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
21 use C4::Biblio; # GetMarcFromKohaField
22 use C4::Koha; # getFacets
24 use C4::Dates qw(format_date);
26 use vars qw($VERSION @ISA @EXPORT @EXPORT_OK %EXPORT_TAGS $DEBUG);
28 # set the version for version checking
31 $DEBUG = ($ENV{DEBUG}) ? 1 : 0;
36 C4::Search - Functions for searching the Koha catalog.
40 see opac/opac-search.pl or catalogue/search.pl for example of usage
44 This module provides the searching facilities for the Koha into a zebra catalog.
62 # make all your functions, whether exported or not;
64 =head2 findseealso($dbh,$fields);
66 C<$dbh> is a link to the DB handler.
69 my $dbh =C4::Context->dbh;
71 C<$fields> is a reference to the fields array
73 This function modify the @$fields array and add related fields to search on.
78 my ( $dbh, $fields ) = @_;
79 my $tagslib = GetMarcStructure( 1 );
80 for ( my $i = 0 ; $i <= $#{$fields} ; $i++ ) {
81 my ($tag) = substr( @$fields[$i], 1, 3 );
82 my ($subfield) = substr( @$fields[$i], 4, 1 );
83 @$fields[$i] .= ',' . $tagslib->{$tag}->{$subfield}->{seealso}
84 if ( $tagslib->{$tag}->{$subfield}->{seealso} );
90 ($biblionumber,$biblionumber,$title) = FindDuplicate($record);
96 my $dbh = C4::Context->dbh;
97 my $result = TransformMarcToKoha( $dbh, $record, '' );
102 my ( $biblionumber, $title );
104 # search duplicate on ISBN, easy and fast..
105 # ... normalize first
106 if ( $result->{isbn} ) {
107 $result->{isbn} =~ s/\(.*$//;
108 $result->{isbn} =~ s/\s+$//;
110 #$search->{'avoidquerylog'}=1;
111 if ( $result->{isbn} ) {
112 $query = "isbn=$result->{isbn}";
115 $result->{title} =~ s /\\//g;
116 $result->{title} =~ s /\"//g;
117 $result->{title} =~ s /\(//g;
118 $result->{title} =~ s /\)//g;
119 # remove valid operators
120 $result->{title} =~ s/(and|or|not)//g;
121 $query = "ti,ext=$result->{title}";
122 $query .= " and itemtype=$result->{itemtype}" if ($result->{itemtype});
123 if ($result->{author}){
124 $result->{author} =~ s /\\//g;
125 $result->{author} =~ s /\"//g;
126 $result->{author} =~ s /\(//g;
127 $result->{author} =~ s /\)//g;
128 # remove valid operators
129 $result->{author} =~ s/(and|or|not)//g;
130 $query .= " and au,ext=$result->{author}";
133 my ($error,$searchresults) =
134 SimpleSearch($query); # FIXME :: hardcoded !
136 foreach my $possible_duplicate_record (@$searchresults) {
138 MARC::Record->new_from_usmarc($possible_duplicate_record);
139 my $result = TransformMarcToKoha( $dbh, $marcrecord, '' );
141 # FIXME :: why 2 $biblionumber ?
143 push @results, $result->{'biblionumber'};
144 push @results, $result->{'title'};
152 ($error,$results) = SimpleSearch($query,@servers);
154 this function performs a simple search on the catalog using zoom.
160 * $query could be a simple keyword or a complete CCL query wich is depending on your ccl file.
161 * @servers is optionnal. default one is read on koha.xml
164 * $error is a string which containt the description error if there is one. Else it's empty.
165 * \@results is an array of marc record.
167 =item C<usage in the script:>
171 my ($error, $marcresults) = SimpleSearch($query);
173 if (defined $error) {
174 $template->param(query_error => $error);
175 warn "error: ".$error;
176 output_html_with_http_headers $input, $cookie, $template->output;
180 my $hits = scalar @$marcresults;
183 for(my $i=0;$i<$hits;$i++) {
185 my $marcrecord = MARC::File::USMARC::decode($marcresults->[$i]);
186 my $biblio = TransformMarcToKoha(C4::Context->dbh,$marcrecord,'');
188 #build the hash for the template.
189 $resultsloop{highlight} = ($i % 2)?(1):(0);
190 $resultsloop{title} = $biblio->{'title'};
191 $resultsloop{subtitle} = $biblio->{'subtitle'};
192 $resultsloop{biblionumber} = $biblio->{'biblionumber'};
193 $resultsloop{author} = $biblio->{'author'};
194 $resultsloop{publishercode} = $biblio->{'publishercode'};
195 $resultsloop{publicationyear} = $biblio->{'publicationyear'};
197 push @results, \%resultsloop;
199 $template->param(result=>\@results);
205 if (C4::Context->preference('NoZebra')) {
206 my $result = NZorder(NZanalyse($query))->{'biblioserver'}->{'RECORDS'};
207 return (undef,$result);
213 return ( "No query entered", undef ) unless $query;
215 #@servers = (C4::Context->config("biblioserver")) unless @servers;
217 ("biblioserver") unless @servers
218 ; # FIXME hardcoded value. See catalog/search.pl & opac-search.pl too.
221 for ( my $i = 0 ; $i < @servers ; $i++ ) {
223 $zconns[$i] = C4::Context->Zconn( $servers[$i], 1 );
226 ->search( new ZOOM::Query::CCL2RPN( $query, $zconns[$i] ) );
228 # getting error message if one occured.
230 $zconns[$i]->errmsg() . " ("
231 . $zconns[$i]->errcode() . ") "
232 . $zconns[$i]->addinfo() . " "
233 . $zconns[$i]->diagset();
235 return ( $error, undef ) if $zconns[$i]->errcode();
238 # caught a ZOOM::Exception
242 . $@->addinfo() . " "
245 return ( $error, undef );
250 while ( ( my $i = ZOOM::event( \@zconns ) ) != 0 ) {
251 $ev = $zconns[ $i - 1 ]->last_event();
252 if ( $ev == ZOOM::Event::ZEND ) {
253 $hits = $tmpresults[ $i - 1 ]->size();
256 for ( my $j = 0 ; $j < $hits ; $j++ ) {
257 my $record = $tmpresults[ $i - 1 ]->record($j)->raw();
258 push @results, $record;
262 return ( undef, \@results );
266 # performs the search
269 $koha_query, $simple_query, $sort_by_ref,
270 $servers_ref, $results_per_page, $offset,
271 $expanded_facet, $branches, $query_type,
274 # warn "Query : $koha_query";
275 my @servers = @$servers_ref;
276 my @sort_by = @$sort_by_ref;
278 # create the zoom connection and query object
282 my $results_hashref = ();
285 my $facets_counter = ();
286 my $facets_info = ();
287 my $facets = getFacets();
289 #### INITIALIZE SOME VARS USED CREATE THE FACETED RESULTS
290 my @facets_loop; # stores the ref to array of hashes for template
291 for ( my $i = 0 ; $i < @servers ; $i++ ) {
292 $zconns[$i] = C4::Context->Zconn( $servers[$i], 1 );
294 # perform the search, create the results objects
295 # if this is a local search, use the $koha-query, if it's a federated one, use the federated-query
297 if ( $servers[$i] =~ /biblioserver/ ) {
298 $query_to_use = $koha_query;
301 $query_to_use = $simple_query;
304 #$query_to_use = $simple_query if $scan;
305 #warn $simple_query if ($scan && $DEBUG);
306 # check if we've got a query_type defined
310 if ( $query_type =~ /^ccl/ ) {
312 s/\:/\=/g; # change : to = last minute (FIXME)
314 # warn "CCL : $query_to_use";
317 new ZOOM::Query::CCL2RPN( $query_to_use, $zconns[$i] )
320 elsif ( $query_type =~ /^cql/ ) {
322 # warn "CQL : $query_to_use";
325 new ZOOM::Query::CQL( $query_to_use, $zconns[$i] ) );
327 elsif ( $query_type =~ /^pqf/ ) {
329 # warn "PQF : $query_to_use";
332 new ZOOM::Query::PQF( $query_to_use, $zconns[$i] ) );
337 # warn "preparing to scan:$query_to_use";
340 new ZOOM::Query::CCL2RPN( $query_to_use, $zconns[$i] )
344 # warn "LAST : $query_to_use";
347 new ZOOM::Query::CCL2RPN( $query_to_use, $zconns[$i] )
353 warn "WARNING: query problem with $query_to_use " . $@;
356 # concatenate the sort_by limits and pass them to the results object
358 foreach my $sort (@sort_by) {
359 if ($sort eq "author_az") {
360 $sort_by.="1=1003 <i ";
362 elsif ($sort eq "author_za") {
363 $sort_by.="1=1003 >i ";
365 elsif ($sort eq "popularity_asc") {
366 $sort_by.="1=9003 <i ";
368 elsif ($sort eq "popularity_dsc") {
369 $sort_by.="1=9003 >i ";
371 elsif ($sort eq "call_number_asc") {
372 $sort_by.="1=20 <i ";
374 elsif ($sort eq "call_number_dsc") {
375 $sort_by.="1=20 >i ";
377 elsif ($sort eq "pubdate_asc") {
378 $sort_by.="1=31 <i ";
380 elsif ($sort eq "pubdate_dsc") {
381 $sort_by.="1=31 >i ";
383 elsif ($sort eq "acqdate_asc") {
384 $sort_by.="1=32 <i ";
386 elsif ($sort eq "acqdate_dsc") {
387 $sort_by.="1=32 >i ";
389 elsif ($sort eq "title_az") {
392 elsif ($sort eq "title_za") {
397 if ( $results[$i]->sort( "yaz", $sort_by ) < 0) {
398 warn "WARNING sort $sort_by failed";
402 while ( ( my $i = ZOOM::event( \@zconns ) ) != 0 ) {
403 my $ev = $zconns[ $i - 1 ]->last_event();
404 if ( $ev == ZOOM::Event::ZEND ) {
405 my $size = $results[ $i - 1 ]->size();
408 #$results_hash->{'server'} = $servers[$i-1];
409 # loop through the results
410 $results_hash->{'hits'} = $size;
412 if ( $offset + $results_per_page <= $size ) {
413 $times = $offset + $results_per_page;
418 for ( my $j = $offset ; $j < $times ; $j++ )
419 { #(($offset+$count<=$size) ? ($offset+$count):$size) ; $j++){
423 ## This is just an index scan
425 my ( $term, $occ ) = $results[ $i - 1 ]->term($j);
426 # here we create a minimal MARC record and hand it off to the
427 # template just like a normal result ... perhaps not ideal, but
429 my $tmprecord = MARC::Record->new();
430 $tmprecord->encoding('UTF-8');
433 # srote the minimal record in author/title (depending on MARC flavour)
434 if ( C4::Context->preference("marcflavour") eq
437 $tmptitle = MARC::Field->new(
444 $tmptitle = MARC::Field->new(
450 $tmprecord->append_fields($tmptitle);
451 $results_hash->{'RECORDS'}[$j] =
452 $tmprecord->as_usmarc();
455 $record = $results[ $i - 1 ]->record($j)->raw();
457 #warn "RECORD $j:".$record;
458 $results_hash->{'RECORDS'}[$j] =
459 $record; # making a reference to a hash
460 # Fill the facets while we're looping
461 $facet_record = MARC::Record->new_from_usmarc($record);
463 #warn $servers[$i-1].$facet_record->title();
464 for ( my $k = 0 ; $k <= @$facets ; $k++ ) {
465 if ( $facets->[$k] ) {
467 for my $tag ( @{ $facets->[$k]->{'tags'} } ) {
468 push @fields, $facet_record->field($tag);
470 for my $field (@fields) {
471 my @subfields = $field->subfields();
472 for my $subfield (@subfields) {
473 my ( $code, $data ) = @$subfield;
475 $facets->[$k]->{'subfield'} )
477 $facets_counter->{ $facets->[$k]
478 ->{'link_value'} }->{$data}++;
482 $facets_info->{ $facets->[$k]->{'link_value'} }
484 $facets->[$k]->{'label_value'};
485 $facets_info->{ $facets->[$k]->{'link_value'} }
486 ->{'expanded'} = $facets->[$k]->{'expanded'};
491 $results_hashref->{ $servers[ $i - 1 ] } = $results_hash;
494 #print "connection ", $i-1, ": $size hits";
495 #print $results[$i-1]->record(0)->render() if $size > 0;
498 sort { $facets_counter->{$b} <=> $facets_counter->{$a} }
499 keys %$facets_counter
503 my $number_of_facets;
504 my @this_facets_array;
507 $facets_counter->{$link_value}
508 ->{$b} <=> $facets_counter->{$link_value}->{$a}
509 } keys %{ $facets_counter->{$link_value} }
513 if ( ( $number_of_facets < 6 )
514 || ( $expanded_facet eq $link_value )
515 || ( $facets_info->{$link_value}->{'expanded'} ) )
518 # sanitize the link value ), ( will cause errors with CCL
519 my $facet_link_value = $one_facet;
520 $facet_link_value =~ s/(\(|\))/ /g;
522 # fix the length that will display in the label
523 my $facet_label_value = $one_facet;
524 $facet_label_value = substr( $one_facet, 0, 20 ) . "..."
525 unless length($facet_label_value) <= 20;
527 # well, if it's a branch, label by the name, not the code
528 if ( $link_value =~ /branch/ ) {
530 $branches->{$one_facet}->{'branchname'};
533 # but we're down with the whole label being in the link's title
534 my $facet_title_value = $one_facet;
536 push @this_facets_array,
540 $facets_counter->{$link_value}->{$one_facet},
541 facet_label_value => $facet_label_value,
542 facet_title_value => $facet_title_value,
543 facet_link_value => $facet_link_value,
544 type_link_value => $link_value,
549 unless ( $facets_info->{$link_value}->{'expanded'} ) {
551 if ( ( $number_of_facets > 6 )
552 && ( $expanded_facet ne $link_value ) );
557 type_link_value => $link_value,
558 type_id => $link_value . "_id",
560 $facets_info->{$link_value}->{'label_value'},
561 facets => \@this_facets_array,
562 expandable => $expandable,
563 expand => $link_value,
569 return ( undef, $results_hashref, \@facets_loop );
573 sub _remove_stopwords {
574 my ($operand,$index) = @_;
575 my @stopwords_removed;
576 # phrase and exact-qualified indexes shouldn't have stopwords removed
577 if ($index!~m/phr|ext/){
578 # remove stopwords from operand : parse all stopwords & remove them (case insensitive)
579 # we use IsAlpha unicode definition, to deal correctly with diacritics.
580 # otherwise, a French word like "leçon" woudl be split into "le" "çon", le
581 # is an empty word, we'd get "çon" and wouldn't find anything...
582 foreach (keys %{C4::Context->stopwords}) {
583 next if ($_ =~/(and|or|not)/); # don't remove operators
584 if ($operand =~ /(\P{IsAlpha}$_\P{IsAlpha}|^$_\P{IsAlpha}|\P{IsAlpha}$_$)/) {
585 $operand=~ s/\P{IsAlpha}$_\P{IsAlpha}/ /gi;
586 $operand=~ s/^$_\P{IsAlpha}/ /gi;
587 $operand=~ s/\P{IsAlpha}$_$/ /gi;
588 push @stopwords_removed, $_;
592 return ($operand, \@stopwords_removed);
596 sub _detect_truncation {
597 my ($operand,$index) = @_;
598 my (@nontruncated,@righttruncated,@lefttruncated,@rightlefttruncated,@regexpr);
600 my @wordlist= split (/\s/,$operand);
601 foreach my $word (@wordlist){
602 if ($word=~s/^\*([^\*]+)\*$/$1/){
603 push @rightlefttruncated,$word;
605 elsif($word=~s/^\*([^\*]+)$/$1/){
606 push @lefttruncated,$word;
608 elsif ($word=~s/^([^\*]+)\*$/$1/){
609 push @righttruncated,$word;
611 elsif (index($word,"*")<0){
612 push @nontruncated,$word;
618 return (\@nontruncated,\@righttruncated,\@lefttruncated,\@rightlefttruncated,\@regexpr);
621 sub _build_stemmed_operand {
624 # FIXME: the locale should be set based on the user's language and/or search choice
625 my $stemmer = Lingua::Stem->new( -locale => 'EN-US' );
626 # FIXME: these should be stored in the db so the librarian can modify the behavior
627 $stemmer->add_exceptions(
635 my @words = split( / /, $operand );
636 my $stems = $stemmer->stem(@words);
637 for my $stem (@$stems) {
638 $stemmed_operand .= "$stem";
639 $stemmed_operand .= "?" unless ( $stem =~ /(and$|or$|not$)/ ) || ( length($stem) < 3 );
640 $stemmed_operand .= " ";
642 #warn "STEMMED OPERAND: $stemmed_operand";
643 return $stemmed_operand;
646 sub _build_weighted_query {
647 # FIELD WEIGHTING - This is largely experimental stuff. What I'm committing works
648 # pretty well but will work much better when we have an actual query parser
649 my ($operand,$stemmed_operand,$index) = @_;
650 my $stemming = C4::Context->preference("QueryStemming") || 0;
651 my $weight_fields = C4::Context->preference("QueryWeightFields") || 0;
652 my $fuzzy_enabled = C4::Context->preference("QueryFuzzy") || 0;
654 my $weighted_query .= "(rk=("; # Specifies that we're applying rank
656 # Keyword, or, no index specified
657 if ( ( $index eq 'kw' ) || ( !$index ) ) {
658 $weighted_query .= "Title-cover,ext,r1=\"$operand\""; # exact title-cover
659 $weighted_query .= " or ti,ext,r2=\"$operand\""; # exact title
660 $weighted_query .= " or ti,phr,r3=\"$operand\""; # phrase title
661 #$weighted_query .= " or any,ext,r4=$operand"; # exact any
662 #$weighted_query .=" or kw,wrdl,r5=\"$operand\""; # word list any
663 $weighted_query .= " or wrd,fuzzy,r8=\"$operand\"" if $fuzzy_enabled; # add fuzzy, word list
664 $weighted_query .= " or wrd,right-Truncation,r9=\"$stemmed_operand\"" if ($stemming and $stemmed_operand); # add stemming, right truncation
665 # embedded sorting: 0 a-z; 1 z-a
666 # $weighted_query .= ") or (sort1,aut=1";
668 elsif ( $index eq 'bc' ) {
669 $weighted_query .= "bc=\"$operand\"";
671 # if the index already has more than one qualifier, just wrap the operand
672 # in quotes and pass it back
673 elsif ($index =~ ',') {
674 $weighted_query .=" $index=\"$operand\"";
676 #TODO: build better cases based on specific search indexes
678 $weighted_query .= " $index,ext,r1=\"$operand\""; # exact index
679 #$weighted_query .= " or (title-sort-az=0 or $index,startswithnt,st-word,r3=$operand #)";
680 $weighted_query .= " or $index,phr,r3=\"$operand\""; # phrase index
681 $weighted_query .= " or $index,rt,wrd,r3=\"$operand\""; # word list index
683 $weighted_query .= "))"; # close rank specification
684 return $weighted_query;
687 # build the query itself
689 my ( $operators, $operands, $indexes, $limits, $sort_by, $scan) = @_;
691 my @operators = @$operators if $operators;
692 my @indexes = @$indexes if $indexes;
693 my @operands = @$operands if $operands;
694 my @limits = @$limits if $limits;
695 my @sort_by = @$sort_by if $sort_by;
697 my $stemming = C4::Context->preference("QueryStemming") || 0;
698 my $auto_truncation = C4::Context->preference("QueryAutoTruncate") || 0;
699 my $weight_fields = C4::Context->preference("QueryWeightFields") || 0;
700 my $fuzzy_enabled = C4::Context->preference("QueryFuzzy") || 0;
701 # no stemming/weight/fuzzy in NoZebra
702 if (C4::Context->preference("NoZebra")) {
707 my $remove_stopwords = C4::Context->preference("QueryRemoveStopwords") || 0;
709 my $query = $operands[0];
710 my $simple_query = $operands[0];
719 my $stopwords_removed;
721 # for handling ccl, cql, pqf queries in diagnostic mode, skip the rest of the steps
723 if ( $query =~ /^ccl=/ ) {
724 return ( undef, $', $', $', $', '', '', '', '', 'ccl' );
726 if ( $query =~ /^cql=/ ) {
727 return ( undef, $', $', $', $', '', '', '', '', 'cql' );
729 if ( $query =~ /^pqf=/ ) {
730 return ( undef, $', $', $', $', '', '', '', '', 'pqf' );
733 # pass nested queries directly
734 if ( $query =~ /(\(|\))/ ) {
735 return ( undef, $query, $simple_query, $query_cgi, $query, $limit, $limit_cgi, $limit_desc, $stopwords_removed, 'ccl' );
738 # form-based queries are limited to non-nested at a specific depth, so we can easily
739 # modify the incoming query operands and indexes to do stemming and field weighting
740 # Once we do so, we'll end up with a value in $query, just like if we had an
741 # incoming $query from the user
743 $query = ""; # clear it out so we can populate properly with field-weighted stemmed query
744 my $previous_operand; # a flag used to keep track if there was a previous query
745 # if there was, we can apply the current operator
747 for ( my $i = 0 ; $i <= @operands ; $i++ ) {
749 # COMBINE OPERANDS, INDEXES AND OPERATORS
750 if ( $operands[$i] ) {
752 # a flag to determine whether or not to add the index to the query
754 # if the user is sophisticated enough to specify an index, turn off some defaults
755 if ($operands[$i] =~ /(:|=)/ || $scan) {
758 $remove_stopwords = 0;
760 my $operand = $operands[$i];
761 my $index = $indexes[$i];
763 # some helpful index modifs
764 my $index_plus = "$index:" if $index;
765 my $index_plus_comma="$index," if $index;
768 if ($remove_stopwords) {
769 ($operand, $stopwords_removed) = _remove_stopwords($operand,$index);
770 warn "OPERAND w/out STOPWORDS: >$operand<" if $DEBUG;
771 warn "REMOVED STOPWORDS: @$stopwords_removed" if ($stopwords_removed && $DEBUG);
775 my ($nontruncated,$righttruncated,$lefttruncated,$rightlefttruncated,$regexpr);
776 my $truncated_operand;
777 ($nontruncated,$righttruncated,$lefttruncated,$rightlefttruncated,$regexpr) = _detect_truncation($operand,$index);
778 warn "TRUNCATION: NON:>@$nontruncated< RIGHT:>@$righttruncated< LEFT:>@$lefttruncated< RIGHTLEFT:>@$rightlefttruncated< REGEX:>@$regexpr<" if $DEBUG;
781 if (scalar(@$righttruncated)+scalar(@$lefttruncated)+scalar(@$rightlefttruncated)>0){
782 # don't field weight or add the index to the query, we do it here
784 undef $weight_fields;
785 my $previous_truncation_operand;
786 if (scalar(@$nontruncated)>0) {
787 $truncated_operand.= "$index_plus @$nontruncated ";
788 $previous_truncation_operand = 1;
790 if (scalar(@$righttruncated)>0){
791 $truncated_operand .= "and " if $previous_truncation_operand;
792 $truncated_operand .= "$index_plus_comma"."rtrn:@$righttruncated ";
793 $previous_truncation_operand = 1;
795 if (scalar(@$lefttruncated)>0){
796 $truncated_operand .= "and " if $previous_truncation_operand;
797 $truncated_operand .= "$index_plus_comma"."ltrn:@$lefttruncated ";
798 $previous_truncation_operand = 1;
800 if (scalar(@$rightlefttruncated)>0){
801 $truncated_operand .= "and " if $previous_truncation_operand;
802 $truncated_operand .= "$index_plus_comma"."rltrn:@$rightlefttruncated ";
803 $previous_truncation_operand = 1;
806 $operand = $truncated_operand if $truncated_operand;
807 warn "TRUNCATED OPERAND: >$truncated_operand<" if $DEBUG;
811 $stemmed_operand = _build_stemmed_operand($operand) if $stemming;
812 warn "STEMMED OPERAND: >$stemmed_operand<" if $DEBUG;
814 # Handle Field Weighting
815 my $weighted_operand;
816 $weighted_operand = _build_weighted_query($operand,$stemmed_operand,$index) if $weight_fields;
817 warn "FIELD WEIGHTED OPERAND: >$weighted_operand<" if $DEBUG;
818 $operand = $weighted_operand if $weight_fields;
819 $indexes_set = 1 if $weight_fields;
821 # If there's a previous operand, we need to add an operator
822 if ($previous_operand) {
824 # user-specified operator
825 if ( $operators[$i-1] ) {
826 $query .= " $operators[$i-1] ";
827 $query .= " $index_plus " unless $indexes_set;
828 $query .= " $operand";
829 $query_cgi .="&op=$operators[$i-1]";
830 $query_cgi .="&idx=$index" if $index;
831 $query_cgi .="&q=$operands[$i]" if $operands[$i];
832 $query_desc .=" $operators[$i-1] $index_plus $operands[$i]";
835 # the default operator is and
838 $query .= "$index_plus " unless $indexes_set;
839 $query .= "$operand";
840 $query_cgi .="&op=and&idx=$index" if $index;
841 $query_cgi .="&q=$operands[$i]" if $operands[$i];
842 $query_desc .= " and $index_plus $operands[$i]";
846 # there isn't a pervious operand, don't need an operator
848 # field-weighted queries already have indexes set
849 $query .=" $index_plus " unless $indexes_set;
851 $query_desc .= " $index_plus $operands[$i]";
852 $query_cgi.="&idx=$index" if $index;
853 $query_cgi.="&q=$operands[$i]" if $operands[$i];
855 $previous_operand = 1;
860 warn "QUERY BEFORE LIMITS: >$query<" if $DEBUG;
864 my $availability_limit;
865 foreach my $this_limit (@limits) {
866 if ( $this_limit =~ /available/ ) {
867 # available is defined as (items.notloan is NULL) and (items.itemlost > 0 or NULL) (last clause handles NULL values for lost in zebra)
868 $availability_limit .="( ( allrecords,AlwaysMatches='' not onloan,AlwaysMatches='') and ((lost,st-numeric gt 0) or ( allrecords,AlwaysMatches='' not lost,AlwaysMatches='')) )";
869 $limit_cgi .= "&limit=available";
873 # these are treated as OR
874 elsif ( $this_limit =~ /mc/ ) {
875 $group_OR_limits .= " or " if $group_OR_limits;
876 $limit_desc .=" or " if $group_OR_limits;
877 $group_OR_limits .= "$this_limit";
878 $limit_cgi .="&limit=$this_limit";
879 $limit_desc .= "$this_limit";
884 $limit .= " and " if $limit || $query;
885 $limit .= "$this_limit";
886 $limit_cgi .="&limit=$this_limit";
887 $limit_desc .=" and $this_limit";
890 if ($group_OR_limits) {
891 $limit.=" and " if ($query || $limit );
892 $limit.="($group_OR_limits)";
894 if ($availability_limit) {
895 $limit.=" not " if ($query || $limit );
896 $limit.="$availability_limit";
898 # normalize the strings
901 for ($query, $query_desc, $limit, $limit_desc) {
902 $_ =~ s/ / /g; # remove extra spaces
903 $_ =~ s/^ //g; # remove any beginning spaces
904 $_ =~ s/ $//g; # remove any ending spaces
905 $_ =~ s/==/=/g; # remove double == from query
908 $query_cgi =~ s/^&//;
910 # append the limit to the query
913 warn "query=$query and limit=$limit" if $DEBUG;
915 warn "QUERY:".$query if $DEBUG;
916 warn "QUERY CGI:".$query_cgi if $DEBUG;
917 warn "QUERY DESC:".$query_desc if $DEBUG;
918 warn "LIMIT:".$limit if $DEBUG;
919 warn "LIMIT CGI:".$limit_cgi if $DEBUG;
920 warn "LIMIT DESC:".$limit_desc if $DEBUG;
922 return ( undef, $query,$simple_query,$query_cgi,$query_desc,$limit,$limit_cgi,$limit_desc,$stopwords_removed,$query_type );
925 # IMO this subroutine is pretty messy still -- it's responsible for
926 # building the HTML output for the template
928 my ( $searchdesc, $hits, $results_per_page, $offset, @marcresults ) = @_;
929 my $dbh = C4::Context->dbh;
933 my $span_terms_hashref;
934 for my $span_term ( split( / /, $searchdesc ) ) {
935 $span_term =~ s/(.*=|\)|\(|\+|\.)//g;
936 $span_terms_hashref->{$span_term}++;
939 #Build branchnames hash
941 #get branch information.....
944 $dbh->prepare("SELECT branchcode,branchname FROM branches")
945 ; # FIXME : use C4::Koha::GetBranches
947 while ( my $bdata = $bsth->fetchrow_hashref ) {
948 $branches{ $bdata->{'branchcode'} } = $bdata->{'branchname'};
952 #find itemtype & itemtype image
955 $dbh->prepare("SELECT itemtype,description,imageurl,summary,notforloan FROM itemtypes");
957 while ( my $bdata = $bsth->fetchrow_hashref ) {
958 $itemtypes{ $bdata->{'itemtype'} }->{description} =
959 $bdata->{'description'};
960 $itemtypes{ $bdata->{'itemtype'} }->{imageurl} = $bdata->{'imageurl'};
961 $itemtypes{ $bdata->{'itemtype'} }->{summary} = $bdata->{'summary'};
962 $itemtypes{ $bdata->{'itemtype'} }->{notforloan} = $bdata->{'notforloan'};
965 #search item field code
966 my $sth = $dbh->prepare("SELECT tagfield FROM marc_subfield_structure WHERE kohafield LIKE 'items.itemnumber'");
968 my ($itemtag) = $sth->fetchrow;
970 ## find column names of items related to MARC
971 my $sth2 = $dbh->prepare("SHOW COLUMNS FROM items");
973 my %subfieldstosearch;
974 while ( ( my $column ) = $sth2->fetchrow ) {
975 my ( $tagfield, $tagsubfield ) =
976 &GetMarcFromKohaField( "items." . $column, "" );
977 $subfieldstosearch{$column} = $tagsubfield;
981 if ( $hits && $offset + $results_per_page <= $hits ) {
982 $times = $offset + $results_per_page;
988 for ( my $i = $offset ; $i <= $times - 1 ; $i++ ) {
990 $marcrecord = MARC::File::USMARC::decode( $marcresults[$i] );
991 my $oldbiblio = TransformMarcToKoha( $dbh, $marcrecord, '' );
992 $oldbiblio->{result_number} = $i+1;
993 # add image url if there is one
994 if ( $itemtypes{ $oldbiblio->{itemtype} }->{imageurl} =~ /^http:/ ) {
995 $oldbiblio->{imageurl} =
996 $itemtypes{ $oldbiblio->{itemtype} }->{imageurl};
997 $oldbiblio->{description} =
998 $itemtypes{ $oldbiblio->{itemtype} }->{description};
1001 $oldbiblio->{imageurl} =
1002 getitemtypeimagesrc() . "/"
1003 . $itemtypes{ $oldbiblio->{itemtype} }->{imageurl}
1004 if ( $itemtypes{ $oldbiblio->{itemtype} }->{imageurl} );
1005 $oldbiblio->{description} =
1006 $itemtypes{ $oldbiblio->{itemtype} }->{description};
1009 # build summary if there is one (the summary is defined in itemtypes table
1011 if ($itemtypes{ $oldbiblio->{itemtype} }->{summary}) {
1012 my $summary = $itemtypes{ $oldbiblio->{itemtype} }->{summary};
1013 my @fields = $marcrecord->fields();
1014 foreach my $field (@fields) {
1015 my $tag = $field->tag();
1016 my $tagvalue = $field->as_string();
1017 $summary =~ s/\[(.?.?.?.?)$tag\*(.*?)]/$1$tagvalue$2\[$1$tag$2]/g;
1019 my @subf = $field->subfields;
1020 for my $i (0..$#subf) {
1021 my $subfieldcode = $subf[$i][0];
1022 my $subfieldvalue = $subf[$i][1];
1023 my $tagsubf = $tag.$subfieldcode;
1024 $summary =~ s/\[(.?.?.?.?)$tagsubf(.*?)]/$1$subfieldvalue$2\[$1$tagsubf$2]/g;
1028 $summary =~ s/\[(.*?)]//g;
1029 $summary =~ s/\n/<br>/g;
1030 $oldbiblio->{summary} = $summary;
1032 # add spans to search term in results for search term highlighting
1033 # save a native author, for the <a href=search.lq=<!--tmpl_var name="author"-->> link
1034 $oldbiblio->{'author_nospan'} = $oldbiblio->{'author'};
1035 foreach my $term ( keys %$span_terms_hashref ) {
1036 my $old_term = $term;
1037 if ( length($term) > 3 ) {
1038 $term =~ s/(.*=|\)|\(|\+|\.|\?|\[|\])//g;
1042 #FIXME: is there a better way to do this?
1043 $oldbiblio->{'title'} =~ s/$term/<span class=\"term\">$&<\/span>/gi;
1044 $oldbiblio->{'subtitle'} =~
1045 s/$term/<span class=\"term\">$&<\/span>/gi;
1047 $oldbiblio->{'author'} =~ s/$term/<span class=\"term\">$&<\/span>/gi;
1048 $oldbiblio->{'publishercode'} =~ s/$term/<span class=\"term\">$&<\/span>/gi;
1049 $oldbiblio->{'place'} =~ s/$term/<span class=\"term\">$&<\/span>/gi;
1050 $oldbiblio->{'pages'} =~ s/$term/<span class=\"term\">$&<\/span>/gi;
1051 $oldbiblio->{'notes'} =~ s/$term/<span class=\"term\">$&<\/span>/gi;
1052 $oldbiblio->{'size'} =~ s/$term/<span class=\"term\">$&<\/span>/gi;
1057 $toggle = "#ffffcc";
1062 $oldbiblio->{'toggle'} = $toggle;
1063 my @fields = $marcrecord->field($itemtag);
1066 my $ordered_count = 0;
1067 my $onloan_count = 0;
1068 my $wthdrawn_count = 0;
1069 my $itemlost_count = 0;
1073 # check the loan status of the item :
1074 # it is not stored in the MARC record, for pref (zebra reindexing)
1075 # reason. Thus, we have to get the status from a specific SQL query
1077 my $sth_issue = $dbh->prepare("
1078 SELECT date_due,returndate
1080 WHERE itemnumber=? AND returndate IS NULL");
1081 my $items_count=scalar(@fields);
1082 foreach my $field (@fields) {
1084 foreach my $code ( keys %subfieldstosearch ) {
1085 $item->{$code} = $field->subfield( $subfieldstosearch{$code} );
1087 $sth_issue->execute($item->{itemnumber});
1088 $item->{due_date} = format_date($sth_issue->fetchrow) if $sth_issue->fetchrow;
1089 $item->{onloan} = 1 if $item->{due_date};
1090 # at least one item can be reserved : suppose no
1092 if ( $item->{wthdrawn} ) {
1094 $items->{ $item->{'homebranch'}.'--'.$item->{'itemcallnumber'} }->{unavailable}=1;
1095 $items->{ $item->{'homebranch'}.'--'.$item->{'itemcallnumber'} }->{wthdrawn}=1;
1097 elsif ( $item->{itemlost} ) {
1099 $items->{ $item->{'homebranch'}.'--'.$item->{'itemcallnumber'} }->{unavailable}=1;
1100 $items->{ $item->{'homebranch'}.'--'.$item->{'itemcallnumber'} }->{itemlost}=1;
1102 unless ( $item->{notforloan}) {
1103 # OK, this one can be issued, so at least one can be reserved
1106 if ( ( $item->{onloan} ) && ( $item->{onloan} != '0000-00-00' ) )
1108 $items->{ $item->{'homebranch'}.'--'.$item->{'itemcallnumber'} }->{unavailable}=1;
1109 $items->{ $item->{'homebranch'}.'--'.$item->{'itemcallnumber'} }->{onloancount} = 1;
1110 $items->{ $item->{'homebranch'}.'--'.$item->{'itemcallnumber'} }->{due_date} = $item->{due_date};
1113 if ( $item->{'homebranch'} ) {
1114 $items->{ $item->{'homebranch'}.'--'.$item->{'itemcallnumber'} }->{count}++;
1118 elsif ( $item->{'holdingbranch'} ) {
1119 $items->{ $item->{'holdingbranch'} }->{count}++;
1121 $items->{ $item->{'homebranch'}.'--'.$item->{'itemcallnumber'} }->{itemcallnumber} = $item->{itemcallnumber};
1122 $items->{ $item->{'homebranch'}.'--'.$item->{'itemcallnumber'} }->{location} = $item->{location};
1123 $items->{ $item->{'homebranch'}.'--'.$item->{'itemcallnumber'} }->{branchcode} = $item->{homebranch};
1124 } # notforloan, item level and biblioitem level
1126 # last check for norequest : if itemtype is notforloan, it can't be reserved either, whatever the items
1127 $norequests = 1 if $itemtypes{$oldbiblio->{itemtype}}->{notforloan};
1129 for my $key ( sort keys %$items ) {
1132 branchname => $branches{$items->{$key}->{branchcode}},
1133 branchcode => $items->{$key}->{branchcode},
1134 count => $items->{$key}->{count},
1135 itemcallnumber => $items->{$key}->{itemcallnumber},
1136 location => $items->{$key}->{location},
1137 onloancount => $items->{$key}->{onloancount},
1138 due_date => $items->{$key}->{due_date},
1139 wthdrawn => $items->{$key}->{wthdrawn},
1140 lost => $items->{$key}->{itemlost},
1142 # only show the number specified by the user
1143 my $maxitems = (C4::Context->preference('maxItemsinSearchResults')) ? C4::Context->preference('maxItemsinSearchResults')- 1 : 1;
1144 push @items_loop, $this_item unless $itemscount > $maxitems;;
1146 $oldbiblio->{norequests} = $norequests;
1147 $oldbiblio->{items_count} = $items_count;
1148 $oldbiblio->{items_loop} = \@items_loop;
1149 $oldbiblio->{onloancount} = $onloan_count;
1150 $oldbiblio->{wthdrawncount} = $wthdrawn_count;
1151 $oldbiblio->{itemlostcount} = $itemlost_count;
1152 $oldbiblio->{orderedcount} = $ordered_count;
1153 $oldbiblio->{isbn} =~ s/-//g; # deleting - in isbn to enable amazon content
1154 push( @newresults, $oldbiblio );
1161 #----------------------------------------------------------------------
1163 # Non-Zebra GetRecords#
1164 #----------------------------------------------------------------------
1168 NZgetRecords has the same API as zera getRecords, even if some parameters are not managed
1172 my ($query,$simple_query,$sort_by_ref,$servers_ref,$results_per_page,$offset,$expanded_facet,$branches,$query_type,$scan) = @_;
1173 my $result = NZanalyse($query);
1174 return (undef,NZorder($result,@$sort_by_ref[0],$results_per_page,$offset),undef);
1179 NZanalyse : get a CQL string as parameter, and returns a list of biblionumber;title,biblionumber;title,...
1180 the list is built from an inverted index in the nozebra SQL table
1181 note that title is here only for convenience : the sorting will be very fast when requested on title
1182 if the sorting is requested on something else, we will have to reread all results, and that may be longer.
1187 my ($string,$server) = @_;
1188 # $server contains biblioserver or authorities, depending on what we search on.
1189 #warn "querying : $string on $server";
1190 $server='biblioserver' unless $server;
1192 # if we have a ", replace the content to discard temporarily any and/or/not inside
1194 if ($string =~/"/) {
1195 $string =~ s/"(.*?)"/__X__/;
1197 warn "commacontent : $commacontent" if $DEBUG;
1199 # split the query string in 3 parts : X AND Y means : $left="X", $operand="AND" and $right="Y"
1200 # then, call again NZanalyse with $left and $right
1201 # (recursive until we find a leaf (=> something without and/or/not)
1202 # delete repeated operator... Would then go in infinite loop
1203 while ($string =~s/( and| or| not| AND| OR| NOT)\1/$1/g){
1205 #process parenthesis before.
1206 if ($string =~ /^\s*\((.*)\)(( and | or | not | AND | OR | NOT )(.*))?/){
1208 # warn "left :".$left;
1210 my $operator = lc($3); # FIXME: and/or/not are operators, not operands
1211 my $leftresult = NZanalyse($left,$server);
1213 my $rightresult = NZanalyse($right,$server);
1214 # OK, we have the results for right and left part of the query
1215 # depending of operand, intersect, union or exclude both lists
1216 # to get a result list
1217 if ($operator eq ' and ') {
1218 my @leftresult = split /;/, $leftresult;
1219 # my @rightresult = split /;/,$leftresult;
1221 # parse the left results, and if the biblionumber exist in the right result, save it in finalresult
1222 # the result is stored twice, to have the same weight for AND than OR.
1223 # example : TWO : 61,61,64,121 (two is twice in the biblio #61) / TOWER : 61,64,130
1224 # result : 61,61,61,61,64,64 for two AND tower : 61 has more weight than 64
1225 foreach (@leftresult) {
1226 if ($rightresult =~ "$_;") {
1227 $finalresult .= "$_;$_;";
1230 return $finalresult;
1231 } elsif ($operator eq ' or ') {
1232 # just merge the 2 strings
1233 return $leftresult.$rightresult;
1234 } elsif ($operator eq ' not ') {
1235 my @leftresult = split /;/, $leftresult;
1236 # my @rightresult = split /;/,$leftresult;
1238 foreach (@leftresult) {
1239 unless ($rightresult =~ "$_;") {
1240 $finalresult .= "$_;";
1243 return $finalresult;
1245 # this error is impossible, because of the regexp that isolate the operand, but just in case...
1251 warn "string :".$string if $DEBUG;
1252 $string =~ /(.*?)( and | or | not | AND | OR | NOT )(.*)/;
1255 my $operand = lc($2); # FIXME: and/or/not are operators, not operands
1256 # it's not a leaf, we have a and/or/not
1258 # reintroduce comma content if needed
1259 $right =~ s/__X__/"$commacontent"/ if $commacontent;
1260 $left =~ s/__X__/"$commacontent"/ if $commacontent;
1261 warn "node : $left / $operand / $right\n" if $DEBUG;
1262 my $leftresult = NZanalyse($left,$server);
1263 my $rightresult = NZanalyse($right,$server);
1264 # OK, we have the results for right and left part of the query
1265 # depending of operand, intersect, union or exclude both lists
1266 # to get a result list
1267 if ($operand eq ' and ') {
1268 my @leftresult = split /;/, $leftresult;
1269 # my @rightresult = split /;/,$leftresult;
1271 # parse the left results, and if the biblionumber exist in the right result, save it in finalresult
1272 # the result is stored twice, to have the same weight for AND than OR.
1273 # example : TWO : 61,61,64,121 (two is twice in the biblio #61) / TOWER : 61,64,130
1274 # result : 61,61,61,61,64,64 for two AND tower : 61 has more weight than 64
1275 foreach (@leftresult) {
1276 if ($rightresult =~ "$_;") {
1277 $finalresult .= "$_;$_;";
1280 return $finalresult;
1281 } elsif ($operand eq ' or ') {
1282 # just merge the 2 strings
1283 return $leftresult.$rightresult;
1284 } elsif ($operand eq ' not ') {
1285 my @leftresult = split /;/, $leftresult;
1286 # my @rightresult = split /;/,$leftresult;
1288 foreach (@leftresult) {
1289 unless ($rightresult =~ "$_;") {
1290 $finalresult .= "$_;";
1293 return $finalresult;
1295 # this error is impossible, because of the regexp that isolate the operand, but just in case...
1296 die "error : operand unknown : $operand for $string";
1298 # it's a leaf, do the real SQL query and return the result
1300 $string =~ s/__X__/"$commacontent"/ if $commacontent;
1301 $string =~ s/-|\.|\?|,|;|!|'|\(|\)|\[|\]|{|}|"|&|\+|\*|\// /g;
1302 warn "leaf:$string" if $DEBUG;
1303 # parse the string in in operator/operand/value again
1304 $string =~ /(.*)(>=|<=)(.*)/;
1308 unless ($operator) {
1309 $string =~ /(.*)(>|<|=)(.*)/;
1315 # automatic replace for short operators
1316 $left='title' if $left =~ '^ti$';
1317 $left='author' if $left =~ '^au$';
1318 $left='publisher' if $left =~ '^pb$';
1319 $left='subject' if $left =~ '^su$';
1320 $left='koha-Auth-Number' if $left =~ '^an$';
1321 $left='keyword' if $left =~ '^kw$';
1322 if ($operator && $left ne 'keyword' ) {
1323 # do a specific search
1324 my $dbh = C4::Context->dbh;
1325 $operator='LIKE' if $operator eq '=' and $right=~ /%/;
1326 my $sth = $dbh->prepare("SELECT biblionumbers,value FROM nozebra WHERE server=? AND indexname=? AND value $operator ?");
1327 warn "$left / $operator / $right\n";
1328 # split each word, query the DB and build the biblionumbers result
1329 #sanitizing leftpart
1330 $left=~s/^\s+|\s+$//;
1331 my ($biblionumbers,$value);
1332 foreach (split / /,$right) {
1334 warn "EXECUTE : $server, $left, $_";
1335 $sth->execute($server, $left, $_) or warn "execute failed: $!";
1336 while (my ($line,$value) = $sth->fetchrow) {
1337 # if we are dealing with a numeric value, use only numeric results (in case of >=, <=, > or <)
1338 # otherwise, fill the result
1339 $biblionumbers .= $line unless ($right =~ /\d/ && $value =~ /\D/);
1340 # warn "result : $value ". ($right =~ /\d/) . "==".(!$value =~ /\d/) ;#= $line";
1342 # do a AND with existing list if there is one, otherwise, use the biblionumbers list as 1st result list
1344 my @leftresult = split /;/, $biblionumbers;
1346 foreach my $entry (@leftresult) { # $_ contains biblionumber,title-weight
1347 # remove weight at the end
1348 my $cleaned = $entry;
1349 $cleaned =~ s/-\d*$//;
1350 # if the entry already in the hash, take it & increase weight
1351 warn "===== $cleaned =====" if $DEBUG;
1352 if ($results =~ "$cleaned") {
1353 $temp .= "$entry;$entry;";
1354 warn "INCLUDING $entry" if $DEBUG;
1359 $results = $biblionumbers;
1363 #do a complete search (all indexes)
1364 my $dbh = C4::Context->dbh;
1365 my $sth = $dbh->prepare("SELECT biblionumbers FROM nozebra WHERE server=? AND value LIKE ?");
1366 # split each word, query the DB and build the biblionumbers result
1367 foreach (split / /,$string) {
1368 next if C4::Context->stopwords->{uc($_)}; # skip if stopword
1369 warn "search on all indexes on $_" if $DEBUG;
1372 $sth->execute($server, $_);
1373 while (my $line = $sth->fetchrow) {
1374 $biblionumbers .= $line;
1376 # do a AND with existing list if there is one, otherwise, use the biblionumbers list as 1st result list
1378 warn "RES for $_ = $biblionumbers" if $DEBUG;
1379 my @leftresult = split /;/, $biblionumbers;
1381 foreach my $entry (@leftresult) { # $_ contains biblionumber,title-weight
1382 # remove weight at the end
1383 my $cleaned = $entry;
1384 $cleaned =~ s/-\d*$//;
1385 # if the entry already in the hash, take it & increase weight
1386 warn "===== $cleaned =====" if $DEBUG;
1387 if ($results =~ "$cleaned") {
1388 $temp .= "$entry;$entry;";
1389 warn "INCLUDING $entry" if $DEBUG;
1394 warn "NEW RES for $_ = $biblionumbers" if $DEBUG;
1395 $results = $biblionumbers;
1399 # warn "return : $results for LEAF : $string" if $DEBUG;
1406 $finalresult = NZorder($biblionumbers, $ordering,$results_per_page,$offset);
1414 my ($biblionumbers, $ordering,$results_per_page,$offset) = @_;
1415 # order title asc by default
1416 # $ordering = '1=36 <i' unless $ordering;
1417 $results_per_page=20 unless $results_per_page;
1418 $offset = 0 unless $offset;
1419 my $dbh = C4::Context->dbh;
1421 # order by POPULARITY
1423 if ($ordering =~ /popularity/) {
1426 # popularity is not in MARC record, it's builded from a specific query
1427 my $sth = $dbh->prepare("select sum(issues) from items where biblionumber=?");
1428 foreach (split /;/,$biblionumbers) {
1429 my ($biblionumber,$title) = split /,/,$_;
1430 $result{$biblionumber}=GetMarcBiblio($biblionumber);
1431 $sth->execute($biblionumber);
1432 my $popularity= $sth->fetchrow ||0;
1433 # hint : the key is popularity.title because we can have
1434 # many results with the same popularity. In this cas, sub-ordering is done by title
1435 # we also have biblionumber to avoid bug for 2 biblios with the same title & popularity
1436 # (un-frequent, I agree, but we won't forget anything that way ;-)
1437 $popularity{sprintf("%10d",$popularity).$title.$biblionumber} = $biblionumber;
1439 # sort the hash and return the same structure as GetRecords (Zebra querying)
1442 if ($ordering eq 'popularity_dsc') { # sort popularity DESC
1443 foreach my $key (sort {$b cmp $a} (keys %popularity)) {
1444 $result_hash->{'RECORDS'}[$numbers++] = $result{$popularity{$key}}->as_usmarc();
1446 } else { # sort popularity ASC
1447 foreach my $key (sort (keys %popularity)) {
1448 $result_hash->{'RECORDS'}[$numbers++] = $result{$popularity{$key}}->as_usmarc();
1452 $result_hash->{'hits'} = $numbers;
1453 $finalresult->{'biblioserver'} = $result_hash;
1454 return $finalresult;
1458 } elsif ($ordering =~/author/){
1460 foreach (split /;/,$biblionumbers) {
1461 my ($biblionumber,$title) = split /,/,$_;
1462 my $record=GetMarcBiblio($biblionumber);
1464 if (C4::Context->preference('marcflavour') eq 'UNIMARC') {
1465 $author=$record->subfield('200','f');
1466 $author=$record->subfield('700','a') unless $author;
1468 $author=$record->subfield('100','a');
1470 # hint : the result is sorted by title.biblionumber because we can have X biblios with the same title
1471 # and we don't want to get only 1 result for each of them !!!
1472 $result{$author.$biblionumber}=$record;
1474 # sort the hash and return the same structure as GetRecords (Zebra querying)
1477 if ($ordering eq 'author_za') { # sort by author desc
1478 foreach my $key (sort { $b cmp $a } (keys %result)) {
1479 $result_hash->{'RECORDS'}[$numbers++] = $result{$key}->as_usmarc();
1481 } else { # sort by author ASC
1482 foreach my $key (sort (keys %result)) {
1483 $result_hash->{'RECORDS'}[$numbers++] = $result{$key}->as_usmarc();
1487 $result_hash->{'hits'} = $numbers;
1488 $finalresult->{'biblioserver'} = $result_hash;
1489 return $finalresult;
1491 # ORDER BY callnumber
1493 } elsif ($ordering =~/callnumber/){
1495 foreach (split /;/,$biblionumbers) {
1496 my ($biblionumber,$title) = split /,/,$_;
1497 my $record=GetMarcBiblio($biblionumber);
1499 my ($callnumber_tag,$callnumber_subfield)=GetMarcFromKohaField($dbh,'items.itemcallnumber');
1500 ($callnumber_tag,$callnumber_subfield)= GetMarcFromKohaField('biblioitems.callnumber') unless $callnumber_tag;
1501 if (C4::Context->preference('marcflavour') eq 'UNIMARC') {
1502 $callnumber=$record->subfield('200','f');
1504 $callnumber=$record->subfield('100','a');
1506 # hint : the result is sorted by title.biblionumber because we can have X biblios with the same title
1507 # and we don't want to get only 1 result for each of them !!!
1508 $result{$callnumber.$biblionumber}=$record;
1510 # sort the hash and return the same structure as GetRecords (Zebra querying)
1513 if ($ordering eq 'call_number_dsc') { # sort by title desc
1514 foreach my $key (sort { $b cmp $a } (keys %result)) {
1515 $result_hash->{'RECORDS'}[$numbers++] = $result{$key}->as_usmarc();
1517 } else { # sort by title ASC
1518 foreach my $key (sort { $a cmp $b } (keys %result)) {
1519 $result_hash->{'RECORDS'}[$numbers++] = $result{$key}->as_usmarc();
1523 $result_hash->{'hits'} = $numbers;
1524 $finalresult->{'biblioserver'} = $result_hash;
1525 return $finalresult;
1526 } elsif ($ordering =~ /pubdate/){ #pub year
1528 foreach (split /;/,$biblionumbers) {
1529 my ($biblionumber,$title) = split /,/,$_;
1530 my $record=GetMarcBiblio($biblionumber);
1531 my ($publicationyear_tag,$publicationyear_subfield)=GetMarcFromKohaField('biblioitems.publicationyear','');
1532 my $publicationyear=$record->subfield($publicationyear_tag,$publicationyear_subfield);
1533 # hint : the result is sorted by title.biblionumber because we can have X biblios with the same title
1534 # and we don't want to get only 1 result for each of them !!!
1535 $result{$publicationyear.$biblionumber}=$record;
1537 # sort the hash and return the same structure as GetRecords (Zebra querying)
1540 if ($ordering eq 'pubdate_dsc') { # sort by pubyear desc
1541 foreach my $key (sort { $b cmp $a } (keys %result)) {
1542 $result_hash->{'RECORDS'}[$numbers++] = $result{$key}->as_usmarc();
1544 } else { # sort by pub year ASC
1545 foreach my $key (sort (keys %result)) {
1546 $result_hash->{'RECORDS'}[$numbers++] = $result{$key}->as_usmarc();
1550 $result_hash->{'hits'} = $numbers;
1551 $finalresult->{'biblioserver'} = $result_hash;
1552 return $finalresult;
1556 } elsif ($ordering =~ /title/) {
1557 # the title is in the biblionumbers string, so we just need to build a hash, sort it and return
1559 foreach (split /;/,$biblionumbers) {
1560 my ($biblionumber,$title) = split /,/,$_;
1561 # hint : the result is sorted by title.biblionumber because we can have X biblios with the same title
1562 # and we don't want to get only 1 result for each of them !!!
1563 # hint & speed improvement : we can order without reading the record
1564 # so order, and read records only for the requested page !
1565 $result{$title.$biblionumber}=$biblionumber;
1567 # sort the hash and return the same structure as GetRecords (Zebra querying)
1570 if ($ordering eq 'title_az') { # sort by title desc
1571 foreach my $key (sort (keys %result)) {
1572 $result_hash->{'RECORDS'}[$numbers++] = $result{$key};
1574 } else { # sort by title ASC
1575 foreach my $key (sort { $b cmp $a } (keys %result)) {
1576 $result_hash->{'RECORDS'}[$numbers++] = $result{$key};
1579 # limit the $results_per_page to result size if it's more
1580 $results_per_page = $numbers-1 if $numbers < $results_per_page;
1581 # for the requested page, replace biblionumber by the complete record
1582 # speed improvement : avoid reading too much things
1583 for (my $counter=$offset;$counter<=$offset+$results_per_page;$counter++) {
1584 $result_hash->{'RECORDS'}[$counter] = GetMarcBiblio($result_hash->{'RECORDS'}[$counter])->as_usmarc;
1587 $result_hash->{'hits'} = $numbers;
1588 $finalresult->{'biblioserver'} = $result_hash;
1589 return $finalresult;
1594 # we need 2 hashes to order by ranking : the 1st one to count the ranking, the 2nd to order by ranking
1597 foreach (split /;/,$biblionumbers) {
1598 my ($biblionumber,$title) = split /,/,$_;
1599 $title =~ /(.*)-(\d)/;
1602 # note that we + the ranking because ranking is calculated on weight of EACH term requested.
1603 # if we ask for "two towers", and "two" has weight 2 in biblio N, and "towers" has weight 4 in biblio N
1604 # biblio N has ranking = 6
1605 $count_ranking{$biblionumber} += $ranking;
1607 # build the result by "inverting" the count_ranking hash
1608 # 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
1610 foreach (keys %count_ranking) {
1611 $result{sprintf("%10d",$count_ranking{$_}).'-'.$_} = $_;
1613 # sort the hash and return the same structure as GetRecords (Zebra querying)
1616 foreach my $key (sort {$b cmp $a} (keys %result)) {
1617 $result_hash->{'RECORDS'}[$numbers++] = $result{$key};
1619 # limit the $results_per_page to result size if it's more
1620 $results_per_page = $numbers-1 if $numbers < $results_per_page;
1621 # for the requested page, replace biblionumber by the complete record
1622 # speed improvement : avoid reading too much things
1623 for (my $counter=$offset;$counter<=$offset+$results_per_page;$counter++) {
1624 $result_hash->{'RECORDS'}[$counter] = GetMarcBiblio($result_hash->{'RECORDS'}[$counter])->as_usmarc if $result_hash->{'RECORDS'}[$counter];
1627 $result_hash->{'hits'} = $numbers;
1628 $finalresult->{'biblioserver'} = $result_hash;
1629 return $finalresult;
1634 ($countchanged,$listunchanged) = ModBiblios($listbiblios, $tagsubfield,$initvalue,$targetvalue,$test);
1636 this function changes all the values $initvalue in subfield $tag$subfield in any record in $listbiblios
1637 test parameter if set donot perform change to records in database.
1643 * $listbiblios is an array ref to marcrecords to be changed
1644 * $tagsubfield is the reference of the subfield to change.
1645 * $initvalue is the value to search the record for
1646 * $targetvalue is the value to set the subfield to
1647 * $test is to be set only not to perform changes in database.
1649 =item C<Output arg:>
1650 * $countchanged counts all the changes performed.
1651 * $listunchanged contains the list of all the biblionumbers of records unchanged.
1653 =item C<usage in the script:>
1657 my ($countchanged, $listunchanged) = EditBiblios($results->{RECORD}, $tagsubfield,$initvalue,$targetvalue);;
1658 #If one wants to display unchanged records, you should get biblios foreach @$listunchanged
1659 $template->param(countchanged => $countchanged, loopunchanged=>$listunchanged);
1664 my ($listbiblios,$tagsubfield,$initvalue,$targetvalue,$test)=@_;
1667 my ($tag,$subfield)=($1,$2) if ($tagsubfield=~/^(\d{1,3})([a-z0-9A-Z@])?$/);
1668 if ((length($tag)<3)&& $subfield=~/0-9/){
1669 $tag=$tag.$subfield;
1672 my ($bntag,$bnsubf) = GetMarcFromKohaField('biblio.biblionumber');
1673 my ($itemtag,$itemsubf) = GetMarcFromKohaField('items.itemnumber');
1674 foreach my $usmarc (@$listbiblios){
1676 $record=eval{MARC::Record->new_from_usmarc($usmarc)};
1679 # usmarc is not a valid usmarc May be a biblionumber
1680 if ($tag eq $itemtag){
1681 my $bib=GetBiblioFromItemNumber($usmarc);
1682 $record=GetMarcItem($bib->{'biblionumber'},$usmarc) ;
1683 $biblionumber=$bib->{'biblionumber'};
1685 $record=GetMarcBiblio($usmarc);
1686 $biblionumber=$usmarc;
1690 $biblionumber = $record->subfield($bntag,$bnsubf);
1692 $biblionumber=$record->field($bntag)->data;
1695 #GetBiblionumber is to be written.
1696 #Could be replaced by TransformMarcToKoha (But Would be longer)
1697 if ($record->field($tag)){
1699 foreach my $field ($record->field($tag)){
1701 if ($field->delete_subfield('code' =>$subfield,'match'=>qr($initvalue))){
1704 $field->update($subfield,$targetvalue) if ($targetvalue);
1708 if ($field->delete_field($field)){
1713 $field->data=$targetvalue if ($field->data=~qr($initvalue));
1717 # warn $record->as_formatted;
1719 ModBiblio($record,$biblionumber,GetFrameworkCode($biblionumber)) unless ($test);
1721 push @unmatched, $biblionumber;
1724 push @unmatched, $biblionumber;
1727 return ($countmatched,\@unmatched);
1730 END { } # module clean-up code here (global destructor)
1737 Koha Developement team <info@koha.org>