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
25 use vars qw($VERSION @ISA @EXPORT @EXPORT_OK %EXPORT_TAGS);
27 # set the version for version checking
28 $VERSION = do { my @v = '$Revision$' =~ /\d+/g;
29 shift(@v) . "." . join( "_", map { sprintf "%03d", $_ } @v );
34 C4::Search - Functions for searching the Koha catalog.
38 see opac/opac-search.pl or catalogue/search.pl for example of usage
42 This module provides the searching facilities for the Koha into a zebra catalog.
60 # make all your functions, whether exported or not;
62 =head2 findseealso($dbh,$fields);
64 C<$dbh> is a link to the DB handler.
67 my $dbh =C4::Context->dbh;
69 C<$fields> is a reference to the fields array
71 This function modify the @$fields array and add related fields to search on.
76 my ( $dbh, $fields ) = @_;
77 my $tagslib = GetMarcStructure( $dbh, 1 );
78 for ( my $i = 0 ; $i <= $#{$fields} ; $i++ ) {
79 my ($tag) = substr( @$fields[$i], 1, 3 );
80 my ($subfield) = substr( @$fields[$i], 4, 1 );
81 @$fields[$i] .= ',' . $tagslib->{$tag}->{$subfield}->{seealso}
82 if ( $tagslib->{$tag}->{$subfield}->{seealso} );
88 ($biblionumber,$biblionumber,$title) = FindDuplicate($record);
95 my $dbh = C4::Context->dbh;
96 my $result = TransformMarcToKoha( $dbh, $record, '' );
101 my ( $biblionumber, $title );
103 # search duplicate on ISBN, easy and fast..
104 #$search->{'avoidquerylog'}=1;
105 if ( $result->{isbn} ) {
106 $query = "isbn=$result->{isbn}";
109 $result->{title} =~ s /\\//g;
110 $result->{title} =~ s /\"//g;
111 $result->{title} =~ s /\(//g;
112 $result->{title} =~ s /\)//g;
113 $query = "ti,ext=$result->{title}";
115 my ($possible_duplicate_record) =
116 C4::Biblio::getRecord( "biblioserver", $query, "usmarc" ); # FIXME :: hardcoded !
117 if ($possible_duplicate_record) {
119 MARC::Record->new_from_usmarc($possible_duplicate_record);
120 my $result = TransformMarcToKoha( $dbh, $marcrecord, '' );
122 # FIXME :: why 2 $biblionumber ?
123 return $result->{'biblionumber'}, $result->{'biblionumber'},
131 ($error,$results) = SimpleSearch($query,@servers);
133 this function performs a simple search on the catalog using zoom.
139 * $query could be a simple keyword or a complete CCL query wich is depending on your ccl file.
140 * @servers is optionnal. default one is read on koha.xml
143 * $error is a string which containt the description error if there is one. Else it's empty.
144 * \@results is an array of marc record.
146 =item C<usage in the script:>
150 my ($error, $marcresults) = SimpleSearch($query);
152 if (defined $error) {
153 $template->param(query_error => $error);
154 warn "error: ".$error;
155 output_html_with_http_headers $input, $cookie, $template->output;
159 my $hits = scalar @$marcresults;
162 for(my $i=0;$i<$hits;$i++) {
164 my $marcrecord = MARC::File::USMARC::decode($marcresults->[$i]);
165 my $biblio = TransformMarcToKoha(C4::Context->dbh,$marcrecord,'');
167 #build the hash for the template.
168 $resultsloop{highlight} = ($i % 2)?(1):(0);
169 $resultsloop{title} = $biblio->{'title'};
170 $resultsloop{subtitle} = $biblio->{'subtitle'};
171 $resultsloop{biblionumber} = $biblio->{'biblionumber'};
172 $resultsloop{author} = $biblio->{'author'};
173 $resultsloop{publishercode} = $biblio->{'publishercode'};
174 $resultsloop{publicationyear} = $biblio->{'publicationyear'};
176 push @results, \%resultsloop;
178 $template->param(result=>\@results);
184 if (C4::Context->preference('NoZebra')) {
185 my $result = NZorder(NZanalyse($query))->{'biblioserver'}->{'RECORDS'};
186 return (undef,$result);
192 return ( "No query entered", undef ) unless $query;
194 #@servers = (C4::Context->config("biblioserver")) unless @servers;
196 ("biblioserver") unless @servers
197 ; # FIXME hardcoded value. See catalog/search.pl & opac-search.pl too.
200 for ( my $i = 0 ; $i < @servers ; $i++ ) {
201 $zconns[$i] = C4::Context->Zconn( $servers[$i], 1 );
204 ->search( new ZOOM::Query::CCL2RPN( $query, $zconns[$i] ) );
206 # getting error message if one occured.
208 $zconns[$i]->errmsg() . " ("
209 . $zconns[$i]->errcode() . ") "
210 . $zconns[$i]->addinfo() . " "
211 . $zconns[$i]->diagset();
213 return ( $error, undef ) if $zconns[$i]->errcode();
217 while ( ( my $i = ZOOM::event( \@zconns ) ) != 0 ) {
218 $ev = $zconns[ $i - 1 ]->last_event();
219 if ( $ev == ZOOM::Event::ZEND ) {
220 $hits = $tmpresults[ $i - 1 ]->size();
223 for ( my $j = 0 ; $j < $hits ; $j++ ) {
224 my $record = $tmpresults[ $i - 1 ]->record($j)->raw();
225 push @results, $record;
229 return ( undef, \@results );
233 # performs the search
236 $koha_query, $federated_query, $sort_by_ref,
237 $servers_ref, $results_per_page, $offset,
238 $expanded_facet, $branches, $query_type,
242 my @servers = @$servers_ref;
243 my @sort_by = @$sort_by_ref;
245 # create the zoom connection and query object
249 my $results_hashref = ();
252 my $facets_counter = ();
253 my $facets_info = ();
254 my $facets = getFacets();
256 #### INITIALIZE SOME VARS USED CREATE THE FACETED RESULTS
257 my @facets_loop; # stores the ref to array of hashes for template
258 for ( my $i = 0 ; $i < @servers ; $i++ ) {
259 $zconns[$i] = C4::Context->Zconn( $servers[$i], 1 );
261 # perform the search, create the results objects
262 # if this is a local search, use the $koha-query, if it's a federated one, use the federated-query
264 if ( $servers[$i] =~ /biblioserver/ ) {
265 $query_to_use = $koha_query;
268 $query_to_use = $federated_query;
271 # check if we've got a query_type defined
275 if ( $query_type =~ /^ccl/ ) {
277 s/\:/\=/g; # change : to = last minute (FIXME)
279 # warn "CCL : $query_to_use";
282 new ZOOM::Query::CCL2RPN( $query_to_use, $zconns[$i] )
285 elsif ( $query_type =~ /^cql/ ) {
287 # warn "CQL : $query_to_use";
290 new ZOOM::Query::CQL( $query_to_use, $zconns[$i] ) );
292 elsif ( $query_type =~ /^pqf/ ) {
294 # warn "PQF : $query_to_use";
297 new ZOOM::Query::PQF( $query_to_use, $zconns[$i] ) );
303 # warn "preparing to scan";
306 new ZOOM::Query::CCL2RPN( $query_to_use, $zconns[$i] )
311 # warn "LAST : $query_to_use";
314 new ZOOM::Query::CCL2RPN( $query_to_use, $zconns[$i] )
320 warn "prob with query toto $query_to_use " . $@;
323 # concatenate the sort_by limits and pass them to the results object
325 foreach my $sort (@sort_by) {
326 $sort_by .= $sort . " "; # used to be $sort,
328 $results[$i]->sort( "yaz", $sort_by ) if $sort_by;
330 while ( ( my $i = ZOOM::event( \@zconns ) ) != 0 ) {
331 my $ev = $zconns[ $i - 1 ]->last_event();
332 if ( $ev == ZOOM::Event::ZEND ) {
333 my $size = $results[ $i - 1 ]->size();
336 #$results_hash->{'server'} = $servers[$i-1];
337 # loop through the results
338 $results_hash->{'hits'} = $size;
340 if ( $offset + $results_per_page <= $size ) {
341 $times = $offset + $results_per_page;
346 for ( my $j = $offset ; $j < $times ; $j++ )
347 { #(($offset+$count<=$size) ? ($offset+$count):$size) ; $j++){
351 ## This is just an index scan
353 my ( $term, $occ ) = $results[ $i - 1 ]->term($j);
355 # here we create a minimal MARC record and hand it off to the
356 # template just like a normal result ... perhaps not ideal, but
358 my $tmprecord = MARC::Record->new();
359 $tmprecord->encoding('UTF-8');
362 # srote the minimal record in author/title (depending on MARC flavour)
363 if ( C4::Context->preference("marcflavour") eq
366 $tmptitle = MARC::Field->new(
373 $tmptitle = MARC::Field->new(
379 $tmprecord->append_fields($tmptitle);
380 $results_hash->{'RECORDS'}[$j] =
381 $tmprecord->as_usmarc();
384 $record = $results[ $i - 1 ]->record($j)->raw();
386 #warn "RECORD $j:".$record;
387 $results_hash->{'RECORDS'}[$j] =
388 $record; # making a reference to a hash
389 # Fill the facets while we're looping
390 $facet_record = MARC::Record->new_from_usmarc($record);
392 #warn $servers[$i-1].$facet_record->title();
393 for ( my $k = 0 ; $k <= @$facets ; $k++ ) {
394 if ( $facets->[$k] ) {
396 for my $tag ( @{ $facets->[$k]->{'tags'} } ) {
397 push @fields, $facet_record->field($tag);
399 for my $field (@fields) {
400 my @subfields = $field->subfields();
401 for my $subfield (@subfields) {
402 my ( $code, $data ) = @$subfield;
404 $facets->[$k]->{'subfield'} )
406 $facets_counter->{ $facets->[$k]
407 ->{'link_value'} }->{$data}++;
411 $facets_info->{ $facets->[$k]->{'link_value'} }
413 $facets->[$k]->{'label_value'};
414 $facets_info->{ $facets->[$k]->{'link_value'} }
415 ->{'expanded'} = $facets->[$k]->{'expanded'};
420 $results_hashref->{ $servers[ $i - 1 ] } = $results_hash;
423 #print "connection ", $i-1, ": $size hits";
424 #print $results[$i-1]->record(0)->render() if $size > 0;
427 sort { $facets_counter->{$b} <=> $facets_counter->{$a} }
428 keys %$facets_counter
432 my $number_of_facets;
433 my @this_facets_array;
436 $facets_counter->{$link_value}
437 ->{$b} <=> $facets_counter->{$link_value}->{$a}
438 } keys %{ $facets_counter->{$link_value} }
442 if ( ( $number_of_facets < 6 )
443 || ( $expanded_facet eq $link_value )
444 || ( $facets_info->{$link_value}->{'expanded'} ) )
447 # sanitize the link value ), ( will cause errors with CCL
448 my $facet_link_value = $one_facet;
449 $facet_link_value =~ s/(\(|\))/ /g;
451 # fix the length that will display in the label
452 my $facet_label_value = $one_facet;
453 $facet_label_value = substr( $one_facet, 0, 20 ) . "..."
454 unless length($facet_label_value) <= 20;
456 # well, if it's a branch, label by the name, not the code
457 if ( $link_value =~ /branch/ ) {
459 $branches->{$one_facet}->{'branchname'};
462 # but we're down with the whole label being in the link's title
463 my $facet_title_value = $one_facet;
465 push @this_facets_array,
469 $facets_counter->{$link_value}->{$one_facet},
470 facet_label_value => $facet_label_value,
471 facet_title_value => $facet_title_value,
472 facet_link_value => $facet_link_value,
473 type_link_value => $link_value,
478 unless ( $facets_info->{$link_value}->{'expanded'} ) {
480 if ( ( $number_of_facets > 6 )
481 && ( $expanded_facet ne $link_value ) );
486 type_link_value => $link_value,
487 type_id => $link_value . "_id",
489 $facets_info->{$link_value}->{'label_value'},
490 facets => \@this_facets_array,
491 expandable => $expandable,
492 expand => $link_value,
498 return ( undef, $results_hashref, \@facets_loop );
501 # build the query itself
503 my ( $query, $operators, $operands, $indexes, $limits, $sort_by ) = @_;
505 my @operators = @$operators if $operators;
506 my @indexes = @$indexes if $indexes;
507 my @operands = @$operands if $operands;
508 my @limits = @$limits if $limits;
509 my @sort_by = @$sort_by if $sort_by;
511 my $human_search_desc; # a human-readable query
512 my $machine_search_desc; #a machine-readable query
513 # FIXME: the locale should be set based on the syspref
514 my $stemmer = Lingua::Stem->new( -locale => 'EN-US' );
516 # FIXME: these should be stored in the db so the librarian can modify the behavior
517 $stemmer->add_exceptions(
525 # STEP I: determine if this is a form-based / simple query or if it's complex (if complex,
526 # we can't handle field weighting, stemming until a formal query parser is written
527 # I'll work on this soon -- JF
528 #if (!$query) { # form-based
529 # check if this is a known query language query, if it is, return immediately:
530 if ( $query =~ /^ccl=/ ) {
531 return ( undef, $', $', $', 'ccl' );
533 if ( $query =~ /^cql=/ ) {
534 return ( undef, $', $', $', 'cql' );
536 if ( $query =~ /^pqf=/ ) {
537 return ( undef, $', $', $', 'pqf' );
539 if ( $query =~ /(\(|\))/ ) { # sorry, too complex
540 return ( undef, $query, $query, $query, 'ccl' );
543 # form-based queries are limited to non-nested a specific depth, so we can easily
544 # modify the incoming query operands and indexes to do stemming and field weighting
545 # Once we do so, we'll end up with a value in $query, just like if we had an
546 # incoming $query from the user
549 ; # clear it out so we can populate properly with field-weighted stemmed query
551 ; # a flag used to keep track if there was a previous query
552 # if there was, we can apply the current operator
553 for ( my $i = 0 ; $i <= @operands ; $i++ ) {
554 my $operand = $operands[$i];
555 my $index = $indexes[$i];
557 my $stemming = C4::Context->parameters("Stemming") || 0;
558 my $weight_fields = C4::Context->parameters("WeightFields") || 0;
560 if ( $operands[$i] ) {
562 # STEMMING FIXME: need to refine the field weighting so stemmed operands don't disrupt the query ranking
564 my @words = split( / /, $operands[$i] );
565 my $stems = $stemmer->stem(@words);
566 foreach my $stem (@$stems) {
567 $stemmed_operand .= "$stem";
568 $stemmed_operand .= "?"
569 unless ( $stem =~ /(and$|or$|not$)/ )
570 || ( length($stem) < 3 );
571 $stemmed_operand .= " ";
573 #warn "STEM: $stemmed_operand";
576 #$operand = $stemmed_operand;
579 # FIELD WEIGHTING - This is largely experimental stuff. What I'm committing works
580 # pretty well but will work much better when we have an actual query parser
582 if ($weight_fields) {
584 " rk=("; # Specifies that we're applying rank
585 # keyword has different weight properties
586 if ( ( $index =~ /kw/ ) || ( !$index ) )
587 { # FIXME: do I need to add right-truncation in the case of stemming?
588 # a simple way to find out if this query uses an index
589 if ( $operand =~ /(\=|\:)/ ) {
590 $weighted_query .= " $operand";
594 " Title-cover,ext,r1=\"$operand\""
595 ; # index label as exact
597 " or ti,ext,r2=$operand"; # index as exact
598 #$weighted_query .= " or ti,phr,r3=$operand"; # index as phrase
599 #$weighted_query .= " or any,ext,r4=$operand"; # index as exact
601 " or kw,wrdl,r5=$operand"; # index as exact
602 $weighted_query .= " or wrd,fuzzy,r9=$operand";
603 $weighted_query .= " or wrd=$stemmed_operand"
607 elsif ( $index =~ /au/ ) {
609 " $index,ext,r1=$operand"; # index label as exact
610 #$weighted_query .= " or (title-sort-az=0 or $index,startswithnt,st-word,r3=$operand #)";
612 " or $index,phr,r3=$operand"; # index as phrase
613 $weighted_query .= " or $index,rt,wrd,r3=$operand";
615 elsif ( $index =~ /ti/ ) {
617 " Title-cover,ext,r1=$operand"; # index label as exact
618 $weighted_query .= " or Title-series,ext,r2=$operand";
620 #$weighted_query .= " or ti,ext,r2=$operand";
621 #$weighted_query .= " or ti,phr,r3=$operand";
622 #$weighted_query .= " or ti,wrd,r3=$operand";
624 " or (title-sort-az=0 or Title-cover,startswithnt,st-word,r3=$operand #)";
626 " or (title-sort-az=0 or Title-cover,phr,r6=$operand)";
628 #$weighted_query .= " or Title-cover,wrd,r5=$operand";
629 #$weighted_query .= " or ti,ext,r6=$operand";
630 #$weighted_query .= " or ti,startswith,phr,r7=$operand";
631 #$weighted_query .= " or ti,phr,r8=$operand";
632 #$weighted_query .= " or ti,wrd,r9=$operand";
634 #$weighted_query .= " or ti,ext,r2=$operand"; # index as exact
635 #$weighted_query .= " or ti,phr,r3=$operand"; # index as phrase
636 #$weighted_query .= " or any,ext,r4=$operand"; # index as exact
637 #$weighted_query .= " or kw,wrd,r5=$operand"; # index as exact
641 " $index,ext,r1=$operand"; # index label as exact
642 #$weighted_query .= " or $index,ext,r2=$operand"; # index as exact
644 " or $index,phr,r3=$operand"; # index as phrase
645 $weighted_query .= " or $index,rt,wrd,r3=$operand";
647 " or $index,wrd,r5=$operand"
648 ; # index as word right-truncated
649 $weighted_query .= " or $index,wrd,fuzzy,r8=$operand";
651 $weighted_query .= ")"; # close rank specification
652 $operand = $weighted_query;
655 # only add an operator if there is a previous operand
656 if ($previous_operand) {
657 if ( $operators[ $i - 1 ] ) {
658 $query .= " $operators[$i-1] $index: $operand";
660 $human_search_desc .=
661 " $operators[$i-1] $operands[$i]";
664 $human_search_desc .=
665 " $operators[$i-1] $index: $operands[$i]";
669 # the default operator is and
671 $query .= " and $index: $operand";
672 $human_search_desc .= " and $index: $operands[$i]";
677 $query .= " $operand";
678 $human_search_desc .= " $operands[$i]";
681 $query .= " $index: $operand";
682 $human_search_desc .= " $index: $operands[$i]";
684 $previous_operand = 1;
692 my $limit_search_desc;
693 foreach my $limit (@limits) {
695 # FIXME: not quite right yet ... will work on this soon -- JF
696 my $type = $1 if $limit =~ m/([^:]+):([^:]*)/;
697 if ( $limit =~ /available/ ) {
699 " (($query and datedue=0000-00-00) or ($query and datedue=0000-00-00 not lost=1) or ($query and datedue=0000-00-00 not lost=2))";
701 #$limit_search_desc.=" and available";
703 elsif ( ($limit_query) && ( index( $limit_query, $type, 0 ) > 0 ) ) {
704 if ( $limit_query !~ /\(/ ) {
706 substr( $limit_query, 0, index( $limit_query, $type, 0 ) )
708 . substr( $limit_query, index( $limit_query, $type, 0 ) )
712 substr( $limit_search_desc, 0,
713 index( $limit_search_desc, $type, 0 ) )
715 . substr( $limit_search_desc,
716 index( $limit_search_desc, $type, 0 ) )
722 chop $limit_search_desc;
723 $limit_query .= " or $limit )" if $limit;
724 $limit_search_desc .= " or $limit )" if $limit;
727 elsif ( ($limit_query) && ( $limit =~ /mc/ ) ) {
728 $limit_query .= " or $limit" if $limit;
729 $limit_search_desc .= " or $limit" if $limit;
732 # these are treated as AND
733 elsif ($limit_query) {
734 if ($limit =~ /branch/){
735 $limit_query .= " ) and ( $limit" if $limit;
736 $limit_search_desc .= " ) and ( $limit" if $limit;
738 $limit_query .= " or $limit" if $limit;
739 $limit_search_desc .= " or $limit" if $limit;
743 # otherwise, there is nothing but the limit
745 $limit_query .= "$limit" if $limit;
746 $limit_search_desc .= "$limit" if $limit;
750 # if there's also a query, we need to AND the limits to it
751 if ( ($limit_query) && ($query) ) {
752 $limit_query = " and (" . $limit_query . ")";
753 $limit_search_desc = " and ($limit_search_desc)" if $limit_search_desc;
756 $query .= $limit_query;
757 $human_search_desc .= $limit_search_desc;
759 # now normalize the strings
760 $query =~ s/ / /g; # remove extra spaces
761 $query =~ s/^ //g; # remove any beginning spaces
762 $query =~ s/:/=/g; # causes probs for server
763 $query =~ s/==/=/g; # remove double == from query
765 my $federated_query = $human_search_desc;
766 $federated_query =~ s/ / /g;
767 $federated_query =~ s/^ //g;
768 $federated_query =~ s/:/=/g;
769 my $federated_query_opensearch = $federated_query;
771 # my $federated_query_RPN = new ZOOM::Query::CCL2RPN( $query , C4::Context->ZConn('biblioserver'));
773 $human_search_desc =~ s/ / /g;
774 $human_search_desc =~ s/^ //g;
775 my $koha_query = $query;
777 #warn "QUERY:".$koha_query;
778 #warn "SEARCHDESC:".$human_search_desc;
779 #warn "FEDERATED QUERY:".$federated_query;
780 return ( undef, $human_search_desc, $koha_query, $federated_query );
783 # IMO this subroutine is pretty messy still -- it's responsible for
784 # building the HTML output for the template
786 my ( $searchdesc, $hits, $results_per_page, $offset, @marcresults ) = @_;
788 my $dbh = C4::Context->dbh;
792 my $span_terms_hashref;
793 for my $span_term ( split( / /, $searchdesc ) ) {
794 $span_term =~ s/(.*=|\)|\(|\+|\.)//g;
795 $span_terms_hashref->{$span_term}++;
798 #Build brancnames hash
800 #get branch information.....
803 $dbh->prepare("SELECT branchcode,branchname FROM branches")
804 ; # FIXME : use C4::Koha::GetBranches
806 while ( my $bdata = $bsth->fetchrow_hashref ) {
807 $branches{ $bdata->{'branchcode'} } = $bdata->{'branchname'};
811 #find itemtype & itemtype image
814 $dbh->prepare("SELECT itemtype,description,imageurl,summary FROM itemtypes");
816 while ( my $bdata = $bsth->fetchrow_hashref ) {
817 $itemtypes{ $bdata->{'itemtype'} }->{description} =
818 $bdata->{'description'};
819 $itemtypes{ $bdata->{'itemtype'} }->{imageurl} = $bdata->{'imageurl'};
820 $itemtypes{ $bdata->{'itemtype'} }->{summary} = $bdata->{'summary'};
823 #search item field code
826 "select tagfield from marc_subfield_structure where kohafield like 'items.itemnumber'"
829 my ($itemtag) = $sth->fetchrow;
831 ## find column names of items related to MARC
832 my $sth2 = $dbh->prepare("SHOW COLUMNS from items");
834 my %subfieldstosearch;
835 while ( ( my $column ) = $sth2->fetchrow ) {
836 my ( $tagfield, $tagsubfield ) =
837 &GetMarcFromKohaField( "items." . $column, "" );
838 $subfieldstosearch{$column} = $tagsubfield;
842 if ( $hits && $offset + $results_per_page <= $hits ) {
843 $times = $offset + $results_per_page;
849 for ( my $i = $offset ; $i <= $times - 1 ; $i++ ) {
851 $marcrecord = MARC::File::USMARC::decode( $marcresults[$i] );
853 my $oldbiblio = TransformMarcToKoha( $dbh, $marcrecord, '' );
855 # add image url if there is one
856 if ( $itemtypes{ $oldbiblio->{itemtype} }->{imageurl} =~ /^http:/ ) {
857 $oldbiblio->{imageurl} =
858 $itemtypes{ $oldbiblio->{itemtype} }->{imageurl};
859 $oldbiblio->{description} =
860 $itemtypes{ $oldbiblio->{itemtype} }->{description};
863 $oldbiblio->{imageurl} =
864 getitemtypeimagesrc() . "/"
865 . $itemtypes{ $oldbiblio->{itemtype} }->{imageurl}
866 if ( $itemtypes{ $oldbiblio->{itemtype} }->{imageurl} );
867 $oldbiblio->{description} =
868 $itemtypes{ $oldbiblio->{itemtype} }->{description};
871 # build summary if there is one (the summary is defined in itemtypes table
873 if ($itemtypes{ $oldbiblio->{itemtype} }->{summary}) {
874 my $summary = $itemtypes{ $oldbiblio->{itemtype} }->{summary};
875 my @fields = $marcrecord->fields();
876 foreach my $field (@fields) {
877 my $tag = $field->tag();
878 my $tagvalue = $field->as_string();
879 $summary =~ s/\[(.?.?.?.?)$tag\*(.*?)]/$1$tagvalue$2\[$1$tag$2]/g;
881 my @subf = $field->subfields;
882 for my $i (0..$#subf) {
883 my $subfieldcode = $subf[$i][0];
884 my $subfieldvalue = $subf[$i][1];
885 my $tagsubf = $tag.$subfieldcode;
886 $summary =~ s/\[(.?.?.?.?)$tagsubf(.*?)]/$1$subfieldvalue$2\[$1$tagsubf$2]/g;
890 $summary =~ s/\[(.*?)]//g;
891 $summary =~ s/\n/<br>/g;
892 $oldbiblio->{summary} = $summary;
894 # add spans to search term in results
895 foreach my $term ( keys %$span_terms_hashref ) {
898 my $old_term = $term;
899 if ( length($term) > 3 ) {
900 $term =~ s/(.*=|\)|\(|\+|\.|\?)//g;
902 #FIXME: is there a better way to do this?
903 $oldbiblio->{'title'} =~ s/$term/<span class=term>$&<\/span>/gi;
904 $oldbiblio->{'subtitle'} =~
905 s/$term/<span class=term>$&<\/span>/gi;
907 $oldbiblio->{'author'} =~ s/$term/<span class=term>$&<\/span>/gi;
908 $oldbiblio->{'publishercode'} =~ s/$term/<span class=term>$&<\/span>/gi;
909 $oldbiblio->{'place'} =~ s/$term/<span class=term>$&<\/span>/gi;
910 $oldbiblio->{'pages'} =~ s/$term/<span class=term>$&<\/span>/gi;
911 $oldbiblio->{'notes'} =~ s/$term/<span class=term>$&<\/span>/gi;
912 $oldbiblio->{'size'} =~ s/$term/<span class=term>$&<\/span>/gi;
922 $oldbiblio->{'toggle'} = $toggle;
923 my @fields = $marcrecord->field($itemtag);
926 my $ordered_count = 0;
927 my $onloan_count = 0;
928 my $wthdrawn_count = 0;
929 my $itemlost_count = 0;
930 my $itembinding_count = 0;
933 foreach my $field (@fields) {
935 foreach my $code ( keys %subfieldstosearch ) {
936 $item->{$code} = $field->subfield( $subfieldstosearch{$code} );
938 if ( $item->{wthdrawn} ) {
941 elsif ( $item->{notforloan} == -1 ) {
945 elsif ( $item->{itemlost} ) {
948 elsif ( $item->{binding} ) {
949 $itembinding_count++;
951 elsif ( ( $item->{onloan} ) && ( $item->{onloan} != '0000-00-00' ) )
958 if ( $item->{'homebranch'} ) {
959 $items->{ $item->{'homebranch'} }->{count}++;
963 elsif ( $item->{'holdingbranch'} ) {
964 $items->{ $item->{'homebranch'} }->{count}++;
966 $items->{ $item->{homebranch} }->{itemcallnumber} =
967 $item->{itemcallnumber};
968 $items->{ $item->{homebranch} }->{location} =
971 } # notforloan, item level and biblioitem level
972 for my $key ( keys %$items ) {
976 branchname => $branches{$key},
978 count => $items->{$key}->{count},
979 itemcallnumber => $items->{$key}->{itemcallnumber},
980 location => $items->{$key}->{location},
982 push @items_loop, $this_item;
984 $oldbiblio->{norequests} = $norequests;
985 $oldbiblio->{items_loop} = \@items_loop;
986 $oldbiblio->{onloancount} = $onloan_count;
987 $oldbiblio->{wthdrawncount} = $wthdrawn_count;
988 $oldbiblio->{itemlostcount} = $itemlost_count;
989 $oldbiblio->{bindingcount} = $itembinding_count;
990 $oldbiblio->{orderedcount} = $ordered_count;
993 # Ugh ... this is ugly, I'll re-write it better above then delete it
994 # my $norequests = 1;
998 # foreach my $itm (@items) {
999 # $norequests = 0 unless $itm->{'itemnotforloan'};
1002 # $oldbiblio->{'noitems'} = $noitems;
1003 # $oldbiblio->{'norequests'} = $norequests;
1004 # $oldbiblio->{'even'} = $even = not $even;
1005 # $oldbiblio->{'itemcount'} = $counts{'total'};
1006 # my $totalitemcounts = 0;
1007 # foreach my $key (keys %counts){
1008 # if ($key ne 'total'){
1009 # $totalitemcounts+= $counts{$key};
1010 # $oldbiblio->{'locationhash'}->{$key}=$counts{$key};
1013 # my ($locationtext, $locationtextonly, $notavailabletext) = ('','','');
1014 # foreach (sort keys %{$oldbiblio->{'locationhash'}}) {
1015 # if ($_ eq 'notavailable') {
1016 # $notavailabletext="Not available";
1017 # my $c=$oldbiblio->{'locationhash'}->{$_};
1018 # $oldbiblio->{'not-available-p'}=$c;
1020 # $locationtext.="$_";
1021 # my $c=$oldbiblio->{'locationhash'}->{$_};
1022 # if ($_ eq 'Item Lost') {
1023 # $oldbiblio->{'lost-p'} = $c;
1024 # } elsif ($_ eq 'Withdrawn') {
1025 # $oldbiblio->{'withdrawn-p'} = $c;
1026 # } elsif ($_ eq 'On Loan') {
1027 # $oldbiblio->{'on-loan-p'} = $c;
1029 # $locationtextonly.= $_;
1030 # $locationtextonly.= " ($c)<br/> " if $totalitemcounts > 1;
1032 # if ($totalitemcounts>1) {
1033 # $locationtext.=" ($c)<br/> ";
1037 # if ($notavailabletext) {
1038 # $locationtext.= $notavailabletext;
1040 # $locationtext=~s/, $//;
1042 # $oldbiblio->{'location'} = $locationtext;
1043 # $oldbiblio->{'location-only'} = $locationtextonly;
1044 # $oldbiblio->{'use-location-flags-p'} = 1;
1046 push( @newresults, $oldbiblio );
1054 ($countchanged,$listunchanged) = EditBiblios($listbiblios, $tagsubfield,$initvalue,$targetvalue,$test);
1056 this function changes all the values $initvalue in subfield $tag$subfield in any record in $listbiblios
1057 test parameter if set donot perform change to records in database.
1063 * $listbiblios is an array ref to marcrecords to be changed
1064 * $tagsubfield is the reference of the subfield to change.
1065 * $initvalue is the value to search the record for
1066 * $targetvalue is the value to set the subfield to
1067 * $test is to be set only not to perform changes in database.
1069 =item C<Output arg:>
1070 * $countchanged counts all the changes performed.
1071 * $listunchanged contains the list of all the biblionumbers of records unchanged.
1073 =item C<usage in the script:>
1077 my ($countchanged, $listunchanged) = EditBiblios($results->{RECORD}, $tagsubfield,$initvalue,$targetvalue);;
1078 #If one wants to display unchanged records, you should get biblios foreach @$listunchanged
1079 $template->param(countchanged => $countchanged, loopunchanged=>$listunchanged);
1083 my ($listbiblios,$tagsubfield,$initvalue,$targetvalue,$test)=@_;
1086 my ($tag,$subfield)=($1,$2) if ($tagsubfield=~/^(\d{1,3})(.)$/);
1087 my ($bntag,$bnsubf) = GetMarcFromKohaField('biblio.biblionumber');
1089 foreach my $usmarc (@$listbiblios){
1090 my $record=MARC::Record->new_from_usmarc($usmarc);
1093 $biblionumber = $record->subfield($bntag,$bnsubf);
1095 $biblionumber=$record->field($bntag)->data;
1097 #GetBiblionumber is to be written.
1098 #Could be replaced by TransformMarcToKoha (But Would be longer)
1099 if ($record->field($tag)){
1100 foreach my $field ($record->field($tag)){
1101 if ($field->delete_subfield('code' =>$subfield,'match'=>qr($initvalue))){
1103 $field->update($subfield,$targetvalue) if ($targetvalue);
1106 # warn $record->as_formatted;
1107 ModBiblio($record,$biblionumber,GetFrameworkCode($biblionumber)) unless ($test);
1109 push @unmatched, $biblionumber;
1112 return ($countmatched,\@unmatched);
1115 #----------------------------------------------------------------------
1117 # Non-Zebra GetRecords#
1118 #----------------------------------------------------------------------
1121 NZgetRecords has the same API as zera getRecords, even if some parameters are not managed
1126 $koha_query, $federated_query, $sort_by_ref,
1127 $servers_ref, $results_per_page, $offset,
1128 $expanded_facet, $branches, $query_type,
1131 my $result = NZanalyse($koha_query);
1133 # warn "==========".@$sort_by_ref[0];
1134 return (undef,NZorder($result,@$sort_by_ref[0],$results_per_page,$offset),undef);
1139 NZanalyse : get a CQL string as parameter, and returns a list of biblionumber;title,biblionumber;title,...
1140 the list is builded from inverted index in nozebra SQL table
1141 note that title is here only for convenience : the sorting will be very fast when requested on title
1142 if the sorting is requested on something else, we will have to reread all results, and that may be longer.
1148 # if we have a ", replace the content to discard temporarily any and/or/not inside
1150 if ($string =~/"/) {
1151 $string =~ s/"(.*?)"/__X__/;
1153 # print "commacontent : $commacontent\n";
1155 # split the query string in 3 parts : X AND Y means : $left="X", $operand="AND" and $right="Y"
1156 # then, call again NZanalyse with $left and $right
1157 # (recursive until we find a leaf (=> something without and/or/not)
1158 $string =~ /(.*)( and | or | not )(.*)/;
1162 # it's not a leaf, we have a and/or/not
1164 # reintroduce comma content if needed
1165 $right =~ s/__X__/"$commacontent"/ if $commacontent;
1166 $left =~ s/__X__/"$commacontent"/ if $commacontent;
1167 # print "noeud : $left / $operand / $right\n";
1168 my $leftresult = NZanalyse($left);
1169 my $rightresult = NZanalyse($right);
1170 # OK, we have the results for right and left part of the query
1171 # depending of operand, intersect, union or exclude both lists
1172 # to get a result list
1173 if ($operand eq ' and ') {
1174 my @leftresult = split /,/, $leftresult;
1175 # my @rightresult = split /,/,$leftresult;
1177 # parse the left results, and if the biblionumber exist in the right result, save it in finalresult
1178 # the result is stored twice, to have the same weight for AND than OR.
1179 # example : TWO : 61,61,64,121 (two is twice in the biblio #61) / TOWER : 61,64,130
1180 # result : 61,61,61,61,64,64 for two AND tower : 61 has more weight than 64
1181 foreach (@leftresult) {
1182 if ($rightresult =~ "$_,") {
1183 $finalresult .= "$_,$_,";
1186 return $finalresult;
1187 } elsif ($operand eq ' or ') {
1188 # just merge the 2 strings
1189 return $leftresult.$rightresult;
1190 } elsif ($operand eq ' not ') {
1191 my @leftresult = split /,/, $leftresult;
1192 # my @rightresult = split /,/,$leftresult;
1194 foreach (@leftresult) {
1195 unless ($rightresult =~ "$_,") {
1196 $finalresult .= "$_,";
1199 return $finalresult;
1201 # this error is impossible, because of the regexp that isolate the operand, but just in case...
1202 die "error : operand unknown : $operand for $string";
1204 # it's a leaf, do the real SQL query and return the result
1206 $string =~ s/__X__/"$commacontent"/ if $commacontent;
1207 $string =~ s/-|\.|\?|,|;|!|'|\(|\)|\[|\]|{|}|"|<|>|&|\+|\*|\// /g;
1208 # print "feuille : $string\n";
1209 # parse the string in in operator/operand/value again
1210 $string =~ /(.*)(=|>|>=|<|<=)(.*)/;
1215 # automatic replace for short operator
1216 $left='title' if $left eq 'ti';
1217 $left='author' if $left eq 'au';
1219 #do a specific search
1220 my $dbh = C4::Context->dbh;
1221 $operator='LIKE' if $operator eq '=' and $right=~ /%/;
1222 my $sth = $dbh->prepare("SELECT biblionumbers FROM nozebra WHERE indexname=? AND value $operator ?");
1223 # print "$left / $operator / $right\n";
1224 # split each word, query the DB and build the biblionumbers result
1225 foreach (split / /,$right) {
1227 $sth->execute($left,$_);
1228 while (my $line = $sth->fetchrow) {
1229 $biblionumbers .= $line;
1231 # do a AND with existing list if there is one, otherwise, use the biblionumbers list as 1st result list
1233 my @leftresult = split /;/, $biblionumbers;
1235 foreach (@leftresult) {
1236 if ($results =~ "$_;") {
1242 $results = $biblionumbers;
1246 #do a complete search (all indexes)
1247 my $dbh = C4::Context->dbh;
1248 my $sth = $dbh->prepare("SELECT biblionumbers FROM nozebra WHERE value LIKE ?");
1249 # split each word, query the DB and build the biblionumbers result
1250 foreach (split / /,$string) {
1253 while (my $line = $sth->fetchrow) {
1254 $biblionumbers .= $line;
1256 # do a AND with existing list if there is one, otherwise, use the biblionumbers list as 1st result list
1258 my @leftresult = split /,/, $biblionumbers;
1260 foreach (@leftresult) {
1261 if ($results =~ "$_;") {
1267 $results = $biblionumbers;
1276 my ($biblionumbers, $ordering,$results_per_page,$offset) = @_;
1277 # order title asc by default
1278 # $ordering = '1=36 <i' unless $ordering;
1279 $results_per_page=20 unless $results_per_page;
1280 $offset = 0 unless $offset;
1281 my $dbh = C4::Context->dbh;
1283 # order by POPULARITY
1285 if ($ordering =~ /1=9523/) {
1288 # popularity is not in MARC record, it's builded from a specific query
1289 my $sth = $dbh->prepare("select sum(issues) from items where biblionumber=?");
1290 foreach (split /;/,$biblionumbers) {
1291 my ($biblionumber,$title) = split /,/,$_;
1292 $result{$biblionumber}=GetMarcBiblio($biblionumber);
1293 $sth->execute($biblionumber);
1294 my $popularity= $sth->fetchrow ||0;
1295 # hint : the key is popularity.title because we can have
1296 # many results with the same popularity. In this cas, sub-ordering is done by title
1297 # we also have biblionumber to avoid bug for 2 biblios with the same title & popularity
1298 # (un-frequent, I agree, but we won't forget anything that way ;-)
1299 $popularity{sprintf("%10d",$popularity).$title.$biblionumber} = $biblionumber;
1301 # sort the hash and return the same structure as GetRecords (Zebra querying)
1304 if ($ordering eq '1=9523 >i') { # sort popularity DESC
1305 foreach my $key (sort {$b <=> $a} (keys %popularity)) {
1306 $result_hash->{'RECORDS'}[$numbers++] = $result{$popularity{$key}}->as_usmarc();
1308 } else { # sort popularity ASC
1309 foreach my $key (sort (keys %popularity)) {
1310 $result_hash->{'RECORDS'}[$numbers++] = $result{$popularity{$key}}->as_usmarc();
1314 $result_hash->{'hits'} = $numbers;
1315 $finalresult->{'biblioserver'} = $result_hash;
1316 return $finalresult;
1320 } elsif ($ordering eq '1=1003 <i'){
1322 foreach (split /;/,$biblionumbers) {
1323 my ($biblionumber,$title) = split /,/,$_;
1324 my $record=GetMarcBiblio($biblionumber);
1326 if (C4::Context->preference('marcflavour') eq 'UNIMARC') {
1327 $author=$record->subfield('200','f');
1328 $author=$record->subfield('700','a') unless $author;
1330 $author=$record->subfield('100','a');
1332 # hint : the result is sorted by title.biblionumber because we can have X biblios with the same title
1333 # and we don't want to get only 1 result for each of them !!!
1334 $result{$author.$biblionumber}=$record;
1336 # sort the hash and return the same structure as GetRecords (Zebra querying)
1339 if ($ordering eq '1=1003 <i') { # sort by title desc
1340 foreach my $key (sort (keys %result)) {
1341 $result_hash->{'RECORDS'}[$numbers++] = $result{$key}->as_usmarc();
1343 } else { # sort by title ASC
1344 foreach my $key (sort { $a <=> $b } (keys %result)) {
1345 $result_hash->{'RECORDS'}[$numbers++] = $result{$key}->as_usmarc();
1349 $result_hash->{'hits'} = $numbers;
1350 $finalresult->{'biblioserver'} = $result_hash;
1351 return $finalresult;
1353 # ORDER BY callnumber
1355 } elsif ($ordering eq '1=20 <i'){
1357 foreach (split /;/,$biblionumbers) {
1358 my ($biblionumber,$title) = split /,/,$_;
1359 my $record=GetMarcBiblio($biblionumber);
1361 my ($callnumber_tag,$callnumber_subfield)=GetMarcFromKohaField($dbh,'items.itemcallnumber');
1362 ($callnumber_tag,$callnumber_subfield)= GetMarcFromKohaField('biblioitems.callnumber') unless $callnumber_tag;
1363 if (C4::Context->preference('marcflavour') eq 'UNIMARC') {
1364 $callnumber=$record->subfield('200','f');
1366 $callnumber=$record->subfield('100','a');
1368 # hint : the result is sorted by title.biblionumber because we can have X biblios with the same title
1369 # and we don't want to get only 1 result for each of them !!!
1370 $result{$callnumber.$biblionumber}=$record;
1372 # sort the hash and return the same structure as GetRecords (Zebra querying)
1375 if ($ordering eq '1=1003 <i') { # sort by title desc
1376 foreach my $key (sort (keys %result)) {
1377 $result_hash->{'RECORDS'}[$numbers++] = $result{$key}->as_usmarc();
1379 } else { # sort by title ASC
1380 foreach my $key (sort { $a <=> $b } (keys %result)) {
1381 $result_hash->{'RECORDS'}[$numbers++] = $result{$key}->as_usmarc();
1385 $result_hash->{'hits'} = $numbers;
1386 $finalresult->{'biblioserver'} = $result_hash;
1387 return $finalresult;
1388 } elsif ($ordering =~ /1=31/){ #pub year
1390 foreach (split /;/,$biblionumbers) {
1391 my ($biblionumber,$title) = split /,/,$_;
1392 my $record=GetMarcBiblio($biblionumber);
1393 my ($publicationyear_tag,$publicationyear_subfield)=GetMarcFromKohaField($dbh,'biblioitems.publicationyear');
1394 my $publicationyear=$record->subfield($publicationyear_tag,$publicationyear_subfield);
1395 # hint : the result is sorted by title.biblionumber because we can have X biblios with the same title
1396 # and we don't want to get only 1 result for each of them !!!
1397 $result{$publicationyear.$biblionumber}=$record;
1399 # sort the hash and return the same structure as GetRecords (Zebra querying)
1402 if ($ordering eq '1=31 <i') { # sort by title desc
1403 foreach my $key (sort (keys %result)) {
1404 $result_hash->{'RECORDS'}[$numbers++] = $result{$key}->as_usmarc();
1406 } else { # sort by title ASC
1407 foreach my $key (sort { $a <=> $b } (keys %result)) {
1408 $result_hash->{'RECORDS'}[$numbers++] = $result{$key}->as_usmarc();
1412 $result_hash->{'hits'} = $numbers;
1413 $finalresult->{'biblioserver'} = $result_hash;
1414 return $finalresult;
1418 } elsif ($ordering =~ /1=36/) {
1419 # the title is in the biblionumbers string, so we just need to build a hash, sort it and return
1421 foreach (split /;/,$biblionumbers) {
1422 my ($biblionumber,$title) = split /,/,$_;
1423 # hint : the result is sorted by title.biblionumber because we can have X biblios with the same title
1424 # and we don't want to get only 1 result for each of them !!!
1425 # hint & speed improvement : we can order without reading the record
1426 # so order, and read records only for the requested page !
1427 $result{$title.$biblionumber}=$biblionumber;
1429 # sort the hash and return the same structure as GetRecords (Zebra querying)
1432 if ($ordering eq '1=36 <i') { # sort by title desc
1433 foreach my $key (sort (keys %result)) {
1434 $result_hash->{'RECORDS'}[$numbers++] = $result{$key};
1436 } else { # sort by title ASC
1437 foreach my $key (sort { $a <=> $b } (keys %result)) {
1438 $result_hash->{'RECORDS'}[$numbers++] = $result{$key};
1441 # limit the $results_per_page to result size if it's more
1442 $results_per_page = $numbers-1 if $numbers < $results_per_page;
1443 # for the requested page, replace biblionumber by the complete record
1444 # speed improvement : avoid reading too much things
1445 for (my $counter=$offset;$counter<=$offset+$results_per_page;$counter++) {
1446 $result_hash->{'RECORDS'}[$counter] = GetMarcBiblio($result_hash->{'RECORDS'}[$counter])->as_usmarc;
1449 $result_hash->{'hits'} = $numbers;
1450 $finalresult->{'biblioserver'} = $result_hash;
1451 return $finalresult;
1456 # we need 2 hashes to order by ranking : the 1st one to count the ranking, the 2nd to order by ranking
1459 foreach (split /;/,$biblionumbers) {
1460 my ($biblionumber,$title) = split /,/,$_;
1461 $title =~ /(.*)-(\d)/;
1464 # note that we + the ranking because ranking is calculated on weight of EACH term requested.
1465 # if we ask for "two towers", and "two" has weight 2 in biblio N, and "towers" has weight 4 in biblio N
1466 # biblio N has ranking = 6
1467 $count_ranking{$biblionumber} =+ $ranking;
1469 # build the result by "inverting" the count_ranking hash
1470 # 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
1472 foreach (keys %count_ranking) {
1473 $result{sprintf("%10d",$count_ranking{$_}).'-'.$_} = $_;
1475 # sort the hash and return the same structure as GetRecords (Zebra querying)
1478 foreach my $key (sort {$b <=> $a} (keys %result)) {
1479 $result_hash->{'RECORDS'}[$numbers++] = $result{$key};
1481 # limit the $results_per_page to result size if it's more
1482 $results_per_page = $numbers-1 if $numbers < $results_per_page;
1483 # for the requested page, replace biblionumber by the complete record
1484 # speed improvement : avoid reading too much things
1485 for (my $counter=$offset;$counter<=$offset+$results_per_page;$counter++) {
1486 $result_hash->{'RECORDS'}[$counter] = GetMarcBiblio($result_hash->{'RECORDS'}[$counter])->as_usmarc;
1489 $result_hash->{'hits'} = $numbers;
1490 $finalresult->{'biblioserver'} = $result_hash;
1491 return $finalresult;
1495 END { } # module clean-up code here (global destructor)
1502 Koha Developement team <info@koha.org>