value = GetAuthValCode('items.notforloan','');
1157 #Build itemtype hash
1158 #find itemtype & itemtype image
1162 "SELECT itemtype,description,imageurl,summary,notforloan FROM itemtypes"
1165 while ( my $bdata = $bsth->fetchrow_hashref ) {
1166 foreach (qw(description imageurl summary notforloan)) {
1167 $itemtypes{ $bdata->{'itemtype'} }->{$_} = $bdata->{$_};
1171 #search item field code
1174 "SELECT tagfield FROM marc_subfield_structure WHERE kohafield LIKE 'items.itemnumber'"
1177 my ($itemtag) = $sth->fetchrow;
1179 ## find column names of items related to MARC
1180 my $sth2 = $dbh->prepare("SHOW COLUMNS FROM items");
1182 my %subfieldstosearch;
1183 while ( ( my $column ) = $sth2->fetchrow ) {
1184 my ( $tagfield, $tagsubfield ) =
1185 &GetMarcFromKohaField( "items." . $column, "" );
1186 $subfieldstosearch{$column} = $tagsubfield;
1189 # handle which records to actually retrieve
1191 if ( $hits && $offset + $results_per_page <= $hits ) {
1192 $times = $offset + $results_per_page;
1195 $times = $hits; # FIXME: if $hits is undefined, why do we want to equal it?
1198 # loop through all of the records we've retrieved
1199 for ( my $i = $offset ; $i <= $times - 1 ; $i++ ) {
1200 my $marcrecord = MARC::File::USMARC::decode( $marcresults[$i] );
1201 my $oldbiblio = TransformMarcToKoha( $dbh, $marcrecord, '' );
1202 $oldbiblio->{subtitle} = C4::Biblio::get_koha_field_from_marc('bibliosubtitle', 'subtitle', $marcrecord, '');
1203 $oldbiblio->{result_number} = $i + 1;
1205 # add imageurl to itemtype if there is one
1206 $oldbiblio->{imageurl} = getitemtypeimagelocation( 'opac', $itemtypes{ $oldbiblio->{itemtype} }->{imageurl} );
1208 $oldbiblio->{'authorised_value_images'} = C4::Items::get_authorised_value_images( C4::Biblio::get_biblio_authorised_values( $oldbiblio->{'biblionumber'}, $marcrecord ) );
1209 (my $aisbn) = $oldbiblio->{isbn} =~ /([\d-]*[X]*)/;
1211 $oldbiblio->{amazonisbn} = $aisbn;
1212 $oldbiblio->{description} = $itemtypes{ $oldbiblio->{itemtype} }->{description};
1213 # Build summary if there is one (the summary is defined in the itemtypes table)
1214 # FIXME: is this used anywhere, I think it can be commented out? -- JF
1215 if ( $itemtypes{ $oldbiblio->{itemtype} }->{summary} ) {
1216 my $summary = $itemtypes{ $oldbiblio->{itemtype} }->{summary};
1217 my @fields = $marcrecord->fields();
1218 foreach my $field (@fields) {
1219 my $tag = $field->tag();
1220 my $tagvalue = $field->as_string();
1221 if (! utf8::is_utf8($tagvalue)) {
1222 utf8::decode($tagvalue);
1226 s/\[(.?.?.?.?)$tag\*(.*?)]/$1$tagvalue$2\[$1$tag$2]/g;
1227 unless ( $tag < 10 ) {
1228 my @subf = $field->subfields;
1229 for my $i ( 0 .. $#subf ) {
1230 my $subfieldcode = $subf[$i][0];
1231 my $subfieldvalue = $subf[$i][1];
1232 if (! utf8::is_utf8($subfieldvalue)) {
1233 utf8::decode($subfieldvalue);
1235 my $tagsubf = $tag . $subfieldcode;
1237 s/\[(.?.?.?.?)$tagsubf(.*?)]/$1$subfieldvalue$2\[$1$tagsubf$2]/g;
1242 $summary =~ s/\[(.*?)]//g;
1243 $summary =~ s/\n/<br\/>/g;
1244 $oldbiblio->{summary} = $summary;
1247 # save an author with no <span> tag, for the <a href=search.pl?q=<!--tmpl_var name="author"-->> link
1248 $oldbiblio->{'author_nospan'} = $oldbiblio->{'author'};
1249 $oldbiblio->{'title_nospan'} = $oldbiblio->{'title'};
1250 $oldbiblio->{'subtitle_nospan'} = $oldbiblio->{'subtitle'};
1251 # Add search-term highlighting to the whole record where they match using <span>s
1252 if (C4::Context->preference("OpacHighlightedWords")){
1253 my $searchhighlightblob;
1254 for my $highlight_field ( $marcrecord->fields ) {
1256 # FIXME: need to skip title, subtitle, author, etc., as they are handled below
1257 next if $highlight_field->tag() =~ /(^00)/; # skip fixed fields
1258 for my $subfield ($highlight_field->subfields()) {
1260 next if $subfield->[0] eq '9';
1261 my $field = $subfield->[1];
1262 for my $term ( keys %$span_terms_hashref ) {
1263 if ( ( $field =~ /$term/i ) && (( length($term) > 3 ) || ($field =~ / $term /i)) ) {
1264 $field =~ s/$term/<span class=\"term\">$&<\/span>/gi;
1268 $searchhighlightblob .= $field . " ... " if $match;
1272 $searchhighlightblob = ' ... '.$searchhighlightblob if $searchhighlightblob;
1273 $oldbiblio->{'searchhighlightblob'} = $searchhighlightblob;
1276 # Add search-term highlighting to the title, subtitle, etc. fields
1277 for my $term ( keys %$span_terms_hashref ) {
1278 my $old_term = $term;
1279 if ( length($term) > 3 ) {
1280 $term =~ s/(.*=|\)|\(|\+|\.|\?|\[|\]|\\|\*)//g;
1281 foreach(qw(title subtitle author publishercode place pages notes size)) {
1282 $oldbiblio->{$_} =~ s/$term/<span class=\"term\">$&<\/span>/gi;
1287 ($i % 2) and $oldbiblio->{'toggle'} = 1;
1289 # Pull out the items fields
1290 my @fields = $marcrecord->field($itemtag);
1292 # Setting item statuses for display
1293 my @available_items_loop;
1294 my @onloan_items_loop;
1295 my @notforloan_items_loop;
1296 my @other_items_loop;
1298 my $available_items;
1300 my $notforloan_items;
1303 my $ordered_count = 0;
1304 my $available_count = 0;
1305 my $onloan_count = 0;
1306 my $notforloan_count = 0;
1307 my $longoverdue_count = 0;
1308 my $other_count = 0;
1309 my $wthdrawn_count = 0;
1310 my $itemlost_count = 0;
1311 my $itembinding_count = 0;
1312 my $itemdamaged_count = 0;
1313 my $item_in_transit_count = 0;
1314 my $can_place_holds = 0;
1315 my $items_count = scalar(@fields);
1317 ( C4::Context->preference('maxItemsinSearchResults') )
1318 ? C4::Context->preference('maxItemsinSearchResults') - 1
1321 # loop through every item
1322 foreach my $field (@fields) {
1325 # populate the items hash
1326 foreach my $code ( keys %subfieldstosearch ) {
1327 $item->{$code} = $field->subfield( $subfieldstosearch{$code} );
1329 my $hbranch = C4::Context->preference('HomeOrHoldingBranch') eq 'homebranch' ? 'homebranch' : 'holdingbranch';
1330 my $otherbranch = C4::Context->preference('HomeOrHoldingBranch') eq 'homebranch' ? 'holdingbranch' : 'homebranch';
1331 # set item's branch name, use HomeOrHoldingBranch syspref first, fall back to the other one
1332 if ($item->{$hbranch}) {
1333 $item->{'branchname'} = $branches{$item->{$hbranch}};
1335 elsif ($item->{$otherbranch}) { # Last resort
1336 $item->{'branchname'} = $branches{$item->{$otherbranch}};
1339 ($item->{'reserved'}) = C4::Reserves::CheckReserves($item->{itemnumber});
1341 my $prefix = $item->{$hbranch} . '--' . $item->{location} . $item->{itype} . $item->{itemcallnumber};
1342 # For each grouping of items (onloan, available, unavailable), we build a key to store relevant info about that item
1343 if ( $item->{onloan} or $item->{reserved} ) {
1345 my $key = $prefix . $item->{onloan} . $item->{barcode};
1346 $onloan_items->{$key}->{due_date} = format_date($item->{onloan});
1347 $onloan_items->{$key}->{count}++ if $item->{$hbranch};
1348 $onloan_items->{$key}->{branchname} = $item->{branchname};
1349 $onloan_items->{$key}->{location} = $shelflocations->{ $item->{location} };
1350 $onloan_items->{$key}->{itemcallnumber} = $item->{itemcallnumber};
1351 $onloan_items->{$key}->{imageurl} = getitemtypeimagelocation( 'opac', $itemtypes{ $item->{itype} }->{imageurl} );
1352 $onloan_items->{$key}->{barcode} = $item->{barcode};
1353 # if something's checked out and lost, mark it as 'long overdue'
1354 if ( $item->{itemlost} ) {
1355 $onloan_items->{$prefix}->{longoverdue}++;
1356 $longoverdue_count++;
1357 } else { # can place holds as long as item isn't lost
1358 $can_place_holds = 1;
1362 # items not on loan, but still unavailable ( lost, withdrawn, damaged )
1366 if ( $item->{notforloan} == -1 ) {
1370 # is item in transit?
1371 my $transfertwhen = '';
1372 my ($transfertfrom, $transfertto);
1374 unless ($item->{wthdrawn}
1375 || $item->{itemlost}
1377 || $item->{notforloan}
1378 || $items_count > 20) {
1380 # A couple heuristics to limit how many times
1381 # we query the database for item transfer information, sacrificing
1382 # accuracy in some cases for speed;
1384 # 1. don't query if item has one of the other statuses
1385 # 2. don't check transit status if the bib has
1386 # more than 20 items
1388 # FIXME: to avoid having the query the database like this, and to make
1389 # the in transit status count as unavailable for search limiting,
1390 # should map transit status to record indexed in Zebra.
1392 ($transfertwhen, $transfertfrom, $transfertto) = C4::Circulation::GetTransfers($item->{itemnumber});
1395 # item is withdrawn, lost or damaged
1396 if ( $item->{wthdrawn}
1397 || $item->{itemlost}
1399 || $item->{notforloan}
1400 || $item->{reserved}
1401 || ($transfertwhen ne ''))
1403 $wthdrawn_count++ if $item->{wthdrawn};
1404 $itemlost_count++ if $item->{itemlost};
1405 $itemdamaged_count++ if $item->{damaged};
1406 $item_in_transit_count++ if $transfertwhen ne '';
1407 $item->{status} = $item->{wthdrawn} . "-" . $item->{itemlost} . "-" . $item->{damaged} . "-" . $item->{notforloan};
1409 my $key = $prefix . $item->{status};
1411 foreach (qw(wthdrawn itemlost damaged branchname itemcallnumber)) {
1412 if($item->{notforloan} == 1){
1413 $notforloan_items->{$key}->{$_} = $item->{$_};
1415 $other_items->{$key}->{$_} = $item->{$_};
1418 if($item->{notforloan} == 1){
1419 $notforloan_count++;
1421 $notforloan_items->{$key}->{intransit} = ($transfertwhen ne '') ? 1 : 0;
1422 $notforloan_items->{$key}->{notforloan} = GetAuthorisedValueDesc('','',$item->{notforloan},'','',$notforloan_authorised_value) if $notforloan_authorised_value;
1423 $notforloan_items->{$key}->{count}++ if $item->{$hbranch};
1424 $notforloan_items->{$key}->{location} = $shelflocations->{ $item->{location} };
1425 $notforloan_items->{$key}->{imageurl} = getitemtypeimagelocation( 'opac', $itemtypes{ $item->{itype} }->{imageurl} );
1426 $notforloan_items->{$key}->{barcode} = $item->{barcode};
1430 $other_items->{$key}->{intransit} = ($transfertwhen ne '') ? 1 : 0;
1431 $other_items->{$key}->{notforloan} = GetAuthorisedValueDesc('','',$item->{notforloan},'','',$notforloan_authorised_value) if $notforloan_authorised_value;
1432 $other_items->{$key}->{count}++ if $item->{$hbranch};
1433 $other_items->{$key}->{location} = $shelflocations->{ $item->{location} };
1434 $other_items->{$key}->{imageurl} = getitemtypeimagelocation( 'opac', $itemtypes{ $item->{itype} }->{imageurl} );
1435 $other_items->{$key}->{barcode} = $item->{barcode};
1441 $can_place_holds = 1;
1443 $available_items->{$prefix}->{count}++ if $item->{$hbranch};
1444 foreach (qw(branchname itemcallnumber barcode)) {
1445 $available_items->{$prefix}->{$_} = $item->{$_};
1447 $available_items->{$prefix}->{location} = $shelflocations->{ $item->{location} };
1448 $available_items->{$prefix}->{imageurl} = getitemtypeimagelocation( 'opac', $itemtypes{ $item->{itype} }->{imageurl} );
1451 } # notforloan, item level and biblioitem level
1452 my ( $availableitemscount, $onloanitemscount, $notforloanitemscount,$otheritemscount );
1454 ( C4::Context->preference('maxItemsinSearchResults') )
1455 ? C4::Context->preference('maxItemsinSearchResults') - 1
1457 for my $key ( sort keys %$onloan_items ) {
1458 (++$onloanitemscount > $maxitems) and last;
1459 push @onloan_items_loop, $onloan_items->{$key};
1461 for my $key ( sort keys %$other_items ) {
1462 (++$otheritemscount > $maxitems) and last;
1463 push @other_items_loop, $other_items->{$key};
1465 for my $key ( sort keys %$notforloan_items ) {
1466 (++$notforloanitemscount > $maxitems) and last;
1467 push @notforloan_items_loop, $notforloan_items->{$key};
1469 for my $key ( sort keys %$available_items ) {
1470 (++$availableitemscount > $maxitems) and last;
1471 push @available_items_loop, $available_items->{$key}
1474 # XSLT processing of some stuff
1475 if (C4::Context->preference("XSLTResultsDisplay") && !$scan) {
1476 $oldbiblio->{XSLTResultsRecord} = XSLTParse4Display(
1477 $oldbiblio->{biblionumber}, $marcrecord, 'Results' );
1480 # last check for norequest : if itemtype is notforloan, it can't be reserved either, whatever the items
1481 $can_place_holds = 0 if $itemtypes{ $oldbiblio->{itemtype} }->{notforloan};
1482 $oldbiblio->{norequests} = 1 unless $can_place_holds;
1483 $oldbiblio->{itemsplural} = 1 if $items_count > 1;
1484 $oldbiblio->{items_count} = $items_count;
1485 $oldbiblio->{available_items_loop} = \@available_items_loop;
1486 $oldbiblio->{notforloan_items_loop}= \@notforloan_items_loop;
1487 $oldbiblio->{onloan_items_loop} = \@onloan_items_loop;
1488 $oldbiblio->{other_items_loop} = \@other_items_loop;
1489 $oldbiblio->{availablecount} = $available_count;
1490 $oldbiblio->{availableplural} = 1 if $available_count > 1;
1491 $oldbiblio->{onloancount} = $onloan_count;
1492 $oldbiblio->{onloanplural} = 1 if $onloan_count > 1;
1493 $oldbiblio->{notforloancount} = $notforloan_count;
1494 $oldbiblio->{othercount} = $other_count;
1495 $oldbiblio->{otherplural} = 1 if $other_count > 1;
1496 $oldbiblio->{wthdrawncount} = $wthdrawn_count;
1497 $oldbiblio->{itemlostcount} = $itemlost_count;
1498 $oldbiblio->{damagedcount} = $itemdamaged_count;
1499 $oldbiblio->{intransitcount} = $item_in_transit_count;
1500 $oldbiblio->{orderedcount} = $ordered_count;
1501 $oldbiblio->{isbn} =~
1502 s/-//g; # deleting - in isbn to enable amazon content
1503 push( @newresults, $oldbiblio );
1508 #----------------------------------------------------------------------
1510 # Non-Zebra GetRecords#
1511 #----------------------------------------------------------------------
1515 NZgetRecords has the same API as zera getRecords, even if some parameters are not managed
1521 $query, $simple_query, $sort_by_ref, $servers_ref,
1522 $results_per_page, $offset, $expanded_facet, $branches,
1525 warn "query =$query" if $DEBUG;
1526 my $result = NZanalyse($query);
1527 warn "results =$result" if $DEBUG;
1529 NZorder( $result, @$sort_by_ref[0], $results_per_page, $offset ),
1535 NZanalyse : get a CQL string as parameter, and returns a list of biblionumber;title,biblionumber;title,...
1536 the list is built from an inverted index in the nozebra SQL table
1537 note that title is here only for convenience : the sorting will be very fast when requested on title
1538 if the sorting is requested on something else, we will have to reread all results, and that may be longer.
1543 my ( $string, $server ) = @_;
1544 # warn "---------" if $DEBUG;
1545 warn " NZanalyse" if $DEBUG;
1546 # warn "---------" if $DEBUG;
1548 # $server contains biblioserver or authorities, depending on what we search on.
1549 #warn "querying : $string on $server";
1550 $server = 'biblioserver' unless $server;
1552 # if we have a ", replace the content to discard temporarily any and/or/not inside
1554 if ( $string =~ /"/ ) {
1555 $string =~ s/"(.*?)"/__X__/;
1557 warn "commacontent : $commacontent" if $DEBUG;
1560 # split the query string in 3 parts : X AND Y means : $left="X", $operand="AND" and $right="Y"
1561 # then, call again NZanalyse with $left and $right
1562 # (recursive until we find a leaf (=> something without and/or/not)
1563 # delete repeated operator... Would then go in infinite loop
1564 while ( $string =~ s/( and| or| not| AND| OR| NOT)\1/$1/g ) {
1567 #process parenthesis before.
1568 if ( $string =~ /^\s*\((.*)\)(( and | or | not | AND | OR | NOT )(.*))?/ ) {
1571 my $operator = lc($3); # FIXME: and/or/not are operators, not operands
1573 "dealing w/parenthesis before recursive sub call. left :$left operator:$operator right:$right"
1575 my $leftresult = NZanalyse( $left, $server );
1577 my $rightresult = NZanalyse( $right, $server );
1579 # OK, we have the results for right and left part of the query
1580 # depending of operand, intersect, union or exclude both lists
1581 # to get a result list
1582 if ( $operator eq ' and ' ) {
1583 return NZoperatorAND($leftresult,$rightresult);
1585 elsif ( $operator eq ' or ' ) {
1587 # just merge the 2 strings
1588 return $leftresult . $rightresult;
1590 elsif ( $operator eq ' not ' ) {
1591 return NZoperatorNOT($leftresult,$rightresult);
1595 # this error is impossible, because of the regexp that isolate the operand, but just in case...
1599 warn "string :" . $string if $DEBUG;
1603 if ($string =~ /(.*?)( and | or | not | AND | OR | NOT )(.*)/) {
1606 $operator = lc($2); # FIXME: and/or/not are operators, not operands
1608 warn "no parenthesis. left : $left operator: $operator right: $right"
1611 # it's not a leaf, we have a and/or/not
1614 # reintroduce comma content if needed
1615 $right =~ s/__X__/"$commacontent"/ if $commacontent;
1616 $left =~ s/__X__/"$commacontent"/ if $commacontent;
1617 warn "node : $left / $operator / $right\n" if $DEBUG;
1618 my $leftresult = NZanalyse( $left, $server );
1619 my $rightresult = NZanalyse( $right, $server );
1620 warn " leftresult : $leftresult" if $DEBUG;
1621 warn " rightresult : $rightresult" if $DEBUG;
1622 # OK, we have the results for right and left part of the query
1623 # depending of operand, intersect, union or exclude both lists
1624 # to get a result list
1625 if ( $operator eq ' and ' ) {
1627 return NZoperatorAND($leftresult,$rightresult);
1629 elsif ( $operator eq ' or ' ) {
1631 # just merge the 2 strings
1632 return $leftresult . $rightresult;
1634 elsif ( $operator eq ' not ' ) {
1635 return NZoperatorNOT($leftresult,$rightresult);
1639 # this error is impossible, because of the regexp that isolate the operand, but just in case...
1640 die "error : operand unknown : $operator for $string";
1643 # it's a leaf, do the real SQL query and return the result
1646 $string =~ s/__X__/"$commacontent"/ if $commacontent;
1647 $string =~ s/-|\.|\?|,|;|!|'|\(|\)|\[|\]|{|}|"|&|\+|\*|\// /g;
1648 #remove trailing blank at the beginning
1650 warn "leaf:$string" if $DEBUG;
1652 # parse the string in in operator/operand/value again
1656 if ($string =~ /(.*)(>=|<=)(.*)/) {
1663 # warn "handling leaf... left:$left operator:$operator right:$right"
1665 unless ($operator) {
1666 if ($string =~ /(.*)(>|<|=)(.*)/) {
1671 "handling unless (operator)... left:$left operator:$operator right:$right"
1679 # strip adv, zebra keywords, currently not handled in nozebra: wrdl, ext, phr...
1682 # automatic replace for short operators
1683 $left = 'title' if $left =~ '^ti$';
1684 $left = 'author' if $left =~ '^au$';
1685 $left = 'publisher' if $left =~ '^pb$';
1686 $left = 'subject' if $left =~ '^su$';
1687 $left = 'koha-Auth-Number' if $left =~ '^an$';
1688 $left = 'keyword' if $left =~ '^kw$';
1689 $left = 'itemtype' if $left =~ '^mc$'; # Fix for Bug 2599 - Search limits not working for NoZebra
1690 warn "handling leaf... left:$left operator:$operator right:$right" if $DEBUG;
1691 my $dbh = C4::Context->dbh;
1692 if ( $operator && $left ne 'keyword' ) {
1693 #do a specific search
1694 $operator = 'LIKE' if $operator eq '=' and $right =~ /%/;
1695 my $sth = $dbh->prepare(
1696 "SELECT biblionumbers,value FROM nozebra WHERE server=? AND indexname=? AND value $operator ?"
1698 warn "$left / $operator / $right\n" if $DEBUG;
1700 # split each word, query the DB and build the biblionumbers result
1701 #sanitizing leftpart
1702 $left =~ s/^\s+|\s+$//;
1703 foreach ( split / /, $right ) {
1705 $_ =~ s/^\s+|\s+$//;
1707 warn "EXECUTE : $server, $left, $_" if $DEBUG;
1708 $sth->execute( $server, $left, $_ )
1709 or warn "execute failed: $!";
1710 while ( my ( $line, $value ) = $sth->fetchrow ) {
1712 # if we are dealing with a numeric value, use only numeric results (in case of >=, <=, > or <)
1713 # otherwise, fill the result
1714 $biblionumbers .= $line
1715 unless ( $right =~ /^\d+$/ && $value =~ /\D/ );
1716 warn "result : $value "
1717 . ( $right =~ /\d/ ) . "=="
1718 . ( $value =~ /\D/?$line:"" ) if $DEBUG; #= $line";
1721 # do a AND with existing list if there is one, otherwise, use the biblionumbers list as 1st result list
1723 warn "NZAND" if $DEBUG;
1724 $results = NZoperatorAND($biblionumbers,$results);
1726 $results = $biblionumbers;
1731 #do a complete search (all indexes), if index='kw' do complete search too.
1732 my $sth = $dbh->prepare(
1733 "SELECT biblionumbers FROM nozebra WHERE server=? AND value LIKE ?"
1736 # split each word, query the DB and build the biblionumbers result
1737 foreach ( split / /, $string ) {
1738 next if C4::Context->stopwords->{ uc($_) }; # skip if stopword
1739 warn "search on all indexes on $_" if $DEBUG;
1742 $sth->execute( $server, $_ );
1743 while ( my $line = $sth->fetchrow ) {
1744 $biblionumbers .= $line;
1747 # do a AND with existing list if there is one, otherwise, use the biblionumbers list as 1st result list
1749 $results = NZoperatorAND($biblionumbers,$results);
1752 warn "NEW RES for $_ = $biblionumbers" if $DEBUG;
1753 $results = $biblionumbers;
1757 warn "return : $results for LEAF : $string" if $DEBUG;
1760 warn "---------\nLeave NZanalyse\n---------" if $DEBUG;
1764 my ($rightresult, $leftresult)=@_;
1766 my @leftresult = split /;/, $leftresult;
1767 warn " @leftresult / $rightresult \n" if $DEBUG;
1769 # my @rightresult = split /;/,$leftresult;
1772 # parse the left results, and if the biblionumber exist in the right result, save it in finalresult
1773 # the result is stored twice, to have the same weight for AND than OR.
1774 # example : TWO : 61,61,64,121 (two is twice in the biblio #61) / TOWER : 61,64,130
1775 # result : 61,61,61,61,64,64 for two AND tower : 61 has more weight than 64
1776 foreach (@leftresult) {
1779 ( $value, $countvalue ) = ( $1, $2 ) if ($value=~/(.*)-(\d+)$/);
1780 if ( $rightresult =~ /\Q$value\E-(\d+);/ ) {
1781 $countvalue = ( $1 > $countvalue ? $countvalue : $1 );
1783 "$value-$countvalue;$value-$countvalue;";
1786 warn "NZAND DONE : $finalresult \n" if $DEBUG;
1787 return $finalresult;
1791 my ($rightresult, $leftresult)=@_;
1792 return $rightresult.$leftresult;
1796 my ($leftresult, $rightresult)=@_;
1798 my @leftresult = split /;/, $leftresult;
1800 # my @rightresult = split /;/,$leftresult;
1802 foreach (@leftresult) {
1804 $value=$1 if $value=~m/(.*)-\d+$/;
1805 unless ($rightresult =~ "$value-") {
1806 $finalresult .= "$_;";
1809 return $finalresult;
1814 $finalresult = NZorder($biblionumbers, $ordering,$results_per_page,$offset);
1821 my ( $biblionumbers, $ordering, $results_per_page, $offset ) = @_;
1822 warn "biblionumbers = $biblionumbers and ordering = $ordering\n" if $DEBUG;
1824 # order title asc by default
1825 # $ordering = '1=36 <i' unless $ordering;
1826 $results_per_page = 20 unless $results_per_page;
1827 $offset = 0 unless $offset;
1828 my $dbh = C4::Context->dbh;
1831 # order by POPULARITY
1833 if ( $ordering =~ /popularity/ ) {
1837 # popularity is not in MARC record, it's builded from a specific query
1839 $dbh->prepare("select sum(issues) from items where biblionumber=?");
1840 foreach ( split /;/, $biblionumbers ) {
1841 my ( $biblionumber, $title ) = split /,/, $_;
1842 $result{$biblionumber} = GetMarcBiblio($biblionumber);
1843 $sth->execute($biblionumber);
1844 my $popularity = $sth->fetchrow || 0;
1846 # hint : the key is popularity.title because we can have
1847 # many results with the same popularity. In this case, sub-ordering is done by title
1848 # we also have biblionumber to avoid bug for 2 biblios with the same title & popularity
1849 # (un-frequent, I agree, but we won't forget anything that way ;-)
1850 $popularity{ sprintf( "%10d", $popularity ) . $title
1851 . $biblionumber } = $biblionumber;
1854 # sort the hash and return the same structure as GetRecords (Zebra querying)
1857 if ( $ordering eq 'popularity_dsc' ) { # sort popularity DESC
1858 foreach my $key ( sort { $b cmp $a } ( keys %popularity ) ) {
1859 $result_hash->{'RECORDS'}[ $numbers++ ] =
1860 $result{ $popularity{$key} }->as_usmarc();
1863 else { # sort popularity ASC
1864 foreach my $key ( sort ( keys %popularity ) ) {
1865 $result_hash->{'RECORDS'}[ $numbers++ ] =
1866 $result{ $popularity{$key} }->as_usmarc();
1869 my $finalresult = ();
1870 $result_hash->{'hits'} = $numbers;
1871 $finalresult->{'biblioserver'} = $result_hash;
1872 return $finalresult;
1878 elsif ( $ordering =~ /author/ ) {
1880 foreach ( split /;/, $biblionumbers ) {
1881 my ( $biblionumber, $title ) = split /,/, $_;
1882 my $record = GetMarcBiblio($biblionumber);
1884 if ( C4::Context->preference('marcflavour') eq 'UNIMARC' ) {
1885 $author = $record->subfield( '200', 'f' );
1886 $author = $record->subfield( '700', 'a' ) unless $author;
1889 $author = $record->subfield( '100', 'a' );
1892 # hint : the result is sorted by title.biblionumber because we can have X biblios with the same title
1893 # and we don't want to get only 1 result for each of them !!!
1894 $result{ $author . $biblionumber } = $record;
1897 # sort the hash and return the same structure as GetRecords (Zebra querying)
1900 if ( $ordering eq 'author_za' ) { # sort by author desc
1901 foreach my $key ( sort { $b cmp $a } ( keys %result ) ) {
1902 $result_hash->{'RECORDS'}[ $numbers++ ] =
1903 $result{$key}->as_usmarc();
1906 else { # sort by author ASC
1907 foreach my $key ( sort ( keys %result ) ) {
1908 $result_hash->{'RECORDS'}[ $numbers++ ] =
1909 $result{$key}->as_usmarc();
1912 my $finalresult = ();
1913 $result_hash->{'hits'} = $numbers;
1914 $finalresult->{'biblioserver'} = $result_hash;
1915 return $finalresult;
1918 # ORDER BY callnumber
1921 elsif ( $ordering =~ /callnumber/ ) {
1923 foreach ( split /;/, $biblionumbers ) {
1924 my ( $biblionumber, $title ) = split /,/, $_;
1925 my $record = GetMarcBiblio($biblionumber);
1927 my ( $callnumber_tag, $callnumber_subfield ) =
1928 GetMarcFromKohaField( 'items.itemcallnumber','' );
1929 ( $callnumber_tag, $callnumber_subfield ) =
1930 GetMarcFromKohaField('biblioitems.callnumber','')
1931 unless $callnumber_tag;
1932 if ( C4::Context->preference('marcflavour') eq 'UNIMARC' ) {
1933 $callnumber = $record->subfield( '200', 'f' );
1936 $callnumber = $record->subfield( '100', 'a' );
1939 # hint : the result is sorted by title.biblionumber because we can have X biblios with the same title
1940 # and we don't want to get only 1 result for each of them !!!
1941 $result{ $callnumber . $biblionumber } = $record;
1944 # sort the hash and return the same structure as GetRecords (Zebra querying)
1947 if ( $ordering eq 'call_number_dsc' ) { # sort by title desc
1948 foreach my $key ( sort { $b cmp $a } ( keys %result ) ) {
1949 $result_hash->{'RECORDS'}[ $numbers++ ] =
1950 $result{$key}->as_usmarc();
1953 else { # sort by title ASC
1954 foreach my $key ( sort { $a cmp $b } ( keys %result ) ) {
1955 $result_hash->{'RECORDS'}[ $numbers++ ] =
1956 $result{$key}->as_usmarc();
1959 my $finalresult = ();
1960 $result_hash->{'hits'} = $numbers;
1961 $finalresult->{'biblioserver'} = $result_hash;
1962 return $finalresult;
1964 elsif ( $ordering =~ /pubdate/ ) { #pub year
1966 foreach ( split /;/, $biblionumbers ) {
1967 my ( $biblionumber, $title ) = split /,/, $_;
1968 my $record = GetMarcBiblio($biblionumber);
1969 my ( $publicationyear_tag, $publicationyear_subfield ) =
1970 GetMarcFromKohaField( 'biblioitems.publicationyear', '' );
1971 my $publicationyear =
1972 $record->subfield( $publicationyear_tag,
1973 $publicationyear_subfield );
1975 # hint : the result is sorted by title.biblionumber because we can have X biblios with the same title
1976 # and we don't want to get only 1 result for each of them !!!
1977 $result{ $publicationyear . $biblionumber } = $record;
1980 # sort the hash and return the same structure as GetRecords (Zebra querying)
1983 if ( $ordering eq 'pubdate_dsc' ) { # sort by pubyear desc
1984 foreach my $key ( sort { $b cmp $a } ( keys %result ) ) {
1985 $result_hash->{'RECORDS'}[ $numbers++ ] =
1986 $result{$key}->as_usmarc();
1989 else { # sort by pub year ASC
1990 foreach my $key ( sort ( keys %result ) ) {
1991 $result_hash->{'RECORDS'}[ $numbers++ ] =
1992 $result{$key}->as_usmarc();
1995 my $finalresult = ();
1996 $result_hash->{'hits'} = $numbers;
1997 $finalresult->{'biblioserver'} = $result_hash;
1998 return $finalresult;
2004 elsif ( $ordering =~ /title/ ) {
2006 # the title is in the biblionumbers string, so we just need to build a hash, sort it and return
2008 foreach ( split /;/, $biblionumbers ) {
2009 my ( $biblionumber, $title ) = split /,/, $_;
2011 # hint : the result is sorted by title.biblionumber because we can have X biblios with the same title
2012 # and we don't want to get only 1 result for each of them !!!
2013 # hint & speed improvement : we can order without reading the record
2014 # so order, and read records only for the requested page !
2015 $result{ $title . $biblionumber } = $biblionumber;
2018 # sort the hash and return the same structure as GetRecords (Zebra querying)
2021 if ( $ordering eq 'title_az' ) { # sort by title desc
2022 foreach my $key ( sort ( keys %result ) ) {
2023 $result_hash->{'RECORDS'}[ $numbers++ ] = $result{$key};
2026 else { # sort by title ASC
2027 foreach my $key ( sort { $b cmp $a } ( keys %result ) ) {
2028 $result_hash->{'RECORDS'}[ $numbers++ ] = $result{$key};
2032 # limit the $results_per_page to result size if it's more
2033 $results_per_page = $numbers - 1 if $numbers < $results_per_page;
2035 # for the requested page, replace biblionumber by the complete record
2036 # speed improvement : avoid reading too much things
2038 my $counter = $offset ;
2039 $counter <= $offset + $results_per_page ;
2043 $result_hash->{'RECORDS'}[$counter] =
2044 GetMarcBiblio( $result_hash->{'RECORDS'}[$counter] )->as_usmarc;
2046 my $finalresult = ();
2047 $result_hash->{'hits'} = $numbers;
2048 $finalresult->{'biblioserver'} = $result_hash;
2049 return $finalresult;
2056 # we need 2 hashes to order by ranking : the 1st one to count the ranking, the 2nd to order by ranking
2059 foreach ( split /;/, $biblionumbers ) {
2060 my ( $biblionumber, $title ) = split /,/, $_;
2061 $title =~ /(.*)-(\d)/;
2066 # note that we + the ranking because ranking is calculated on weight of EACH term requested.
2067 # if we ask for "two towers", and "two" has weight 2 in biblio N, and "towers" has weight 4 in biblio N
2068 # biblio N has ranking = 6
2069 $count_ranking{$biblionumber} += $ranking;
2072 # build the result by "inverting" the count_ranking hash
2073 # 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
2075 foreach ( keys %count_ranking ) {
2076 $result{ sprintf( "%10d", $count_ranking{$_} ) . '-' . $_ } = $_;
2079 # sort the hash and return the same structure as GetRecords (Zebra querying)
2082 foreach my $key ( sort { $b cmp $a } ( keys %result ) ) {
2083 $result_hash->{'RECORDS'}[ $numbers++ ] = $result{$key};
2086 # limit the $results_per_page to result size if it's more
2087 $results_per_page = $numbers - 1 if $numbers < $results_per_page;
2089 # for the requested page, replace biblionumber by the complete record
2090 # speed improvement : avoid reading too much things
2092 my $counter = $offset ;
2093 $counter <= $offset + $results_per_page ;
2097 $result_hash->{'RECORDS'}[$counter] =
2098 GetMarcBiblio( $result_hash->{'RECORDS'}[$counter] )->as_usmarc
2099 if $result_hash->{'RECORDS'}[$counter];
2101 my $finalresult = ();
2102 $result_hash->{'hits'} = $numbers;
2103 $finalresult->{'biblioserver'} = $result_hash;
2104 return $finalresult;
2108 END { } # module clean-up code here (global destructor)
2115 Koha Developement team <info@koha.org>