C4/Search.pm

   1 package C4::Search;
   2
   3 # This file is part of Koha.
   4 #
   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
   8 # version.
   9 #
  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.
  13 #
  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
  17
  18 use strict;
  19 require Exporter;
  20 use C4::Context;
  21 use C4::Biblio;    # GetMarcFromKohaField
  22 use C4::Koha;      # getFacets
  23 use Lingua::Stem;
  24 use C4::Dates qw(format_date);
  25
  26 use vars qw($VERSION @ISA @EXPORT @EXPORT_OK %EXPORT_TAGS $DEBUG);
  27
  28 # set the version for version checking
  29 BEGIN {
  30     $VERSION = 3.01;
  31     $DEBUG = ($ENV{DEBUG}) ? 1 : 0;
  32 }
  33
  34 =head1 NAME
  35
  36 C4::Search - Functions for searching the Koha catalog.
  37
  38 =head1 SYNOPSIS
  39
  40 See opac/opac-search.pl or catalogue/search.pl for example of usage
  41
  42 =head1 DESCRIPTION
  43
  44 This module provides searching functions for Koha's bibliographic databases
  45
  46 =head1 FUNCTIONS
  47
  48 =cut
  49
  50 @ISA    = qw(Exporter);
  51 @EXPORT = qw(
  52   &findseealso
  53   &FindDuplicate
  54   &SimpleSearch
  55   &searchResults
  56   &getRecords
  57   &buildQuery
  58   &NZgetRecords
  59   &ModBiblios
  60 );
  61
  62 # make all your functions, whether exported or not;
  63
  64 =head2 findseealso($dbh,$fields);
  65
  66 C<$dbh> is a link to the DB handler.
  67
  68 use C4::Context;
  69 my $dbh =C4::Context->dbh;
  70
  71 C<$fields> is a reference to the fields array
  72
  73 This function modifies the @$fields array and adds related fields to search on.
  74
  75 FIXME: this function is probably deprecated in Koha 3
  76
  77 =cut
  78
  79 sub findseealso {
  80     my ( $dbh, $fields ) = @_;
  81     my $tagslib = GetMarcStructure(1);
  82     for ( my $i = 0 ; $i <= $#{$fields} ; $i++ ) {
  83         my ($tag)      = substr( @$fields[$i], 1, 3 );
  84         my ($subfield) = substr( @$fields[$i], 4, 1 );
  85         @$fields[$i] .= ',' . $tagslib->{$tag}->{$subfield}->{seealso}
  86           if ( $tagslib->{$tag}->{$subfield}->{seealso} );
  87     }
  88 }
  89
  90 =head2 FindDuplicate
  91
  92 ($biblionumber,$biblionumber,$title) = FindDuplicate($record);
  93
  94 This function attempts to find duplicate records using a hard-coded, fairly simplistic algorithm
  95
  96 =cut
  97
  98 sub FindDuplicate {
  99     my ($record) = @_;
 100     my $dbh = C4::Context->dbh;
 101     my $result = TransformMarcToKoha( $dbh, $record, '' );
 102     my $sth;
 103     my $query;
 104     my $search;
 105     my $type;
 106     my ( $biblionumber, $title );
 107
 108     # search duplicate on ISBN, easy and fast..
 109     # ... normalize first
 110     if ( $result->{isbn} ) {
 111         $result->{isbn} =~ s/\(.*$//;
 112         $result->{isbn} =~ s/\s+$//;
 113         $query = "isbn=$result->{isbn}";
 114     }
 115     else {
 116         $result->{title} =~ s /\\//g;
 117         $result->{title} =~ s /\"//g;
 118         $result->{title} =~ s /\(//g;
 119         $result->{title} =~ s /\)//g;
 120
 121         # FIXME: instead of removing operators, could just do
 122         # quotes around the value
 123         $result->{title} =~ s/(and|or|not)//g;
 124         $query = "ti,ext=$result->{title}";
 125         $query .= " and itemtype=$result->{itemtype}"
 126           if ( $result->{itemtype} );
 127         if   ( $result->{author} ) {
 128             $result->{author} =~ s /\\//g;
 129             $result->{author} =~ s /\"//g;
 130             $result->{author} =~ s /\(//g;
 131             $result->{author} =~ s /\)//g;
 132
 133             # remove valid operators
 134             $result->{author} =~ s/(and|or|not)//g;
 135             $query .= " and au,ext=$result->{author}";
 136         }
 137     }
 138
 139     # FIXME: add error handling
 140     my ( $error, $searchresults ) = SimpleSearch($query); # FIXME :: hardcoded !
 141     my @results;
 142     foreach my $possible_duplicate_record (@$searchresults) {
 143         my $marcrecord =
 144           MARC::Record->new_from_usmarc($possible_duplicate_record);
 145         my $result = TransformMarcToKoha( $dbh, $marcrecord, '' );
 146
 147         # FIXME :: why 2 $biblionumber ?
 148         if ($result) {
 149             push @results, $result->{'biblionumber'};
 150             push @results, $result->{'title'};
 151         }
 152     }
 153     return @results;
 154 }
 155
 156 =head2 SimpleSearch
 157
 158 ($error,$results) = SimpleSearch($query,@servers);
 159
 160 This function provides a simple search API on the bibliographic catalog
 161
 162 =over 2
 163
 164 =item C<input arg:>
 165
 166     * $query can be a simple keyword or a complete CCL query
 167     * @servers is optional. Defaults to biblioserver as found in koha-conf.xml
 168
 169 =item C<Output arg:>
 170     * $error is a empty unless an error is detected
 171     * \@results is an array of records.
 172
 173 =item C<usage in the script:>
 174
 175 =back
 176
 177 my ($error, $marcresults) = SimpleSearch($query);
 178
 179 if (defined $error) {
 180     $template->param(query_error => $error);
 181     warn "error: ".$error;
 182     output_html_with_http_headers $input, $cookie, $template->output;
 183     exit;
 184 }
 185
 186 my $hits = scalar @$marcresults;
 187 my @results;
 188
 189 for(my $i=0;$i<$hits;$i++) {
 190     my %resultsloop;
 191     my $marcrecord = MARC::File::USMARC::decode($marcresults->[$i]);
 192     my $biblio = TransformMarcToKoha(C4::Context->dbh,$marcrecord,'');
 193
 194     #build the hash for the template.
 195     $resultsloop{highlight}       = ($i % 2)?(1):(0);
 196     $resultsloop{title}           = $biblio->{'title'};
 197     $resultsloop{subtitle}        = $biblio->{'subtitle'};
 198     $resultsloop{biblionumber}    = $biblio->{'biblionumber'};
 199     $resultsloop{author}          = $biblio->{'author'};
 200     $resultsloop{publishercode}   = $biblio->{'publishercode'};
 201     $resultsloop{publicationyear} = $biblio->{'publicationyear'};
 202
 203     push @results, \%resultsloop;
 204 }
 205
 206 $template->param(result=>\@results);
 207
 208 =cut
 209
 210 sub SimpleSearch {
 211     my $query = shift;
 212     if ( C4::Context->preference('NoZebra') ) {
 213         my $result = NZorder( NZanalyse($query) )->{'biblioserver'};
 214         my $search_result =
 215           (      $result->{hits}
 216               && $result->{hits} > 0 ? $result->{'RECORDS'} : [] );
 217         return ( undef, $search_result );
 218     }
 219     else {
 220         my @servers = @_;
 221         my @results;
 222         my @tmpresults;
 223         my @zconns;
 224         return ( "No query entered", undef ) unless $query;
 225
 226         # FIXME hardcoded value. See catalog/search.pl & opac-search.pl too.
 227         @servers = ("biblioserver") unless @servers;
 228
 229         # Initialize & Search Zebra
 230         for ( my $i = 0 ; $i < @servers ; $i++ ) {
 231             eval {
 232                 $zconns[$i] = C4::Context->Zconn( $servers[$i], 1 );
 233                 $tmpresults[$i] =
 234                   $zconns[$i]
 235                   ->search( new ZOOM::Query::CCL2RPN( $query, $zconns[$i] ) );
 236
 237                 # error handling
 238                 my $error =
 239                     $zconns[$i]->errmsg() . " ("
 240                   . $zconns[$i]->errcode() . ") "
 241                   . $zconns[$i]->addinfo() . " "
 242                   . $zconns[$i]->diagset();
 243
 244                 return ( $error, undef ) if $zconns[$i]->errcode();
 245             };
 246             if ($@) {
 247
 248                 # caught a ZOOM::Exception
 249                 my $error =
 250                     $@->message() . " ("
 251                   . $@->code() . ") "
 252                   . $@->addinfo() . " "
 253                   . $@->diagset();
 254                 warn $error;
 255                 return ( $error, undef );
 256             }
 257         }
 258         my $hits;
 259         my $ev;
 260         while ( ( my $i = ZOOM::event( \@zconns ) ) != 0 ) {
 261             $ev = $zconns[ $i - 1 ]->last_event();
 262             if ( $ev == ZOOM::Event::ZEND ) {
 263                 $hits = $tmpresults[ $i - 1 ]->size();
 264             }
 265             if ( $hits > 0 ) {
 266                 for ( my $j = 0 ; $j < $hits ; $j++ ) {
 267                     my $record = $tmpresults[ $i - 1 ]->record($j)->raw();
 268                     push @results, $record;
 269                 }
 270             }
 271         }
 272         return ( undef, \@results );
 273     }
 274 }
 275
 276 =head2 getRecords
 277
 278 ( undef, $results_hashref, \@facets_loop ) = getRecords (
 279
 280         $koha_query,       $simple_query, $sort_by_ref,    $servers_ref,
 281         $results_per_page, $offset,       $expanded_facet, $branches,
 282         $query_type,       $scan
 283     );
 284
 285 The all singing, all dancing, multi-server, asynchronous, scanning,
 286 searching, record nabbing, facet-building
 287
 288 See verbse embedded documentation.
 289
 290 =cut
 291
 292 sub getRecords {
 293     my (
 294         $koha_query,       $simple_query, $sort_by_ref,    $servers_ref,
 295         $results_per_page, $offset,       $expanded_facet, $branches,
 296         $query_type,       $scan
 297     ) = @_;
 298
 299     my @servers = @$servers_ref;
 300     my @sort_by = @$sort_by_ref;
 301
 302     # Initialize variables for the ZOOM connection and results object
 303     my $zconn;
 304     my @zconns;
 305     my @results;
 306     my $results_hashref = ();
 307
 308     # Initialize variables for the faceted results objects
 309     my $facets_counter = ();
 310     my $facets_info    = ();
 311     my $facets         = getFacets();
 312
 313     my @facets_loop
 314       ;    # stores the ref to array of hashes for template facets loop
 315
 316     ### LOOP THROUGH THE SERVERS
 317     for ( my $i = 0 ; $i < @servers ; $i++ ) {
 318         $zconns[$i] = C4::Context->Zconn( $servers[$i], 1 );
 319
 320 # perform the search, create the results objects
 321 # if this is a local search, use the $koha-query, if it's a federated one, use the federated-query
 322         my $query_to_use;
 323         if ( $servers[$i] =~ /biblioserver/ ) {
 324             $query_to_use = $koha_query;
 325         }
 326         else {
 327             $query_to_use = $simple_query;
 328         }
 329
 330         #$query_to_use = $simple_query if $scan;
 331         warn $simple_query if ( $scan and $DEBUG );
 332
 333         # Check if we've got a query_type defined, if so, use it
 334         eval {
 335             if ($query_type)
 336             {
 337                 if ( $query_type =~ /^ccl/ ) {
 338                     $query_to_use =~
 339                       s/\:/\=/g;    # change : to = last minute (FIXME)
 340                     $results[$i] =
 341                       $zconns[$i]->search(
 342                         new ZOOM::Query::CCL2RPN( $query_to_use, $zconns[$i] )
 343                       );
 344                 }
 345                 elsif ( $query_type =~ /^cql/ ) {
 346                     $results[$i] =
 347                       $zconns[$i]->search(
 348                         new ZOOM::Query::CQL( $query_to_use, $zconns[$i] ) );
 349                 }
 350                 elsif ( $query_type =~ /^pqf/ ) {
 351                     $results[$i] =
 352                       $zconns[$i]->search(
 353                         new ZOOM::Query::PQF( $query_to_use, $zconns[$i] ) );
 354                 }
 355             }
 356             else {
 357                 if ($scan) {
 358                     $results[$i] =
 359                       $zconns[$i]->scan(
 360                         new ZOOM::Query::CCL2RPN( $query_to_use, $zconns[$i] )
 361                       );
 362                 }
 363                 else {
 364                     $results[$i] =
 365                       $zconns[$i]->search(
 366                         new ZOOM::Query::CCL2RPN( $query_to_use, $zconns[$i] )
 367                       );
 368                 }
 369             }
 370         };
 371         if ($@) {
 372             warn "WARNING: query problem with $query_to_use " . $@;
 373         }
 374
 375         # Concatenate the sort_by limits and pass them to the results object
 376         # Note: sort will override rank
 377         my $sort_by;
 378         foreach my $sort (@sort_by) {
 379             if ( $sort eq "author_az" ) {
 380                 $sort_by .= "1=1003 <i ";
 381             }
 382             elsif ( $sort eq "author_za" ) {
 383                 $sort_by .= "1=1003 >i ";
 384             }
 385             elsif ( $sort eq "popularity_asc" ) {
 386                 $sort_by .= "1=9003 <i ";
 387             }
 388             elsif ( $sort eq "popularity_dsc" ) {
 389                 $sort_by .= "1=9003 >i ";
 390             }
 391             elsif ( $sort eq "call_number_asc" ) {
 392                 $sort_by .= "1=20  <i ";
 393             }
 394             elsif ( $sort eq "call_number_dsc" ) {
 395                 $sort_by .= "1=20 >i ";
 396             }
 397             elsif ( $sort eq "pubdate_asc" ) {
 398                 $sort_by .= "1=31 <i ";
 399             }
 400             elsif ( $sort eq "pubdate_dsc" ) {
 401                 $sort_by .= "1=31 >i ";
 402             }
 403             elsif ( $sort eq "acqdate_asc" ) {
 404                 $sort_by .= "1=32 <i ";
 405             }
 406             elsif ( $sort eq "acqdate_dsc" ) {
 407                 $sort_by .= "1=32 >i ";
 408             }
 409             elsif ( $sort eq "title_az" ) {
 410                 $sort_by .= "1=4 <i ";
 411             }
 412             elsif ( $sort eq "title_za" ) {
 413                 $sort_by .= "1=4 >i ";
 414             }
 415         }
 416         if ($sort_by) {
 417             if ( $results[$i]->sort( "yaz", $sort_by ) < 0 ) {
 418                 warn "WARNING sort $sort_by failed";
 419             }
 420         }
 421     }    # finished looping through servers
 422
 423     # The big moment: asynchronously retrieve results from all servers
 424     while ( ( my $i = ZOOM::event( \@zconns ) ) != 0 ) {
 425         my $ev = $zconns[ $i - 1 ]->last_event();
 426         if ( $ev == ZOOM::Event::ZEND ) {
 427             next unless $results[ $i - 1 ];
 428             my $size = $results[ $i - 1 ]->size();
 429             if ( $size > 0 ) {
 430                 my $results_hash;
 431
 432                 # loop through the results
 433                 $results_hash->{'hits'} = $size;
 434                 my $times;
 435                 if ( $offset + $results_per_page <= $size ) {
 436                     $times = $offset + $results_per_page;
 437                 }
 438                 else {
 439                     $times = $size;
 440                 }
 441                 for ( my $j = $offset ; $j < $times ; $j++ ) {
 442                     my $records_hash;
 443                     my $record;
 444                     my $facet_record;
 445
 446                     ## Check if it's an index scan
 447                     if ($scan) {
 448                         my ( $term, $occ ) = $results[ $i - 1 ]->term($j);
 449
 450                  # here we create a minimal MARC record and hand it off to the
 451                  # template just like a normal result ... perhaps not ideal, but
 452                  # it works for now
 453                         my $tmprecord = MARC::Record->new();
 454                         $tmprecord->encoding('UTF-8');
 455                         my $tmptitle;
 456                         my $tmpauthor;
 457
 458                 # the minimal record in author/title (depending on MARC flavour)
 459                         if ( C4::Context->preference("marcflavour") eq
 460                             "UNIMARC" )
 461                         {
 462                             $tmptitle = MARC::Field->new(
 463                                 '200', ' ', ' ',
 464                                 a => $term,
 465                                 f => $occ
 466                             );
 467                         }
 468                         else {
 469                             $tmptitle =
 470                               MARC::Field->new( '245', ' ', ' ', a => $term, );
 471                             $tmpauthor =
 472                               MARC::Field->new( '100', ' ', ' ', a => $occ, );
 473                         }
 474                         $tmprecord->append_fields($tmptitle);
 475                         $tmprecord->append_fields($tmpauthor);
 476                         $results_hash->{'RECORDS'}[$j] =
 477                           $tmprecord->as_usmarc();
 478                     }
 479
 480                     # not an index scan
 481                     else {
 482                         $record = $results[ $i - 1 ]->record($j)->raw();
 483
 484                         # warn "RECORD $j:".$record;
 485                         $results_hash->{'RECORDS'}[$j] = $record;
 486
 487             # Fill the facets while we're looping, but only for the biblioserver
 488                         $facet_record = MARC::Record->new_from_usmarc($record)
 489                           if $servers[ $i - 1 ] =~ /biblioserver/;
 490
 491                     #warn $servers[$i-1]."\n".$record; #.$facet_record->title();
 492                         if ($facet_record) {
 493                             for ( my $k = 0 ; $k <= @$facets ; $k++ ) {
 494
 495                                 if ( $facets->[$k] ) {
 496                                     my @fields;
 497                                     for my $tag ( @{ $facets->[$k]->{'tags'} } )
 498                                     {
 499                                         push @fields,
 500                                           $facet_record->field($tag);
 501                                     }
 502                                     for my $field (@fields) {
 503                                         my @subfields = $field->subfields();
 504                                         for my $subfield (@subfields) {
 505                                             my ( $code, $data ) = @$subfield;
 506                                             if ( $code eq
 507                                                 $facets->[$k]->{'subfield'} )
 508                                             {
 509                                                 $facets_counter->{ $facets->[$k]
 510                                                       ->{'link_value'} }
 511                                                   ->{$data}++;
 512                                             }
 513                                         }
 514                                     }
 515                                     $facets_info->{ $facets->[$k]
 516                                           ->{'link_value'} }->{'label_value'} =
 517                                       $facets->[$k]->{'label_value'};
 518                                     $facets_info->{ $facets->[$k]
 519                                           ->{'link_value'} }->{'expanded'} =
 520                                       $facets->[$k]->{'expanded'};
 521                                 }
 522                             }
 523                         }
 524                     }
 525                 }
 526                 $results_hashref->{ $servers[ $i - 1 ] } = $results_hash;
 527             }
 528
 529             # warn "connection ", $i-1, ": $size hits";
 530             # warn $results[$i-1]->record(0)->render() if $size > 0;
 531
 532             # BUILD FACETS
 533             if ( $servers[ $i - 1 ] =~ /biblioserver/ ) {
 534                 for my $link_value (
 535                     sort { $facets_counter->{$b} <=> $facets_counter->{$a} }
 536                     keys %$facets_counter )
 537                 {
 538                     my $expandable;
 539                     my $number_of_facets;
 540                     my @this_facets_array;
 541                     for my $one_facet (
 542                         sort {
 543                             $facets_counter->{$link_value}
 544                               ->{$b} <=> $facets_counter->{$link_value}->{$a}
 545                         } keys %{ $facets_counter->{$link_value} }
 546                       )
 547                     {
 548                         $number_of_facets++;
 549                         if (   ( $number_of_facets < 6 )
 550                             || ( $expanded_facet eq $link_value )
 551                             || ( $facets_info->{$link_value}->{'expanded'} ) )
 552                         {
 553
 554                       # Sanitize the link value ), ( will cause errors with CCL,
 555                             my $facet_link_value = $one_facet;
 556                             $facet_link_value =~ s/(\(|\))/ /g;
 557
 558                             # fix the length that will display in the label,
 559                             my $facet_label_value = $one_facet;
 560                             $facet_label_value =
 561                               substr( $one_facet, 0, 20 ) . "..."
 562                               unless length($facet_label_value) <= 20;
 563
 564                             # if it's a branch, label by the name, not the code,
 565                             if ( $link_value =~ /branch/ ) {
 566                                 $facet_label_value =
 567                                   $branches->{$one_facet}->{'branchname'};
 568                             }
 569
 570                 # but we're down with the whole label being in the link's title.
 571                             my $facet_title_value = $one_facet;
 572
 573                             push @this_facets_array,
 574                               (
 575                                 {
 576                                     facet_count =>
 577                                       $facets_counter->{$link_value}
 578                                       ->{$one_facet},
 579                                     facet_label_value => $facet_label_value,
 580                                     facet_title_value => $facet_title_value,
 581                                     facet_link_value  => $facet_link_value,
 582                                     type_link_value   => $link_value,
 583                                 },
 584                               );
 585                         }
 586                     }
 587
 588                     # handle expanded option
 589                     unless ( $facets_info->{$link_value}->{'expanded'} ) {
 590                         $expandable = 1
 591                           if ( ( $number_of_facets > 6 )
 592                             && ( $expanded_facet ne $link_value ) );
 593                     }
 594                     push @facets_loop,
 595                       (
 596                         {
 597                             type_link_value => $link_value,
 598                             type_id         => $link_value . "_id",
 599                             type_label =>
 600                               $facets_info->{$link_value}->{'label_value'},
 601                             facets     => \@this_facets_array,
 602                             expandable => $expandable,
 603                             expand     => $link_value,
 604                         }
 605                       );
 606                 }
 607             }
 608         }
 609     }
 610     return ( undef, $results_hashref, \@facets_loop );
 611 }
 612
 613 # STOPWORDS
 614 sub _remove_stopwords {
 615     my ( $operand, $index ) = @_;
 616     my @stopwords_removed;
 617
 618     # phrase and exact-qualified indexes shouldn't have stopwords removed
 619     if ( $index !~ m/phr|ext/ ) {
 620
 621 # remove stopwords from operand : parse all stopwords & remove them (case insensitive)
 622 #       we use IsAlpha unicode definition, to deal correctly with diacritics.
 623 #       otherwise, a French word like "leçon" woudl be split into "le" "çon", "le"
 624 #       is a stopword, we'd get "çon" and wouldn't find anything...
 625         foreach ( keys %{ C4::Context->stopwords } ) {
 626             next if ( $_ =~ /(and|or|not)/ );    # don't remove operators
 627             if ( $operand =~
 628                 /(\P{IsAlpha}$_\P{IsAlpha}|^$_\P{IsAlpha}|\P{IsAlpha}$_$)/ )
 629             {
 630                 $operand =~ s/\P{IsAlpha}$_\P{IsAlpha}/ /gi;
 631                 $operand =~ s/^$_\P{IsAlpha}/ /gi;
 632                 $operand =~ s/\P{IsAlpha}$_$/ /gi;
 633                 push @stopwords_removed, $_;
 634             }
 635         }
 636     }
 637     return ( $operand, \@stopwords_removed );
 638 }
 639
 640 # TRUNCATION
 641 sub _detect_truncation {
 642     my ( $operand, $index ) = @_;
 643     my ( @nontruncated, @righttruncated, @lefttruncated, @rightlefttruncated,
 644         @regexpr );
 645     $operand =~ s/^ //g;
 646     my @wordlist = split( /\s/, $operand );
 647     foreach my $word (@wordlist) {
 648         if ( $word =~ s/^\*([^\*]+)\*$/$1/ ) {
 649             push @rightlefttruncated, $word;
 650         }
 651         elsif ( $word =~ s/^\*([^\*]+)$/$1/ ) {
 652             push @lefttruncated, $word;
 653         }
 654         elsif ( $word =~ s/^([^\*]+)\*$/$1/ ) {
 655             push @righttruncated, $word;
 656         }
 657         elsif ( index( $word, "*" ) < 0 ) {
 658             push @nontruncated, $word;
 659         }
 660         else {
 661             push @regexpr, $word;
 662         }
 663     }
 664     return (
 665         \@nontruncated,       \@righttruncated, \@lefttruncated,
 666         \@rightlefttruncated, \@regexpr
 667     );
 668 }
 669
 670 # STEMMING
 671 sub _build_stemmed_operand {
 672     my ($operand) = @_;
 673     my $stemmed_operand;
 674
 675 # FIXME: the locale should be set based on the user's language and/or search choice
 676     my $stemmer = Lingua::Stem->new( -locale => 'EN-US' );
 677
 678 # FIXME: these should be stored in the db so the librarian can modify the behavior
 679     $stemmer->add_exceptions(
 680         {
 681             'and' => 'and',
 682             'or'  => 'or',
 683             'not' => 'not',
 684         }
 685     );
 686     my @words = split( / /, $operand );
 687     my $stems = $stemmer->stem(@words);
 688     for my $stem (@$stems) {
 689         $stemmed_operand .= "$stem";
 690         $stemmed_operand .= "?"
 691           unless ( $stem =~ /(and$|or$|not$)/ ) || ( length($stem) < 3 );
 692         $stemmed_operand .= " ";
 693     }
 694     warn "STEMMED OPERAND: $stemmed_operand" if $DEBUG;
 695     return $stemmed_operand;
 696 }
 697
 698 # FIELD WEIGHTING
 699 sub _build_weighted_query {
 700
 701 # FIELD WEIGHTING - This is largely experimental stuff. What I'm committing works
 702 # pretty well but could work much better if we had a smarter query parser
 703     my ( $operand, $stemmed_operand, $index ) = @_;
 704     my $stemming      = C4::Context->preference("QueryStemming")     || 0;
 705     my $weight_fields = C4::Context->preference("QueryWeightFields") || 0;
 706     my $fuzzy_enabled = C4::Context->preference("QueryFuzzy")        || 0;
 707
 708     my $weighted_query .= "(rk=(";    # Specifies that we're applying rank
 709
 710     # Keyword, or, no index specified
 711     if ( ( $index eq 'kw' ) || ( !$index ) ) {
 712         $weighted_query .=
 713           "Title-cover,ext,r1=\"$operand\"";    # exact title-cover
 714         $weighted_query .= " or ti,ext,r2=\"$operand\"";    # exact title
 715         $weighted_query .= " or ti,phr,r3=\"$operand\"";    # phrase title
 716           #$weighted_query .= " or any,ext,r4=$operand";               # exact any
 717           #$weighted_query .=" or kw,wrdl,r5=\"$operand\"";            # word list any
 718         $weighted_query .= " or wrdl,fuzzy,r8=\"$operand\""
 719           if $fuzzy_enabled;    # add fuzzy, word list
 720         $weighted_query .= " or wrdl,right-Truncation,r9=\"$stemmed_operand\""
 721           if ( $stemming and $stemmed_operand )
 722           ;                     # add stemming, right truncation
 723         $weighted_query .= " or wrdl,r9=\"$operand\"";
 724
 725         # embedded sorting: 0 a-z; 1 z-a
 726         # $weighted_query .= ") or (sort1,aut=1";
 727     }
 728
 729     # Barcode searches should skip this process
 730     elsif ( $index eq 'bc' ) {
 731         $weighted_query .= "bc=\"$operand\"";
 732     }
 733
 734     # Authority-number searches should skip this process
 735     elsif ( $index eq 'an' ) {
 736         $weighted_query .= "an=\"$operand\"";
 737     }
 738
 739     # If the index already has more than one qualifier, wrap the operand
 740     # in quotes and pass it back (assumption is that the user knows what they
 741     # are doing and won't appreciate us mucking up their query
 742     elsif ( $index =~ ',' ) {
 743         $weighted_query .= " $index=\"$operand\"";
 744     }
 745
 746     #TODO: build better cases based on specific search indexes
 747     else {
 748         $weighted_query .= " $index,ext,r1=\"$operand\"";    # exact index
 749           #$weighted_query .= " or (title-sort-az=0 or $index,startswithnt,st-word,r3=$operand #)";
 750         $weighted_query .= " or $index,phr,r3=\"$operand\"";    # phrase index
 751         $weighted_query .=
 752           " or $index,rt,wrdl,r3=\"$operand\"";    # word list index
 753     }
 754
 755     $weighted_query .= "))";                       # close rank specification
 756     return $weighted_query;
 757 }
 758
 759 =head2 buildQuery
 760
 761 ( $error, $query,
 762 $simple_query, $query_cgi,
 763 $query_desc, $limit,
 764 $limit_cgi, $limit_desc,
 765 $stopwords_removed, $query_type ) = getRecords ( $operators, $operands, $indexes, $limits, $sort_by, $scan);
 766
 767 Build queries and limits in CCL, CGI, Human,
 768 handle truncation, stemming, field weighting, stopwords, fuzziness, etc.
 769
 770 See verbose embedded documentation.
 771
 772
 773 =cut
 774
 775 sub buildQuery {
 776     my ( $operators, $operands, $indexes, $limits, $sort_by, $scan ) = @_;
 777
 778     warn "---------"        if $DEBUG;
 779     warn "Enter buildQuery" if $DEBUG;
 780     warn "---------"        if $DEBUG;
 781
 782     # dereference
 783     my @operators = @$operators if $operators;
 784     my @indexes   = @$indexes   if $indexes;
 785     my @operands  = @$operands  if $operands;
 786     my @limits    = @$limits    if $limits;
 787     my @sort_by   = @$sort_by   if $sort_by;
 788
 789     my $stemming         = C4::Context->preference("QueryStemming")        || 0;
 790     my $auto_truncation  = C4::Context->preference("QueryAutoTruncate")    || 0;
 791     my $weight_fields    = C4::Context->preference("QueryWeightFields")    || 0;
 792     my $fuzzy_enabled    = C4::Context->preference("QueryFuzzy")           || 0;
 793     my $remove_stopwords = C4::Context->preference("QueryRemoveStopwords") || 0;
 794
 795     # no stemming/weight/fuzzy in NoZebra
 796     if ( C4::Context->preference("NoZebra") ) {
 797         $stemming      = 0;
 798         $weight_fields = 0;
 799         $fuzzy_enabled = 0;
 800     }
 801
 802     my $query        = $operands[0];
 803     my $simple_query = $operands[0];
 804
 805     # initialize the variables we're passing back
 806     my $query_cgi;
 807     my $query_desc;
 808     my $query_type;
 809
 810     my $limit;
 811     my $limit_cgi;
 812     my $limit_desc;
 813
 814     my $stopwords_removed;    # flag to determine if stopwords have been removed
 815
 816 # for handling ccl, cql, pqf queries in diagnostic mode, skip the rest of the steps
 817 # DIAGNOSTIC ONLY!!
 818     if ( $query =~ /^ccl=/ ) {
 819         return ( undef, $', $', $', $', '', '', '', '', 'ccl' );
 820     }
 821     if ( $query =~ /^cql=/ ) {
 822         return ( undef, $', $', $', $', '', '', '', '', 'cql' );
 823     }
 824     if ( $query =~ /^pqf=/ ) {
 825         return ( undef, $', $', $', $', '', '', '', '', 'pqf' );
 826     }
 827
 828     # pass nested queries directly
 829     # FIXME: need better handling of some of these variables in this case
 830     if ( $query =~ /(\(|\))/ ) {
 831         return (
 832             undef,              $query, $simple_query, $query_cgi,
 833             $query,             $limit, $limit_cgi,    $limit_desc,
 834             $stopwords_removed, 'ccl'
 835         );
 836     }
 837
 838 # Form-based queries are non-nested and fixed depth, so we can easily modify the incoming
 839 # query operands and indexes and add stemming, truncation, field weighting, etc.
 840 # Once we do so, we'll end up with a value in $query, just like if we had an
 841 # incoming $query from the user
 842     else {
 843         $query = ""
 844           ; # clear it out so we can populate properly with field-weighted, stemmed, etc. query
 845         my $previous_operand
 846           ;    # a flag used to keep track if there was a previous query
 847                # if there was, we can apply the current operator
 848                # for every operand
 849         for ( my $i = 0 ; $i <= @operands ; $i++ ) {
 850
 851             # COMBINE OPERANDS, INDEXES AND OPERATORS
 852             if ( $operands[$i] ) {
 853
 854               # A flag to determine whether or not to add the index to the query
 855                 my $indexes_set;
 856
 857 # If the user is sophisticated enough to specify an index, turn off field weighting, stemming, and stopword handling
 858                 if ( $operands[$i] =~ /(:|=)/ || $scan ) {
 859                     $weight_fields    = 0;
 860                     $stemming         = 0;
 861                     $remove_stopwords = 0;
 862                 }
 863                 my $operand = $operands[$i];
 864                 my $index   = $indexes[$i];
 865
 866                 # Add index-specific attributes
 867                 # Date of Publication
 868                 if ( $index eq 'yr' ) {
 869                     $index .= ",st-numeric";
 870                     $indexes_set++;
 871                     (
 872                         $stemming,      $auto_truncation,
 873                         $weight_fields, $fuzzy_enabled,
 874                         $remove_stopwords
 875                     ) = ( 0, 0, 0, 0, 0 );
 876                 }
 877
 878                 # Date of Acquisition
 879                 elsif ( $index eq 'acqdate' ) {
 880                     $index .= ",st-date-normalized";
 881                     $indexes_set++;
 882                     (
 883                         $stemming,      $auto_truncation,
 884                         $weight_fields, $fuzzy_enabled,
 885                         $remove_stopwords
 886                     ) = ( 0, 0, 0, 0, 0 );
 887                 }
 888
 889                 # Set default structure attribute (word list)
 890                 my $struct_attr;
 891                 unless ( !$index || $index =~ /(st-|phr|ext|wrdl)/ ) {
 892                     $struct_attr = ",wrdl";
 893                 }
 894
 895                 # Some helpful index variants
 896                 my $index_plus       = $index . $struct_attr . ":" if $index;
 897                 my $index_plus_comma = $index . $struct_attr . "," if $index;
 898
 899                 # Remove Stopwords
 900                 if ($remove_stopwords) {
 901                     ( $operand, $stopwords_removed ) =
 902                       _remove_stopwords( $operand, $index );
 903                     warn "OPERAND w/out STOPWORDS: >$operand<" if $DEBUG;
 904                     warn "REMOVED STOPWORDS: @$stopwords_removed"
 905                       if ( $stopwords_removed && $DEBUG );
 906                 }
 907
 908                 # Detect Truncation
 909                 my ( $nontruncated, $righttruncated, $lefttruncated,
 910                     $rightlefttruncated, $regexpr );
 911                 my $truncated_operand;
 912                 (
 913                     $nontruncated, $righttruncated, $lefttruncated,
 914                     $rightlefttruncated, $regexpr
 915                 ) = _detect_truncation( $operand, $index );
 916                 warn
 917 "TRUNCATION: NON:>@$nontruncated< RIGHT:>@$righttruncated< LEFT:>@$lefttruncated< RIGHTLEFT:>@$rightlefttruncated< REGEX:>@$regexpr<"
 918                   if $DEBUG;
 919
 920                 # Apply Truncation
 921                 if (
 922                     scalar(@$righttruncated) + scalar(@$lefttruncated) +
 923                     scalar(@$rightlefttruncated) > 0 )
 924                 {
 925
 926                # Don't field weight or add the index to the query, we do it here
 927                     $indexes_set = 1;
 928                     undef $weight_fields;
 929                     my $previous_truncation_operand;
 930                     if ( scalar(@$nontruncated) > 0 ) {
 931                         $truncated_operand .= "$index_plus @$nontruncated ";
 932                         $previous_truncation_operand = 1;
 933                     }
 934                     if ( scalar(@$righttruncated) > 0 ) {
 935                         $truncated_operand .= "and "
 936                           if $previous_truncation_operand;
 937                         $truncated_operand .=
 938                           "$index_plus_comma" . "rtrn:@$righttruncated ";
 939                         $previous_truncation_operand = 1;
 940                     }
 941                     if ( scalar(@$lefttruncated) > 0 ) {
 942                         $truncated_operand .= "and "
 943                           if $previous_truncation_operand;
 944                         $truncated_operand .=
 945                           "$index_plus_comma" . "ltrn:@$lefttruncated ";
 946                         $previous_truncation_operand = 1;
 947                     }
 948                     if ( scalar(@$rightlefttruncated) > 0 ) {
 949                         $truncated_operand .= "and "
 950                           if $previous_truncation_operand;
 951                         $truncated_operand .=
 952                           "$index_plus_comma" . "rltrn:@$rightlefttruncated ";
 953                         $previous_truncation_operand = 1;
 954                     }
 955                 }
 956                 $operand = $truncated_operand if $truncated_operand;
 957                 warn "TRUNCATED OPERAND: >$truncated_operand<" if $DEBUG;
 958
 959                 # Handle Stemming
 960                 my $stemmed_operand;
 961                 $stemmed_operand = _build_stemmed_operand($operand)
 962                   if $stemming;
 963                 warn "STEMMED OPERAND: >$stemmed_operand<" if $DEBUG;
 964
 965                 # Handle Field Weighting
 966                 my $weighted_operand;
 967                 $weighted_operand =
 968                   _build_weighted_query( $operand, $stemmed_operand, $index )
 969                   if $weight_fields;
 970                 warn "FIELD WEIGHTED OPERAND: >$weighted_operand<" if $DEBUG;
 971                 $operand = $weighted_operand if $weight_fields;
 972                 $indexes_set = 1 if $weight_fields;
 973
 974                 # If there's a previous operand, we need to add an operator
 975                 if ($previous_operand) {
 976
 977                     # User-specified operator
 978                     if ( $operators[ $i - 1 ] ) {
 979                         $query     .= " $operators[$i-1] ";
 980                         $query     .= " $index_plus " unless $indexes_set;
 981                         $query     .= " $operand";
 982                         $query_cgi .= "&op=$operators[$i-1]";
 983                         $query_cgi .= "&idx=$index" if $index;
 984                         $query_cgi .= "&q=$operands[$i]" if $operands[$i];
 985                         $query_desc .=
 986                           " $operators[$i-1] $index_plus $operands[$i]";
 987                     }
 988
 989                     # Default operator is and
 990                     else {
 991                         $query      .= " and ";
 992                         $query      .= "$index_plus " unless $indexes_set;
 993                         $query      .= "$operand";
 994                         $query_cgi  .= "&op=and&idx=$index" if $index;
 995                         $query_cgi  .= "&q=$operands[$i]" if $operands[$i];
 996                         $query_desc .= " and $index_plus $operands[$i]";
 997                     }
 998                 }
 999
1000                 # There isn't a pervious operand, don't need an operator
1001                 else {
1002
1003                     # Field-weighted queries already have indexes set
1004                     $query .= " $index_plus " unless $indexes_set;
1005                     $query .= $operand;
1006                     $query_desc .= " $index_plus $operands[$i]";
1007                     $query_cgi  .= "&idx=$index" if $index;
1008                     $query_cgi  .= "&q=$operands[$i]" if $operands[$i];
1009                     $previous_operand = 1;
1010                 }
1011             }    #/if $operands
1012         }    # /for
1013     }
1014     warn "QUERY BEFORE LIMITS: >$query<" if $DEBUG;
1015
1016     # add limits
1017     my $group_OR_limits;
1018     my $availability_limit;
1019     foreach my $this_limit (@limits) {
1020         if ( $this_limit =~ /available/ ) {
1021
1022 # 'available' is defined as (items.onloan is NULL) and (items.itemlost = 0)
1023 # In English:
1024 # all records not indexed in the onloan register (zebra) and all records with a value of lost equal to 0
1025             $availability_limit .=
1026 "( ( allrecords,AlwaysMatches='' not onloan,AlwaysMatches='') and (lost,st-numeric=0) )"; #or ( allrecords,AlwaysMatches='' not lost,AlwaysMatches='')) )";
1027             $limit_cgi  .= "&limit=available";
1028             $limit_desc .= "";
1029         }
1030
1031         # group_OR_limits, prefixed by mc-
1032         # OR every member of the group
1033         elsif ( $this_limit =~ /mc/ ) {
1034             $group_OR_limits .= " or " if $group_OR_limits;
1035             $limit_desc      .= " or " if $group_OR_limits;
1036             $group_OR_limits .= "$this_limit";
1037             $limit_cgi       .= "&limit=$this_limit";
1038             $limit_desc      .= " $this_limit";
1039         }
1040
1041         # Regular old limits
1042         else {
1043             $limit .= " and " if $limit || $query;
1044             $limit      .= "$this_limit";
1045             $limit_cgi  .= "&limit=$this_limit";
1046             $limit_desc .= " $this_limit";
1047         }
1048     }
1049     if ($group_OR_limits) {
1050         $limit .= " and " if ( $query || $limit );
1051         $limit .= "($group_OR_limits)";
1052     }
1053     if ($availability_limit) {
1054         $limit .= " and " if ( $query || $limit );
1055         $limit .= "($availability_limit)";
1056     }
1057
1058     # Normalize the query and limit strings
1059     $query =~ s/:/=/g;
1060     $limit =~ s/:/=/g;
1061     for ( $query, $query_desc, $limit, $limit_desc ) {
1062         $_ =~ s/  / /g;    # remove extra spaces
1063         $_ =~ s/^ //g;     # remove any beginning spaces
1064         $_ =~ s/ $//g;     # remove any ending spaces
1065         $_ =~ s/==/=/g;    # remove double == from query
1066     }
1067     $query_cgi =~ s/^&//; # remove unnecessary & from beginning of the query cgi
1068
1069     for ($query_cgi,$simple_query) {
1070         $_ =~ s/"//g;
1071     }
1072     # append the limit to the query
1073     $query .= " " . $limit;
1074
1075     # Warnings if DEBUG
1076     if ($DEBUG) {
1077         warn "QUERY:" . $query;
1078         warn "QUERY CGI:" . $query_cgi;
1079         warn "QUERY DESC:" . $query_desc;
1080         warn "LIMIT:" . $limit;
1081         warn "LIMIT CGI:" . $limit_cgi;
1082         warn "LIMIT DESC:" . $limit_desc;
1083         warn "---------";
1084         warn "Leave buildQuery";
1085         warn "---------";
1086     }
1087     return (
1088         undef,              $query, $simple_query, $query_cgi,
1089         $query_desc,        $limit, $limit_cgi,    $limit_desc,
1090         $stopwords_removed, $query_type
1091     );
1092 }
1093
1094 =head2 searchResults
1095
1096 Format results in a form suitable for passing to the template
1097
1098 =cut
1099
1100 # IMO this subroutine is pretty messy still -- it's responsible for
1101 # building the HTML output for the template
1102 sub searchResults {
1103     my ( $searchdesc, $hits, $results_per_page, $offset, @marcresults ) = @_;
1104     my $dbh = C4::Context->dbh;
1105     my $toggle;
1106     my $even = 1;
1107     my @newresults;
1108
1109     # add search-term highlighting via <span>s on the search terms
1110     my $span_terms_hashref;
1111     for my $span_term ( split( / /, $searchdesc ) ) {
1112         $span_term =~ s/(.*=|\)|\(|\+|\.|\*)//g;
1113         $span_terms_hashref->{$span_term}++;
1114     }
1115
1116     #Build branchnames hash
1117     #find branchname
1118     #get branch information.....
1119     my %branches;
1120     my $bsth =
1121       $dbh->prepare("SELECT branchcode,branchname FROM branches")
1122       ;    # FIXME : use C4::Koha::GetBranches
1123     $bsth->execute();
1124     while ( my $bdata = $bsth->fetchrow_hashref ) {
1125         $branches{ $bdata->{'branchcode'} } = $bdata->{'branchname'};
1126     }
1127     my %locations;
1128     my $lsch =
1129       $dbh->prepare(
1130 "SELECT authorised_value,lib FROM authorised_values WHERE category = 'LOC'"
1131       );
1132     $lsch->execute();
1133     while ( my $ldata = $lsch->fetchrow_hashref ) {
1134         $locations{ $ldata->{'authorised_value'} } = $ldata->{'lib'};
1135     }
1136
1137     #Build itemtype hash
1138     #find itemtype & itemtype image
1139     my %itemtypes;
1140     $bsth =
1141       $dbh->prepare(
1142         "SELECT itemtype,description,imageurl,summary,notforloan FROM itemtypes"
1143       );
1144     $bsth->execute();
1145     while ( my $bdata = $bsth->fetchrow_hashref ) {
1146         $itemtypes{ $bdata->{'itemtype'} }->{description} =
1147           $bdata->{'description'};
1148         $itemtypes{ $bdata->{'itemtype'} }->{imageurl} = $bdata->{'imageurl'};
1149         $itemtypes{ $bdata->{'itemtype'} }->{summary}  = $bdata->{'summary'};
1150         $itemtypes{ $bdata->{'itemtype'} }->{notforloan} =
1151           $bdata->{'notforloan'};
1152     }
1153
1154     #search item field code
1155     my $sth =
1156       $dbh->prepare(
1157 "SELECT tagfield FROM marc_subfield_structure WHERE kohafield LIKE 'items.itemnumber'"
1158       );
1159     $sth->execute;
1160     my ($itemtag) = $sth->fetchrow;
1161
1162     # get notforloan authorised value list
1163     $sth =
1164       $dbh->prepare(
1165 "SELECT authorised_value FROM `marc_subfield_structure` WHERE kohafield = 'items.notforloan' AND frameworkcode=''"
1166       );
1167     $sth->execute;
1168     my ($notforloan_authorised_value) = $sth->fetchrow;
1169
1170     ## find column names of items related to MARC
1171     my $sth2 = $dbh->prepare("SHOW COLUMNS FROM items");
1172     $sth2->execute;
1173     my %subfieldstosearch;
1174     while ( ( my $column ) = $sth2->fetchrow ) {
1175         my ( $tagfield, $tagsubfield ) =
1176           &GetMarcFromKohaField( "items." . $column, "" );
1177         $subfieldstosearch{$column} = $tagsubfield;
1178     }
1179
1180     # handle which records to actually retrieve
1181     my $times;
1182     if ( $hits && $offset + $results_per_page <= $hits ) {
1183         $times = $offset + $results_per_page;
1184     }
1185     else {
1186         $times = $hits;
1187     }
1188
1189     # loop through all of the records we've retrieved
1190     for ( my $i = $offset ; $i <= $times - 1 ; $i++ ) {
1191         my $marcrecord;
1192         $marcrecord = MARC::File::USMARC::decode( $marcresults[$i] );
1193         my $oldbiblio = TransformMarcToKoha( $dbh, $marcrecord, '' );
1194         $oldbiblio->{result_number} = $i + 1;
1195
1196         # add imageurl to itemtype if there is one
1197         if ( $itemtypes{ $oldbiblio->{itemtype} }->{imageurl} =~ /^http:/ ) {
1198             $oldbiblio->{imageurl} =
1199               $itemtypes{ $oldbiblio->{itemtype} }->{imageurl};
1200             $oldbiblio->{description} =
1201               $itemtypes{ $oldbiblio->{itemtype} }->{description};
1202         }
1203         else {
1204             $oldbiblio->{imageurl} =
1205               getitemtypeimagesrc() . "/"
1206               . $itemtypes{ $oldbiblio->{itemtype} }->{imageurl}
1207               if ( $itemtypes{ $oldbiblio->{itemtype} }->{imageurl} );
1208             $oldbiblio->{description} =
1209               $itemtypes{ $oldbiblio->{itemtype} }->{description};
1210         }
1211
1212  # Build summary if there is one (the summary is defined in the itemtypes table)
1213  # FIXME: is this used anywhere, I think it can be commented out? -- JF
1214         if ( $itemtypes{ $oldbiblio->{itemtype} }->{summary} ) {
1215             my $summary = $itemtypes{ $oldbiblio->{itemtype} }->{summary};
1216             my @fields  = $marcrecord->fields();
1217             foreach my $field (@fields) {
1218                 my $tag      = $field->tag();
1219                 my $tagvalue = $field->as_string();
1220                 $summary =~
1221                   s/\[(.?.?.?.?)$tag\*(.*?)]/$1$tagvalue$2\[$1$tag$2]/g;
1222                 unless ( $tag < 10 ) {
1223                     my @subf = $field->subfields;
1224                     for my $i ( 0 .. $#subf ) {
1225                         my $subfieldcode  = $subf[$i][0];
1226                         my $subfieldvalue = $subf[$i][1];
1227                         my $tagsubf       = $tag . $subfieldcode;
1228                         $summary =~
1229 s/\[(.?.?.?.?)$tagsubf(.*?)]/$1$subfieldvalue$2\[$1$tagsubf$2]/g;
1230                     }
1231                 }
1232             }
1233             # FIXME: yuk
1234             $summary =~ s/\[(.*?)]//g;
1235             $summary =~ s/\n/<br>/g;
1236             $oldbiblio->{summary} = $summary;
1237         }
1238
1239 # Add search-term highlighting to the whole record where they match using <span>s
1240         my $searchhighlightblob;
1241         for my $highlight_field ( $marcrecord->fields ) {
1242
1243   # FIXME: need to skip title, subtitle, author, etc., as they are handled below
1244             next if $highlight_field->tag() =~ /(^00)/;    # skip fixed fields
1245             for my $subfield ($highlight_field->subfields()) {
1246                 my $match;
1247                 next if $subfield->[0] eq '9';
1248                 my $field = $subfield->[1];
1249                 for my $term ( keys %$span_terms_hashref ) {
1250                     if ( ( $field =~ /$term/i ) && (( length($term) > 3 ) || ($field =~ / $term /i)) ) {
1251                         $field =~ s/$term/<span class=\"term\">$&<\/span>/gi;
1252                     $match++;
1253                     }
1254                 }
1255                 $searchhighlightblob .= $field . " ... " if $match;
1256             }
1257
1258         }
1259         $searchhighlightblob = ' ... '.$searchhighlightblob if $searchhighlightblob;
1260         $oldbiblio->{'searchhighlightblob'} = $searchhighlightblob;
1261
1262 # save an author with no <span> tag, for the <a href=search.pl?q=<!--tmpl_var name="author"-->> link
1263         $oldbiblio->{'author_nospan'} = $oldbiblio->{'author'};
1264
1265         # Add search-term highlighting to the title, subtitle, etc. fields
1266         for my $term ( keys %$span_terms_hashref ) {
1267             my $old_term = $term;
1268             if ( length($term) > 3 ) {
1269                 $term =~ s/(.*=|\)|\(|\+|\.|\?|\[|\]|\\|\*)//g;
1270                 $oldbiblio->{'title'} =~
1271                   s/$term/<span class=\"term\">$&<\/span>/gi;
1272                 $oldbiblio->{'subtitle'} =~
1273                   s/$term/<span class=\"term\">$&<\/span>/gi;
1274                 $oldbiblio->{'author'} =~
1275                   s/$term/<span class=\"term\">$&<\/span>/gi;
1276                 $oldbiblio->{'publishercode'} =~
1277                   s/$term/<span class=\"term\">$&<\/span>/gi;
1278                 $oldbiblio->{'place'} =~
1279                   s/$term/<span class=\"term\">$&<\/span>/gi;
1280                 $oldbiblio->{'pages'} =~
1281                   s/$term/<span class=\"term\">$&<\/span>/gi;
1282                 $oldbiblio->{'notes'} =~
1283                   s/$term/<span class=\"term\">$&<\/span>/gi;
1284                 $oldbiblio->{'size'} =~
1285                   s/$term/<span class=\"term\">$&<\/span>/gi;
1286             }
1287         }
1288
1289         # FIXME:
1290         # surely there's a better way to handle this
1291         if ( $i % 2 ) {
1292             $toggle = "#ffffcc";
1293         }
1294         else {
1295             $toggle = "white";
1296         }
1297         $oldbiblio->{'toggle'} = $toggle;
1298
1299         # Pull out the items fields
1300         my @fields = $marcrecord->field($itemtag);
1301
1302         # Setting item statuses for display
1303         my @available_items_loop;
1304         my @onloan_items_loop;
1305         my @other_items_loop;
1306
1307         my $available_items;
1308         my $onloan_items;
1309         my $other_items;
1310
1311         my $ordered_count     = 0;
1312         my $available_count   = 0;
1313         my $onloan_count      = 0;
1314         my $longoverdue_count = 0;
1315         my $other_count       = 0;
1316         my $wthdrawn_count    = 0;
1317         my $itemlost_count    = 0;
1318         my $itembinding_count = 0;
1319         my $itemdamaged_count = 0;
1320         my $can_place_holds   = 0;
1321         my $items_count       = scalar(@fields);
1322         my $items_counter;
1323         my $maxitems =
1324           ( C4::Context->preference('maxItemsinSearchResults') )
1325           ? C4::Context->preference('maxItemsinSearchResults') - 1
1326           : 1;
1327
1328         # loop through every item
1329         foreach my $field (@fields) {
1330             my $item;
1331             $items_counter++;
1332
1333             # populate the items hash
1334             foreach my $code ( keys %subfieldstosearch ) {
1335                 $item->{$code} = $field->subfield( $subfieldstosearch{$code} );
1336             }
1337
1338       # set item's branch name, use homebranch first, fall back to holdingbranch
1339             if ( $item->{'homebranch'} ) {
1340                 $item->{'branchname'} = $branches{ $item->{homebranch} };
1341             }
1342
1343             # Last resort
1344             elsif ( $item->{'holdingbranch'} ) {
1345                 $item->{'branchname'} = $branches{ $item->{holdingbranch} };
1346             }
1347
1348 # For each grouping of items (onloan, available, unavailable), we build a key to store relevant info about that item
1349             if ( $item->{onloan} ) {
1350                 $onloan_count++;
1351                 $onloan_items->{ $item->{'homebranch'} . '--' . $item->{location} . $item->{'itype'} . $item->{'itemcallnumber'} . $item->{due_date}  }->{due_date} = format_date( $item->{onloan} );
1352                 $onloan_items->{ $item->{'homebranch'} . '--' . $item->{location} . $item->{'itype'} . $item->{'itemcallnumber'} . $item->{due_date}  }->{count}++ if $item->{'homebranch'};
1353                 $onloan_items->{ $item->{'homebranch'} . '--' . $item->{location} . $item->{'itype'} . $item->{'itemcallnumber'} . $item->{due_date}  }->{branchname} = $item->{'branchname'};
1354                 $onloan_items->{ $item->{'homebranch'} . '--' . $item->{location} . $item->{'itype'} . $item->{'itemcallnumber'} . $item->{due_date}  }->{location} = $locations{ $item->{location} };
1355                 $onloan_items->{ $item->{'homebranch'} . '--' . $item->{location} . $item->{'itype'} . $item->{'itemcallnumber'} . $item->{due_date}  }->{itemcallnumber} = $item->{itemcallnumber};
1356         $onloan_items->{ $item->{'homebranch'} . '--' . $item->{location} . $item->{'itype'} . $item->{'itemcallnumber'} . $item->{due_date}  }->{imageurl} = getitemtypeimagesrc() . "/" . $itemtypes{ $item->{itype} }->{imageurl};
1357                 # if something's checked out and lost, mark it as 'long overdue'
1358                 if ( $item->{itemlost} ) {
1359                     $onloan_items->{ $item->{'homebranch'} . '--' . $item->{location} . $item->{'itype'} . $item->{'itemcallnumber'} . $item->{due_date} }->{longoverdue}++;
1360                     $longoverdue_count++;
1361                 }
1362
1363                 # can place holds as long as this item isn't lost
1364                 else {
1365                     $can_place_holds = 1;
1366                 }
1367             }
1368
1369          # items not on loan, but still unavailable ( lost, withdrawn, damaged )
1370             else {
1371
1372                 # item is on order
1373                 if ( $item->{notforloan} == -1 ) {
1374                     $ordered_count++;
1375                 }
1376
1377                 # item is withdrawn, lost or damaged
1378                 if (   $item->{wthdrawn}
1379                     || $item->{itemlost}
1380                     || $item->{damaged}
1381                     || $item->{notforloan} )
1382                 {
1383                     $wthdrawn_count++    if $item->{wthdrawn};
1384                     $itemlost_count++    if $item->{itemlost};
1385                     $itemdamaged_count++ if $item->{damaged};
1386                     $item->{status} = $item->{wthdrawn} . "-" . $item->{itemlost} . "-" . $item->{damaged} . "-" . $item->{notforloan};
1387                     $other_count++;
1388
1389                     $other_items->{ $item->{'homebranch'} . '--' . $item->{location} . $item->{'itype'} . $item->{'itemcallnumber'} . $item->{status} }->{wthdrawn} = $item->{wthdrawn};
1390                     $other_items->{ $item->{'homebranch'} . '--' . $item->{location} . $item->{'itype'} . $item->{'itemcallnumber'} . $item->{status} }->{itemlost} = $item->{itemlost};
1391                     $other_items->{ $item->{'homebranch'} . '--' . $item->{location} . $item->{'itype'} . $item->{'itemcallnumber'} . $item->{status} }->{damaged} = $item->{damaged};
1392                     $other_items->{ $item->{'homebranch'} . '--' . $item->{location} . $item->{'itype'} . $item->{'itemcallnumber'} . $item->{status} }->{notforloan} = GetAuthorisedValueDesc( '', '', $item->{notforloan}, '', '', $notforloan_authorised_value ) if $notforloan_authorised_value;
1393                     $other_items->{ $item->{'homebranch'} . '--' . $item->{location} . $item->{'itype'} . $item->{'itemcallnumber'} . $item->{status} }->{count}++ if $item->{'homebranch'};
1394                     $other_items->{ $item->{'homebranch'} . '--' . $item->{location} . $item->{'itype'} . $item->{'itemcallnumber'} . $item->{status} }->{branchname} = $item->{'branchname'};
1395                     $other_items->{ $item->{'homebranch'} . '--' . $item->{location} . $item->{'itype'} . $item->{'itemcallnumber'} . $item->{status} }->{location} = $locations{ $item->{location} };
1396                     $other_items->{ $item->{'homebranch'} . '--' . $item->{location} . $item->{'itype'} . $item->{'itemcallnumber'} . $item->{status} }->{itemcallnumber} = $item->{itemcallnumber};
1397             $other_items->{ $item->{'homebranch'} . '--' . $item->{location} . $item->{'itype'} . $item->{'itemcallnumber'} . $item->{status} }->{imageurl} = getitemtypeimagesrc() . "/" . $itemtypes{ $item->{itype} }->{imageurl};
1398                 }
1399
1400                 # item is available
1401                 else {
1402                     $can_place_holds = 1;
1403                     $available_count++;
1404                     $available_items->{ $item->{'homebranch'} . '--' . $item->{location} . $item->{'itype'} . $item->{'itemcallnumber'} }->{count}++ if $item->{'homebranch'};
1405                     $available_items->{ $item->{'homebranch'} . '--' . $item->{location} . $item->{'itype'} . $item->{'itemcallnumber'} }->{branchname} = $item->{'branchname'};
1406                     $available_items->{ $item->{'homebranch'} . '--' . $item->{location} . $item->{'itype'} . $item->{'itemcallnumber'} }->{location} = $locations{ $item->{location} };
1407                     $available_items->{ $item->{'homebranch'} . '--' . $item->{location} . $item->{'itype'} . $item->{'itemcallnumber'} }->{itemcallnumber} = $item->{itemcallnumber};
1408             $available_items->{ $item->{'homebranch'} . '--' . $item->{location} . $item->{'itype'} . $item->{'itemcallnumber'} }->{imageurl} = getitemtypeimagesrc() . "/" . $itemtypes{ $item->{itype} }->{imageurl};
1409                 }
1410             }
1411         }    # notforloan, item level and biblioitem level
1412         my ( $availableitemscount, $onloanitemscount, $otheritemscount );
1413         $maxitems =
1414           ( C4::Context->preference('maxItemsinSearchResults') )
1415           ? C4::Context->preference('maxItemsinSearchResults') - 1
1416           : 1;
1417         for my $key ( sort keys %$onloan_items ) {
1418             $onloanitemscount++;
1419             push @onloan_items_loop, $onloan_items->{$key}
1420               unless $onloanitemscount > $maxitems;
1421         }
1422         for my $key ( sort keys %$other_items ) {
1423             $otheritemscount++;
1424             push @other_items_loop, $other_items->{$key}
1425               unless $otheritemscount > $maxitems;
1426         }
1427         for my $key ( sort keys %$available_items ) {
1428             $availableitemscount++;
1429             push @available_items_loop, $available_items->{$key}
1430               unless $availableitemscount > $maxitems;
1431         }
1432
1433 # last check for norequest : if itemtype is notforloan, it can't be reserved either, whatever the items
1434         $can_place_holds = 0
1435           if $itemtypes{ $oldbiblio->{itemtype} }->{notforloan};
1436         $oldbiblio->{norequests} = 1 unless $can_place_holds;
1437         $oldbiblio->{itemsplural}          = 1 if $items_count > 1;
1438         $oldbiblio->{items_count}          = $items_count;
1439         $oldbiblio->{available_items_loop} = \@available_items_loop;
1440         $oldbiblio->{onloan_items_loop}    = \@onloan_items_loop;
1441         $oldbiblio->{other_items_loop}     = \@other_items_loop;
1442         $oldbiblio->{availablecount}       = $available_count;
1443         $oldbiblio->{availableplural}      = 1 if $available_count > 1;
1444         $oldbiblio->{onloancount}          = $onloan_count;
1445         $oldbiblio->{onloanplural}         = 1 if $onloan_count > 1;
1446         $oldbiblio->{othercount}           = $other_count;
1447         $oldbiblio->{otherplural}          = 1 if $other_count > 1;
1448         $oldbiblio->{wthdrawncount}        = $wthdrawn_count;
1449         $oldbiblio->{itemlostcount}        = $itemlost_count;
1450         $oldbiblio->{damagedcount}         = $itemdamaged_count;
1451         $oldbiblio->{orderedcount}         = $ordered_count;
1452         $oldbiblio->{isbn} =~
1453           s/-//g;    # deleting - in isbn to enable amazon content
1454         push( @newresults, $oldbiblio );
1455     }
1456     return @newresults;
1457 }
1458
1459 #----------------------------------------------------------------------
1460 #
1461 # Non-Zebra GetRecords#
1462 #----------------------------------------------------------------------
1463
1464 =head2 NZgetRecords
1465
1466   NZgetRecords has the same API as zera getRecords, even if some parameters are not managed
1467
1468 =cut
1469
1470 sub NZgetRecords {
1471     my (
1472         $query,            $simple_query, $sort_by_ref,    $servers_ref,
1473         $results_per_page, $offset,       $expanded_facet, $branches,
1474         $query_type,       $scan
1475     ) = @_;
1476     warn "query =$query" if $DEBUG;
1477     my $result = NZanalyse($query);
1478     warn "results =$result" if $DEBUG;
1479     return ( undef,
1480         NZorder( $result, @$sort_by_ref[0], $results_per_page, $offset ),
1481         undef );
1482 }
1483
1484 =head2 NZanalyse
1485
1486   NZanalyse : get a CQL string as parameter, and returns a list of biblionumber;title,biblionumber;title,...
1487   the list is built from an inverted index in the nozebra SQL table
1488   note that title is here only for convenience : the sorting will be very fast when requested on title
1489   if the sorting is requested on something else, we will have to reread all results, and that may be longer.
1490
1491 =cut
1492
1493 sub NZanalyse {
1494     my ( $string, $server ) = @_;
1495 #     warn "---------"       if $DEBUG;
1496     warn " NZanalyse" if $DEBUG;
1497 #     warn "---------"       if $DEBUG;
1498
1499  # $server contains biblioserver or authorities, depending on what we search on.
1500  #warn "querying : $string on $server";
1501     $server = 'biblioserver' unless $server;
1502
1503 # if we have a ", replace the content to discard temporarily any and/or/not inside
1504     my $commacontent;
1505     if ( $string =~ /"/ ) {
1506         $string =~ s/"(.*?)"/__X__/;
1507         $commacontent = $1;
1508         warn "commacontent : $commacontent" if $DEBUG;
1509     }
1510
1511 # split the query string in 3 parts : X AND Y means : $left="X", $operand="AND" and $right="Y"
1512 # then, call again NZanalyse with $left and $right
1513 # (recursive until we find a leaf (=> something without and/or/not)
1514 # delete repeated operator... Would then go in infinite loop
1515     while ( $string =~ s/( and| or| not| AND| OR| NOT)\1/$1/g ) {
1516     }
1517
1518     #process parenthesis before.
1519     if ( $string =~ /^\s*\((.*)\)(( and | or | not | AND | OR | NOT )(.*))?/ ) {
1520         my $left     = $1;
1521         my $right    = $4;
1522         my $operator = lc($3);   # FIXME: and/or/not are operators, not operands
1523         warn
1524 "dealing w/parenthesis before recursive sub call. left :$left operator:$operator right:$right"
1525           if $DEBUG;
1526         my $leftresult = NZanalyse( $left, $server );
1527         if ($operator) {
1528             my $rightresult = NZanalyse( $right, $server );
1529
1530             # OK, we have the results for right and left part of the query
1531             # depending of operand, intersect, union or exclude both lists
1532             # to get a result list
1533             if ( $operator eq ' and ' ) {
1534                 return NZoperatorAND($leftresult,$rightresult);
1535             }
1536             elsif ( $operator eq ' or ' ) {
1537
1538                 # just merge the 2 strings
1539                 return $leftresult . $rightresult;
1540             }
1541             elsif ( $operator eq ' not ' ) {
1542                 return NZoperatorNOT($leftresult,$rightresult);
1543             }
1544         }
1545         else {
1546 # this error is impossible, because of the regexp that isolate the operand, but just in case...
1547             return $leftresult;
1548         }
1549     }
1550     warn "string :" . $string if $DEBUG;
1551     $string =~ /(.*?)( and | or | not | AND | OR | NOT )(.*)/;
1552     my $left     = $1;
1553     my $right    = $3;
1554     my $operator = lc($2);    # FIXME: and/or/not are operators, not operands
1555     warn "no parenthesis. left : $left operator: $operator right: $right"
1556       if $DEBUG;
1557
1558     # it's not a leaf, we have a and/or/not
1559     if ($operator) {
1560
1561         # reintroduce comma content if needed
1562         $right =~ s/__X__/"$commacontent"/ if $commacontent;
1563         $left  =~ s/__X__/"$commacontent"/ if $commacontent;
1564         warn "node : $left / $operator / $right\n" if $DEBUG;
1565         my $leftresult  = NZanalyse( $left,  $server );
1566         my $rightresult = NZanalyse( $right, $server );
1567         warn " leftresult : $leftresult" if $DEBUG;
1568         warn " rightresult : $rightresult" if $DEBUG;
1569         # OK, we have the results for right and left part of the query
1570         # depending of operand, intersect, union or exclude both lists
1571         # to get a result list
1572         if ( $operator eq ' and ' ) {
1573             warn "NZAND";
1574             return NZoperatorAND($leftresult,$rightresult);
1575         }
1576         elsif ( $operator eq ' or ' ) {
1577
1578             # just merge the 2 strings
1579             return $leftresult . $rightresult;
1580         }
1581         elsif ( $operator eq ' not ' ) {
1582             return NZoperatorNOT($leftresult,$rightresult);
1583         }
1584         else {
1585
1586 # this error is impossible, because of the regexp that isolate the operand, but just in case...
1587             die "error : operand unknown : $operator for $string";
1588         }
1589
1590         # it's a leaf, do the real SQL query and return the result
1591     }
1592     else {
1593         $string =~ s/__X__/"$commacontent"/ if $commacontent;
1594         $string =~ s/-|\.|\?|,|;|!|'|\(|\)|\[|\]|{|}|"|&|\+|\*|\// /g;
1595         warn "leaf:$string" if $DEBUG;
1596
1597         # parse the string in in operator/operand/value again
1598         $string =~ /(.*)(>=|<=)(.*)/;
1599         my $left     = $1;
1600         my $operator = $2;
1601         my $right    = $3;
1602 #         warn "handling leaf... left:$left operator:$operator right:$right"
1603 #           if $DEBUG;
1604         unless ($operator) {
1605             $string =~ /(.*)(>|<|=)(.*)/;
1606             $left     = $1;
1607             $operator = $2;
1608             $right    = $3;
1609 #             warn
1610 # "handling unless (operator)... left:$left operator:$operator right:$right"
1611 #               if $DEBUG;
1612         }
1613         my $results;
1614
1615 # strip adv, zebra keywords, currently not handled in nozebra: wrdl, ext, phr...
1616         $left =~ s/[, ].*$//;
1617
1618         # automatic replace for short operators
1619         $left = 'title'            if $left =~ '^ti$';
1620         $left = 'author'           if $left =~ '^au$';
1621         $left = 'publisher'        if $left =~ '^pb$';
1622         $left = 'subject'          if $left =~ '^su$';
1623         $left = 'koha-Auth-Number' if $left =~ '^an$';
1624         $left = 'keyword'          if $left =~ '^kw$';
1625         warn "handling leaf... left:$left operator:$operator right:$right";
1626         if ( $operator && $left ne 'keyword' ) {
1627
1628             #do a specific search
1629             my $dbh = C4::Context->dbh;
1630             $operator = 'LIKE' if $operator eq '=' and $right =~ /%/;
1631             my $sth =
1632               $dbh->prepare(
1633 "SELECT biblionumbers,value FROM nozebra WHERE server=? AND indexname=? AND value $operator ?"
1634               );
1635             warn "$left / $operator / $right\n";
1636
1637             # split each word, query the DB and build the biblionumbers result
1638             #sanitizing leftpart
1639             $left =~ s/^\s+|\s+$//;
1640             foreach ( split / /, $right ) {
1641                 my $biblionumbers;
1642                 $_ =~ s/^\s+|\s+$//;
1643                 next unless $_;
1644                 warn "EXECUTE : $server, $left, $_";
1645                 $sth->execute( $server, $left, $_ )
1646                   or warn "execute failed: $!";
1647                 while ( my ( $line, $value ) = $sth->fetchrow ) {
1648
1649 # if we are dealing with a numeric value, use only numeric results (in case of >=, <=, > or <)
1650 # otherwise, fill the result
1651                     $biblionumbers .= $line
1652                       unless ( $right =~ /^\d+$/ && $value =~ /\D/ );
1653                     warn "result : $value "
1654                       . ( $right  =~ /\d/ ) . "=="
1655                       . ( $value =~ /\D/?$line:"" );         #= $line";
1656                 }
1657
1658 # do a AND with existing list if there is one, otherwise, use the biblionumbers list as 1st result list
1659                 if ($results) {
1660                     warn "NZAND";
1661                     $results = NZoperatorAND($biblionumbers,$results);
1662                 }
1663                 else {
1664                     $results = $biblionumbers;
1665                 }
1666             }
1667         }
1668         else {
1669
1670       #do a complete search (all indexes), if index='kw' do complete search too.
1671             my $dbh = C4::Context->dbh;
1672             my $sth =
1673               $dbh->prepare(
1674 "SELECT biblionumbers FROM nozebra WHERE server=? AND value LIKE ?"
1675               );
1676
1677             # split each word, query the DB and build the biblionumbers result
1678             foreach ( split / /, $string ) {
1679                 next if C4::Context->stopwords->{ uc($_) };   # skip if stopword
1680                 warn "search on all indexes on $_" if $DEBUG;
1681                 my $biblionumbers;
1682                 next unless $_;
1683                 $sth->execute( $server, $_ );
1684                 while ( my $line = $sth->fetchrow ) {
1685                     $biblionumbers .= $line;
1686                 }
1687
1688 # do a AND with existing list if there is one, otherwise, use the biblionumbers list as 1st result list
1689                 if ($results) {
1690                     $results = NZoperatorAND($biblionumbers,$results);
1691                 }
1692                 else {
1693                     warn "NEW RES for $_ = $biblionumbers" if $DEBUG;
1694                     $results = $biblionumbers;
1695                 }
1696             }
1697         }
1698         warn "return : $results for LEAF : $string" if $DEBUG;
1699         return $results;
1700     }
1701     warn "---------"       if $DEBUG;
1702     warn "Leave NZanalyse" if $DEBUG;
1703     warn "---------"       if $DEBUG;
1704 }
1705
1706 sub NZoperatorAND{
1707     my ($rightresult, $leftresult)=@_;
1708
1709     my @leftresult = split /;/, $leftresult;
1710     warn " @leftresult / $rightresult \n" if $DEBUG;
1711
1712     #             my @rightresult = split /;/,$leftresult;
1713     my $finalresult;
1714
1715 # parse the left results, and if the biblionumber exist in the right result, save it in finalresult
1716 # the result is stored twice, to have the same weight for AND than OR.
1717 # example : TWO : 61,61,64,121 (two is twice in the biblio #61) / TOWER : 61,64,130
1718 # result : 61,61,61,61,64,64 for two AND tower : 61 has more weight than 64
1719     foreach (@leftresult) {
1720         my $value = $_;
1721         my $countvalue;
1722         ( $value, $countvalue ) = ( $1, $2 ) if ($value=~/(.*)-(\d+)$/);
1723         if ( $rightresult =~ /$value-(\d+);/ ) {
1724             $countvalue = ( $1 > $countvalue ? $countvalue : $1 );
1725             $finalresult .=
1726                 "$value-$countvalue;$value-$countvalue;";
1727         }
1728     }
1729     warn " $finalresult \n" if $DEBUG;
1730     return $finalresult;
1731 }
1732
1733 sub NZoperatorOR{
1734     my ($rightresult, $leftresult)=@_;
1735     return $rightresult.$leftresult;
1736 }
1737
1738 sub NZoperatorNOT{
1739     my ($rightresult, $leftresult)=@_;
1740
1741     my @leftresult = split /;/, $leftresult;
1742
1743     #             my @rightresult = split /;/,$leftresult;
1744     my $finalresult;
1745     foreach (@leftresult) {
1746         my $value=$_;
1747         $value=$1 if $value=~m/(.*)-\d+$/;
1748         unless ($rightresult =~ "$value-") {
1749             $finalresult .= "$_;";
1750         }
1751     }
1752     return $finalresult;
1753 }
1754
1755 =head2 NZorder
1756
1757   $finalresult = NZorder($biblionumbers, $ordering,$results_per_page,$offset);
1758
1759   TODO :: Description
1760
1761 =cut
1762
1763 sub NZorder {
1764     my ( $biblionumbers, $ordering, $results_per_page, $offset ) = @_;
1765     warn "biblionumbers = $biblionumbers and ordering = $ordering\n" if $DEBUG;
1766
1767     # order title asc by default
1768     #     $ordering = '1=36 <i' unless $ordering;
1769     $results_per_page = 20 unless $results_per_page;
1770     $offset           = 0  unless $offset;
1771     my $dbh = C4::Context->dbh;
1772
1773     #
1774     # order by POPULARITY
1775     #
1776     if ( $ordering =~ /popularity/ ) {
1777         my %result;
1778         my %popularity;
1779
1780         # popularity is not in MARC record, it's builded from a specific query
1781         my $sth =
1782           $dbh->prepare("select sum(issues) from items where biblionumber=?");
1783         foreach ( split /;/, $biblionumbers ) {
1784             my ( $biblionumber, $title ) = split /,/, $_;
1785             $result{$biblionumber} = GetMarcBiblio($biblionumber);
1786             $sth->execute($biblionumber);
1787             my $popularity = $sth->fetchrow || 0;
1788
1789 # hint : the key is popularity.title because we can have
1790 # many results with the same popularity. In this cas, sub-ordering is done by title
1791 # we also have biblionumber to avoid bug for 2 biblios with the same title & popularity
1792 # (un-frequent, I agree, but we won't forget anything that way ;-)
1793             $popularity{ sprintf( "%10d", $popularity ) . $title
1794                   . $biblionumber } = $biblionumber;
1795         }
1796
1797     # sort the hash and return the same structure as GetRecords (Zebra querying)
1798         my $result_hash;
1799         my $numbers = 0;
1800         if ( $ordering eq 'popularity_dsc' ) {    # sort popularity DESC
1801             foreach my $key ( sort { $b cmp $a } ( keys %popularity ) ) {
1802                 $result_hash->{'RECORDS'}[ $numbers++ ] =
1803                   $result{ $popularity{$key} }->as_usmarc();
1804             }
1805         }
1806         else {                                    # sort popularity ASC
1807             foreach my $key ( sort ( keys %popularity ) ) {
1808                 $result_hash->{'RECORDS'}[ $numbers++ ] =
1809                   $result{ $popularity{$key} }->as_usmarc();
1810             }
1811         }
1812         my $finalresult = ();
1813         $result_hash->{'hits'}         = $numbers;
1814         $finalresult->{'biblioserver'} = $result_hash;
1815         return $finalresult;
1816
1817         #
1818         # ORDER BY author
1819         #
1820     }
1821     elsif ( $ordering =~ /author/ ) {
1822         my %result;
1823         foreach ( split /;/, $biblionumbers ) {
1824             my ( $biblionumber, $title ) = split /,/, $_;
1825             my $record = GetMarcBiblio($biblionumber);
1826             my $author;
1827             if ( C4::Context->preference('marcflavour') eq 'UNIMARC' ) {
1828                 $author = $record->subfield( '200', 'f' );
1829                 $author = $record->subfield( '700', 'a' ) unless $author;
1830             }
1831             else {
1832                 $author = $record->subfield( '100', 'a' );
1833             }
1834
1835 # hint : the result is sorted by title.biblionumber because we can have X biblios with the same title
1836 # and we don't want to get only 1 result for each of them !!!
1837             $result{ $author . $biblionumber } = $record;
1838         }
1839
1840     # sort the hash and return the same structure as GetRecords (Zebra querying)
1841         my $result_hash;
1842         my $numbers = 0;
1843         if ( $ordering eq 'author_za' ) {    # sort by author desc
1844             foreach my $key ( sort { $b cmp $a } ( keys %result ) ) {
1845                 $result_hash->{'RECORDS'}[ $numbers++ ] =
1846                   $result{$key}->as_usmarc();
1847             }
1848         }
1849         else {                               # sort by author ASC
1850             foreach my $key ( sort ( keys %result ) ) {
1851                 $result_hash->{'RECORDS'}[ $numbers++ ] =
1852                   $result{$key}->as_usmarc();
1853             }
1854         }
1855         my $finalresult = ();
1856         $result_hash->{'hits'}         = $numbers;
1857         $finalresult->{'biblioserver'} = $result_hash;
1858         return $finalresult;
1859
1860         #
1861         # ORDER BY callnumber
1862         #
1863     }
1864     elsif ( $ordering =~ /callnumber/ ) {
1865         my %result;
1866         foreach ( split /;/, $biblionumbers ) {
1867             my ( $biblionumber, $title ) = split /,/, $_;
1868             my $record = GetMarcBiblio($biblionumber);
1869             my $callnumber;
1870             my ( $callnumber_tag, $callnumber_subfield ) =
1871               GetMarcFromKohaField( $dbh, 'items.itemcallnumber' );
1872             ( $callnumber_tag, $callnumber_subfield ) =
1873               GetMarcFromKohaField('biblioitems.callnumber')
1874               unless $callnumber_tag;
1875             if ( C4::Context->preference('marcflavour') eq 'UNIMARC' ) {
1876                 $callnumber = $record->subfield( '200', 'f' );
1877             }
1878             else {
1879                 $callnumber = $record->subfield( '100', 'a' );
1880             }
1881
1882 # hint : the result is sorted by title.biblionumber because we can have X biblios with the same title
1883 # and we don't want to get only 1 result for each of them !!!
1884             $result{ $callnumber . $biblionumber } = $record;
1885         }
1886
1887     # sort the hash and return the same structure as GetRecords (Zebra querying)
1888         my $result_hash;
1889         my $numbers = 0;
1890         if ( $ordering eq 'call_number_dsc' ) {    # sort by title desc
1891             foreach my $key ( sort { $b cmp $a } ( keys %result ) ) {
1892                 $result_hash->{'RECORDS'}[ $numbers++ ] =
1893                   $result{$key}->as_usmarc();
1894             }
1895         }
1896         else {                                     # sort by title ASC
1897             foreach my $key ( sort { $a cmp $b } ( keys %result ) ) {
1898                 $result_hash->{'RECORDS'}[ $numbers++ ] =
1899                   $result{$key}->as_usmarc();
1900             }
1901         }
1902         my $finalresult = ();
1903         $result_hash->{'hits'}         = $numbers;
1904         $finalresult->{'biblioserver'} = $result_hash;
1905         return $finalresult;
1906     }
1907     elsif ( $ordering =~ /pubdate/ ) {             #pub year
1908         my %result;
1909         foreach ( split /;/, $biblionumbers ) {
1910             my ( $biblionumber, $title ) = split /,/, $_;
1911             my $record = GetMarcBiblio($biblionumber);
1912             my ( $publicationyear_tag, $publicationyear_subfield ) =
1913               GetMarcFromKohaField( 'biblioitems.publicationyear', '' );
1914             my $publicationyear =
1915               $record->subfield( $publicationyear_tag,
1916                 $publicationyear_subfield );
1917
1918 # hint : the result is sorted by title.biblionumber because we can have X biblios with the same title
1919 # and we don't want to get only 1 result for each of them !!!
1920             $result{ $publicationyear . $biblionumber } = $record;
1921         }
1922
1923     # sort the hash and return the same structure as GetRecords (Zebra querying)
1924         my $result_hash;
1925         my $numbers = 0;
1926         if ( $ordering eq 'pubdate_dsc' ) {    # sort by pubyear desc
1927             foreach my $key ( sort { $b cmp $a } ( keys %result ) ) {
1928                 $result_hash->{'RECORDS'}[ $numbers++ ] =
1929                   $result{$key}->as_usmarc();
1930             }
1931         }
1932         else {                                 # sort by pub year ASC
1933             foreach my $key ( sort ( keys %result ) ) {
1934                 $result_hash->{'RECORDS'}[ $numbers++ ] =
1935                   $result{$key}->as_usmarc();
1936             }
1937         }
1938         my $finalresult = ();
1939         $result_hash->{'hits'}         = $numbers;
1940         $finalresult->{'biblioserver'} = $result_hash;
1941         return $finalresult;
1942
1943         #
1944         # ORDER BY title
1945         #
1946     }
1947     elsif ( $ordering =~ /title/ ) {
1948
1949 # the title is in the biblionumbers string, so we just need to build a hash, sort it and return
1950         my %result;
1951         foreach ( split /;/, $biblionumbers ) {
1952             my ( $biblionumber, $title ) = split /,/, $_;
1953
1954 # hint : the result is sorted by title.biblionumber because we can have X biblios with the same title
1955 # and we don't want to get only 1 result for each of them !!!
1956 # hint & speed improvement : we can order without reading the record
1957 # so order, and read records only for the requested page !
1958             $result{ $title . $biblionumber } = $biblionumber;
1959         }
1960
1961     # sort the hash and return the same structure as GetRecords (Zebra querying)
1962         my $result_hash;
1963         my $numbers = 0;
1964         if ( $ordering eq 'title_az' ) {    # sort by title desc
1965             foreach my $key ( sort ( keys %result ) ) {
1966                 $result_hash->{'RECORDS'}[ $numbers++ ] = $result{$key};
1967             }
1968         }
1969         else {                              # sort by title ASC
1970             foreach my $key ( sort { $b cmp $a } ( keys %result ) ) {
1971                 $result_hash->{'RECORDS'}[ $numbers++ ] = $result{$key};
1972             }
1973         }
1974
1975         # limit the $results_per_page to result size if it's more
1976         $results_per_page = $numbers - 1 if $numbers < $results_per_page;
1977
1978         # for the requested page, replace biblionumber by the complete record
1979         # speed improvement : avoid reading too much things
1980         for (
1981             my $counter = $offset ;
1982             $counter <= $offset + $results_per_page ;
1983             $counter++
1984           )
1985         {
1986             $result_hash->{'RECORDS'}[$counter] =
1987               GetMarcBiblio( $result_hash->{'RECORDS'}[$counter] )->as_usmarc;
1988         }
1989         my $finalresult = ();
1990         $result_hash->{'hits'}         = $numbers;
1991         $finalresult->{'biblioserver'} = $result_hash;
1992         return $finalresult;
1993     }
1994     else {
1995
1996 #
1997 # order by ranking
1998 #
1999 # we need 2 hashes to order by ranking : the 1st one to count the ranking, the 2nd to order by ranking
2000         my %result;
2001         my %count_ranking;
2002         foreach ( split /;/, $biblionumbers ) {
2003             my ( $biblionumber, $title ) = split /,/, $_;
2004             $title =~ /(.*)-(\d)/;
2005
2006             # get weight
2007             my $ranking = $2;
2008
2009 # note that we + the ranking because ranking is calculated on weight of EACH term requested.
2010 # if we ask for "two towers", and "two" has weight 2 in biblio N, and "towers" has weight 4 in biblio N
2011 # biblio N has ranking = 6
2012             $count_ranking{$biblionumber} += $ranking;
2013         }
2014
2015 # build the result by "inverting" the count_ranking hash
2016 # 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
2017 #         warn "counting";
2018         foreach ( keys %count_ranking ) {
2019             $result{ sprintf( "%10d", $count_ranking{$_} ) . '-' . $_ } = $_;
2020         }
2021
2022     # sort the hash and return the same structure as GetRecords (Zebra querying)
2023         my $result_hash;
2024         my $numbers = 0;
2025         foreach my $key ( sort { $b cmp $a } ( keys %result ) ) {
2026             $result_hash->{'RECORDS'}[ $numbers++ ] = $result{$key};
2027         }
2028
2029         # limit the $results_per_page to result size if it's more
2030         $results_per_page = $numbers - 1 if $numbers < $results_per_page;
2031
2032         # for the requested page, replace biblionumber by the complete record
2033         # speed improvement : avoid reading too much things
2034         for (
2035             my $counter = $offset ;
2036             $counter <= $offset + $results_per_page ;
2037             $counter++
2038           )
2039         {
2040             $result_hash->{'RECORDS'}[$counter] =
2041               GetMarcBiblio( $result_hash->{'RECORDS'}[$counter] )->as_usmarc
2042               if $result_hash->{'RECORDS'}[$counter];
2043         }
2044         my $finalresult = ();
2045         $result_hash->{'hits'}         = $numbers;
2046         $finalresult->{'biblioserver'} = $result_hash;
2047         return $finalresult;
2048     }
2049 }
2050
2051 =head2 ModBiblios
2052
2053 ($countchanged,$listunchanged) = ModBiblios($listbiblios, $tagsubfield,$initvalue,$targetvalue,$test);
2054
2055 this function changes all the values $initvalue in subfield $tag$subfield in any record in $listbiblios
2056 test parameter if set donot perform change to records in database.
2057
2058 =over 2
2059
2060 =item C<input arg:>
2061
2062     * $listbiblios is an array ref to marcrecords to be changed
2063     * $tagsubfield is the reference of the subfield to change.
2064     * $initvalue is the value to search the record for
2065     * $targetvalue is the value to set the subfield to
2066     * $test is to be set only not to perform changes in database.
2067
2068 =item C<Output arg:>
2069     * $countchanged counts all the changes performed.
2070     * $listunchanged contains the list of all the biblionumbers of records unchanged.
2071
2072 =item C<usage in the script:>
2073
2074 =back
2075
2076 my ($countchanged, $listunchanged) = EditBiblios($results->{RECORD}, $tagsubfield,$initvalue,$targetvalue);;
2077 #If one wants to display unchanged records, you should get biblios foreach @$listunchanged
2078 $template->param(countchanged => $countchanged, loopunchanged=>$listunchanged);
2079
2080 =cut
2081
2082 sub ModBiblios {
2083     my ( $listbiblios, $tagsubfield, $initvalue, $targetvalue, $test ) = @_;
2084     my $countmatched;
2085     my @unmatched;
2086     my ( $tag, $subfield ) = ( $1, $2 )
2087       if ( $tagsubfield =~ /^(\d{1,3})([a-z0-9A-Z@])?$/ );
2088     if ( ( length($tag) < 3 ) && $subfield =~ /0-9/ ) {
2089         $tag = $tag . $subfield;
2090         undef $subfield;
2091     }
2092     my ( $bntag,   $bnsubf )   = GetMarcFromKohaField('biblio.biblionumber');
2093     my ( $itemtag, $itemsubf ) = GetMarcFromKohaField('items.itemnumber');
2094     if ($tag eq $itemtag) {
2095         # do not allow the embedded item tag to be
2096         # edited from here
2097         warn "Attempting to edit item tag via C4::Search::ModBiblios -- not allowed";
2098         return (0, []);
2099     }
2100     foreach my $usmarc (@$listbiblios) {
2101         my $record;
2102         $record = eval { MARC::Record->new_from_usmarc($usmarc) };
2103         my $biblionumber;
2104         if ($@) {
2105
2106             # usmarc is not a valid usmarc May be a biblionumber
2107             # FIXME - sorry, please let's figure out whether
2108             #         this function is to be passed a list of
2109             #         record numbers or a list of MARC::Record
2110             #         objects.  The former is probably better
2111             #         because the MARC records supplied by Zebra
2112             #         may be not current.
2113             $record       = GetMarcBiblio($usmarc);
2114             $biblionumber = $usmarc;
2115         }
2116         else {
2117             if ( $bntag >= 010 ) {
2118                 $biblionumber = $record->subfield( $bntag, $bnsubf );
2119             }
2120             else {
2121                 $biblionumber = $record->field($bntag)->data;
2122             }
2123         }
2124
2125         #GetBiblionumber is to be written.
2126         #Could be replaced by TransformMarcToKoha (But Would be longer)
2127         if ( $record->field($tag) ) {
2128             my $modify = 0;
2129             foreach my $field ( $record->field($tag) ) {
2130                 if ($subfield) {
2131                     if (
2132                         $field->delete_subfield(
2133                             'code'  => $subfield,
2134                             'match' => qr($initvalue)
2135                         )
2136                       )
2137                     {
2138                         $countmatched++;
2139                         $modify = 1;
2140                         $field->update( $subfield, $targetvalue )
2141                           if ($targetvalue);
2142                     }
2143                 }
2144                 else {
2145                     if ( $tag >= 010 ) {
2146                         if ( $field->delete_field($field) ) {
2147                             $countmatched++;
2148                             $modify = 1;
2149                         }
2150                     }
2151                     else {
2152                         $field->data = $targetvalue
2153                           if ( $field->data =~ qr($initvalue) );
2154                     }
2155                 }
2156             }
2157
2158             #       warn $record->as_formatted;
2159             if ($modify) {
2160                 ModBiblio( $record, $biblionumber,
2161                     GetFrameworkCode($biblionumber) )
2162                   unless ($test);
2163             }
2164             else {
2165                 push @unmatched, $biblionumber;
2166             }
2167         }
2168         else {
2169             push @unmatched, $biblionumber;
2170         }
2171     }
2172     return ( $countmatched, \@unmatched );
2173 }
2174
2175 END { }    # module clean-up code here (global destructor)
2176
2177 1;
2178 __END__
2179
2180 =head1 AUTHOR
2181
2182 Koha Developement team <info@koha.org>
2183
2184 =cut