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
26 use vars qw($VERSION @ISA @EXPORT @EXPORT_OK %EXPORT_TAGS);
28 # set the version for version checking
33 C4::Search - Functions for searching the Koha catalog.
37 see opac/opac-search.pl or catalogue/search.pl for example of usage
41 This module provides the searching facilities for the Koha into a zebra catalog.
59 # make all your functions, whether exported or not;
61 =head2 findseealso($dbh,$fields);
63 C<$dbh> is a link to the DB handler.
66 my $dbh =C4::Context->dbh;
68 C<$fields> is a reference to the fields array
70 This function modify the @$fields array and add related fields to search on.
75 my ( $dbh, $fields ) = @_;
76 my $tagslib = GetMarcStructure( 1 );
77 for ( my $i = 0 ; $i <= $#{$fields} ; $i++ ) {
78 my ($tag) = substr( @$fields[$i], 1, 3 );
79 my ($subfield) = substr( @$fields[$i], 4, 1 );
80 @$fields[$i] .= ',' . $tagslib->{$tag}->{$subfield}->{seealso}
81 if ( $tagslib->{$tag}->{$subfield}->{seealso} );
87 ($biblionumber,$biblionumber,$title) = FindDuplicate($record);
93 my $dbh = C4::Context->dbh;
94 my $result = TransformMarcToKoha( $dbh, $record, '' );
99 my ( $biblionumber, $title );
101 # search duplicate on ISBN, easy and fast..
102 # ... normalize first
103 if ( $result->{isbn} ) {
104 $result->{isbn} =~ s/\(.*$//;
105 $result->{isbn} =~ s/\s+$//;
107 #$search->{'avoidquerylog'}=1;
108 if ( $result->{isbn} ) {
109 $query = "isbn=$result->{isbn}";
112 $result->{title} =~ s /\\//g;
113 $result->{title} =~ s /\"//g;
114 $result->{title} =~ s /\(//g;
115 $result->{title} =~ s /\)//g;
116 # remove valid operators
117 $result->{title} =~ s/(and|or|not)//g;
118 $query = "ti,ext=$result->{title}";
119 $query .= " and mt=$result->{itemtype}" if ($result->{itemtype});
120 if ($result->{author}){
121 $result->{author} =~ s /\\//g;
122 $result->{author} =~ s /\"//g;
123 $result->{author} =~ s /\(//g;
124 $result->{author} =~ s /\)//g;
125 # remove valid operators
126 $result->{author} =~ s/(and|or|not)//g;
127 $query .= " and au,ext=$result->{author}";
130 my ($error,$searchresults) =
131 SimpleSearch($query); # FIXME :: hardcoded !
133 foreach my $possible_duplicate_record (@$searchresults) {
135 MARC::Record->new_from_usmarc($possible_duplicate_record);
136 my $result = TransformMarcToKoha( $dbh, $marcrecord, '' );
138 # FIXME :: why 2 $biblionumber ?
140 push @results, $result->{'biblionumber'};
141 push @results, $result->{'title'};
149 ($error,$results) = SimpleSearch($query,@servers);
151 this function performs a simple search on the catalog using zoom.
157 * $query could be a simple keyword or a complete CCL query wich is depending on your ccl file.
158 * @servers is optionnal. default one is read on koha.xml
161 * $error is a string which containt the description error if there is one. Else it's empty.
162 * \@results is an array of marc record.
164 =item C<usage in the script:>
168 my ($error, $marcresults) = SimpleSearch($query);
170 if (defined $error) {
171 $template->param(query_error => $error);
172 warn "error: ".$error;
173 output_html_with_http_headers $input, $cookie, $template->output;
177 my $hits = scalar @$marcresults;
180 for(my $i=0;$i<$hits;$i++) {
182 my $marcrecord = MARC::File::USMARC::decode($marcresults->[$i]);
183 my $biblio = TransformMarcToKoha(C4::Context->dbh,$marcrecord,'');
185 #build the hash for the template.
186 $resultsloop{highlight} = ($i % 2)?(1):(0);
187 $resultsloop{title} = $biblio->{'title'};
188 $resultsloop{subtitle} = $biblio->{'subtitle'};
189 $resultsloop{biblionumber} = $biblio->{'biblionumber'};
190 $resultsloop{author} = $biblio->{'author'};
191 $resultsloop{publishercode} = $biblio->{'publishercode'};
192 $resultsloop{publicationyear} = $biblio->{'publicationyear'};
194 push @results, \%resultsloop;
196 $template->param(result=>\@results);
202 if (C4::Context->preference('NoZebra')) {
203 my $result = NZorder(NZanalyse($query))->{'biblioserver'}->{'RECORDS'};
204 return (undef,$result);
210 return ( "No query entered", undef ) unless $query;
212 #@servers = (C4::Context->config("biblioserver")) unless @servers;
214 ("biblioserver") unless @servers
215 ; # FIXME hardcoded value. See catalog/search.pl & opac-search.pl too.
218 for ( my $i = 0 ; $i < @servers ; $i++ ) {
219 $zconns[$i] = C4::Context->Zconn( $servers[$i], 1 );
222 ->search( new ZOOM::Query::CCL2RPN( $query, $zconns[$i] ) );
224 # getting error message if one occured.
226 $zconns[$i]->errmsg() . " ("
227 . $zconns[$i]->errcode() . ") "
228 . $zconns[$i]->addinfo() . " "
229 . $zconns[$i]->diagset();
231 return ( $error, undef ) if $zconns[$i]->errcode();
235 while ( ( my $i = ZOOM::event( \@zconns ) ) != 0 ) {
236 $ev = $zconns[ $i - 1 ]->last_event();
237 if ( $ev == ZOOM::Event::ZEND ) {
238 $hits = $tmpresults[ $i - 1 ]->size();
241 for ( my $j = 0 ; $j < $hits ; $j++ ) {
242 my $record = $tmpresults[ $i - 1 ]->record($j)->raw();
243 push @results, $record;
247 return ( undef, \@results );
251 # performs the search
254 $koha_query, $federated_query, $sort_by_ref,
255 $servers_ref, $results_per_page, $offset,
256 $expanded_facet, $branches, $query_type,
259 # warn "Query : $koha_query";
260 my @servers = @$servers_ref;
261 my @sort_by = @$sort_by_ref;
263 # create the zoom connection and query object
267 my $results_hashref = ();
270 my $facets_counter = ();
271 my $facets_info = ();
272 my $facets = getFacets();
274 #### INITIALIZE SOME VARS USED CREATE THE FACETED RESULTS
275 my @facets_loop; # stores the ref to array of hashes for template
276 for ( my $i = 0 ; $i < @servers ; $i++ ) {
277 $zconns[$i] = C4::Context->Zconn( $servers[$i], 1 );
279 # perform the search, create the results objects
280 # if this is a local search, use the $koha-query, if it's a federated one, use the federated-query
282 if ( $servers[$i] =~ /biblioserver/ ) {
283 $query_to_use = $koha_query;
286 $query_to_use = $federated_query;
289 # check if we've got a query_type defined
293 if ( $query_type =~ /^ccl/ ) {
295 s/\:/\=/g; # change : to = last minute (FIXME)
297 # warn "CCL : $query_to_use";
300 new ZOOM::Query::CCL2RPN( $query_to_use, $zconns[$i] )
303 elsif ( $query_type =~ /^cql/ ) {
305 # warn "CQL : $query_to_use";
308 new ZOOM::Query::CQL( $query_to_use, $zconns[$i] ) );
310 elsif ( $query_type =~ /^pqf/ ) {
312 # warn "PQF : $query_to_use";
315 new ZOOM::Query::PQF( $query_to_use, $zconns[$i] ) );
321 # warn "preparing to scan";
324 new ZOOM::Query::CCL2RPN( $query_to_use, $zconns[$i] )
329 # warn "LAST : $query_to_use";
332 new ZOOM::Query::CCL2RPN( $query_to_use, $zconns[$i] )
338 warn "WARNING: query problem with $query_to_use " . $@;
341 # concatenate the sort_by limits and pass them to the results object
343 foreach my $sort (@sort_by) {
344 if ($sort eq "author_az") {
345 $sort_by.="1=1003 <i ";
347 elsif ($sort eq "author_za") {
348 $sort_by.="1=1003 >i ";
350 elsif ($sort eq "popularity_asc") {
351 $sort_by.="1=9003 <i ";
353 elsif ($sort eq "popularity_dsc") {
354 $sort_by.="1=9003 >i ";
356 elsif ($sort eq "call_number_asc") {
357 $sort_by.="1=20 <i ";
359 elsif ($sort eq "call_number_dsc") {
360 $sort_by.="1=20 >i ";
362 elsif ($sort eq "pubdate_asc") {
363 $sort_by.="1=31 <i ";
365 elsif ($sort eq "pubdate_dsc") {
366 $sort_by.="1=31 >i ";
368 elsif ($sort eq "acqdate_asc") {
369 $sort_by.="1=32 <i ";
371 elsif ($sort eq "acqdate_dsc") {
372 $sort_by.="1=32 >i ";
374 elsif ($sort eq "title_az") {
377 elsif ($sort eq "title_za") {
382 if ( $results[$i]->sort( "yaz", $sort_by ) < 0) {
383 warn "WARNING sort $sort_by failed";
387 while ( ( my $i = ZOOM::event( \@zconns ) ) != 0 ) {
388 my $ev = $zconns[ $i - 1 ]->last_event();
389 if ( $ev == ZOOM::Event::ZEND ) {
390 my $size = $results[ $i - 1 ]->size();
393 #$results_hash->{'server'} = $servers[$i-1];
394 # loop through the results
395 $results_hash->{'hits'} = $size;
397 if ( $offset + $results_per_page <= $size ) {
398 $times = $offset + $results_per_page;
403 for ( my $j = $offset ; $j < $times ; $j++ )
404 { #(($offset+$count<=$size) ? ($offset+$count):$size) ; $j++){
408 ## This is just an index scan
410 my ( $term, $occ ) = $results[ $i - 1 ]->term($j);
412 # here we create a minimal MARC record and hand it off to the
413 # template just like a normal result ... perhaps not ideal, but
415 my $tmprecord = MARC::Record->new();
416 $tmprecord->encoding('UTF-8');
419 # srote the minimal record in author/title (depending on MARC flavour)
420 if ( C4::Context->preference("marcflavour") eq
423 $tmptitle = MARC::Field->new(
430 $tmptitle = MARC::Field->new(
436 $tmprecord->append_fields($tmptitle);
437 $results_hash->{'RECORDS'}[$j] =
438 $tmprecord->as_usmarc();
441 $record = $results[ $i - 1 ]->record($j)->raw();
443 #warn "RECORD $j:".$record;
444 $results_hash->{'RECORDS'}[$j] =
445 $record; # making a reference to a hash
446 # Fill the facets while we're looping
447 $facet_record = MARC::Record->new_from_usmarc($record);
449 #warn $servers[$i-1].$facet_record->title();
450 for ( my $k = 0 ; $k <= @$facets ; $k++ ) {
451 if ( $facets->[$k] ) {
453 for my $tag ( @{ $facets->[$k]->{'tags'} } ) {
454 push @fields, $facet_record->field($tag);
456 for my $field (@fields) {
457 my @subfields = $field->subfields();
458 for my $subfield (@subfields) {
459 my ( $code, $data ) = @$subfield;
461 $facets->[$k]->{'subfield'} )
463 $facets_counter->{ $facets->[$k]
464 ->{'link_value'} }->{$data}++;
468 $facets_info->{ $facets->[$k]->{'link_value'} }
470 $facets->[$k]->{'label_value'};
471 $facets_info->{ $facets->[$k]->{'link_value'} }
472 ->{'expanded'} = $facets->[$k]->{'expanded'};
477 $results_hashref->{ $servers[ $i - 1 ] } = $results_hash;
480 #print "connection ", $i-1, ": $size hits";
481 #print $results[$i-1]->record(0)->render() if $size > 0;
484 sort { $facets_counter->{$b} <=> $facets_counter->{$a} }
485 keys %$facets_counter
489 my $number_of_facets;
490 my @this_facets_array;
493 $facets_counter->{$link_value}
494 ->{$b} <=> $facets_counter->{$link_value}->{$a}
495 } keys %{ $facets_counter->{$link_value} }
499 if ( ( $number_of_facets < 6 )
500 || ( $expanded_facet eq $link_value )
501 || ( $facets_info->{$link_value}->{'expanded'} ) )
504 # sanitize the link value ), ( will cause errors with CCL
505 my $facet_link_value = $one_facet;
506 $facet_link_value =~ s/(\(|\))/ /g;
508 # fix the length that will display in the label
509 my $facet_label_value = $one_facet;
510 $facet_label_value = substr( $one_facet, 0, 20 ) . "..."
511 unless length($facet_label_value) <= 20;
513 # well, if it's a branch, label by the name, not the code
514 if ( $link_value =~ /branch/ ) {
516 $branches->{$one_facet}->{'branchname'};
519 # but we're down with the whole label being in the link's title
520 my $facet_title_value = $one_facet;
522 push @this_facets_array,
526 $facets_counter->{$link_value}->{$one_facet},
527 facet_label_value => $facet_label_value,
528 facet_title_value => $facet_title_value,
529 facet_link_value => $facet_link_value,
530 type_link_value => $link_value,
535 unless ( $facets_info->{$link_value}->{'expanded'} ) {
537 if ( ( $number_of_facets > 6 )
538 && ( $expanded_facet ne $link_value ) );
543 type_link_value => $link_value,
544 type_id => $link_value . "_id",
546 $facets_info->{$link_value}->{'label_value'},
547 facets => \@this_facets_array,
548 expandable => $expandable,
549 expand => $link_value,
555 return ( undef, $results_hashref, \@facets_loop );
558 sub _remove_stopwords {
559 my ($operand,$index) = @_;
560 # if the index contains more than one qualifier, but not phrase:
561 if ($index!~m/phr|ext/){
562 # operand may be a wordlist deleting stopwords
563 # remove stopwords from operand : parse all stopwords & remove them (case insensitive)
564 # we use IsAlpha unicode definition, to deal correctly with diacritics.
565 # otherwise, a french word like "leçon" is splitted in "le" "çon", le is an empty word, we get "çon"
566 # and don't find anything...
567 foreach (keys %{C4::Context->stopwords}) {
568 next if ($_ =~/(and|or|not)/); # don't remove operators
569 $operand=~ s/\P{IsAlpha}$_\P{IsAlpha}/ /i;
570 $operand=~ s/^$_\P{IsAlpha}/ /i;
571 $operand=~ s/\P{IsAlpha}$_$/ /i;
577 sub _add_truncation {
578 my ($operand,$index) = @_;
579 my (@nontruncated,@righttruncated,@lefttruncated,@rightlefttruncated,@regexpr);
580 # if the index contains more than one qualifier, but not phrase, add truncation qualifiers
581 #if (index($index,"phr")<0 && index($index,",")>0){
582 # warn "ADDING TRUNCATION QUALIFIERS";
584 my @wordlist= split (/\s/,$operand);
585 foreach my $word (@wordlist){
586 if ($word=~s/^\*([^\*]+)\*$/$1/){
587 push @rightlefttruncated,$word;
589 elsif($word=~s/^\*([^\*]+)$/$1/){
590 push @lefttruncated,$word;
593 elsif ($word=~s/^([^\*]+)\*$/$1/){
594 push @righttruncated,$word;
596 elsif (index($word,"*")<0){
597 push @nontruncated,$word;
605 return (\@nontruncated,\@righttruncated,\@lefttruncated,\@rightlefttruncated,\@regexpr);
608 sub _build_stemmed_operand {
611 #$operand =~ s/^(and |or |not )//i;
612 # STEMMING FIXME: may need to refine the field weighting so stemmed operands don't
613 # disrupt the query ranking, this needs more testing
614 # FIXME: the locale should be set based on the user's language and/or search choice
615 my $stemmer = Lingua::Stem->new( -locale => 'EN-US' );
616 # FIXME: these should be stored in the db so the librarian can modify the behavior
617 $stemmer->add_exceptions(
625 my @words = split( / /, $operand );
626 my $stems = $stemmer->stem(@words);
627 foreach my $stem (@$stems) {
628 $stemmed_operand .= "$stem";
629 $stemmed_operand .= "?" unless ( $stem =~ /(and$|or$|not$)/ ) || ( length($stem) < 3 );
630 $stemmed_operand .= " ";
632 #warn "STEMMED OPERAND: $stemmed_operand";
633 return $stemmed_operand;
636 sub _build_weighted_query {
637 # FIELD WEIGHTING - This is largely experimental stuff. What I'm committing works
638 # pretty well but will work much better when we have an actual query parser
639 my ($operand,$stemmed_operand,$index) = @_;
640 my $stemming = C4::Context->preference("QueryStemming") || 0;
641 my $weight_fields = C4::Context->preference("QueryWeightFields") || 0;
642 my $fuzzy_enabled = C4::Context->preference("QueryFuzzy") || 0;
644 my $weighted_query .= " (rk=("; # Specifies that we're applying rank
645 # keyword has different weight properties
646 if ( ( $index =~ /kw/ ) || ( !$index ) ) {
647 # a simple way to find out if this query uses an index
648 if ( $operand =~ /(\=|\:)/ ) {
649 $weighted_query .= " $operand";
652 $weighted_query .=" Title-cover,ext,r1=\"$operand\""; # title cover as exact
653 $weighted_query .=" or ti,ext,r2=\"$operand\""; # exact title elsewhere
654 $weighted_query .= " or ti,phr,r3=\"$operand\""; # index as phrase
655 #$weighted_query .= " or any,ext,r4=$operand"; # index as exact
656 #$weighted_query .=" or kw,wrdl,r5=\"$operand\""; # all the words in the query (wordlist)
657 $weighted_query .= " or wrd,fuzzy,r8=\"$operand\"" if $fuzzy_enabled; # add fuzzy
658 $weighted_query .= " or wrd,right-Truncation,r9=\"$stemmed_operand\"" if ($stemming and $stemmed_operand); # add stemming
659 # embedded sorting: 0 a-z; 1 z-a
660 #$weighted_query .= ") or (sort1,aut=1";
664 #TODO: build better cases based on specific search indexes
665 #elsif ( $index =~ /au/ ) {
666 # $weighted_query .=" $index,ext,r1=$operand"; # index label as exact
667 # #$weighted_query .= " or (title-sort-az=0 or $index,startswithnt,st-word,r3=$operand #)";
668 # $weighted_query .=" or $index,phr,r3=$operand"; # index as phrase
669 # $weighted_query .= " or $index,rt,wrd,r3=$operand";
671 #elsif ( $index =~ /ti/ ) {
672 # $weighted_query .=" Title-cover,ext,r1=$operand"; # index label as exact
673 # $weighted_query .= " or Title-series,ext,r2=$operand";
674 # #$weighted_query .= " or ti,ext,r2=$operand";
675 # #$weighted_query .= " or ti,phr,r3=$operand";
676 # #$weighted_query .= " or ti,wrd,r3=$operand";
677 # $weighted_query .=" or (title-sort-az=0 or Title-cover,startswithnt,st-word,r3=$operand #)";
678 # $weighted_query .=" or (title-sort-az=0 or Title-cover,phr,r6=$operand)";
679 #$weighted_query .= " or Title-cover,wrd,r5=$operand";
680 #$weighted_query .= " or ti,ext,r6=$operand";
681 #$weighted_query .= " or ti,startswith,phr,r7=$operand";
682 #$weighted_query .= " or ti,phr,r8=$operand";
683 #$weighted_query .= " or ti,wrd,r9=$operand";
684 #$weighted_query .= " or ti,ext,r2=$operand"; # index as exact
685 #$weighted_query .= " or ti,phr,r3=$operand"; # index as phrase
686 #$weighted_query .= " or any,ext,r4=$operand"; # index as exact
687 #$weighted_query .= " or kw,wrd,r5=$operand"; # index as exact
690 warn "WEIGHT GENERIC";
691 $weighted_query .=" $index=$operand";
692 #$weighted_query .=" $index,ext,r1=$operand"; # index label as exact
693 #$weighted_query .= " or $index,ext,r2=$operand"; # index as exact
694 #$weighted_query .=" or $index,phr,r3=$operand"; # index as phrase
695 #$weighted_query .= " or $index,rt,wrd,r3=$operand";
696 #$weighted_query .=" or $index,wrd,r5=$operand"; # index as word right-truncated
697 #$weighted_query .= " or $index,wrd,fuzzy,r8=$operand" if $fuzzy_enabled;
699 $weighted_query .= "))"; # close rank specification
700 return $weighted_query;
703 # build the query itself
705 my ( $operators, $operands, $indexes, $limits, $sort_by ) = @_;
707 my @operators = @$operators if $operators;
708 my @indexes = @$indexes if $indexes;
709 my @operands = @$operands if $operands;
710 my @limits = @$limits if $limits;
711 my @sort_by = @$sort_by if $sort_by;
713 my $stemming = C4::Context->preference("QueryStemming") || 0;
714 my $weight_fields = C4::Context->preference("QueryWeightFields") || 0;
715 my $fuzzy_enabled = C4::Context->preference("QueryFuzzy") || 0;
717 my $human_search_desc; # a human-readable query
718 my $machine_search_desc; #a machine-readable query
719 #warn "OPERATORS: >@operators< INDEXES: >@indexes< OPERANDS: >@operands< LIMITS: >@limits< SORTS: >@sort_by<";
720 my $query = $operands[0];
721 # STEP I: determine if this is a form-based / simple query or if it's complex (if complex,
722 # we can't handle field weighting, stemming until a formal query parser is written
724 # check if this is a known query language query, if it is, return immediately,
725 # the user is responsible for constructing valid syntax:
726 if ( $query =~ /^ccl=/ ) {
727 return ( undef, $', $', $', 'ccl' );
729 if ( $query =~ /^cql=/ ) {
730 return ( undef, $', $', $', 'cql' );
732 if ( $query =~ /^pqf=/ ) {
733 return ( undef, $', $', $', 'pqf' );
735 if ( $query =~ /(\(|\))/ ) { # sorry, too complex, assume CCL
736 return ( undef, $query, $query, $query, 'ccl' );
739 # form-based queries are limited to non-nested at a specific depth, so we can easily
740 # modify the incoming query operands and indexes to do stemming and field weighting
741 # Once we do so, we'll end up with a value in $query, just like if we had an
742 # incoming $query from the user
744 $query = ""; # clear it out so we can populate properly with field-weighted stemmed query
745 my $previous_operand; # a flag used to keep track if there was a previous query
746 # if there was, we can apply the current operator
748 for ( my $i = 0 ; $i <= @operands ; $i++ ) {
750 # COMBINE OPERANDS, INDEXES AND OPERATORS
751 if ( $operands[$i] ) {
752 my $operand = $operands[$i];
753 my $index = $indexes[$i];
755 # if there's no index, don't use one, it will throw a CCL error
756 my $index_plus;# $index_plus = "$index:" if $index;
757 my $index_plus_comma;# $index_plus_comma="$index," if $index;
760 $operand = _remove_stopwords($operand,$index);
761 #warn "OP_SW: $operand";
763 my ($nontruncated,$righttruncated,$lefttruncated,$rightlefttruncated,$regexpr);
764 ($nontruncated,$righttruncated,$lefttruncated,$rightlefttruncated,$regexpr) = _add_truncation($operand,$index);
765 #warn "TRUNCATION: NON:@$nontruncated RIGHT:@$righttruncated LEFT:@$lefttruncated RIGHTLEFT:@$rightlefttruncated REGEX:@$regexpr";
769 $stemmed_operand = _build_stemmed_operand($operand) if $stemming;
771 # Handle Field Weighting
772 my $weighted_operand;
773 $weighted_operand = _build_weighted_query($operand,$stemmed_operand,$index) if $weight_fields;
775 # proves we're operating in multi-leaf mode
776 # $weighted_operand = "$weighted_operand and $weighted_operand";
777 $operand = $weighted_operand if $weight_fields;
779 # If there's a previous operand, we need to add an operator
780 if ($previous_operand) {
781 # user-specified operator
782 if ( $operators[$i-1] ) {
783 $human_search_desc .=" $operators[$i-1] $index_plus $operands[$i]";
784 $query .= " $operators[$i-1] $index_plus $operand";
786 # the default operator is and
788 $query .= " and $index_plus $operand";
789 $human_search_desc .= " and $index_plus $operands[$i]";
792 # There's no previous operand - FIXME: completely ignoring our $query, no field weighting, no stemming
793 # FIXME: also, doesn't preserve original order
795 # if there are terms to fit with truncation
796 if (scalar(@$righttruncated)+scalar(@$lefttruncated)+scalar(@$rightlefttruncated)>0){
797 # add the non-truncated ones first
798 $query.= "$index_plus @$nontruncated " if (scalar(@$nontruncated)>0);
799 if (scalar(@$righttruncated)>0){
800 $query .= "and $index_plus_comma"."rtrn:@$righttruncated ";
802 if (scalar(@$lefttruncated)>0){
803 $query .= "and $index_plus_comma"."ltrn:@$lefttruncated ";
805 if (scalar(@$rightlefttruncated)>0){
806 $query .= "and $index_plus_comma"."rltrn:@$rightlefttruncated ";
808 $query=~s/^and//; # FIXME: this is cheating :-)
809 $human_search_desc .= $query;
811 $query .= " $index_plus $operand";
812 $human_search_desc .= " $index_plus $operands[$i]";
814 $previous_operand = 1;
819 #warn "QUERY:".$query;
822 my $limit_search_desc;
823 foreach my $limit (@limits) {
825 # FIXME: not quite right yet ... will work on this soon -- JF
826 my $type = $1 if $limit =~ m/([^:]+):([^:]*)/;
827 if ( $limit =~ /available/ ) {
828 $limit_query .= " (($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))";
829 #$limit_search_desc.=" and available";
831 elsif ( ($limit_query) && ( index( $limit_query, $type, 0 ) > 0 ) ) {
832 if ( $limit_query !~ /\(/ ) {
834 substr( $limit_query, 0, index( $limit_query, $type, 0 ) )
836 . substr( $limit_query, index( $limit_query, $type, 0 ) )
840 substr( $limit_search_desc, 0,
841 index( $limit_search_desc, $type, 0 ) )
843 . substr( $limit_search_desc,
844 index( $limit_search_desc, $type, 0 ) )
850 chop $limit_search_desc;
851 $limit_query .= " or $limit )" if $limit;
852 $limit_search_desc .= " or $limit )" if $limit;
855 elsif ( ($limit_query) && ( $limit =~ /mc/ ) ) {
856 $limit_query .= " or $limit" if $limit;
857 $limit_search_desc .= " or $limit" if $limit;
860 # these are treated as AND
861 elsif ($limit_query) {
862 if ($limit =~ /branch/){
863 $limit_query .= " ) and ( $limit" if $limit;
864 $limit_search_desc .= " ) and ( $limit" if $limit;
866 $limit_query .= " or $limit" if $limit;
867 $limit_search_desc .= " or $limit" if $limit;
871 # otherwise, there is nothing but the limit
873 $limit_query .= "$limit" if $limit;
874 $limit_search_desc .= "$limit" if $limit;
878 # if there's also a query, we need to AND the limits to it
879 if ( ($limit_query) && ($query) ) {
880 $limit_query = " and (" . $limit_query . ")";
881 $limit_search_desc = " and ($limit_search_desc)" if $limit_search_desc;
884 #warn "LIMIT: $limit_query";
885 $query .= $limit_query;
886 $human_search_desc .= $limit_search_desc;
888 # now normalize the strings
889 $query =~ s/ / /g; # remove extra spaces
890 $query =~ s/^ //g; # remove any beginning spaces
891 $query =~ s/:/=/g; # causes probs for server
892 $query =~ s/==/=/g; # remove double == from query
894 my $federated_query = $human_search_desc;
895 $federated_query =~ s/ / /g;
896 $federated_query =~ s/^ //g;
897 $federated_query =~ s/:/=/g;
898 my $federated_query_opensearch = $federated_query;
900 # my $federated_query_RPN = new ZOOM::Query::CCL2RPN( $query , C4::Context->ZConn('biblioserver'));
902 $human_search_desc =~ s/ / /g;
903 $human_search_desc =~ s/^ //g;
904 my $koha_query = $query;
906 #warn "QUERY:".$koha_query;
907 #warn "SEARCHDESC:".$human_search_desc;
908 #warn "FEDERATED QUERY:".$federated_query;
909 return ( undef, $human_search_desc, $koha_query, $federated_query );
912 # IMO this subroutine is pretty messy still -- it's responsible for
913 # building the HTML output for the template
915 my ( $searchdesc, $hits, $results_per_page, $offset, @marcresults ) = @_;
917 my $dbh = C4::Context->dbh;
921 my $span_terms_hashref;
922 for my $span_term ( split( / /, $searchdesc ) ) {
923 $span_term =~ s/(.*=|\)|\(|\+|\.)//g;
924 $span_terms_hashref->{$span_term}++;
927 #Build brancnames hash
929 #get branch information.....
932 $dbh->prepare("SELECT branchcode,branchname FROM branches")
933 ; # FIXME : use C4::Koha::GetBranches
935 while ( my $bdata = $bsth->fetchrow_hashref ) {
936 $branches{ $bdata->{'branchcode'} } = $bdata->{'branchname'};
940 #find itemtype & itemtype image
943 $dbh->prepare("SELECT itemtype,description,imageurl,summary,notforloan FROM itemtypes");
945 while ( my $bdata = $bsth->fetchrow_hashref ) {
946 $itemtypes{ $bdata->{'itemtype'} }->{description} =
947 $bdata->{'description'};
948 $itemtypes{ $bdata->{'itemtype'} }->{imageurl} = $bdata->{'imageurl'};
949 $itemtypes{ $bdata->{'itemtype'} }->{summary} = $bdata->{'summary'};
950 $itemtypes{ $bdata->{'itemtype'} }->{notforloan} = $bdata->{'notforloan'};
953 #search item field code
956 "select tagfield from marc_subfield_structure where kohafield like 'items.itemnumber'"
959 my ($itemtag) = $sth->fetchrow;
961 ## find column names of items related to MARC
962 my $sth2 = $dbh->prepare("SHOW COLUMNS from items");
964 my %subfieldstosearch;
965 while ( ( my $column ) = $sth2->fetchrow ) {
966 my ( $tagfield, $tagsubfield ) =
967 &GetMarcFromKohaField( "items." . $column, "" );
968 $subfieldstosearch{$column} = $tagsubfield;
972 if ( $hits && $offset + $results_per_page <= $hits ) {
973 $times = $offset + $results_per_page;
979 for ( my $i = $offset ; $i <= $times - 1 ; $i++ ) {
981 $marcrecord = MARC::File::USMARC::decode( $marcresults[$i] );
982 my $oldbiblio = TransformMarcToKoha( $dbh, $marcrecord, '' );
983 # add image url if there is one
984 if ( $itemtypes{ $oldbiblio->{itemtype} }->{imageurl} =~ /^http:/ ) {
985 $oldbiblio->{imageurl} =
986 $itemtypes{ $oldbiblio->{itemtype} }->{imageurl};
987 $oldbiblio->{description} =
988 $itemtypes{ $oldbiblio->{itemtype} }->{description};
991 $oldbiblio->{imageurl} =
992 getitemtypeimagesrc() . "/"
993 . $itemtypes{ $oldbiblio->{itemtype} }->{imageurl}
994 if ( $itemtypes{ $oldbiblio->{itemtype} }->{imageurl} );
995 $oldbiblio->{description} =
996 $itemtypes{ $oldbiblio->{itemtype} }->{description};
999 # build summary if there is one (the summary is defined in itemtypes table
1001 if ($itemtypes{ $oldbiblio->{itemtype} }->{summary}) {
1002 my $summary = $itemtypes{ $oldbiblio->{itemtype} }->{summary};
1003 my @fields = $marcrecord->fields();
1004 foreach my $field (@fields) {
1005 my $tag = $field->tag();
1006 my $tagvalue = $field->as_string();
1007 $summary =~ s/\[(.?.?.?.?)$tag\*(.*?)]/$1$tagvalue$2\[$1$tag$2]/g;
1009 my @subf = $field->subfields;
1010 for my $i (0..$#subf) {
1011 my $subfieldcode = $subf[$i][0];
1012 my $subfieldvalue = $subf[$i][1];
1013 my $tagsubf = $tag.$subfieldcode;
1014 $summary =~ s/\[(.?.?.?.?)$tagsubf(.*?)]/$1$subfieldvalue$2\[$1$tagsubf$2]/g;
1018 $summary =~ s/\[(.*?)]//g;
1019 $summary =~ s/\n/<br>/g;
1020 $oldbiblio->{summary} = $summary;
1022 # add spans to search term in results for search term highlighting
1023 foreach my $term ( keys %$span_terms_hashref ) {
1024 my $old_term = $term;
1025 if ( length($term) > 3 ) {
1026 $term =~ s/(.*=|\)|\(|\+|\.|\?|\[|\])//g;
1030 #FIXME: is there a better way to do this?
1031 $oldbiblio->{'title'} =~ s/$term/<span class=term>$&<\/span>/gi;
1032 $oldbiblio->{'subtitle'} =~
1033 s/$term/<span class=term>$&<\/span>/gi;
1035 $oldbiblio->{'author'} =~ s/$term/<span class=term>$&<\/span>/gi;
1036 $oldbiblio->{'publishercode'} =~ s/$term/<span class=term>$&<\/span>/gi;
1037 $oldbiblio->{'place'} =~ s/$term/<span class=term>$&<\/span>/gi;
1038 $oldbiblio->{'pages'} =~ s/$term/<span class=term>$&<\/span>/gi;
1039 $oldbiblio->{'notes'} =~ s/$term/<span class=term>$&<\/span>/gi;
1040 $oldbiblio->{'size'} =~ s/$term/<span class=term>$&<\/span>/gi;
1045 $toggle = "#ffffcc";
1050 $oldbiblio->{'toggle'} = $toggle;
1051 my @fields = $marcrecord->field($itemtag);
1054 my $ordered_count = 0;
1055 my $onloan_count = 0;
1056 my $wthdrawn_count = 0;
1057 my $itemlost_count = 0;
1061 # check the loan status of the item :
1062 # it is not stored in the MARC record, for pref (zebra reindexing)
1063 # reason. Thus, we have to get the status from a specific SQL query
1065 my $sth_issue = $dbh->prepare("
1066 SELECT date_due,returndate
1068 WHERE itemnumber=? AND returndate IS NULL");
1069 my $items_count=scalar(@fields);
1070 foreach my $field (@fields) {
1072 foreach my $code ( keys %subfieldstosearch ) {
1073 $item->{$code} = $field->subfield( $subfieldstosearch{$code} );
1075 $sth_issue->execute($item->{itemnumber});
1076 $item->{due_date} = format_date($sth_issue->fetchrow);
1077 $item->{onloan} = 1 if $item->{due_date};
1078 # at least one item can be reserved : suppose no
1080 if ( $item->{wthdrawn} ) {
1082 $items->{ $item->{'homebranch'}.'--'.$item->{'itemcallnumber'} }->{unavailable}=1;
1083 $items->{ $item->{'homebranch'}.'--'.$item->{'itemcallnumber'} }->{wthdrawn}=1;
1085 elsif ( $item->{itemlost} ) {
1087 $items->{ $item->{'homebranch'}.'--'.$item->{'itemcallnumber'} }->{unavailable}=1;
1088 $items->{ $item->{'homebranch'}.'--'.$item->{'itemcallnumber'} }->{itemlost}=1;
1090 unless ( $item->{notforloan}) {
1091 # OK, this one can be issued, so at least one can be reserved
1094 if ( ( $item->{onloan} ) && ( $item->{onloan} != '0000-00-00' ) )
1096 $items->{ $item->{'homebranch'}.'--'.$item->{'itemcallnumber'} }->{unavailable}=1;
1097 $items->{ $item->{'homebranch'}.'--'.$item->{'itemcallnumber'} }->{onloancount} = 1;
1098 $items->{ $item->{'homebranch'}.'--'.$item->{'itemcallnumber'} }->{due_date} = $item->{due_date};
1101 if ( $item->{'homebranch'} ) {
1102 $items->{ $item->{'homebranch'}.'--'.$item->{'itemcallnumber'} }->{count}++;
1106 elsif ( $item->{'holdingbranch'} ) {
1107 $items->{ $item->{'holdingbranch'} }->{count}++;
1109 $items->{ $item->{'homebranch'}.'--'.$item->{'itemcallnumber'} }->{itemcallnumber} = $item->{itemcallnumber};
1110 $items->{ $item->{'homebranch'}.'--'.$item->{'itemcallnumber'} }->{location} = $item->{location};
1111 $items->{ $item->{'homebranch'}.'--'.$item->{'itemcallnumber'} }->{branchcode} = $item->{homebranch};
1112 } # notforloan, item level and biblioitem level
1114 # last check for norequest : if itemtype is notforloan, it can't be reserved either, whatever the items
1115 $norequests = 1 if $itemtypes{$oldbiblio->{itemtype}}->{notforloan};
1117 for my $key ( sort keys %$items ) {
1119 branchname => $branches{$items->{$key}->{branchcode}},
1120 branchcode => $items->{$key}->{branchcode},
1121 count => $items->{$key}->{count},
1122 itemcallnumber => $items->{$key}->{itemcallnumber},
1123 location => $items->{$key}->{location},
1124 onloancount => $items->{$key}->{onloancount},
1125 due_date => $items->{$key}->{due_date},
1126 wthdrawn => $items->{$key}->{wthdrawn},
1127 lost => $items->{$key}->{itemlost},
1129 push @items_loop, $this_item;
1131 $oldbiblio->{norequests} = $norequests;
1132 $oldbiblio->{items_count} = $items_count;
1133 $oldbiblio->{items_loop} = \@items_loop;
1134 $oldbiblio->{onloancount} = $onloan_count;
1135 $oldbiblio->{wthdrawncount} = $wthdrawn_count;
1136 $oldbiblio->{itemlostcount} = $itemlost_count;
1137 $oldbiblio->{orderedcount} = $ordered_count;
1138 $oldbiblio->{isbn} =~ s/-//g; # deleting - in isbn to enable amazon content
1139 push( @newresults, $oldbiblio );
1146 #----------------------------------------------------------------------
1148 # Non-Zebra GetRecords#
1149 #----------------------------------------------------------------------
1153 NZgetRecords has the same API as zera getRecords, even if some parameters are not managed
1159 $koha_query, $federated_query, $sort_by_ref,
1160 $servers_ref, $results_per_page, $offset,
1161 $expanded_facet, $branches, $query_type,
1164 my $result = NZanalyse($koha_query);
1165 return (undef,NZorder($result,@$sort_by_ref[0],$results_per_page,$offset),undef);
1170 NZanalyse : get a CQL string as parameter, and returns a list of biblionumber;title,biblionumber;title,...
1171 the list is builded from inverted index in nozebra SQL table
1172 note that title is here only for convenience : the sorting will be very fast when requested on title
1173 if the sorting is requested on something else, we will have to reread all results, and that may be longer.
1178 my ($string,$server) = @_;
1179 # $server contains biblioserver or authorities, depending on what we search on.
1180 #warn "querying : $string on $server";
1181 $server='biblioserver' unless $server;
1182 # if we have a ", replace the content to discard temporarily any and/or/not inside
1184 if ($string =~/"/) {
1185 $string =~ s/"(.*?)"/__X__/;
1187 # print "commacontent : $commacontent\n";
1189 # split the query string in 3 parts : X AND Y means : $left="X", $operand="AND" and $right="Y"
1190 # then, call again NZanalyse with $left and $right
1191 # (recursive until we find a leaf (=> something without and/or/not)
1192 $string =~ /(.*)( and | or | not | AND | OR | NOT )(.*)/;
1195 my $operand = lc($2);
1196 # it's not a leaf, we have a and/or/not
1198 # reintroduce comma content if needed
1199 $right =~ s/__X__/"$commacontent"/ if $commacontent;
1200 $left =~ s/__X__/"$commacontent"/ if $commacontent;
1201 # warn "node : $left / $operand / $right\n";
1202 my $leftresult = NZanalyse($left,$server);
1203 my $rightresult = NZanalyse($right,$server);
1204 # OK, we have the results for right and left part of the query
1205 # depending of operand, intersect, union or exclude both lists
1206 # to get a result list
1207 if ($operand eq ' and ') {
1208 my @leftresult = split /;/, $leftresult;
1209 # my @rightresult = split /;/,$leftresult;
1211 # parse the left results, and if the biblionumber exist in the right result, save it in finalresult
1212 # the result is stored twice, to have the same weight for AND than OR.
1213 # example : TWO : 61,61,64,121 (two is twice in the biblio #61) / TOWER : 61,64,130
1214 # result : 61,61,61,61,64,64 for two AND tower : 61 has more weight than 64
1215 foreach (@leftresult) {
1216 if ($rightresult =~ "$_;") {
1217 $finalresult .= "$_;$_;";
1220 return $finalresult;
1221 } elsif ($operand eq ' or ') {
1222 # just merge the 2 strings
1223 return $leftresult.$rightresult;
1224 } elsif ($operand eq ' not ') {
1225 my @leftresult = split /;/, $leftresult;
1226 # my @rightresult = split /;/,$leftresult;
1228 foreach (@leftresult) {
1229 unless ($rightresult =~ "$_;") {
1230 $finalresult .= "$_;";
1233 return $finalresult;
1235 # this error is impossible, because of the regexp that isolate the operand, but just in case...
1236 die "error : operand unknown : $operand for $string";
1238 # it's a leaf, do the real SQL query and return the result
1240 $string =~ s/__X__/"$commacontent"/ if $commacontent;
1241 $string =~ s/-|\.|\?|,|;|!|'|\(|\)|\[|\]|{|}|"|&|\+|\*|\// /g;
1242 # warn "leaf : $string\n";
1243 # parse the string in in operator/operand/value again
1244 $string =~ /(.*)(>=|<=)(.*)/;
1248 unless ($operator) {
1249 $string =~ /(.*)(>|<|=)(.*)/;
1255 # automatic replace for short operators
1256 $left='title' if $left =~ '^ti';
1257 $left='author' if $left =~ '^au';
1258 $left='publisher' if $left =~ '^pb';
1259 $left='subject' if $left =~ '^su';
1260 $left='koha-Auth-Number' if $left =~ '^an';
1261 $left='keyword' if $left =~ '^kw';
1263 #do a specific search
1264 my $dbh = C4::Context->dbh;
1265 $operator='LIKE' if $operator eq '=' and $right=~ /%/;
1266 my $sth = $dbh->prepare("SELECT biblionumbers,value FROM nozebra WHERE server=? AND indexname=? AND value $operator ?");
1267 warn "$left / $operator / $right\n";
1268 # split each word, query the DB and build the biblionumbers result
1269 foreach (split / /,$right) {
1270 my ($biblionumbers,$value);
1272 warn "EXECUTE : $server, $left, $_";
1273 $sth->execute($server, $left, $_);
1274 while (my ($line,$value) = $sth->fetchrow) {
1275 # if we are dealing with a numeric value, use only numeric results (in case of >=, <=, > or <)
1276 # otherwise, fill the result
1277 $biblionumbers .= $line unless ($right =~ /\d/ && $value =~ /\D/);
1278 warn "result : $value ". ($right =~ /\d/) . "==".(!$value =~ /\d/) ;#= $line";
1280 # do a AND with existing list if there is one, otherwise, use the biblionumbers list as 1st result list
1282 my @leftresult = split /;/, $biblionumbers;
1284 foreach my $entry (@leftresult) { # $_ contains biblionumber,title-weight
1285 # remove weight at the end
1286 my $cleaned = $entry;
1287 $cleaned =~ s/-\d*$//;
1288 # if the entry already in the hash, take it & increase weight
1289 # warn "===== $cleaned =====";
1290 if ($results =~ "$cleaned") {
1291 $temp .= "$entry;$entry;";
1292 # warn "INCLUDING $entry";
1297 $results = $biblionumbers;
1301 #do a complete search (all indexes)
1302 my $dbh = C4::Context->dbh;
1303 my $sth = $dbh->prepare("SELECT biblionumbers FROM nozebra WHERE server=? AND value LIKE ?");
1304 # split each word, query the DB and build the biblionumbers result
1305 foreach (split / /,$string) {
1306 next if C4::Context->stopwords->{uc($_)}; # skip if stopword
1307 #warn "search on all indexes on $_";
1310 $sth->execute($server, $_);
1311 while (my $line = $sth->fetchrow) {
1312 $biblionumbers .= $line;
1314 # do a AND with existing list if there is one, otherwise, use the biblionumbers list as 1st result list
1316 # warn "RES for $_ = $biblionumbers";
1317 my @leftresult = split /;/, $biblionumbers;
1319 foreach my $entry (@leftresult) { # $_ contains biblionumber,title-weight
1320 # remove weight at the end
1321 my $cleaned = $entry;
1322 $cleaned =~ s/-\d*$//;
1323 # if the entry already in the hash, take it & increase weight
1324 # warn "===== $cleaned =====";
1325 if ($results =~ "$cleaned") {
1326 $temp .= "$entry;$entry;";
1327 # warn "INCLUDING $entry";
1332 # warn "NEW RES for $_ = $biblionumbers";
1333 $results = $biblionumbers;
1337 # warn "return : $results for LEAF : $string";
1344 $finalresult = NZorder($biblionumbers, $ordering,$results_per_page,$offset);
1352 my ($biblionumbers, $ordering,$results_per_page,$offset) = @_;
1353 # order title asc by default
1354 # $ordering = '1=36 <i' unless $ordering;
1355 $results_per_page=20 unless $results_per_page;
1356 $offset = 0 unless $offset;
1357 my $dbh = C4::Context->dbh;
1359 # order by POPULARITY
1361 if ($ordering =~ /popularity/) {
1364 # popularity is not in MARC record, it's builded from a specific query
1365 my $sth = $dbh->prepare("select sum(issues) from items where biblionumber=?");
1366 foreach (split /;/,$biblionumbers) {
1367 my ($biblionumber,$title) = split /,/,$_;
1368 $result{$biblionumber}=GetMarcBiblio($biblionumber);
1369 $sth->execute($biblionumber);
1370 my $popularity= $sth->fetchrow ||0;
1371 # hint : the key is popularity.title because we can have
1372 # many results with the same popularity. In this cas, sub-ordering is done by title
1373 # we also have biblionumber to avoid bug for 2 biblios with the same title & popularity
1374 # (un-frequent, I agree, but we won't forget anything that way ;-)
1375 $popularity{sprintf("%10d",$popularity).$title.$biblionumber} = $biblionumber;
1377 # sort the hash and return the same structure as GetRecords (Zebra querying)
1380 if ($ordering eq 'popularity_dsc') { # sort popularity DESC
1381 foreach my $key (sort {$b cmp $a} (keys %popularity)) {
1382 $result_hash->{'RECORDS'}[$numbers++] = $result{$popularity{$key}}->as_usmarc();
1384 } else { # sort popularity ASC
1385 foreach my $key (sort (keys %popularity)) {
1386 $result_hash->{'RECORDS'}[$numbers++] = $result{$popularity{$key}}->as_usmarc();
1390 $result_hash->{'hits'} = $numbers;
1391 $finalresult->{'biblioserver'} = $result_hash;
1392 return $finalresult;
1396 } elsif ($ordering =~/author/){
1398 foreach (split /;/,$biblionumbers) {
1399 my ($biblionumber,$title) = split /,/,$_;
1400 my $record=GetMarcBiblio($biblionumber);
1402 if (C4::Context->preference('marcflavour') eq 'UNIMARC') {
1403 $author=$record->subfield('200','f');
1404 $author=$record->subfield('700','a') unless $author;
1406 $author=$record->subfield('100','a');
1408 # hint : the result is sorted by title.biblionumber because we can have X biblios with the same title
1409 # and we don't want to get only 1 result for each of them !!!
1410 $result{$author.$biblionumber}=$record;
1412 # sort the hash and return the same structure as GetRecords (Zebra querying)
1415 if ($ordering eq 'author_za') { # sort by author desc
1416 foreach my $key (sort (keys %result)) {
1417 $result_hash->{'RECORDS'}[$numbers++] = $result{$key}->as_usmarc();
1419 } else { # sort by author ASC
1420 foreach my $key (sort { $a cmp $b } (keys %result)) {
1421 $result_hash->{'RECORDS'}[$numbers++] = $result{$key}->as_usmarc();
1425 $result_hash->{'hits'} = $numbers;
1426 $finalresult->{'biblioserver'} = $result_hash;
1427 return $finalresult;
1429 # ORDER BY callnumber
1431 } elsif ($ordering =~/callnumber/){
1433 foreach (split /;/,$biblionumbers) {
1434 my ($biblionumber,$title) = split /,/,$_;
1435 my $record=GetMarcBiblio($biblionumber);
1437 my ($callnumber_tag,$callnumber_subfield)=GetMarcFromKohaField($dbh,'items.itemcallnumber');
1438 ($callnumber_tag,$callnumber_subfield)= GetMarcFromKohaField('biblioitems.callnumber') unless $callnumber_tag;
1439 if (C4::Context->preference('marcflavour') eq 'UNIMARC') {
1440 $callnumber=$record->subfield('200','f');
1442 $callnumber=$record->subfield('100','a');
1444 # hint : the result is sorted by title.biblionumber because we can have X biblios with the same title
1445 # and we don't want to get only 1 result for each of them !!!
1446 $result{$callnumber.$biblionumber}=$record;
1448 # sort the hash and return the same structure as GetRecords (Zebra querying)
1451 if ($ordering eq 'call_number_dsc') { # sort by title desc
1452 foreach my $key (sort (keys %result)) {
1453 $result_hash->{'RECORDS'}[$numbers++] = $result{$key}->as_usmarc();
1455 } else { # sort by title ASC
1456 foreach my $key (sort { $a cmp $b } (keys %result)) {
1457 $result_hash->{'RECORDS'}[$numbers++] = $result{$key}->as_usmarc();
1461 $result_hash->{'hits'} = $numbers;
1462 $finalresult->{'biblioserver'} = $result_hash;
1463 return $finalresult;
1464 } elsif ($ordering =~ /pubdate/){ #pub year
1466 foreach (split /;/,$biblionumbers) {
1467 my ($biblionumber,$title) = split /,/,$_;
1468 my $record=GetMarcBiblio($biblionumber);
1469 my ($publicationyear_tag,$publicationyear_subfield)=GetMarcFromKohaField($dbh,'biblioitems.publicationyear');
1470 my $publicationyear=$record->subfield($publicationyear_tag,$publicationyear_subfield);
1471 # hint : the result is sorted by title.biblionumber because we can have X biblios with the same title
1472 # and we don't want to get only 1 result for each of them !!!
1473 $result{$publicationyear.$biblionumber}=$record;
1475 # sort the hash and return the same structure as GetRecords (Zebra querying)
1478 if ($ordering eq 'pubdate_dsc') { # sort by pubyear desc
1479 foreach my $key (sort (keys %result)) {
1480 $result_hash->{'RECORDS'}[$numbers++] = $result{$key}->as_usmarc();
1482 } else { # sort by pub year ASC
1483 foreach my $key (sort { $b cmp $a } (keys %result)) {
1484 $result_hash->{'RECORDS'}[$numbers++] = $result{$key}->as_usmarc();
1488 $result_hash->{'hits'} = $numbers;
1489 $finalresult->{'biblioserver'} = $result_hash;
1490 return $finalresult;
1494 } elsif ($ordering =~ /title/) {
1495 # the title is in the biblionumbers string, so we just need to build a hash, sort it and return
1497 foreach (split /;/,$biblionumbers) {
1498 my ($biblionumber,$title) = split /,/,$_;
1499 # hint : the result is sorted by title.biblionumber because we can have X biblios with the same title
1500 # and we don't want to get only 1 result for each of them !!!
1501 # hint & speed improvement : we can order without reading the record
1502 # so order, and read records only for the requested page !
1503 $result{$title.$biblionumber}=$biblionumber;
1505 # sort the hash and return the same structure as GetRecords (Zebra querying)
1508 if ($ordering eq 'title_za') { # sort by title desc
1509 foreach my $key (sort (keys %result)) {
1510 $result_hash->{'RECORDS'}[$numbers++] = $result{$key};
1512 } else { # sort by title ASC
1513 foreach my $key (sort { $b cmp $a } (keys %result)) {
1514 $result_hash->{'RECORDS'}[$numbers++] = $result{$key};
1517 # limit the $results_per_page to result size if it's more
1518 $results_per_page = $numbers-1 if $numbers < $results_per_page;
1519 # for the requested page, replace biblionumber by the complete record
1520 # speed improvement : avoid reading too much things
1521 for (my $counter=$offset;$counter<=$offset+$results_per_page;$counter++) {
1522 $result_hash->{'RECORDS'}[$counter] = GetMarcBiblio($result_hash->{'RECORDS'}[$counter])->as_usmarc;
1525 $result_hash->{'hits'} = $numbers;
1526 $finalresult->{'biblioserver'} = $result_hash;
1527 return $finalresult;
1532 # we need 2 hashes to order by ranking : the 1st one to count the ranking, the 2nd to order by ranking
1535 foreach (split /;/,$biblionumbers) {
1536 my ($biblionumber,$title) = split /,/,$_;
1537 $title =~ /(.*)-(\d)/;
1540 # note that we + the ranking because ranking is calculated on weight of EACH term requested.
1541 # if we ask for "two towers", and "two" has weight 2 in biblio N, and "towers" has weight 4 in biblio N
1542 # biblio N has ranking = 6
1543 $count_ranking{$biblionumber} += $ranking;
1545 # build the result by "inverting" the count_ranking hash
1546 # 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
1548 foreach (keys %count_ranking) {
1549 $result{sprintf("%10d",$count_ranking{$_}).'-'.$_} = $_;
1551 # sort the hash and return the same structure as GetRecords (Zebra querying)
1554 foreach my $key (sort {$b cmp $a} (keys %result)) {
1555 $result_hash->{'RECORDS'}[$numbers++] = $result{$key};
1557 # limit the $results_per_page to result size if it's more
1558 $results_per_page = $numbers-1 if $numbers < $results_per_page;
1559 # for the requested page, replace biblionumber by the complete record
1560 # speed improvement : avoid reading too much things
1561 for (my $counter=$offset;$counter<=$offset+$results_per_page;$counter++) {
1562 $result_hash->{'RECORDS'}[$counter] = GetMarcBiblio($result_hash->{'RECORDS'}[$counter])->as_usmarc;
1565 $result_hash->{'hits'} = $numbers;
1566 $finalresult->{'biblioserver'} = $result_hash;
1567 return $finalresult;
1572 ($countchanged,$listunchanged) = ModBiblios($listbiblios, $tagsubfield,$initvalue,$targetvalue,$test);
1574 this function changes all the values $initvalue in subfield $tag$subfield in any record in $listbiblios
1575 test parameter if set donot perform change to records in database.
1581 * $listbiblios is an array ref to marcrecords to be changed
1582 * $tagsubfield is the reference of the subfield to change.
1583 * $initvalue is the value to search the record for
1584 * $targetvalue is the value to set the subfield to
1585 * $test is to be set only not to perform changes in database.
1587 =item C<Output arg:>
1588 * $countchanged counts all the changes performed.
1589 * $listunchanged contains the list of all the biblionumbers of records unchanged.
1591 =item C<usage in the script:>
1595 my ($countchanged, $listunchanged) = EditBiblios($results->{RECORD}, $tagsubfield,$initvalue,$targetvalue);;
1596 #If one wants to display unchanged records, you should get biblios foreach @$listunchanged
1597 $template->param(countchanged => $countchanged, loopunchanged=>$listunchanged);
1602 my ($listbiblios,$tagsubfield,$initvalue,$targetvalue,$test)=@_;
1605 my ($tag,$subfield)=($1,$2) if ($tagsubfield=~/^(\d{1,3})([a-z0-9A-Z@])?$/);
1606 if ((length($tag)<3)&& $subfield=~/0-9/){
1607 $tag=$tag.$subfield;
1610 my ($bntag,$bnsubf) = GetMarcFromKohaField('biblio.biblionumber');
1611 my ($itemtag,$itemsubf) = GetMarcFromKohaField('items.itemnumber');
1612 foreach my $usmarc (@$listbiblios){
1614 $record=eval{MARC::Record->new_from_usmarc($usmarc)};
1617 # usmarc is not a valid usmarc May be a biblionumber
1618 if ($tag eq $itemtag){
1619 my $bib=GetBiblioFromItemNumber($usmarc);
1620 $record=GetMarcItem($bib->{'biblionumber'},$usmarc) ;
1621 $biblionumber=$bib->{'biblionumber'};
1623 $record=GetMarcBiblio($usmarc);
1624 $biblionumber=$usmarc;
1628 $biblionumber = $record->subfield($bntag,$bnsubf);
1630 $biblionumber=$record->field($bntag)->data;
1633 #GetBiblionumber is to be written.
1634 #Could be replaced by TransformMarcToKoha (But Would be longer)
1635 if ($record->field($tag)){
1637 foreach my $field ($record->field($tag)){
1639 if ($field->delete_subfield('code' =>$subfield,'match'=>qr($initvalue))){
1642 $field->update($subfield,$targetvalue) if ($targetvalue);
1646 if ($field->delete_field($field)){
1651 $field->data=$targetvalue if ($field->data=~qr($initvalue));
1655 # warn $record->as_formatted;
1657 ModBiblio($record,$biblionumber,GetFrameworkCode($biblionumber)) unless ($test);
1659 push @unmatched, $biblionumber;
1662 push @unmatched, $biblionumber;
1665 return ($countmatched,\@unmatched);
1668 END { } # module clean-up code here (global destructor)
1675 Koha Developement team <info@koha.org>