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-conf.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'};
207 my $search_result = ( $result->{hits} && $result->{hits} > 0 ? $result->{'RECORDS'} : [] );
208 return (undef,$search_result);
214 return ( "No query entered", undef ) unless $query;
216 #@servers = (C4::Context->config("biblioserver")) unless @servers;
218 ("biblioserver") unless @servers
219 ; # FIXME hardcoded value. See catalog/search.pl & opac-search.pl too.
222 for ( my $i = 0 ; $i < @servers ; $i++ ) {
224 $zconns[$i] = C4::Context->Zconn( $servers[$i], 1 );
227 ->search( new ZOOM::Query::CCL2RPN( $query, $zconns[$i] ) );
229 # getting error message if one occured.
231 $zconns[$i]->errmsg() . " ("
232 . $zconns[$i]->errcode() . ") "
233 . $zconns[$i]->addinfo() . " "
234 . $zconns[$i]->diagset();
236 return ( $error, undef ) if $zconns[$i]->errcode();
239 # caught a ZOOM::Exception
243 . $@->addinfo() . " "
246 return ( $error, undef );
251 while ( ( my $i = ZOOM::event( \@zconns ) ) != 0 ) {
252 $ev = $zconns[ $i - 1 ]->last_event();
253 if ( $ev == ZOOM::Event::ZEND ) {
254 $hits = $tmpresults[ $i - 1 ]->size();
257 for ( my $j = 0 ; $j < $hits ; $j++ ) {
258 my $record = $tmpresults[ $i - 1 ]->record($j)->raw();
259 push @results, $record;
263 return ( undef, \@results );
267 # performs the search
270 $koha_query, $simple_query, $sort_by_ref,
271 $servers_ref, $results_per_page, $offset,
272 $expanded_facet, $branches, $query_type,
275 # warn "Query : $koha_query";
276 my @servers = @$servers_ref;
277 my @sort_by = @$sort_by_ref;
279 # create the zoom connection and query object
283 my $results_hashref = ();
286 my $facets_counter = ();
287 my $facets_info = ();
288 my $facets = getFacets();
290 #### INITIALIZE SOME VARS USED CREATE THE FACETED RESULTS
291 my @facets_loop; # stores the ref to array of hashes for template
292 for ( my $i = 0 ; $i < @servers ; $i++ ) {
293 $zconns[$i] = C4::Context->Zconn( $servers[$i], 1 );
295 # perform the search, create the results objects
296 # if this is a local search, use the $koha-query, if it's a federated one, use the federated-query
298 if ( $servers[$i] =~ /biblioserver/ ) {
299 $query_to_use = $koha_query;
302 $query_to_use = $simple_query;
305 #$query_to_use = $simple_query if $scan;
306 #warn $simple_query if ($scan && $DEBUG);
307 # check if we've got a query_type defined
311 if ( $query_type =~ /^ccl/ ) {
313 s/\:/\=/g; # change : to = last minute (FIXME)
315 # warn "CCL : $query_to_use";
318 new ZOOM::Query::CCL2RPN( $query_to_use, $zconns[$i] )
321 elsif ( $query_type =~ /^cql/ ) {
323 # warn "CQL : $query_to_use";
326 new ZOOM::Query::CQL( $query_to_use, $zconns[$i] ) );
328 elsif ( $query_type =~ /^pqf/ ) {
330 # warn "PQF : $query_to_use";
333 new ZOOM::Query::PQF( $query_to_use, $zconns[$i] ) );
338 # warn "preparing to scan:$query_to_use";
341 new ZOOM::Query::CCL2RPN( $query_to_use, $zconns[$i] )
345 # warn "LAST : $query_to_use";
348 new ZOOM::Query::CCL2RPN( $query_to_use, $zconns[$i] )
354 warn "WARNING: query problem with $query_to_use " . $@;
357 # concatenate the sort_by limits and pass them to the results object
359 foreach my $sort (@sort_by) {
360 if ($sort eq "author_az") {
361 $sort_by.="1=1003 <i ";
363 elsif ($sort eq "author_za") {
364 $sort_by.="1=1003 >i ";
366 elsif ($sort eq "popularity_asc") {
367 $sort_by.="1=9003 <i ";
369 elsif ($sort eq "popularity_dsc") {
370 $sort_by.="1=9003 >i ";
372 elsif ($sort eq "call_number_asc") {
373 $sort_by.="1=20 <i ";
375 elsif ($sort eq "call_number_dsc") {
376 $sort_by.="1=20 >i ";
378 elsif ($sort eq "pubdate_asc") {
379 $sort_by.="1=31 <i ";
381 elsif ($sort eq "pubdate_dsc") {
382 $sort_by.="1=31 >i ";
384 elsif ($sort eq "acqdate_asc") {
385 $sort_by.="1=32 <i ";
387 elsif ($sort eq "acqdate_dsc") {
388 $sort_by.="1=32 >i ";
390 elsif ($sort eq "title_az") {
393 elsif ($sort eq "title_za") {
398 if ( $results[$i]->sort( "yaz", $sort_by ) < 0) {
399 warn "WARNING sort $sort_by failed";
403 while ( ( my $i = ZOOM::event( \@zconns ) ) != 0 ) {
404 my $ev = $zconns[ $i - 1 ]->last_event();
405 if ( $ev == ZOOM::Event::ZEND ) {
406 next unless $results[ $i - 1 ];
407 my $size = $results[ $i - 1 ]->size();
410 #$results_hash->{'server'} = $servers[$i-1];
411 # loop through the results
412 $results_hash->{'hits'} = $size;
414 if ( $offset + $results_per_page <= $size ) {
415 $times = $offset + $results_per_page;
420 for ( my $j = $offset ; $j < $times ; $j++ )
421 { #(($offset+$count<=$size) ? ($offset+$count):$size) ; $j++){
425 ## This is just an index scan
427 my ( $term, $occ ) = $results[ $i - 1 ]->term($j);
428 # here we create a minimal MARC record and hand it off to the
429 # template just like a normal result ... perhaps not ideal, but
431 my $tmprecord = MARC::Record->new();
432 $tmprecord->encoding('UTF-8');
435 # the minimal record in author/title (depending on MARC flavour)
436 if ( C4::Context->preference("marcflavour") eq
439 $tmptitle = MARC::Field->new(
446 $tmptitle = MARC::Field->new('245', ' ', ' ',a => $term,);
447 $tmpauthor = MARC::Field->new('100', ' ', ' ',a => $occ,);
449 $tmprecord->append_fields($tmptitle);
450 $tmprecord->append_fields($tmpauthor);
451 $results_hash->{'RECORDS'}[$j] = $tmprecord->as_usmarc();
454 $record = $results[ $i - 1 ]->record($j)->raw();
456 #warn "RECORD $j:".$record;
457 $results_hash->{'RECORDS'}[$j] =
458 $record; # making a reference to a hash
459 # Fill the facets while we're looping
460 $facet_record = MARC::Record->new_from_usmarc($record);
462 #warn $servers[$i-1].$facet_record->title();
463 for ( my $k = 0 ; $k <= @$facets ; $k++ ) {
464 if ( $facets->[$k] ) {
466 for my $tag ( @{ $facets->[$k]->{'tags'} } ) {
467 push @fields, $facet_record->field($tag);
469 for my $field (@fields) {
470 my @subfields = $field->subfields();
471 for my $subfield (@subfields) {
472 my ( $code, $data ) = @$subfield;
474 $facets->[$k]->{'subfield'} )
476 $facets_counter->{ $facets->[$k]
477 ->{'link_value'} }->{$data}++;
481 $facets_info->{ $facets->[$k]->{'link_value'} }
483 $facets->[$k]->{'label_value'};
484 $facets_info->{ $facets->[$k]->{'link_value'} }
485 ->{'expanded'} = $facets->[$k]->{'expanded'};
490 $results_hashref->{ $servers[ $i - 1 ] } = $results_hash;
493 #print "connection ", $i-1, ": $size hits";
494 #print $results[$i-1]->record(0)->render() if $size > 0;
497 sort { $facets_counter->{$b} <=> $facets_counter->{$a} }
498 keys %$facets_counter
502 my $number_of_facets;
503 my @this_facets_array;
506 $facets_counter->{$link_value}
507 ->{$b} <=> $facets_counter->{$link_value}->{$a}
508 } keys %{ $facets_counter->{$link_value} }
512 if ( ( $number_of_facets < 6 )
513 || ( $expanded_facet eq $link_value )
514 || ( $facets_info->{$link_value}->{'expanded'} ) )
517 # sanitize the link value ), ( will cause errors with CCL
518 my $facet_link_value = $one_facet;
519 $facet_link_value =~ s/(\(|\))/ /g;
521 # fix the length that will display in the label
522 my $facet_label_value = $one_facet;
523 $facet_label_value = substr( $one_facet, 0, 20 ) . "..."
524 unless length($facet_label_value) <= 20;
526 # well, if it's a branch, label by the name, not the code
527 if ( $link_value =~ /branch/ ) {
529 $branches->{$one_facet}->{'branchname'};
532 # but we're down with the whole label being in the link's title
533 my $facet_title_value = $one_facet;
535 push @this_facets_array,
539 $facets_counter->{$link_value}->{$one_facet},
540 facet_label_value => $facet_label_value,
541 facet_title_value => $facet_title_value,
542 facet_link_value => $facet_link_value,
543 type_link_value => $link_value,
548 unless ( $facets_info->{$link_value}->{'expanded'} ) {
550 if ( ( $number_of_facets > 6 )
551 && ( $expanded_facet ne $link_value ) );
556 type_link_value => $link_value,
557 type_id => $link_value . "_id",
559 $facets_info->{$link_value}->{'label_value'},
560 facets => \@this_facets_array,
561 expandable => $expandable,
562 expand => $link_value,
568 return ( undef, $results_hashref, \@facets_loop );
572 sub _remove_stopwords {
573 my ($operand,$index) = @_;
574 my @stopwords_removed;
575 # phrase and exact-qualified indexes shouldn't have stopwords removed
576 if ($index!~m/phr|ext/){
577 # remove stopwords from operand : parse all stopwords & remove them (case insensitive)
578 # we use IsAlpha unicode definition, to deal correctly with diacritics.
579 # otherwise, a French word like "leçon" woudl be split into "le" "çon", le
580 # is an empty word, we'd get "çon" and wouldn't find anything...
581 foreach (keys %{C4::Context->stopwords}) {
582 next if ($_ =~/(and|or|not)/); # don't remove operators
583 if ($operand =~ /(\P{IsAlpha}$_\P{IsAlpha}|^$_\P{IsAlpha}|\P{IsAlpha}$_$)/) {
584 $operand=~ s/\P{IsAlpha}$_\P{IsAlpha}/ /gi;
585 $operand=~ s/^$_\P{IsAlpha}/ /gi;
586 $operand=~ s/\P{IsAlpha}$_$/ /gi;
587 push @stopwords_removed, $_;
591 return ($operand, \@stopwords_removed);
595 sub _detect_truncation {
596 my ($operand,$index) = @_;
597 my (@nontruncated,@righttruncated,@lefttruncated,@rightlefttruncated,@regexpr);
599 my @wordlist= split (/\s/,$operand);
600 foreach my $word (@wordlist){
601 if ($word=~s/^\*([^\*]+)\*$/$1/){
602 push @rightlefttruncated,$word;
604 elsif($word=~s/^\*([^\*]+)$/$1/){
605 push @lefttruncated,$word;
607 elsif ($word=~s/^([^\*]+)\*$/$1/){
608 push @righttruncated,$word;
610 elsif (index($word,"*")<0){
611 push @nontruncated,$word;
617 return (\@nontruncated,\@righttruncated,\@lefttruncated,\@rightlefttruncated,\@regexpr);
620 sub _build_stemmed_operand {
623 # FIXME: the locale should be set based on the user's language and/or search choice
624 my $stemmer = Lingua::Stem->new( -locale => 'EN-US' );
625 # FIXME: these should be stored in the db so the librarian can modify the behavior
626 $stemmer->add_exceptions(
634 my @words = split( / /, $operand );
635 my $stems = $stemmer->stem(@words);
636 for my $stem (@$stems) {
637 $stemmed_operand .= "$stem";
638 $stemmed_operand .= "?" unless ( $stem =~ /(and$|or$|not$)/ ) || ( length($stem) < 3 );
639 $stemmed_operand .= " ";
641 #warn "STEMMED OPERAND: $stemmed_operand";
642 return $stemmed_operand;
645 sub _build_weighted_query {
646 # FIELD WEIGHTING - This is largely experimental stuff. What I'm committing works
647 # pretty well but will work much better when we have an actual query parser
648 my ($operand,$stemmed_operand,$index) = @_;
649 my $stemming = C4::Context->preference("QueryStemming") || 0;
650 my $weight_fields = C4::Context->preference("QueryWeightFields") || 0;
651 my $fuzzy_enabled = C4::Context->preference("QueryFuzzy") || 0;
653 my $weighted_query .= "(rk=("; # Specifies that we're applying rank
655 # Keyword, or, no index specified
656 if ( ( $index eq 'kw' ) || ( !$index ) ) {
657 $weighted_query .= "Title-cover,ext,r1=\"$operand\""; # exact title-cover
658 $weighted_query .= " or ti,ext,r2=\"$operand\""; # exact title
659 $weighted_query .= " or ti,phr,r3=\"$operand\""; # phrase title
660 #$weighted_query .= " or any,ext,r4=$operand"; # exact any
661 #$weighted_query .=" or kw,wrdl,r5=\"$operand\""; # word list any
662 $weighted_query .= " or wrdl,fuzzy,r8=\"$operand\"" if $fuzzy_enabled; # add fuzzy, word list
663 $weighted_query .= " or wrdl,right-Truncation,r9=\"$stemmed_operand\"" if ($stemming and $stemmed_operand); # add stemming, right truncation
664 $weighted_query .= " or wrdl,r9=\"$operand\"";
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 $index,rt,wrdl,r3=\"$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 warn "---------" if $DEBUG;
692 warn "Enter buildQuery" if $DEBUG;
693 warn "---------" if $DEBUG;
695 my @operators = @$operators if $operators;
696 my @indexes = @$indexes if $indexes;
697 my @operands = @$operands if $operands;
698 my @limits = @$limits if $limits;
699 my @sort_by = @$sort_by if $sort_by;
701 my $stemming = C4::Context->preference("QueryStemming") || 0;
702 my $auto_truncation = C4::Context->preference("QueryAutoTruncate") || 0;
703 my $weight_fields = C4::Context->preference("QueryWeightFields") || 0;
704 my $fuzzy_enabled = C4::Context->preference("QueryFuzzy") || 0;
705 # no stemming/weight/fuzzy in NoZebra
706 if (C4::Context->preference("NoZebra")) {
711 my $remove_stopwords = C4::Context->preference("QueryRemoveStopwords") || 0;
713 my $query = $operands[0];
714 my $simple_query = $operands[0];
723 my $stopwords_removed;
725 # for handling ccl, cql, pqf queries in diagnostic mode, skip the rest of the steps
727 if ( $query =~ /^ccl=/ ) {
728 return ( undef, $', $', $', $', '', '', '', '', 'ccl' );
730 if ( $query =~ /^cql=/ ) {
731 return ( undef, $', $', $', $', '', '', '', '', 'cql' );
733 if ( $query =~ /^pqf=/ ) {
734 return ( undef, $', $', $', $', '', '', '', '', 'pqf' );
737 # pass nested queries directly
738 if ( $query =~ /(\(|\))/ ) {
739 return ( undef, $query, $simple_query, $query_cgi, $query, $limit, $limit_cgi, $limit_desc, $stopwords_removed, 'ccl' );
742 # form-based queries are limited to non-nested at a specific depth, so we can easily
743 # modify the incoming query operands and indexes to do stemming and field weighting
744 # Once we do so, we'll end up with a value in $query, just like if we had an
745 # incoming $query from the user
747 $query = ""; # clear it out so we can populate properly with field-weighted stemmed query
748 my $previous_operand; # a flag used to keep track if there was a previous query
749 # if there was, we can apply the current operator
751 for ( my $i = 0 ; $i <= @operands ; $i++ ) {
753 # COMBINE OPERANDS, INDEXES AND OPERATORS
754 if ( $operands[$i] ) {
756 # a flag to determine whether or not to add the index to the query
759 # if the user is sophisticated enough to specify an index, turn off field weighting, stemming, and stopword handling
760 if ($operands[$i] =~ /(:|=)/ || $scan) {
763 $remove_stopwords = 0;
765 my $operand = $operands[$i];
766 my $index = $indexes[$i];
768 # add some attributes for certain index types
769 # Date of Publication
770 if ($index eq 'yr') {
771 $index .=",st-numeric";
773 ($stemming,$auto_truncation,$weight_fields, $fuzzy_enabled, $remove_stopwords) = (0,0,0,0,0);
775 # Date of Acquisition
776 elsif ($index eq 'acqdate') {
777 $index.=",st-date-normalized";
779 ($stemming,$auto_truncation,$weight_fields, $fuzzy_enabled, $remove_stopwords) = (0,0,0,0,0);
783 # set default structure attribute (word list)
785 unless (!$index || $index =~ /(st-|phr|ext|wrdl)/) {
786 $struct_attr = ",wrdl";
788 # some helpful index modifs
789 my $index_plus = $index.$struct_attr.":" if $index;
790 my $index_plus_comma=$index.$struct_attr."," if $index;
793 if ($remove_stopwords) {
794 ($operand, $stopwords_removed) = _remove_stopwords($operand,$index);
795 warn "OPERAND w/out STOPWORDS: >$operand<" if $DEBUG;
796 warn "REMOVED STOPWORDS: @$stopwords_removed" if ($stopwords_removed && $DEBUG);
800 my ($nontruncated,$righttruncated,$lefttruncated,$rightlefttruncated,$regexpr);
801 my $truncated_operand;
802 ($nontruncated,$righttruncated,$lefttruncated,$rightlefttruncated,$regexpr) = _detect_truncation($operand,$index);
803 warn "TRUNCATION: NON:>@$nontruncated< RIGHT:>@$righttruncated< LEFT:>@$lefttruncated< RIGHTLEFT:>@$rightlefttruncated< REGEX:>@$regexpr<" if $DEBUG;
806 if (scalar(@$righttruncated)+scalar(@$lefttruncated)+scalar(@$rightlefttruncated)>0){
807 # don't field weight or add the index to the query, we do it here
809 undef $weight_fields;
810 my $previous_truncation_operand;
811 if (scalar(@$nontruncated)>0) {
812 $truncated_operand.= "$index_plus @$nontruncated ";
813 $previous_truncation_operand = 1;
815 if (scalar(@$righttruncated)>0){
816 $truncated_operand .= "and " if $previous_truncation_operand;
817 $truncated_operand .= "$index_plus_comma"."rtrn:@$righttruncated ";
818 $previous_truncation_operand = 1;
820 if (scalar(@$lefttruncated)>0){
821 $truncated_operand .= "and " if $previous_truncation_operand;
822 $truncated_operand .= "$index_plus_comma"."ltrn:@$lefttruncated ";
823 $previous_truncation_operand = 1;
825 if (scalar(@$rightlefttruncated)>0){
826 $truncated_operand .= "and " if $previous_truncation_operand;
827 $truncated_operand .= "$index_plus_comma"."rltrn:@$rightlefttruncated ";
828 $previous_truncation_operand = 1;
831 $operand = $truncated_operand if $truncated_operand;
832 warn "TRUNCATED OPERAND: >$truncated_operand<" if $DEBUG;
836 $stemmed_operand = _build_stemmed_operand($operand) if $stemming;
837 warn "STEMMED OPERAND: >$stemmed_operand<" if $DEBUG;
839 # Handle Field Weighting
840 my $weighted_operand;
841 $weighted_operand = _build_weighted_query($operand,$stemmed_operand,$index) if $weight_fields;
842 warn "FIELD WEIGHTED OPERAND: >$weighted_operand<" if $DEBUG;
843 $operand = $weighted_operand if $weight_fields;
844 $indexes_set = 1 if $weight_fields;
846 # If there's a previous operand, we need to add an operator
847 if ($previous_operand) {
849 # user-specified operator
850 if ( $operators[$i-1] ) {
851 $query .= " $operators[$i-1] ";
852 $query .= " $index_plus " unless $indexes_set;
853 $query .= " $operand";
854 $query_cgi .="&op=$operators[$i-1]";
855 $query_cgi .="&idx=$index" if $index;
856 $query_cgi .="&q=$operands[$i]" if $operands[$i];
857 $query_desc .=" $operators[$i-1] $index_plus $operands[$i]";
860 # the default operator is and
863 $query .= "$index_plus " unless $indexes_set;
864 $query .= "$operand";
865 $query_cgi .="&op=and&idx=$index" if $index;
866 $query_cgi .="&q=$operands[$i]" if $operands[$i];
867 $query_desc .= " and $index_plus $operands[$i]";
871 # there isn't a pervious operand, don't need an operator
873 # field-weighted queries already have indexes set
874 $query .=" $index_plus " unless $indexes_set;
876 $query_desc .= " $index_plus $operands[$i]";
877 $query_cgi.="&idx=$index" if $index;
878 $query_cgi.="&q=$operands[$i]" if $operands[$i];
880 $previous_operand = 1;
885 warn "QUERY BEFORE LIMITS: >$query<" if $DEBUG;
890 my $availability_limit;
891 foreach my $this_limit (@limits) {
892 if ( $this_limit =~ /available/ ) {
893 # available is defined as (items.notloan is NULL) and (items.itemlost > 0 or NULL) (last clause handles NULL values for lost in zebra)
894 # all records not indexed in the onloan register and allrecords not indexed in the lost register, or where the value of lost is equal to or less than 0
895 $availability_limit .="( ( allrecords,AlwaysMatches='' not onloan,AlwaysMatches='') and ((lost,st-numeric <= 0) or ( allrecords,AlwaysMatches='' not lost,AlwaysMatches='')) )";
896 $limit_cgi .= "&limit=available";
900 # these are treated as OR
901 elsif ( $this_limit =~ /mc/ ) {
902 $group_OR_limits .= " or " if $group_OR_limits;
903 $limit_desc .=" or " if $group_OR_limits;
904 $group_OR_limits .= "$this_limit";
905 $limit_cgi .="&limit=$this_limit";
906 $limit_desc .= " $this_limit";
910 $limit .= " and " if $limit || $query;
911 $limit .= "$this_limit";
912 $limit_cgi .="&limit=$this_limit";
913 $limit_desc .=" $this_limit";
916 if ($group_OR_limits) {
917 $limit.=" and " if ($query || $limit );
918 $limit.="($group_OR_limits)";
920 if ($availability_limit) {
921 $limit.=" and " if ($query || $limit );
922 $limit.="($availability_limit)";
924 # normalize the strings
927 for ($query, $query_desc, $limit, $limit_desc) {
928 $_ =~ s/ / /g; # remove extra spaces
929 $_ =~ s/^ //g; # remove any beginning spaces
930 $_ =~ s/ $//g; # remove any ending spaces
931 $_ =~ s/==/=/g; # remove double == from query
934 $query_cgi =~ s/^&//;
936 # append the limit to the query
939 warn "query=$query and limit=$limit" if $DEBUG;
941 warn "QUERY:".$query if $DEBUG;
942 warn "QUERY CGI:".$query_cgi if $DEBUG;
943 warn "QUERY DESC:".$query_desc if $DEBUG;
944 warn "LIMIT:".$limit if $DEBUG;
945 warn "LIMIT CGI:".$limit_cgi if $DEBUG;
946 warn "LIMIT DESC:".$limit_desc if $DEBUG;
947 warn "---------" if $DEBUG;
948 warn "Leave buildQuery" if $DEBUG;
949 warn "---------" if $DEBUG;
950 return ( undef, $query,$simple_query,$query_cgi,$query_desc,$limit,$limit_cgi,$limit_desc,$stopwords_removed,$query_type );
953 # IMO this subroutine is pretty messy still -- it's responsible for
954 # building the HTML output for the template
956 my ( $searchdesc, $hits, $results_per_page, $offset, @marcresults ) = @_;
957 my $dbh = C4::Context->dbh;
961 my $span_terms_hashref;
962 for my $span_term ( split( / /, $searchdesc ) ) {
963 $span_term =~ s/(.*=|\)|\(|\+|\.|\*)//g;
964 $span_terms_hashref->{$span_term}++;
967 #Build branchnames hash
969 #get branch information.....
972 $dbh->prepare("SELECT branchcode,branchname FROM branches")
973 ; # FIXME : use C4::Koha::GetBranches
975 while ( my $bdata = $bsth->fetchrow_hashref ) {
976 $branches{ $bdata->{'branchcode'} } = $bdata->{'branchname'};
979 my $lsch = $dbh->prepare("SELECT authorised_value,lib FROM authorised_values WHERE category = 'LOC'");
981 while (my $ldata = $lsch->fetchrow_hashref ) {
982 $locations{ $ldata->{'authorised_value'} } = $ldata->{'lib'};
986 #find itemtype & itemtype image
989 $dbh->prepare("SELECT itemtype,description,imageurl,summary,notforloan FROM itemtypes");
991 while ( my $bdata = $bsth->fetchrow_hashref ) {
992 $itemtypes{ $bdata->{'itemtype'} }->{description} =
993 $bdata->{'description'};
994 $itemtypes{ $bdata->{'itemtype'} }->{imageurl} = $bdata->{'imageurl'};
995 $itemtypes{ $bdata->{'itemtype'} }->{summary} = $bdata->{'summary'};
996 $itemtypes{ $bdata->{'itemtype'} }->{notforloan} = $bdata->{'notforloan'};
999 #search item field code
1000 my $sth = $dbh->prepare("SELECT tagfield FROM marc_subfield_structure WHERE kohafield LIKE 'items.itemnumber'");
1002 my ($itemtag) = $sth->fetchrow;
1004 # get notforloan authorised value list
1005 $sth = $dbh->prepare("SELECT authorised_value FROM `marc_subfield_structure` WHERE kohafield = 'items.notforloan' AND frameworkcode=''");
1007 my ($notforloan_authorised_value) = $sth->fetchrow;
1008 ## find column names of items related to MARC
1009 my $sth2 = $dbh->prepare("SHOW COLUMNS FROM items");
1011 my %subfieldstosearch;
1012 while ( ( my $column ) = $sth2->fetchrow ) {
1013 my ( $tagfield, $tagsubfield ) =
1014 &GetMarcFromKohaField( "items." . $column, "" );
1015 $subfieldstosearch{$column} = $tagsubfield;
1019 if ( $hits && $offset + $results_per_page <= $hits ) {
1020 $times = $offset + $results_per_page;
1026 for ( my $i = $offset ; $i <= $times - 1 ; $i++ ) {
1028 $marcrecord = MARC::File::USMARC::decode( $marcresults[$i] );
1029 my $oldbiblio = TransformMarcToKoha( $dbh, $marcrecord, '' );
1030 $oldbiblio->{result_number} = $i+1;
1031 # add image url if there is one
1032 if ( $itemtypes{ $oldbiblio->{itemtype} }->{imageurl} =~ /^http:/ ) {
1033 $oldbiblio->{imageurl} =
1034 $itemtypes{ $oldbiblio->{itemtype} }->{imageurl};
1035 $oldbiblio->{description} =
1036 $itemtypes{ $oldbiblio->{itemtype} }->{description};
1039 $oldbiblio->{imageurl} =
1040 getitemtypeimagesrc() . "/"
1041 . $itemtypes{ $oldbiblio->{itemtype} }->{imageurl}
1042 if ( $itemtypes{ $oldbiblio->{itemtype} }->{imageurl} );
1043 $oldbiblio->{description} =
1044 $itemtypes{ $oldbiblio->{itemtype} }->{description};
1047 # build summary if there is one (the summary is defined in itemtypes table
1049 if ($itemtypes{ $oldbiblio->{itemtype} }->{summary}) {
1050 my $summary = $itemtypes{ $oldbiblio->{itemtype} }->{summary};
1051 my @fields = $marcrecord->fields();
1052 foreach my $field (@fields) {
1053 my $tag = $field->tag();
1054 my $tagvalue = $field->as_string();
1055 $summary =~ s/\[(.?.?.?.?)$tag\*(.*?)]/$1$tagvalue$2\[$1$tag$2]/g;
1057 my @subf = $field->subfields;
1058 for my $i (0..$#subf) {
1059 my $subfieldcode = $subf[$i][0];
1060 my $subfieldvalue = $subf[$i][1];
1061 my $tagsubf = $tag.$subfieldcode;
1062 $summary =~ s/\[(.?.?.?.?)$tagsubf(.*?)]/$1$subfieldvalue$2\[$1$tagsubf$2]/g;
1066 $summary =~ s/\[(.*?)]//g;
1067 $summary =~ s/\n/<br>/g;
1068 $oldbiblio->{summary} = $summary;
1070 # add spans to search term in results for search term highlighting
1071 my $searchhighlightblob;
1072 for my $highlight_field ($marcrecord->fields) {
1073 next if $highlight_field->tag() =~ /(^00)/; # skip fixed fields
1075 my $field = $highlight_field->as_string();
1076 for my $term ( keys %$span_terms_hashref ) {
1077 if (($field =~ /$term/i) && (length($term) > 3)) {
1078 $field =~ s/$term/<span class=\"term\">$&<\/span>/gi;
1082 $searchhighlightblob .= $field." ... " if $match;
1084 $oldbiblio->{'searchhighlightblob'} = $searchhighlightblob;
1085 # save an author with no <span> tag, for the <a href=search.pl?q=<!--tmpl_var name="author"-->> link
1086 $oldbiblio->{'author_nospan'} = $oldbiblio->{'author'};
1087 for my $term ( keys %$span_terms_hashref ) {
1088 my $old_term = $term;
1089 if ( length($term) > 3 ) {
1090 $term =~ s/(.*=|\)|\(|\+|\.|\?|\[|\]|\\|\*)//g;
1091 $oldbiblio->{'title'} =~ s/$term/<span class=\"term\">$&<\/span>/gi;
1092 $oldbiblio->{'subtitle'} =~ s/$term/<span class=\"term\">$&<\/span>/gi;
1093 $oldbiblio->{'author'} =~ s/$term/<span class=\"term\">$&<\/span>/gi;
1094 $oldbiblio->{'publishercode'} =~ s/$term/<span class=\"term\">$&<\/span>/gi;
1095 $oldbiblio->{'place'} =~ s/$term/<span class=\"term\">$&<\/span>/gi;
1096 $oldbiblio->{'pages'} =~ s/$term/<span class=\"term\">$&<\/span>/gi;
1097 $oldbiblio->{'notes'} =~ s/$term/<span class=\"term\">$&<\/span>/gi;
1098 $oldbiblio->{'size'} =~ s/$term/<span class=\"term\">$&<\/span>/gi;
1103 $toggle = "#ffffcc";
1108 $oldbiblio->{'toggle'} = $toggle;
1109 my @fields = $marcrecord->field($itemtag);
1111 # Setting item statuses for display
1112 my @available_items_loop;
1113 my @onloan_items_loop;
1114 my @other_items_loop;
1116 my $available_items;
1120 my $ordered_count = 0;
1121 my $available_count = 0;
1122 my $onloan_count = 0;
1123 my $longoverdue_count = 0;
1124 my $other_count = 0;
1125 my $wthdrawn_count = 0;
1126 my $itemlost_count = 0;
1127 my $itembinding_count = 0;
1128 my $itemdamaged_count = 0;
1129 my $can_place_holds = 0;
1130 my $items_count=scalar(@fields);
1132 my $maxitems = (C4::Context->preference('maxItemsinSearchResults')) ? C4::Context->preference('maxItemsinSearchResults')- 1 : 1;
1133 foreach my $field (@fields) {
1137 # populate the items hash
1138 foreach my $code ( keys %subfieldstosearch ) {
1139 $item->{$code} = $field->subfield( $subfieldstosearch{$code} );
1141 # set item's branch name, use homebranch first, fall back to holdingbranch
1142 if ($item->{'homebranch'}) {
1143 $item->{'branchname'} = $branches{$item->{homebranch}};
1146 elsif ($item->{'holdingbranch'}) {
1147 $item->{'branchname'} = $branches{$item->{holdingbranch}};
1149 # key for items results is built from branchcode . coded location qualifier . itemcallnumber
1150 if ($item->{onloan}) {
1152 $onloan_items->{ $item->{'homebranch'}.'--'.$item->{location}.$item->{'itemcallnumber'}.$item->{due_date} }->{due_date} = format_date($item->{onloan});
1153 $onloan_items->{ $item->{'homebranch'}.'--'.$item->{location}.$item->{'itemcallnumber'}.$item->{due_date} }->{count}++ if $item->{'homebranch'};
1154 $onloan_items->{ $item->{'homebranch'}.'--'.$item->{location}.$item->{'itemcallnumber'}.$item->{due_date} }->{branchname} = $item->{'branchname'};
1155 $onloan_items->{ $item->{'homebranch'}.'--'.$item->{location}.$item->{'itemcallnumber'}.$item->{due_date} }->{location} = $locations{$item->{location}};
1156 $onloan_items->{ $item->{'homebranch'}.'--'.$item->{location}.$item->{'itemcallnumber'}.$item->{due_date} }->{itemcallnumber} = $item->{itemcallnumber};
1158 # if something's checked out and lost, mark it as 'long overdue'
1159 if ( $item->{itemlost} ) {
1160 $onloan_items->{ $item->{'homebranch'}.'--'.$item->{location}.$item->{'itemcallnumber'}.$item->{due_date} }->{longoverdue}++;
1161 $longoverdue_count++;
1163 # can place holds as long as this item isn't lost
1165 $can_place_holds = 1;
1169 # items not on loan, but still unavailable ( lost, withdrawn, damaged )
1172 if ( $item->{notforloan} == -1) {
1176 # item is withdrawn, lost or damaged
1177 if ( $item->{wthdrawn} || $item->{itemlost} || $item->{damaged} || $item->{notforloan} ) {
1178 $wthdrawn_count++ if $item->{wthdrawn};
1179 $itemlost_count++ if $item->{itemlost};
1180 $itemdamaged_count++ if $item->{damaged};
1181 $item->{status} = $item->{wthdrawn}."-".$item->{itemlost}."-".$item->{damaged}."-".$item->{notforloan};
1184 $other_items->{ $item->{'homebranch'}.'--'.$item->{location}.$item->{'itemcallnumber'}.$item->{status} }->{wthdrawn} = $item->{wthdrawn};
1185 $other_items->{ $item->{'homebranch'}.'--'.$item->{location}.$item->{'itemcallnumber'}.$item->{status} }->{itemlost} = $item->{itemlost};
1186 $other_items->{ $item->{'homebranch'}.'--'.$item->{location}.$item->{'itemcallnumber'}.$item->{status} }->{damaged} = $item->{damaged};
1187 $other_items->{ $item->{'homebranch'}.'--'.$item->{location}.$item->{'itemcallnumber'}.$item->{status} }->{notforloan} = GetAuthorisedValueDesc('','',$item->{notforloan},'','',$notforloan_authorised_value) if $notforloan_authorised_value;
1189 $other_items->{ $item->{'homebranch'}.'--'.$item->{location}.$item->{'itemcallnumber'}.$item->{status} }->{count}++ if $item->{'homebranch'};
1190 $other_items->{ $item->{'homebranch'}.'--'.$item->{location}.$item->{'itemcallnumber'}.$item->{status} }->{branchname} = $item->{'branchname'};
1191 $other_items->{ $item->{'homebranch'}.'--'.$item->{location}.$item->{'itemcallnumber'}.$item->{status} }->{location} = $locations{$item->{location}};
1192 $other_items->{ $item->{'homebranch'}.'--'.$item->{location}.$item->{'itemcallnumber'}.$item->{status} }->{itemcallnumber} = $item->{itemcallnumber};
1197 $can_place_holds = 1;
1199 $available_items->{ $item->{'homebranch'}.'--'.$item->{location}.$item->{'itemcallnumber'} }->{count}++ if $item->{'homebranch'};
1200 $available_items->{ $item->{'homebranch'}.'--'.$item->{location}.$item->{'itemcallnumber'} }->{branchname} = $item->{'branchname'};
1201 $available_items->{ $item->{'homebranch'}.'--'.$item->{location}.$item->{'itemcallnumber'} }->{location} = $locations{$item->{location}};
1202 $available_items->{ $item->{'homebranch'}.'--'.$item->{location}.$item->{'itemcallnumber'} }->{itemcallnumber} = $item->{itemcallnumber};
1205 } # notforloan, item level and biblioitem level
1206 my ($availableitemscount, $onloanitemscount, $otheritemscount);
1207 my $maxitems = (C4::Context->preference('maxItemsinSearchResults')) ? C4::Context->preference('maxItemsinSearchResults')- 1 : 1;
1208 for my $key ( sort keys %$onloan_items ) {
1209 $onloanitemscount++;
1210 push @onloan_items_loop, $onloan_items->{$key} unless $onloanitemscount > $maxitems;
1212 for my $key ( sort keys %$other_items ) {
1214 push @other_items_loop, $other_items->{$key} unless $otheritemscount > $maxitems;
1216 for my $key ( sort keys %$available_items ) {
1217 $availableitemscount++;
1218 push @available_items_loop, $available_items->{$key} unless $availableitemscount > $maxitems;
1220 # last check for norequest : if itemtype is notforloan, it can't be reserved either, whatever the items
1221 $can_place_holds = 0 if $itemtypes{$oldbiblio->{itemtype}}->{notforloan};
1222 $oldbiblio->{norequests} = 1 unless $can_place_holds;
1223 $oldbiblio->{itemsplural} = 1 if $items_count>1;
1224 $oldbiblio->{items_count} = $items_count;
1225 $oldbiblio->{available_items_loop} = \@available_items_loop;
1226 $oldbiblio->{onloan_items_loop} = \@onloan_items_loop;
1227 $oldbiblio->{other_items_loop} = \@other_items_loop;
1228 $oldbiblio->{availablecount} = $available_count;
1229 $oldbiblio->{availableplural} = 1 if $available_count>1;
1230 $oldbiblio->{onloancount} = $onloan_count;
1231 $oldbiblio->{onloanplural} = 1 if $onloan_count>1;
1232 $oldbiblio->{othercount} = $other_count;
1233 $oldbiblio->{otherplural} = 1 if $other_count>1;
1234 $oldbiblio->{wthdrawncount} = $wthdrawn_count;
1235 $oldbiblio->{itemlostcount} = $itemlost_count;
1236 $oldbiblio->{damagedcount} = $itemdamaged_count;
1237 $oldbiblio->{orderedcount} = $ordered_count;
1238 $oldbiblio->{isbn} =~ s/-//g; # deleting - in isbn to enable amazon content
1239 push( @newresults, $oldbiblio );
1246 #----------------------------------------------------------------------
1248 # Non-Zebra GetRecords#
1249 #----------------------------------------------------------------------
1253 NZgetRecords has the same API as zera getRecords, even if some parameters are not managed
1257 my ($query,$simple_query,$sort_by_ref,$servers_ref,$results_per_page,$offset,$expanded_facet,$branches,$query_type,$scan) = @_;
1258 warn "query =$query" if $DEBUG;
1259 my $result = NZanalyse($query);
1260 warn "results =$result" if $DEBUG;
1261 return (undef,NZorder($result,@$sort_by_ref[0],$results_per_page,$offset),undef);
1266 NZanalyse : get a CQL string as parameter, and returns a list of biblionumber;title,biblionumber;title,...
1267 the list is built from an inverted index in the nozebra SQL table
1268 note that title is here only for convenience : the sorting will be very fast when requested on title
1269 if the sorting is requested on something else, we will have to reread all results, and that may be longer.
1274 my ($string,$server) = @_;
1275 warn "---------" if $DEBUG;
1276 warn "Enter NZanalyse" if $DEBUG;
1277 warn "---------" if $DEBUG;
1279 # $server contains biblioserver or authorities, depending on what we search on.
1280 #warn "querying : $string on $server";
1281 $server='biblioserver' unless $server;
1283 # if we have a ", replace the content to discard temporarily any and/or/not inside
1285 if ($string =~/"/) {
1286 $string =~ s/"(.*?)"/__X__/;
1288 warn "commacontent : $commacontent" if $DEBUG;
1290 # split the query string in 3 parts : X AND Y means : $left="X", $operand="AND" and $right="Y"
1291 # then, call again NZanalyse with $left and $right
1292 # (recursive until we find a leaf (=> something without and/or/not)
1293 # delete repeated operator... Would then go in infinite loop
1294 while ($string =~s/( and| or| not| AND| OR| NOT)\1/$1/g){
1296 #process parenthesis before.
1297 if ($string =~ /^\s*\((.*)\)(( and | or | not | AND | OR | NOT )(.*))?/){
1300 my $operator = lc($3); # FIXME: and/or/not are operators, not operands
1301 warn "dealing w/parenthesis before recursive sub call. left :$left operator:$operator right:$right" if $DEBUG;
1302 my $leftresult = NZanalyse($left,$server);
1304 my $rightresult = NZanalyse($right,$server);
1305 # OK, we have the results for right and left part of the query
1306 # depending of operand, intersect, union or exclude both lists
1307 # to get a result list
1308 if ($operator eq ' and ') {
1309 my @leftresult = split /;/, $leftresult;
1310 warn " @leftresult / $rightresult \n" if $DEBUG;
1311 # my @rightresult = split /;/,$leftresult;
1313 # parse the left results, and if the biblionumber exist in the right result, save it in finalresult
1314 # the result is stored twice, to have the same weight for AND than OR.
1315 # example : TWO : 61,61,64,121 (two is twice in the biblio #61) / TOWER : 61,64,130
1316 # result : 61,61,61,61,64,64 for two AND tower : 61 has more weight than 64
1317 foreach (@leftresult) {
1320 ($value,$countvalue)=($1,$2) if $value=~m/(.*)-(\d+)$/;
1321 if ($rightresult =~ /$value-(\d+);/) {
1322 $countvalue=($1>$countvalue?$countvalue:$1);
1323 $finalresult .= "$value-$countvalue;$value-$countvalue;";
1326 warn " $finalresult \n" if $DEBUG;
1327 return $finalresult;
1328 } elsif ($operator eq ' or ') {
1329 # just merge the 2 strings
1330 return $leftresult.$rightresult;
1331 } elsif ($operator eq ' not ') {
1332 my @leftresult = split /;/, $leftresult;
1333 # my @rightresult = split /;/,$leftresult;
1335 foreach (@leftresult) {
1337 $value=$1 if $value=~m/(.*)-\d+$/;
1338 unless ($rightresult =~ "$value-") {
1341 return $finalresult;
1343 # this error is impossible, because of the regexp that isolate the operand, but just in case...
1349 warn "string :".$string if $DEBUG;
1350 $string =~ /(.*?)( and | or | not | AND | OR | NOT )(.*)/;
1353 my $operator = lc($2); # FIXME: and/or/not are operators, not operands
1354 warn "dealing w/parenthesis. left :$left operator:$operator right:$right" if $DEBUG;
1355 # it's not a leaf, we have a and/or/not
1357 # reintroduce comma content if needed
1358 $right =~ s/__X__/"$commacontent"/ if $commacontent;
1359 $left =~ s/__X__/"$commacontent"/ if $commacontent;
1360 warn "node : $left / $operator / $right\n" if $DEBUG;
1361 my $leftresult = NZanalyse($left,$server);
1362 my $rightresult = NZanalyse($right,$server);
1363 # OK, we have the results for right and left part of the query
1364 # depending of operand, intersect, union or exclude both lists
1365 # to get a result list
1366 if ($operator eq ' and ') {
1367 my @leftresult = split /;/, $leftresult;
1368 # my @rightresult = split /;/,$leftresult;
1370 # parse the left results, and if the biblionumber exist in the right result, save it in finalresult
1371 # the result is stored twice, to have the same weight for AND than OR.
1372 # example : TWO : 61,61,64,121 (two is twice in the biblio #61) / TOWER : 61,64,130
1373 # result : 61,61,61,61,64,64 for two AND tower : 61 has more weight than 64
1374 foreach (@leftresult) {
1375 if ($rightresult =~ "$_;") {
1376 $finalresult .= "$_;$_;";
1379 return $finalresult;
1380 } elsif ($operator eq ' or ') {
1381 # just merge the 2 strings
1382 return $leftresult.$rightresult;
1383 } elsif ($operator eq ' not ') {
1384 my @leftresult = split /;/, $leftresult;
1385 # my @rightresult = split /;/,$leftresult;
1387 foreach (@leftresult) {
1388 unless ($rightresult =~ "$_;") {
1389 $finalresult .= "$_;";
1392 return $finalresult;
1394 # this error is impossible, because of the regexp that isolate the operand, but just in case...
1395 die "error : operand unknown : $operator for $string";
1397 # it's a leaf, do the real SQL query and return the result
1399 $string =~ s/__X__/"$commacontent"/ if $commacontent;
1400 $string =~ s/-|\.|\?|,|;|!|'|\(|\)|\[|\]|{|}|"|&|\+|\*|\// /g;
1401 warn "leaf:$string" if $DEBUG;
1402 # parse the string in in operator/operand/value again
1403 $string =~ /(.*)(>=|<=)(.*)/;
1407 warn "handling leaf... left:$left operator:$operator right:$right" if $DEBUG;
1408 unless ($operator) {
1409 $string =~ /(.*)(>|<|=)(.*)/;
1413 warn "handling unless (operator)... left:$left operator:$operator right:$right" if $DEBUG;
1416 # strip adv, zebra keywords, currently not handled in nozebra: wrdl, ext, phr...
1417 $left =~ s/[ ,].*$//;
1418 # automatic replace for short operators
1419 $left='title' if $left =~ '^ti$';
1420 $left='author' if $left =~ '^au$';
1421 $left='publisher' if $left =~ '^pb$';
1422 $left='subject' if $left =~ '^su$';
1423 $left='koha-Auth-Number' if $left =~ '^an$';
1424 $left='keyword' if $left =~ '^kw$';
1425 if ($operator && $left ne 'keyword' ) {
1426 #do a specific search
1427 my $dbh = C4::Context->dbh;
1428 $operator='LIKE' if $operator eq '=' and $right=~ /%/;
1429 my $sth = $dbh->prepare("SELECT biblionumbers,value FROM nozebra WHERE server=? AND indexname=? AND value $operator ?");
1430 warn "$left / $operator / $right\n";
1431 # split each word, query the DB and build the biblionumbers result
1432 #sanitizing leftpart
1433 $left=~s/^\s+|\s+$//;
1434 foreach (split / /,$right) {
1438 warn "EXECUTE : $server, $left, $_";
1439 $sth->execute($server, $left, $_) or warn "execute failed: $!";
1440 while (my ($line,$value) = $sth->fetchrow) {
1441 # if we are dealing with a numeric value, use only numeric results (in case of >=, <=, > or <)
1442 # otherwise, fill the result
1443 $biblionumbers .= $line unless ($right =~ /^\d+$/ && $value =~ /\D/);
1444 warn "result : $value ". ($right =~ /\d/) . "==".(!$value =~ /\d/) ;#= $line";
1446 # do a AND with existing list if there is one, otherwise, use the biblionumbers list as 1st result list
1448 my @leftresult = split /;/, $biblionumbers;
1450 foreach my $entry (@leftresult) { # $_ contains biblionumber,title-weight
1451 # remove weight at the end
1452 my $cleaned = $entry;
1453 $cleaned =~ s/-\d*$//;
1454 # if the entry already in the hash, take it & increase weight
1455 warn "===== $cleaned =====" if $DEBUG;
1456 if ($results =~ "$cleaned") {
1457 $temp .= "$entry;$entry;";
1458 warn "INCLUDING $entry" if $DEBUG;
1463 $results = $biblionumbers;
1467 #do a complete search (all indexes), if index='kw' do complete search too.
1468 my $dbh = C4::Context->dbh;
1469 my $sth = $dbh->prepare("SELECT biblionumbers FROM nozebra WHERE server=? AND value LIKE ?");
1470 # split each word, query the DB and build the biblionumbers result
1471 foreach (split / /,$string) {
1472 next if C4::Context->stopwords->{uc($_)}; # skip if stopword
1473 warn "search on all indexes on $_" if $DEBUG;
1476 $sth->execute($server, $_);
1477 while (my $line = $sth->fetchrow) {
1478 $biblionumbers .= $line;
1480 # do a AND with existing list if there is one, otherwise, use the biblionumbers list as 1st result list
1482 warn "RES for $_ = $biblionumbers" if $DEBUG;
1483 my @leftresult = split /;/, $biblionumbers;
1485 foreach my $entry (@leftresult) { # $_ contains biblionumber,title-weight
1486 # remove weight at the end
1487 my $cleaned = $entry;
1488 $cleaned =~ s/-\d*$//;
1489 # if the entry already in the hash, take it & increase weight
1490 # warn "===== $cleaned =====" if $DEBUG;
1491 if ($results =~ "$cleaned") {
1492 $temp .= "$entry;$entry;";
1493 # warn "INCLUDING $entry" if $DEBUG;
1498 warn "NEW RES for $_ = $biblionumbers" if $DEBUG;
1499 $results = $biblionumbers;
1503 warn "return : $results for LEAF : $string" if $DEBUG;
1506 warn "---------" if $DEBUG;
1507 warn "Leave NZanalyse" if $DEBUG;
1508 warn "---------" if $DEBUG;
1513 $finalresult = NZorder($biblionumbers, $ordering,$results_per_page,$offset);
1521 my ($biblionumbers, $ordering,$results_per_page,$offset) = @_;
1522 warn "biblionumbers = $biblionumbers and ordering = $ordering\n" if $DEBUG;
1523 # order title asc by default
1524 # $ordering = '1=36 <i' unless $ordering;
1525 $results_per_page=20 unless $results_per_page;
1526 $offset = 0 unless $offset;
1527 my $dbh = C4::Context->dbh;
1529 # order by POPULARITY
1531 if ($ordering =~ /popularity/) {
1534 # popularity is not in MARC record, it's builded from a specific query
1535 my $sth = $dbh->prepare("select sum(issues) from items where biblionumber=?");
1536 foreach (split /;/,$biblionumbers) {
1537 my ($biblionumber,$title) = split /,/,$_;
1538 $result{$biblionumber}=GetMarcBiblio($biblionumber);
1539 $sth->execute($biblionumber);
1540 my $popularity= $sth->fetchrow ||0;
1541 # hint : the key is popularity.title because we can have
1542 # many results with the same popularity. In this cas, sub-ordering is done by title
1543 # we also have biblionumber to avoid bug for 2 biblios with the same title & popularity
1544 # (un-frequent, I agree, but we won't forget anything that way ;-)
1545 $popularity{sprintf("%10d",$popularity).$title.$biblionumber} = $biblionumber;
1547 # sort the hash and return the same structure as GetRecords (Zebra querying)
1550 if ($ordering eq 'popularity_dsc') { # sort popularity DESC
1551 foreach my $key (sort {$b cmp $a} (keys %popularity)) {
1552 $result_hash->{'RECORDS'}[$numbers++] = $result{$popularity{$key}}->as_usmarc();
1554 } else { # sort popularity ASC
1555 foreach my $key (sort (keys %popularity)) {
1556 $result_hash->{'RECORDS'}[$numbers++] = $result{$popularity{$key}}->as_usmarc();
1560 $result_hash->{'hits'} = $numbers;
1561 $finalresult->{'biblioserver'} = $result_hash;
1562 return $finalresult;
1566 } elsif ($ordering =~/author/){
1568 foreach (split /;/,$biblionumbers) {
1569 my ($biblionumber,$title) = split /,/,$_;
1570 my $record=GetMarcBiblio($biblionumber);
1572 if (C4::Context->preference('marcflavour') eq 'UNIMARC') {
1573 $author=$record->subfield('200','f');
1574 $author=$record->subfield('700','a') unless $author;
1576 $author=$record->subfield('100','a');
1578 # hint : the result is sorted by title.biblionumber because we can have X biblios with the same title
1579 # and we don't want to get only 1 result for each of them !!!
1580 $result{$author.$biblionumber}=$record;
1582 # sort the hash and return the same structure as GetRecords (Zebra querying)
1585 if ($ordering eq 'author_za') { # sort by author desc
1586 foreach my $key (sort { $b cmp $a } (keys %result)) {
1587 $result_hash->{'RECORDS'}[$numbers++] = $result{$key}->as_usmarc();
1589 } else { # sort by author ASC
1590 foreach my $key (sort (keys %result)) {
1591 $result_hash->{'RECORDS'}[$numbers++] = $result{$key}->as_usmarc();
1595 $result_hash->{'hits'} = $numbers;
1596 $finalresult->{'biblioserver'} = $result_hash;
1597 return $finalresult;
1599 # ORDER BY callnumber
1601 } elsif ($ordering =~/callnumber/){
1603 foreach (split /;/,$biblionumbers) {
1604 my ($biblionumber,$title) = split /,/,$_;
1605 my $record=GetMarcBiblio($biblionumber);
1607 my ($callnumber_tag,$callnumber_subfield)=GetMarcFromKohaField($dbh,'items.itemcallnumber');
1608 ($callnumber_tag,$callnumber_subfield)= GetMarcFromKohaField('biblioitems.callnumber') unless $callnumber_tag;
1609 if (C4::Context->preference('marcflavour') eq 'UNIMARC') {
1610 $callnumber=$record->subfield('200','f');
1612 $callnumber=$record->subfield('100','a');
1614 # hint : the result is sorted by title.biblionumber because we can have X biblios with the same title
1615 # and we don't want to get only 1 result for each of them !!!
1616 $result{$callnumber.$biblionumber}=$record;
1618 # sort the hash and return the same structure as GetRecords (Zebra querying)
1621 if ($ordering eq 'call_number_dsc') { # sort by title desc
1622 foreach my $key (sort { $b cmp $a } (keys %result)) {
1623 $result_hash->{'RECORDS'}[$numbers++] = $result{$key}->as_usmarc();
1625 } else { # sort by title ASC
1626 foreach my $key (sort { $a cmp $b } (keys %result)) {
1627 $result_hash->{'RECORDS'}[$numbers++] = $result{$key}->as_usmarc();
1631 $result_hash->{'hits'} = $numbers;
1632 $finalresult->{'biblioserver'} = $result_hash;
1633 return $finalresult;
1634 } elsif ($ordering =~ /pubdate/){ #pub year
1636 foreach (split /;/,$biblionumbers) {
1637 my ($biblionumber,$title) = split /,/,$_;
1638 my $record=GetMarcBiblio($biblionumber);
1639 my ($publicationyear_tag,$publicationyear_subfield)=GetMarcFromKohaField('biblioitems.publicationyear','');
1640 my $publicationyear=$record->subfield($publicationyear_tag,$publicationyear_subfield);
1641 # hint : the result is sorted by title.biblionumber because we can have X biblios with the same title
1642 # and we don't want to get only 1 result for each of them !!!
1643 $result{$publicationyear.$biblionumber}=$record;
1645 # sort the hash and return the same structure as GetRecords (Zebra querying)
1648 if ($ordering eq 'pubdate_dsc') { # sort by pubyear desc
1649 foreach my $key (sort { $b cmp $a } (keys %result)) {
1650 $result_hash->{'RECORDS'}[$numbers++] = $result{$key}->as_usmarc();
1652 } else { # sort by pub year ASC
1653 foreach my $key (sort (keys %result)) {
1654 $result_hash->{'RECORDS'}[$numbers++] = $result{$key}->as_usmarc();
1658 $result_hash->{'hits'} = $numbers;
1659 $finalresult->{'biblioserver'} = $result_hash;
1660 return $finalresult;
1664 } elsif ($ordering =~ /title/) {
1665 # the title is in the biblionumbers string, so we just need to build a hash, sort it and return
1667 foreach (split /;/,$biblionumbers) {
1668 my ($biblionumber,$title) = split /,/,$_;
1669 # hint : the result is sorted by title.biblionumber because we can have X biblios with the same title
1670 # and we don't want to get only 1 result for each of them !!!
1671 # hint & speed improvement : we can order without reading the record
1672 # so order, and read records only for the requested page !
1673 $result{$title.$biblionumber}=$biblionumber;
1675 # sort the hash and return the same structure as GetRecords (Zebra querying)
1678 if ($ordering eq 'title_az') { # sort by title desc
1679 foreach my $key (sort (keys %result)) {
1680 $result_hash->{'RECORDS'}[$numbers++] = $result{$key};
1682 } else { # sort by title ASC
1683 foreach my $key (sort { $b cmp $a } (keys %result)) {
1684 $result_hash->{'RECORDS'}[$numbers++] = $result{$key};
1687 # limit the $results_per_page to result size if it's more
1688 $results_per_page = $numbers-1 if $numbers < $results_per_page;
1689 # for the requested page, replace biblionumber by the complete record
1690 # speed improvement : avoid reading too much things
1691 for (my $counter=$offset;$counter<=$offset+$results_per_page;$counter++) {
1692 $result_hash->{'RECORDS'}[$counter] = GetMarcBiblio($result_hash->{'RECORDS'}[$counter])->as_usmarc;
1695 $result_hash->{'hits'} = $numbers;
1696 $finalresult->{'biblioserver'} = $result_hash;
1697 return $finalresult;
1702 # we need 2 hashes to order by ranking : the 1st one to count the ranking, the 2nd to order by ranking
1705 foreach (split /;/,$biblionumbers) {
1706 my ($biblionumber,$title) = split /,/,$_;
1707 $title =~ /(.*)-(\d)/;
1710 # note that we + the ranking because ranking is calculated on weight of EACH term requested.
1711 # if we ask for "two towers", and "two" has weight 2 in biblio N, and "towers" has weight 4 in biblio N
1712 # biblio N has ranking = 6
1713 $count_ranking{$biblionumber} += $ranking;
1715 # build the result by "inverting" the count_ranking hash
1716 # 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
1718 foreach (keys %count_ranking) {
1719 $result{sprintf("%10d",$count_ranking{$_}).'-'.$_} = $_;
1721 # sort the hash and return the same structure as GetRecords (Zebra querying)
1724 foreach my $key (sort {$b cmp $a} (keys %result)) {
1725 $result_hash->{'RECORDS'}[$numbers++] = $result{$key};
1727 # limit the $results_per_page to result size if it's more
1728 $results_per_page = $numbers-1 if $numbers < $results_per_page;
1729 # for the requested page, replace biblionumber by the complete record
1730 # speed improvement : avoid reading too much things
1731 for (my $counter=$offset;$counter<=$offset+$results_per_page;$counter++) {
1732 $result_hash->{'RECORDS'}[$counter] = GetMarcBiblio($result_hash->{'RECORDS'}[$counter])->as_usmarc if $result_hash->{'RECORDS'}[$counter];
1735 $result_hash->{'hits'} = $numbers;
1736 $finalresult->{'biblioserver'} = $result_hash;
1737 return $finalresult;
1742 ($countchanged,$listunchanged) = ModBiblios($listbiblios, $tagsubfield,$initvalue,$targetvalue,$test);
1744 this function changes all the values $initvalue in subfield $tag$subfield in any record in $listbiblios
1745 test parameter if set donot perform change to records in database.
1751 * $listbiblios is an array ref to marcrecords to be changed
1752 * $tagsubfield is the reference of the subfield to change.
1753 * $initvalue is the value to search the record for
1754 * $targetvalue is the value to set the subfield to
1755 * $test is to be set only not to perform changes in database.
1757 =item C<Output arg:>
1758 * $countchanged counts all the changes performed.
1759 * $listunchanged contains the list of all the biblionumbers of records unchanged.
1761 =item C<usage in the script:>
1765 my ($countchanged, $listunchanged) = EditBiblios($results->{RECORD}, $tagsubfield,$initvalue,$targetvalue);;
1766 #If one wants to display unchanged records, you should get biblios foreach @$listunchanged
1767 $template->param(countchanged => $countchanged, loopunchanged=>$listunchanged);
1772 my ($listbiblios,$tagsubfield,$initvalue,$targetvalue,$test)=@_;
1775 my ($tag,$subfield)=($1,$2) if ($tagsubfield=~/^(\d{1,3})([a-z0-9A-Z@])?$/);
1776 if ((length($tag)<3)&& $subfield=~/0-9/){
1777 $tag=$tag.$subfield;
1780 my ($bntag,$bnsubf) = GetMarcFromKohaField('biblio.biblionumber');
1781 my ($itemtag,$itemsubf) = GetMarcFromKohaField('items.itemnumber');
1782 foreach my $usmarc (@$listbiblios){
1784 $record=eval{MARC::Record->new_from_usmarc($usmarc)};
1787 # usmarc is not a valid usmarc May be a biblionumber
1788 if ($tag eq $itemtag){
1789 my $bib=GetBiblioFromItemNumber($usmarc);
1790 $record=GetMarcItem($bib->{'biblionumber'},$usmarc) ;
1791 $biblionumber=$bib->{'biblionumber'};
1793 $record=GetMarcBiblio($usmarc);
1794 $biblionumber=$usmarc;
1798 $biblionumber = $record->subfield($bntag,$bnsubf);
1800 $biblionumber=$record->field($bntag)->data;
1803 #GetBiblionumber is to be written.
1804 #Could be replaced by TransformMarcToKoha (But Would be longer)
1805 if ($record->field($tag)){
1807 foreach my $field ($record->field($tag)){
1809 if ($field->delete_subfield('code' =>$subfield,'match'=>qr($initvalue))){
1812 $field->update($subfield,$targetvalue) if ($targetvalue);
1816 if ($field->delete_field($field)){
1821 $field->data=$targetvalue if ($field->data=~qr($initvalue));
1825 # warn $record->as_formatted;
1827 ModBiblio($record,$biblionumber,GetFrameworkCode($biblionumber)) unless ($test);
1829 push @unmatched, $biblionumber;
1832 push @unmatched, $biblionumber;
1835 return ($countmatched,\@unmatched);
1838 END { } # module clean-up code here (global destructor)
1845 Koha Developement team <info@koha.org>