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 # use warnings; # FIXME
  20 require Exporter;
  21 use C4::Context;
  22 use C4::Biblio;    # GetMarcFromKohaField, GetBiblioData
  23 use C4::Koha;      # getFacets
  24 use Lingua::Stem;
  25 use C4::Search::PazPar2;
  26 use XML::Simple;
  27 use C4::Dates qw(format_date);
  28 use C4::XSLT;
  29 use C4::Branch;
  30 use C4::Debug;
  31 use YAML;
  32 use URI::Escape;
  33
  34 use vars qw($VERSION @ISA @EXPORT @EXPORT_OK %EXPORT_TAGS $DEBUG);
  35
  36 # set the version for version checking
  37 BEGIN {
  38     $VERSION = 3.01;
  39     $DEBUG = ($ENV{DEBUG}) ? 1 : 0;
  40 }
  41
  42 =head1 NAME
  43
  44 C4::Search - Functions for searching the Koha catalog.
  45
  46 =head1 SYNOPSIS
  47
  48 See opac/opac-search.pl or catalogue/search.pl for example of usage
  49
  50 =head1 DESCRIPTION
  51
  52 This module provides searching functions for Koha's bibliographic databases
  53
  54 =head1 FUNCTIONS
  55
  56 =cut
  57
  58 @ISA    = qw(Exporter);
  59 @EXPORT = qw(
  60   &FindDuplicate
  61   &SimpleSearch
  62   &searchResults
  63   &getRecords
  64   &buildQuery
  65   &NZgetRecords
  66   &AddSearchHistory
  67   &GetDistinctValues
  68   &BiblioAddAuthorities
  69 );
  70 #FIXME: i had to add BiblioAddAuthorities here because in Biblios.pm it caused circular dependencies (C4::Search uses C4::Biblio, and BiblioAddAuthorities uses SimpleSearch from C4::Search)
  71
  72 # make all your functions, whether exported or not;
  73
  74 =head2 FindDuplicate
  75
  76 ($biblionumber,$biblionumber,$title) = FindDuplicate($record);
  77
  78 This function attempts to find duplicate records using a hard-coded, fairly simplistic algorithm
  79
  80 =cut
  81
  82 sub FindDuplicate {
  83     my ($record) = @_;
  84     my $dbh = C4::Context->dbh;
  85     my $result = TransformMarcToKoha( $dbh, $record, '' );
  86     my $sth;
  87     my $query;
  88     my $search;
  89     my $type;
  90     my ( $biblionumber, $title );
  91
  92     # search duplicate on ISBN, easy and fast..
  93     # ... normalize first
  94     if ( $result->{isbn} ) {
  95         $result->{isbn} =~ s/\(.*$//;
  96         $result->{isbn} =~ s/\s+$//;
  97         $query = "isbn=$result->{isbn}";
  98     }
  99     else {
 100         $result->{title} =~ s /\\//g;
 101         $result->{title} =~ s /\"//g;
 102         $result->{title} =~ s /\(//g;
 103         $result->{title} =~ s /\)//g;
 104
 105         # FIXME: instead of removing operators, could just do
 106         # quotes around the value
 107         $result->{title} =~ s/(and|or|not)//g;
 108         $query = "ti,ext=$result->{title}";
 109         $query .= " and itemtype=$result->{itemtype}"
 110           if ( $result->{itemtype} );
 111         if   ( $result->{author} ) {
 112             $result->{author} =~ s /\\//g;
 113             $result->{author} =~ s /\"//g;
 114             $result->{author} =~ s /\(//g;
 115             $result->{author} =~ s /\)//g;
 116
 117             # remove valid operators
 118             $result->{author} =~ s/(and|or|not)//g;
 119             $query .= " and au,ext=$result->{author}";
 120         }
 121     }
 122
 123     # FIXME: add error handling
 124     my ( $error, $searchresults ) = SimpleSearch($query); # FIXME :: hardcoded !
 125     my @results;
 126     foreach my $possible_duplicate_record (@$searchresults) {
 127         my $marcrecord =
 128           MARC::Record->new_from_usmarc($possible_duplicate_record);
 129         my $result = TransformMarcToKoha( $dbh, $marcrecord, '' );
 130
 131         # FIXME :: why 2 $biblionumber ?
 132         if ($result) {
 133             push @results, $result->{'biblionumber'};
 134             push @results, $result->{'title'};
 135         }
 136     }
 137     return @results;
 138 }
 139
 140 =head2 SimpleSearch
 141
 142 ( $error, $results, $total_hits ) = SimpleSearch( $query, $offset, $max_results, [@servers] );
 143
 144 This function provides a simple search API on the bibliographic catalog
 145
 146 =over 2
 147
 148 =item C<input arg:>
 149
 150     * $query can be a simple keyword or a complete CCL query
 151     * @servers is optional. Defaults to biblioserver as found in koha-conf.xml
 152     * $offset - If present, represents the number of records at the beggining to omit. Defaults to 0
 153     * $max_results - if present, determines the maximum number of records to fetch. undef is All. defaults to undef.
 154
 155
 156 =item C<Output:>
 157
 158     * $error is a empty unless an error is detected
 159     * \@results is an array of records.
 160     * $total_hits is the number of hits that would have been returned with no limit
 161
 162 =item C<usage in the script:>
 163
 164 =back
 165
 166 my ( $error, $marcresults, $total_hits ) = SimpleSearch($query);
 167
 168 if (defined $error) {
 169     $template->param(query_error => $error);
 170     warn "error: ".$error;
 171     output_html_with_http_headers $input, $cookie, $template->output;
 172     exit;
 173 }
 174
 175 my $hits = scalar @$marcresults;
 176 my @results;
 177
 178 for my $i (0..$hits) {
 179     my %resultsloop;
 180     my $marcrecord = MARC::File::USMARC::decode($marcresults->[$i]);
 181     my $biblio = TransformMarcToKoha(C4::Context->dbh,$marcrecord,'');
 182
 183     #build the hash for the template.
 184     $resultsloop{title}           = $biblio->{'title'};
 185     $resultsloop{subtitle}        = $biblio->{'subtitle'};
 186     $resultsloop{biblionumber}    = $biblio->{'biblionumber'};
 187     $resultsloop{author}          = $biblio->{'author'};
 188     $resultsloop{publishercode}   = $biblio->{'publishercode'};
 189     $resultsloop{publicationyear} = $biblio->{'publicationyear'};
 190
 191     push @results, \%resultsloop;
 192 }
 193
 194 $template->param(result=>\@results);
 195
 196 =cut
 197
 198 sub SimpleSearch {
 199     my ( $query, $offset, $max_results, $servers )  = @_;
 200
 201     if ( C4::Context->preference('NoZebra') ) {
 202         my $result = NZorder( NZanalyse($query) )->{'biblioserver'};
 203         my $search_result =
 204           (      $result->{hits}
 205               && $result->{hits} > 0 ? $result->{'RECORDS'} : [] );
 206         return ( undef, $search_result, scalar($result->{hits}) );
 207     }
 208     else {
 209         # FIXME hardcoded value. See catalog/search.pl & opac-search.pl too.
 210         my @servers = defined ( $servers ) ? @$servers : ( "biblioserver" );
 211         my @results;
 212         my @zoom_queries;
 213         my @tmpresults;
 214         my @zconns;
 215         my $total_hits;
 216         return ( "No query entered", undef, undef ) unless $query;
 217
 218         # Initialize & Search Zebra
 219         for ( my $i = 0 ; $i < @servers ; $i++ ) {
 220             eval {
 221                 $zconns[$i] = C4::Context->Zconn( $servers[$i], 1 );
 222                 $zoom_queries[$i] = new ZOOM::Query::CCL2RPN( $query, $zconns[$i]);
 223                 $tmpresults[$i] = $zconns[$i]->search( $zoom_queries[$i] );
 224
 225                 # error handling
 226                 my $error =
 227                     $zconns[$i]->errmsg() . " ("
 228                   . $zconns[$i]->errcode() . ") "
 229                   . $zconns[$i]->addinfo() . " "
 230                   . $zconns[$i]->diagset();
 231
 232                 return ( $error, undef, undef ) if $zconns[$i]->errcode();
 233             };
 234             if ($@) {
 235
 236                 # caught a ZOOM::Exception
 237                 my $error =
 238                     $@->message() . " ("
 239                   . $@->code() . ") "
 240                   . $@->addinfo() . " "
 241                   . $@->diagset();
 242                 warn $error;
 243                 return ( $error, undef, undef );
 244             }
 245         }
 246         while ( ( my $i = ZOOM::event( \@zconns ) ) != 0 ) {
 247             my $event = $zconns[ $i - 1 ]->last_event();
 248             if ( $event == ZOOM::Event::ZEND ) {
 249
 250                 my $first_record = defined( $offset ) ? $offset+1 : 1;
 251                 my $hits = $tmpresults[ $i - 1 ]->size();
 252                 $total_hits += $hits;
 253                 my $last_record = $hits;
 254                 if ( defined $max_results && $offset + $max_results < $hits ) {
 255                     $last_record  = $offset + $max_results;
 256                 }
 257
 258                 for my $j ( $first_record..$last_record ) {
 259                     my $record = $tmpresults[ $i - 1 ]->record( $j-1 )->raw(); # 0 indexed
 260                     push @results, $record;
 261                 }
 262             }
 263         }
 264
 265         foreach my $result (@tmpresults) {
 266             $result->destroy();
 267         }
 268         foreach my $zoom_query (@zoom_queries) {
 269             $zoom_query->destroy();
 270         }
 271
 272         return ( undef, \@results, $total_hits );
 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;    # stores the ref to array of hashes for template facets loop
 314
 315     ### LOOP THROUGH THE SERVERS
 316     for ( my $i = 0 ; $i < @servers ; $i++ ) {
 317         $zconns[$i] = C4::Context->Zconn( $servers[$i], 1 );
 318
 319 # perform the search, create the results objects
 320 # if this is a local search, use the $koha-query, if it's a federated one, use the federated-query
 321         my $query_to_use = ($servers[$i] =~ /biblioserver/) ? $koha_query : $simple_query;
 322
 323         #$query_to_use = $simple_query if $scan;
 324         warn $simple_query if ( $scan and $DEBUG );
 325
 326         # Check if we've got a query_type defined, if so, use it
 327         eval {
 328             if ($query_type) {
 329                 if ($query_type =~ /^ccl/) {
 330                     $query_to_use =~ s/\:/\=/g;    # change : to = last minute (FIXME)
 331                     $results[$i] = $zconns[$i]->search(new ZOOM::Query::CCL2RPN($query_to_use, $zconns[$i]));
 332                 } elsif ($query_type =~ /^cql/) {
 333                     $results[$i] = $zconns[$i]->search(new ZOOM::Query::CQL($query_to_use, $zconns[$i]));
 334                 } elsif ($query_type =~ /^pqf/) {
 335                     $results[$i] = $zconns[$i]->search(new ZOOM::Query::PQF($query_to_use, $zconns[$i]));
 336                 } else {
 337                     warn "Unknown query_type '$query_type'.  Results undetermined.";
 338                 }
 339             } elsif ($scan) {
 340                     $results[$i] = $zconns[$i]->scan(  new ZOOM::Query::CCL2RPN($query_to_use, $zconns[$i]));
 341             } else {
 342                     $results[$i] = $zconns[$i]->search(new ZOOM::Query::CCL2RPN($query_to_use, $zconns[$i]));
 343             }
 344         };
 345         if ($@) {
 346             warn "WARNING: query problem with $query_to_use " . $@;
 347         }
 348
 349         # Concatenate the sort_by limits and pass them to the results object
 350         # Note: sort will override rank
 351         my $sort_by;
 352         foreach my $sort (@sort_by) {
 353             if ( $sort eq "author_az" ) {
 354                 $sort_by .= "1=1003 <i ";
 355             }
 356             elsif ( $sort eq "author_za" ) {
 357                 $sort_by .= "1=1003 >i ";
 358             }
 359             elsif ( $sort eq "popularity_asc" ) {
 360                 $sort_by .= "1=9003 <i ";
 361             }
 362             elsif ( $sort eq "popularity_dsc" ) {
 363                 $sort_by .= "1=9003 >i ";
 364             }
 365             elsif ( $sort eq "call_number_asc" ) {
 366                 $sort_by .= "1=20  <i ";
 367             }
 368             elsif ( $sort eq "call_number_dsc" ) {
 369                 $sort_by .= "1=20 >i ";
 370             }
 371             elsif ( $sort eq "pubdate_asc" ) {
 372                 $sort_by .= "1=31 <i ";
 373             }
 374             elsif ( $sort eq "pubdate_dsc" ) {
 375                 $sort_by .= "1=31 >i ";
 376             }
 377             elsif ( $sort eq "acqdate_asc" ) {
 378                 $sort_by .= "1=32 <i ";
 379             }
 380             elsif ( $sort eq "acqdate_dsc" ) {
 381                 $sort_by .= "1=32 >i ";
 382             }
 383             elsif ( $sort eq "title_az" ) {
 384                 $sort_by .= "1=4 <i ";
 385             }
 386             elsif ( $sort eq "title_za" ) {
 387                 $sort_by .= "1=4 >i ";
 388             }
 389             else {
 390                 warn "Ignoring unrecognized sort '$sort' requested" if $sort_by;
 391             }
 392         }
 393         if ($sort_by) {
 394             if ( $results[$i]->sort( "yaz", $sort_by ) < 0 ) {
 395                 warn "WARNING sort $sort_by failed";
 396             }
 397         }
 398     }    # finished looping through servers
 399
 400     # The big moment: asynchronously retrieve results from all servers
 401     while ( ( my $i = ZOOM::event( \@zconns ) ) != 0 ) {
 402         my $ev = $zconns[ $i - 1 ]->last_event();
 403         if ( $ev == ZOOM::Event::ZEND ) {
 404             next unless $results[ $i - 1 ];
 405             my $size = $results[ $i - 1 ]->size();
 406             if ( $size > 0 ) {
 407                 my $results_hash;
 408
 409                 # loop through the results
 410                 $results_hash->{'hits'} = $size;
 411                 my $times;
 412                 if ( $offset + $results_per_page <= $size ) {
 413                     $times = $offset + $results_per_page;
 414                 }
 415                 else {
 416                     $times = $size;
 417                 }
 418                 for ( my $j = $offset ; $j < $times ; $j++ ) {
 419                     my $records_hash;
 420                     my $record;
 421                     my $facet_record;
 422
 423                     ## Check if it's an index scan
 424                     if ($scan) {
 425                         my ( $term, $occ ) = $results[ $i - 1 ]->term($j);
 426
 427                  # here we create a minimal MARC record and hand it off to the
 428                  # template just like a normal result ... perhaps not ideal, but
 429                  # it works for now
 430                         my $tmprecord = MARC::Record->new();
 431                         $tmprecord->encoding('UTF-8');
 432                         my $tmptitle;
 433                         my $tmpauthor;
 434
 435                 # the minimal record in author/title (depending on MARC flavour)
 436                         if (C4::Context->preference("marcflavour") eq "UNIMARC") {
 437                             $tmptitle = MARC::Field->new('200',' ',' ', a => $term, f => $occ);
 438                             $tmprecord->append_fields($tmptitle);
 439                         } else {
 440                             $tmptitle  = MARC::Field->new('245',' ',' ', a => $term,);
 441                             $tmpauthor = MARC::Field->new('100',' ',' ', a => $occ,);
 442                             $tmprecord->append_fields($tmptitle);
 443                             $tmprecord->append_fields($tmpauthor);
 444                         }
 445                         $results_hash->{'RECORDS'}[$j] = $tmprecord->as_usmarc();
 446                     }
 447
 448                     # not an index scan
 449                     else {
 450                         $record = $results[ $i - 1 ]->record($j)->raw();
 451
 452                         # warn "RECORD $j:".$record;
 453                         $results_hash->{'RECORDS'}[$j] = $record;
 454
 455             # Fill the facets while we're looping, but only for the biblioserver
 456                         $facet_record = MARC::Record->new_from_usmarc($record)
 457                           if $servers[ $i - 1 ] =~ /biblioserver/;
 458
 459                     #warn $servers[$i-1]."\n".$record; #.$facet_record->title();
 460                         if ($facet_record) {
 461                             for ( my $k = 0 ; $k <= @$facets ; $k++ ) {
 462                                 ($facets->[$k]) or next;
 463                                 my @fields = map {$facet_record->field($_)} @{$facets->[$k]->{'tags'}} ;
 464                                 for my $field (@fields) {
 465                                     my @subfields = $field->subfields();
 466                                     for my $subfield (@subfields) {
 467                                         my ( $code, $data ) = @$subfield;
 468                                         ($code eq $facets->[$k]->{'subfield'}) or next;
 469                                         $facets_counter->{ $facets->[$k]->{'link_value'} }->{$data}++;
 470                                     }
 471                                 }
 472                                 $facets_info->{ $facets->[$k]->{'link_value'} }->{'label_value'} =
 473                                     $facets->[$k]->{'label_value'};
 474                                 $facets_info->{ $facets->[$k]->{'link_value'} }->{'expanded'} =
 475                                     $facets->[$k]->{'expanded'};
 476                             }
 477                         }
 478                     }
 479                 }
 480                 $results_hashref->{ $servers[ $i - 1 ] } = $results_hash;
 481             }
 482
 483             # warn "connection ", $i-1, ": $size hits";
 484             # warn $results[$i-1]->record(0)->render() if $size > 0;
 485
 486             # BUILD FACETS
 487             if ( $servers[ $i - 1 ] =~ /biblioserver/ ) {
 488                 for my $link_value (
 489                     sort { $facets_counter->{$b} <=> $facets_counter->{$a} }
 490                         keys %$facets_counter )
 491                 {
 492                     my $expandable;
 493                     my $number_of_facets;
 494                     my @this_facets_array;
 495                     for my $one_facet (
 496                         sort {
 497                              $facets_counter->{$link_value}->{$b}
 498                          <=> $facets_counter->{$link_value}->{$a}
 499                         } keys %{ $facets_counter->{$link_value} }
 500                       )
 501                     {
 502                         $number_of_facets++;
 503                         if (   ( $number_of_facets < 6 )
 504                             || ( $expanded_facet eq $link_value )
 505                             || ( $facets_info->{$link_value}->{'expanded'} ) )
 506                         {
 507
 508                       # Sanitize the link value ), ( will cause errors with CCL,
 509                             my $facet_link_value = $one_facet;
 510                             $facet_link_value =~ s/(\(|\))/ /g;
 511
 512                             # fix the length that will display in the label,
 513                             my $facet_label_value = $one_facet;
 514                             $facet_label_value =
 515                               substr( $one_facet, 0, 20 ) . "..."
 516                               unless length($facet_label_value) <= 20;
 517
 518                             # if it's a branch, label by the name, not the code,
 519                             if ( $link_value =~ /branch/ ) {
 520                                                                 if (defined $branches
 521                                                                         && ref($branches) eq "HASH"
 522                                                                         && defined $branches->{$one_facet}
 523                                                                         && ref ($branches->{$one_facet}) eq "HASH")
 524                                                                 {
 525                                         $facet_label_value =
 526                                                 $branches->{$one_facet}->{'branchname'};
 527                                                                 }
 528                                                                 else {
 529                                                                         $facet_label_value = "*";
 530                                                                 }
 531                             }
 532
 533                             # but we're down with the whole label being in the link's title.
 534                             push @this_facets_array, {
 535                                 facet_count       => $facets_counter->{$link_value}->{$one_facet},
 536                                 facet_label_value => $facet_label_value,
 537                                 facet_title_value => $one_facet,
 538                                 facet_link_value  => $facet_link_value,
 539                                 type_link_value   => $link_value,
 540                             };
 541                         }
 542                     }
 543
 544                     # handle expanded option
 545                     unless ( $facets_info->{$link_value}->{'expanded'} ) {
 546                         $expandable = 1
 547                           if ( ( $number_of_facets > 6 )
 548                             && ( $expanded_facet ne $link_value ) );
 549                     }
 550                     push @facets_loop, {
 551                         type_link_value => $link_value,
 552                         type_id         => $link_value . "_id",
 553                         "type_label_" . $facets_info->{$link_value}->{'label_value'} => 1,
 554                         facets     => \@this_facets_array,
 555                         expandable => $expandable,
 556                         expand     => $link_value,
 557                     } unless ( ($facets_info->{$link_value}->{'label_value'} =~ /Libraries/) and (C4::Context->preference('singleBranchMode')) );
 558                 }
 559             }
 560         }
 561     }
 562     return ( undef, $results_hashref, \@facets_loop );
 563 }
 564
 565 sub pazGetRecords {
 566     my (
 567         $koha_query,       $simple_query, $sort_by_ref,    $servers_ref,
 568         $results_per_page, $offset,       $expanded_facet, $branches,
 569         $query_type,       $scan
 570     ) = @_;
 571
 572     my $paz = C4::Search::PazPar2->new(C4::Context->config('pazpar2url'));
 573     $paz->init();
 574     $paz->search($simple_query);
 575     sleep 1;   # FIXME: WHY?
 576
 577     # do results
 578     my $results_hashref = {};
 579     my $stats = XMLin($paz->stat);
 580     my $results = XMLin($paz->show($offset, $results_per_page, 'work-title:1'), forcearray => 1);
 581
 582     # for a grouped search result, the number of hits
 583     # is the number of groups returned; 'bib_hits' will have
 584     # the total number of bibs.
 585     $results_hashref->{'biblioserver'}->{'hits'} = $results->{'merged'}->[0];
 586     $results_hashref->{'biblioserver'}->{'bib_hits'} = $stats->{'hits'};
 587
 588     HIT: foreach my $hit (@{ $results->{'hit'} }) {
 589         my $recid = $hit->{recid}->[0];
 590
 591         my $work_title = $hit->{'md-work-title'}->[0];
 592         my $work_author;
 593         if (exists $hit->{'md-work-author'}) {
 594             $work_author = $hit->{'md-work-author'}->[0];
 595         }
 596         my $group_label = (defined $work_author) ? "$work_title / $work_author" : $work_title;
 597
 598         my $result_group = {};
 599         $result_group->{'group_label'} = $group_label;
 600         $result_group->{'group_merge_key'} = $recid;
 601
 602         my $count = 1;
 603         if (exists $hit->{count}) {
 604             $count = $hit->{count}->[0];
 605         }
 606         $result_group->{'group_count'} = $count;
 607
 608         for (my $i = 0; $i < $count; $i++) {
 609             # FIXME -- may need to worry about diacritics here
 610             my $rec = $paz->record($recid, $i);
 611             push @{ $result_group->{'RECORDS'} }, $rec;
 612         }
 613
 614         push @{ $results_hashref->{'biblioserver'}->{'GROUPS'} }, $result_group;
 615     }
 616
 617     # pass through facets
 618     my $termlist_xml = $paz->termlist('author,subject');
 619     my $terms = XMLin($termlist_xml, forcearray => 1);
 620     my @facets_loop = ();
 621     #die Dumper($results);
 622 #    foreach my $list (sort keys %{ $terms->{'list'} }) {
 623 #        my @facets = ();
 624 #        foreach my $facet (sort @{ $terms->{'list'}->{$list}->{'term'} } ) {
 625 #            push @facets, {
 626 #                facet_label_value => $facet->{'name'}->[0],
 627 #            };
 628 #        }
 629 #        push @facets_loop, ( {
 630 #            type_label => $list,
 631 #            facets => \@facets,
 632 #        } );
 633 #    }
 634
 635     return ( undef, $results_hashref, \@facets_loop );
 636 }
 637
 638 # STOPWORDS
 639 sub _remove_stopwords {
 640     my ( $operand, $index ) = @_;
 641     my @stopwords_removed;
 642
 643     # phrase and exact-qualified indexes shouldn't have stopwords removed
 644     if ( $index !~ m/phr|ext/ ) {
 645
 646 # remove stopwords from operand : parse all stopwords & remove them (case insensitive)
 647 #       we use IsAlpha unicode definition, to deal correctly with diacritics.
 648 #       otherwise, a French word like "leçon" woudl be split into "le" "çon", "le"
 649 #       is a stopword, we'd get "çon" and wouldn't find anything...
 650 #
 651                 foreach ( keys %{ C4::Context->stopwords } ) {
 652                         next if ( $_ =~ /(and|or|not)/ );    # don't remove operators
 653                         $debug && warn "$_ Dump($operand)";
 654                         if ( my ($matched) = ($operand =~
 655                                 /([^\X\p{isAlnum}]\Q$_\E[^\X\p{isAlnum}]|[^\X\p{isAlnum}]\Q$_\E$|^\Q$_\E[^\X\p{isAlnum}])/gi))
 656                         {
 657                                 $operand =~ s/\Q$matched\E/ /gi;
 658                                 push @stopwords_removed, $_;
 659                         }
 660                 }
 661         }
 662     return ( $operand, \@stopwords_removed );
 663 }
 664
 665 # TRUNCATION
 666 sub _detect_truncation {
 667     my ( $operand, $index ) = @_;
 668     my ( @nontruncated, @righttruncated, @lefttruncated, @rightlefttruncated,
 669         @regexpr );
 670     $operand =~ s/^ //g;
 671     my @wordlist = split( /\s/, $operand );
 672     foreach my $word (@wordlist) {
 673         if ( $word =~ s/^\*([^\*]+)\*$/$1/ ) {
 674             push @rightlefttruncated, $word;
 675         }
 676         elsif ( $word =~ s/^\*([^\*]+)$/$1/ ) {
 677             push @lefttruncated, $word;
 678         }
 679         elsif ( $word =~ s/^([^\*]+)\*$/$1/ ) {
 680             push @righttruncated, $word;
 681         }
 682         elsif ( index( $word, "*" ) < 0 ) {
 683             push @nontruncated, $word;
 684         }
 685         else {
 686             push @regexpr, $word;
 687         }
 688     }
 689     return (
 690         \@nontruncated,       \@righttruncated, \@lefttruncated,
 691         \@rightlefttruncated, \@regexpr
 692     );
 693 }
 694
 695 # STEMMING
 696 sub _build_stemmed_operand {
 697     my ($operand,$lang) = @_;
 698     require Lingua::Stem::Snowball ;
 699     my $stemmed_operand;
 700
 701     # If operand contains a digit, it is almost certainly an identifier, and should
 702     # not be stemmed.  This is particularly relevant for ISBNs and ISSNs, which
 703     # can contain the letter "X" - for example, _build_stemmend_operand would reduce
 704     # "014100018X" to "x ", which for a MARC21 database would bring up irrelevant
 705     # results (e.g., "23 x 29 cm." from the 300$c).  Bug 2098.
 706     return $operand if $operand =~ /\d/;
 707
 708 # FIXME: the locale should be set based on the user's language and/or search choice
 709     #warn "$lang";
 710     my $stemmer = Lingua::Stem::Snowball->new( lang => $lang,
 711                                                encoding => "UTF-8" );
 712
 713     my @words = split( / /, $operand );
 714     my @stems = $stemmer->stem(\@words);
 715     for my $stem (@stems) {
 716         $stemmed_operand .= "$stem";
 717         $stemmed_operand .= "?"
 718           unless ( $stem =~ /(and$|or$|not$)/ ) || ( length($stem) < 3 );
 719         $stemmed_operand .= " ";
 720     }
 721     warn "STEMMED OPERAND: $stemmed_operand" if $DEBUG;
 722     return $stemmed_operand;
 723 }
 724
 725 # FIELD WEIGHTING
 726 sub _build_weighted_query {
 727
 728 # FIELD WEIGHTING - This is largely experimental stuff. What I'm committing works
 729 # pretty well but could work much better if we had a smarter query parser
 730     my ( $operand, $stemmed_operand, $index ) = @_;
 731     my $stemming      = C4::Context->preference("QueryStemming")     || 0;
 732     my $weight_fields = C4::Context->preference("QueryWeightFields") || 0;
 733     my $fuzzy_enabled = C4::Context->preference("QueryFuzzy")        || 0;
 734
 735     my $weighted_query .= "(rk=(";    # Specifies that we're applying rank
 736
 737     # Keyword, or, no index specified
 738     if ( ( $index eq 'kw' ) || ( !$index ) ) {
 739         $weighted_query .=
 740           "Title-cover,ext,r1=\"$operand\"";    # exact title-cover
 741         $weighted_query .= " or ti,ext,r2=\"$operand\"";    # exact title
 742         $weighted_query .= " or ti,phr,r3=\"$operand\"";    # phrase title
 743           #$weighted_query .= " or any,ext,r4=$operand";               # exact any
 744           #$weighted_query .=" or kw,wrdl,r5=\"$operand\"";            # word list any
 745         $weighted_query .= " or wrdl,fuzzy,r8=\"$operand\""
 746           if $fuzzy_enabled;    # add fuzzy, word list
 747         $weighted_query .= " or wrdl,right-Truncation,r9=\"$stemmed_operand\""
 748           if ( $stemming and $stemmed_operand )
 749           ;                     # add stemming, right truncation
 750         $weighted_query .= " or wrdl,r9=\"$operand\"";
 751
 752         # embedded sorting: 0 a-z; 1 z-a
 753         # $weighted_query .= ") or (sort1,aut=1";
 754     }
 755
 756     # Barcode searches should skip this process
 757     elsif ( $index eq 'bc' ) {
 758         $weighted_query .= "bc=\"$operand\"";
 759     }
 760
 761     # Authority-number searches should skip this process
 762     elsif ( $index eq 'an' ) {
 763         $weighted_query .= "an=\"$operand\"";
 764     }
 765
 766     # If the index already has more than one qualifier, wrap the operand
 767     # in quotes and pass it back (assumption is that the user knows what they
 768     # are doing and won't appreciate us mucking up their query
 769     elsif ( $index =~ ',' ) {
 770         $weighted_query .= " $index=\"$operand\"";
 771     }
 772
 773     #TODO: build better cases based on specific search indexes
 774     else {
 775         $weighted_query .= " $index,ext,r1=\"$operand\"";    # exact index
 776           #$weighted_query .= " or (title-sort-az=0 or $index,startswithnt,st-word,r3=$operand #)";
 777         $weighted_query .= " or $index,phr,r3=\"$operand\"";    # phrase index
 778         $weighted_query .=
 779           " or $index,rt,wrdl,r3=\"$operand\"";    # word list index
 780     }
 781
 782     $weighted_query .= "))";                       # close rank specification
 783     return $weighted_query;
 784 }
 785
 786 =head2 buildQuery
 787
 788 ( $error, $query,
 789 $simple_query, $query_cgi,
 790 $query_desc, $limit,
 791 $limit_cgi, $limit_desc,
 792 $stopwords_removed, $query_type ) = buildQuery ( $operators, $operands, $indexes, $limits, $sort_by, $scan, $lang);
 793
 794 Build queries and limits in CCL, CGI, Human,
 795 handle truncation, stemming, field weighting, stopwords, fuzziness, etc.
 796
 797 See verbose embedded documentation.
 798
 799
 800 =cut
 801
 802 sub buildQuery {
 803     my ( $operators, $operands, $indexes, $limits, $sort_by, $scan, $lang) = @_;
 804
 805     warn "---------\nEnter buildQuery\n---------" if $DEBUG;
 806
 807     # dereference
 808     my @operators = $operators ? @$operators : ();
 809     my @indexes   = $indexes   ? @$indexes   : ();
 810     my @operands  = $operands  ? @$operands  : ();
 811     my @limits    = $limits    ? @$limits    : ();
 812     my @sort_by   = $sort_by   ? @$sort_by   : ();
 813
 814     my $stemming         = C4::Context->preference("QueryStemming")        || 0;
 815     my $auto_truncation  = C4::Context->preference("QueryAutoTruncate")    || 0;
 816     my $weight_fields    = C4::Context->preference("QueryWeightFields")    || 0;
 817     my $fuzzy_enabled    = C4::Context->preference("QueryFuzzy")           || 0;
 818     my $remove_stopwords = C4::Context->preference("QueryRemoveStopwords") || 0;
 819
 820     # no stemming/weight/fuzzy in NoZebra
 821     if ( C4::Context->preference("NoZebra") ) {
 822         $stemming      = 0;
 823         $weight_fields = 0;
 824         $fuzzy_enabled = 0;
 825     }
 826
 827     my $query        = $operands[0];
 828     my $simple_query = $operands[0];
 829
 830     # initialize the variables we're passing back
 831     my $query_cgi;
 832     my $query_desc;
 833     my $query_type;
 834
 835     my $limit;
 836     my $limit_cgi;
 837     my $limit_desc;
 838
 839     my $stopwords_removed;    # flag to determine if stopwords have been removed
 840
 841 # for handling ccl, cql, pqf queries in diagnostic mode, skip the rest of the steps
 842 # DIAGNOSTIC ONLY!!
 843     if ( $query =~ /^ccl=/ ) {
 844         return ( undef, $', $', "q=ccl=$'", $', '', '', '', '', 'ccl' );
 845     }
 846     if ( $query =~ /^cql=/ ) {
 847         return ( undef, $', $', "q=cql=$'", $', '', '', '', '', 'cql' );
 848     }
 849     if ( $query =~ /^pqf=/ ) {
 850         return ( undef, $', $', "q=pqf=$'", $', '', '', '', '', 'pqf' );
 851     }
 852
 853     # pass nested queries directly
 854     # FIXME: need better handling of some of these variables in this case
 855     if ( $query =~ /(\(|\))/ ) {
 856         return (
 857             undef,              $query, $simple_query, $query_cgi,
 858             $query,             $limit, $limit_cgi,    $limit_desc,
 859             $stopwords_removed, 'ccl'
 860         );
 861     }
 862
 863 # Form-based queries are non-nested and fixed depth, so we can easily modify the incoming
 864 # query operands and indexes and add stemming, truncation, field weighting, etc.
 865 # Once we do so, we'll end up with a value in $query, just like if we had an
 866 # incoming $query from the user
 867     else {
 868         $query = ""
 869           ; # clear it out so we can populate properly with field-weighted, stemmed, etc. query
 870         my $previous_operand
 871           ;    # a flag used to keep track if there was a previous query
 872                # if there was, we can apply the current operator
 873                # for every operand
 874         for ( my $i = 0 ; $i <= @operands ; $i++ ) {
 875
 876             # COMBINE OPERANDS, INDEXES AND OPERATORS
 877             if ( $operands[$i] ) {
 878
 879               # A flag to determine whether or not to add the index to the query
 880                 my $indexes_set;
 881
 882 # If the user is sophisticated enough to specify an index, turn off field weighting, stemming, and stopword handling
 883                 if ( $operands[$i] =~ /(:|=)/ || $scan ) {
 884                     $weight_fields    = 0;
 885                     $stemming         = 0;
 886                     $remove_stopwords = 0;
 887                 }
 888                 my $operand = $operands[$i];
 889                 my $index   = $indexes[$i];
 890
 891                 # Add index-specific attributes
 892                 # Date of Publication
 893                 if ( $index eq 'yr' ) {
 894                     $index .= ",st-numeric";
 895                     $indexes_set++;
 896                                         $stemming = $auto_truncation = $weight_fields = $fuzzy_enabled = $remove_stopwords = 0;
 897                 }
 898
 899                 # Date of Acquisition
 900                 elsif ( $index eq 'acqdate' ) {
 901                     $index .= ",st-date-normalized";
 902                     $indexes_set++;
 903                                         $stemming = $auto_truncation = $weight_fields = $fuzzy_enabled = $remove_stopwords = 0;
 904                 }
 905                 # ISBN,ISSN,Standard Number, don't need special treatment
 906                 elsif ( $index eq 'nb' || $index eq 'ns' ) {
 907                     $indexes_set++;
 908                     (
 909                         $stemming,      $auto_truncation,
 910                         $weight_fields, $fuzzy_enabled,
 911                         $remove_stopwords
 912                     ) = ( 0, 0, 0, 0, 0 );
 913
 914                 }
 915                 # Set default structure attribute (word list)
 916                 my $struct_attr;
 917                 unless ( $indexes_set || !$index || $index =~ /(st-|phr|ext|wrdl)/ ) {
 918                     $struct_attr = ",wrdl";
 919                 }
 920
 921                 # Some helpful index variants
 922                 my $index_plus       = $index . $struct_attr . ":" if $index;
 923                 my $index_plus_comma = $index . $struct_attr . "," if $index;
 924
 925                 # Remove Stopwords
 926                 if ($remove_stopwords) {
 927                     ( $operand, $stopwords_removed ) =
 928                       _remove_stopwords( $operand, $index );
 929                     warn "OPERAND w/out STOPWORDS: >$operand<" if $DEBUG;
 930                     warn "REMOVED STOPWORDS: @$stopwords_removed"
 931                       if ( $stopwords_removed && $DEBUG );
 932                 }
 933
 934                 if ($auto_truncation){
 935                                         $operand=~join(" ",map{ "$_*" }split (/\s+/,$operand));
 936                                 }
 937
 938                 # Detect Truncation
 939                 my $truncated_operand;
 940                 my( $nontruncated, $righttruncated, $lefttruncated,
 941                     $rightlefttruncated, $regexpr
 942                 ) = _detect_truncation( $operand, $index );
 943                 warn
 944 "TRUNCATION: NON:>@$nontruncated< RIGHT:>@$righttruncated< LEFT:>@$lefttruncated< RIGHTLEFT:>@$rightlefttruncated< REGEX:>@$regexpr<"
 945                   if $DEBUG;
 946
 947                 # Apply Truncation
 948                 if (
 949                     scalar(@$righttruncated) + scalar(@$lefttruncated) +
 950                     scalar(@$rightlefttruncated) > 0 )
 951                 {
 952
 953                # Don't field weight or add the index to the query, we do it here
 954                     $indexes_set = 1;
 955                     undef $weight_fields;
 956                     my $previous_truncation_operand;
 957                     if (scalar @$nontruncated) {
 958                         $truncated_operand .= "$index_plus @$nontruncated ";
 959                         $previous_truncation_operand = 1;
 960                     }
 961                     if (scalar @$righttruncated) {
 962                         $truncated_operand .= "and " if $previous_truncation_operand;
 963                         $truncated_operand .= $index_plus_comma . "rtrn:@$righttruncated ";
 964                         $previous_truncation_operand = 1;
 965                     }
 966                     if (scalar @$lefttruncated) {
 967                         $truncated_operand .= "and " if $previous_truncation_operand;
 968                         $truncated_operand .= $index_plus_comma . "ltrn:@$lefttruncated ";
 969                         $previous_truncation_operand = 1;
 970                     }
 971                     if (scalar @$rightlefttruncated) {
 972                         $truncated_operand .= "and " if $previous_truncation_operand;
 973                         $truncated_operand .= $index_plus_comma . "rltrn:@$rightlefttruncated ";
 974                         $previous_truncation_operand = 1;
 975                     }
 976                 }
 977                 $operand = $truncated_operand if $truncated_operand;
 978                 warn "TRUNCATED OPERAND: >$truncated_operand<" if $DEBUG;
 979
 980                 # Handle Stemming
 981                 my $stemmed_operand;
 982                 $stemmed_operand = _build_stemmed_operand($operand, $lang)
 983                                                                                 if $stemming;
 984
 985                 warn "STEMMED OPERAND: >$stemmed_operand<" if $DEBUG;
 986
 987                 # Handle Field Weighting
 988                 my $weighted_operand;
 989                 if ($weight_fields) {
 990                     $weighted_operand = _build_weighted_query( $operand, $stemmed_operand, $index );
 991                     $operand = $weighted_operand;
 992                     $indexes_set = 1;
 993                 }
 994
 995                 warn "FIELD WEIGHTED OPERAND: >$weighted_operand<" if $DEBUG;
 996
 997                 # If there's a previous operand, we need to add an operator
 998                 if ($previous_operand) {
 999
1000                     # User-specified operator
1001                     if ( $operators[ $i - 1 ] ) {
1002                         $query     .= " $operators[$i-1] ";
1003                         $query     .= " $index_plus " unless $indexes_set;
1004                         $query     .= " $operand";
1005                         $query_cgi .= "&op=$operators[$i-1]";
1006                         $query_cgi .= "&idx=$index" if $index;
1007                         $query_cgi .= "&q=$operands[$i]" if $operands[$i];
1008                         $query_desc .=
1009                           " $operators[$i-1] $index_plus $operands[$i]";
1010                     }
1011
1012                     # Default operator is and
1013                     else {
1014                         $query      .= " and ";
1015                         $query      .= "$index_plus " unless $indexes_set;
1016                         $query      .= "$operand";
1017                         $query_cgi  .= "&op=and&idx=$index" if $index;
1018                         $query_cgi  .= "&q=$operands[$i]" if $operands[$i];
1019                         $query_desc .= " and $index_plus $operands[$i]";
1020                     }
1021                 }
1022
1023                 # There isn't a pervious operand, don't need an operator
1024                 else {
1025
1026                     # Field-weighted queries already have indexes set
1027                     $query .= " $index_plus " unless $indexes_set;
1028                     $query .= $operand;
1029                     $query_desc .= " $index_plus $operands[$i]";
1030                     $query_cgi  .= "&idx=$index" if $index;
1031                     $query_cgi  .= "&q=$operands[$i]" if $operands[$i];
1032                     $previous_operand = 1;
1033                 }
1034             }    #/if $operands
1035         }    # /for
1036     }
1037     warn "QUERY BEFORE LIMITS: >$query<" if $DEBUG;
1038
1039     # add limits
1040     my $group_OR_limits;
1041     my $availability_limit;
1042     foreach my $this_limit (@limits) {
1043 #        if ( $this_limit =~ /available/ ) {
1044 #
1045 ## 'available' is defined as (items.onloan is NULL) and (items.itemlost = 0)
1046 ## In English:
1047 ## all records not indexed in the onloan register (zebra) and all records with a value of lost equal to 0
1048 #            $availability_limit .=
1049 #"( ( allrecords,AlwaysMatches='' not onloan,AlwaysMatches='') and (lost,st-numeric=0) )"; #or ( allrecords,AlwaysMatches='' not lost,AlwaysMatches='')) )";
1050 #            $limit_cgi  .= "&limit=available";
1051 #            $limit_desc .= "";
1052 #        }
1053 #
1054         # group_OR_limits, prefixed by mc-
1055         # OR every member of the group
1056 #        elsif ( $this_limit =~ /mc/ ) {
1057         if ( $this_limit =~ /mc/ ) {
1058             $group_OR_limits .= " or " if $group_OR_limits;
1059             $limit_desc      .= " or " if $group_OR_limits;
1060             $group_OR_limits .= "$this_limit";
1061             $limit_cgi       .= "&limit=$this_limit";
1062             $limit_desc      .= " $this_limit";
1063         }
1064
1065         # Regular old limits
1066         else {
1067             $limit .= " and " if $limit || $query;
1068             $limit      .= "$this_limit";
1069             $limit_cgi  .= "&limit=$this_limit";
1070             if ($this_limit =~ /^branch:(.+)/) {
1071                 my $branchcode = $1;
1072                 my $branchname = GetBranchName($branchcode);
1073                 if (defined $branchname) {
1074                     $limit_desc .= " branch:$branchname";
1075                 } else {
1076                     $limit_desc .= " $this_limit";
1077                 }
1078             } else {
1079                 $limit_desc .= " $this_limit";
1080             }
1081         }
1082     }
1083     if ($group_OR_limits) {
1084         $limit .= " and " if ( $query || $limit );
1085         $limit .= "($group_OR_limits)";
1086     }
1087     if ($availability_limit) {
1088         $limit .= " and " if ( $query || $limit );
1089         $limit .= "($availability_limit)";
1090     }
1091
1092     # Normalize the query and limit strings
1093     $query =~ s/:/=/g;
1094     $limit =~ s/:/=/g;
1095     for ( $query, $query_desc, $limit, $limit_desc ) {
1096         s/  / /g;    # remove extra spaces
1097         s/^ //g;     # remove any beginning spaces
1098         s/ $//g;     # remove any ending spaces
1099         s/==/=/g;    # remove double == from query
1100     }
1101     $query_cgi =~ s/^&//; # remove unnecessary & from beginning of the query cgi
1102
1103     for ($query_cgi,$simple_query) {
1104         s/"//g;
1105     }
1106     # append the limit to the query
1107     $query .= " " . $limit;
1108
1109     # Warnings if DEBUG
1110     if ($DEBUG) {
1111         warn "QUERY:" . $query;
1112         warn "QUERY CGI:" . $query_cgi;
1113         warn "QUERY DESC:" . $query_desc;
1114         warn "LIMIT:" . $limit;
1115         warn "LIMIT CGI:" . $limit_cgi;
1116         warn "LIMIT DESC:" . $limit_desc;
1117         warn "---------\nLeave buildQuery\n---------";
1118     }
1119     return (
1120         undef,              $query, $simple_query, $query_cgi,
1121         $query_desc,        $limit, $limit_cgi,    $limit_desc,
1122         $stopwords_removed, $query_type
1123     );
1124 }
1125
1126 =head2 searchResults
1127
1128 Format results in a form suitable for passing to the template
1129
1130 =cut
1131
1132 # IMO this subroutine is pretty messy still -- it's responsible for
1133 # building the HTML output for the template
1134 sub searchResults {
1135     my ( $searchdesc, $hits, $results_per_page, $offset, $scan, @marcresults, $hidelostitems ) = @_;
1136     my $dbh = C4::Context->dbh;
1137     my @newresults;
1138
1139     #Build branchnames hash
1140     #find branchname
1141     #get branch information.....
1142     my %branches;
1143     my $bsth =$dbh->prepare("SELECT branchcode,branchname FROM branches"); # FIXME : use C4::Branch::GetBranches
1144     $bsth->execute();
1145     while ( my $bdata = $bsth->fetchrow_hashref ) {
1146         $branches{ $bdata->{'branchcode'} } = $bdata->{'branchname'};
1147     }
1148 # FIXME - We build an authorised values hash here, using the default framework
1149 # though it is possible to have different authvals for different fws.
1150
1151     my $shelflocations =GetKohaAuthorisedValues('items.location','');
1152
1153     # get notforloan authorised value list (see $shelflocations  FIXME)
1154     my $notforloan_authorised_value = GetAuthValCode('items.notforloan','');
1155
1156     #Build itemtype hash
1157     #find itemtype & itemtype image
1158     my %itemtypes;
1159     $bsth =
1160       $dbh->prepare(
1161         "SELECT itemtype,description,imageurl,summary,notforloan FROM itemtypes"
1162       );
1163     $bsth->execute();
1164     while ( my $bdata = $bsth->fetchrow_hashref ) {
1165                 foreach (qw(description imageurl summary notforloan)) {
1166                 $itemtypes{ $bdata->{'itemtype'} }->{$_} = $bdata->{$_};
1167                 }
1168     }
1169
1170     #search item field code
1171     my $sth =
1172       $dbh->prepare(
1173 "SELECT tagfield FROM marc_subfield_structure WHERE kohafield LIKE 'items.itemnumber'"
1174       );
1175     $sth->execute;
1176     my ($itemtag) = $sth->fetchrow;
1177
1178     ## find column names of items related to MARC
1179     my $sth2 = $dbh->prepare("SHOW COLUMNS FROM items");
1180     $sth2->execute;
1181     my %subfieldstosearch;
1182     while ( ( my $column ) = $sth2->fetchrow ) {
1183         my ( $tagfield, $tagsubfield ) =
1184           &GetMarcFromKohaField( "items." . $column, "" );
1185         $subfieldstosearch{$column} = $tagsubfield;
1186     }
1187
1188     # handle which records to actually retrieve
1189     my $times;
1190     if ( $hits && $offset + $results_per_page <= $hits ) {
1191         $times = $offset + $results_per_page;
1192     }
1193     else {
1194         $times = $hits;  # FIXME: if $hits is undefined, why do we want to equal it?
1195     }
1196
1197         my $marcflavour = C4::Context->preference("marcflavour");
1198     # We get the biblionumber position in MARC
1199     my ($bibliotag,$bibliosubf)=GetMarcFromKohaField('biblio.biblionumber','');
1200     my $fw;
1201
1202     # loop through all of the records we've retrieved
1203     for ( my $i = $offset ; $i <= $times - 1 ; $i++ ) {
1204         my $marcrecord = MARC::File::USMARC::decode( $marcresults[$i] );
1205
1206         if ($bibliotag<10){
1207             $fw = GetFrameworkCode($marcrecord->field($bibliotag)->data);
1208         }else{
1209             $fw = GetFrameworkCode($marcrecord->subfield($bibliotag,$bibliosubf));
1210         }
1211
1212         my $oldbiblio = TransformMarcToKoha( $dbh, $marcrecord, $fw );
1213         $oldbiblio->{subtitle} = GetRecordValue('subtitle', $marcrecord, $fw);
1214         $oldbiblio->{result_number} = $i + 1;
1215
1216         # add imageurl to itemtype if there is one
1217         $oldbiblio->{imageurl} = getitemtypeimagelocation( 'opac', $itemtypes{ $oldbiblio->{itemtype} }->{imageurl} );
1218
1219         $oldbiblio->{'authorised_value_images'}  = C4::Items::get_authorised_value_images( C4::Biblio::get_biblio_authorised_values( $oldbiblio->{'biblionumber'}, $marcrecord ) );
1220                 $oldbiblio->{normalized_upc}  = GetNormalizedUPC(       $marcrecord,$marcflavour);
1221                 $oldbiblio->{normalized_ean}  = GetNormalizedEAN(       $marcrecord,$marcflavour);
1222                 $oldbiblio->{normalized_oclc} = GetNormalizedOCLCNumber($marcrecord,$marcflavour);
1223                 $oldbiblio->{normalized_isbn} = GetNormalizedISBN(undef,$marcrecord,$marcflavour);
1224                 $oldbiblio->{content_identifier_exists} = 1 if ($oldbiblio->{normalized_isbn} or $oldbiblio->{normalized_oclc} or $oldbiblio->{normalized_ean} or $oldbiblio->{normalized_upc});
1225
1226                 # edition information, if any
1227         $oldbiblio->{edition} = $oldbiblio->{editionstatement};
1228                 $oldbiblio->{description} = $itemtypes{ $oldbiblio->{itemtype} }->{description};
1229  # Build summary if there is one (the summary is defined in the itemtypes table)
1230  # FIXME: is this used anywhere, I think it can be commented out? -- JF
1231         if ( $itemtypes{ $oldbiblio->{itemtype} }->{summary} ) {
1232             my $summary = $itemtypes{ $oldbiblio->{itemtype} }->{summary};
1233             my @fields  = $marcrecord->fields();
1234
1235             my $newsummary;
1236             foreach my $line ( "$summary\n" =~ /(.*)\n/g ){
1237                 my $tags = {};
1238                 foreach my $tag ( $line =~ /\[(\d{3}[\w|\d])\]/ ) {
1239                     $tag =~ /(.{3})(.)/;
1240                     if($marcrecord->field($1)){
1241                         my @abc = $marcrecord->field($1)->subfield($2);
1242                         $tags->{$tag} = $#abc + 1 ;
1243                     }
1244                 }
1245
1246                 # We catch how many times to repeat this line
1247                 my $max = 0;
1248                 foreach my $tag (keys(%$tags)){
1249                     $max = $tags->{$tag} if($tags->{$tag} > $max);
1250                  }
1251
1252                 # we replace, and repeat each line
1253                 for (my $i = 0 ; $i < $max ; $i++){
1254                     my $newline = $line;
1255
1256                     foreach my $tag ( $newline =~ /\[(\d{3}[\w|\d])\]/g ) {
1257                         $tag =~ /(.{3})(.)/;
1258
1259                         if($marcrecord->field($1)){
1260                             my @repl = $marcrecord->field($1)->subfield($2);
1261                             my $subfieldvalue = $repl[$i];
1262
1263                             if (! utf8::is_utf8($subfieldvalue)) {
1264                                 utf8::decode($subfieldvalue);
1265                             }
1266
1267                              $newline =~ s/\[$tag\]/$subfieldvalue/g;
1268                         }
1269                     }
1270                     $newsummary .= "$newline\n";
1271                 }
1272             }
1273
1274             $newsummary =~ s/\[(.*?)]//g;
1275             $newsummary =~ s/\n/<br\/>/g;
1276             $oldbiblio->{summary} = $newsummary;
1277         }
1278
1279         # Pull out the items fields
1280         my @fields = $marcrecord->field($itemtag);
1281
1282         # Setting item statuses for display
1283         my @available_items_loop;
1284         my @onloan_items_loop;
1285         my @other_items_loop;
1286
1287         my $available_items;
1288         my $onloan_items;
1289         my $other_items;
1290
1291         my $ordered_count         = 0;
1292         my $available_count       = 0;
1293         my $onloan_count          = 0;
1294         my $longoverdue_count     = 0;
1295         my $other_count           = 0;
1296         my $wthdrawn_count        = 0;
1297         my $itemlost_count        = 0;
1298         my $itembinding_count     = 0;
1299         my $itemdamaged_count     = 0;
1300         my $item_in_transit_count = 0;
1301         my $can_place_holds       = 0;
1302         my $items_count           = scalar(@fields);
1303         my $maxitems =
1304           ( C4::Context->preference('maxItemsinSearchResults') )
1305           ? C4::Context->preference('maxItemsinSearchResults') - 1
1306           : 1;
1307
1308         # loop through every item
1309         foreach my $field (@fields) {
1310             my $item;
1311
1312             # populate the items hash
1313             foreach my $code ( keys %subfieldstosearch ) {
1314                 $item->{$code} = $field->subfield( $subfieldstosearch{$code} );
1315             }
1316
1317                         my $hbranch     = C4::Context->preference('HomeOrHoldingBranch') eq 'homebranch' ? 'homebranch'    : 'holdingbranch';
1318                         my $otherbranch = C4::Context->preference('HomeOrHoldingBranch') eq 'homebranch' ? 'holdingbranch' : 'homebranch';
1319             # set item's branch name, use HomeOrHoldingBranch syspref first, fall back to the other one
1320             if ($item->{$hbranch}) {
1321                 $item->{'branchname'} = $branches{$item->{$hbranch}};
1322             }
1323             elsif ($item->{$otherbranch}) {     # Last resort
1324                 $item->{'branchname'} = $branches{$item->{$otherbranch}};
1325             }
1326
1327                         my $prefix = $item->{$hbranch} . '--' . $item->{location} . $item->{itype} . $item->{itemcallnumber};
1328 # For each grouping of items (onloan, available, unavailable), we build a key to store relevant info about that item
1329             if ( $item->{onloan} ) {
1330                 $onloan_count++;
1331                                 my $key = $prefix . $item->{onloan} . $item->{barcode};
1332                                 $onloan_items->{$key}->{due_date} = format_date($item->{onloan});
1333                                 $onloan_items->{$key}->{count}++ if $item->{$hbranch};
1334                                 $onloan_items->{$key}->{branchname} = $item->{branchname};
1335                                 $onloan_items->{$key}->{location} = $shelflocations->{ $item->{location} };
1336                                 $onloan_items->{$key}->{itemcallnumber} = $item->{itemcallnumber};
1337                                 $onloan_items->{$key}->{imageurl} = getitemtypeimagelocation( 'opac', $itemtypes{ $item->{itype} }->{imageurl} );
1338                 # if something's checked out and lost, mark it as 'long overdue'
1339                 if ( $item->{itemlost} ) {
1340                     $onloan_items->{$prefix}->{longoverdue}++;
1341                     $longoverdue_count++;
1342                 } else {        # can place holds as long as item isn't lost
1343                     $can_place_holds = 1;
1344                 }
1345             }
1346
1347          # items not on loan, but still unavailable ( lost, withdrawn, damaged )
1348             else {
1349
1350                 # item is on order
1351                 if ( $item->{notforloan} == -1 ) {
1352                     $ordered_count++;
1353                 }
1354
1355                 # is item in transit?
1356                 my $transfertwhen = '';
1357                 my ($transfertfrom, $transfertto);
1358
1359                 unless ($item->{wthdrawn}
1360                         || $item->{itemlost}
1361                         || $item->{damaged}
1362                         || $item->{notforloan}
1363                         || $items_count > 20) {
1364
1365                     # A couple heuristics to limit how many times
1366                     # we query the database for item transfer information, sacrificing
1367                     # accuracy in some cases for speed;
1368                     #
1369                     # 1. don't query if item has one of the other statuses
1370                     # 2. don't check transit status if the bib has
1371                     #    more than 20 items
1372                     #
1373                     # FIXME: to avoid having the query the database like this, and to make
1374                     #        the in transit status count as unavailable for search limiting,
1375                     #        should map transit status to record indexed in Zebra.
1376                     #
1377                     ($transfertwhen, $transfertfrom, $transfertto) = C4::Circulation::GetTransfers($item->{itemnumber});
1378                 }
1379
1380                 # item is withdrawn, lost or damaged
1381                 if (   $item->{wthdrawn}
1382                     || $item->{itemlost}
1383                     || $item->{damaged}
1384                     || $item->{notforloan}
1385                     || ($transfertwhen ne ''))
1386                 {
1387                     $wthdrawn_count++        if $item->{wthdrawn};
1388                     $itemlost_count++        if $item->{itemlost};
1389                     $itemdamaged_count++     if $item->{damaged};
1390                     $item_in_transit_count++ if $transfertwhen ne '';
1391                     $item->{status} = $item->{wthdrawn} . "-" . $item->{itemlost} . "-" . $item->{damaged} . "-" . $item->{notforloan};
1392                     $other_count++;
1393
1394                                         my $key = $prefix . $item->{status};
1395                                         foreach (qw(wthdrawn itemlost damaged branchname itemcallnumber)) {
1396                         $other_items->{$key}->{$_} = $item->{$_};
1397                                         }
1398                     $other_items->{$key}->{intransit} = ($transfertwhen ne '') ? 1 : 0;
1399                                         $other_items->{$key}->{notforloan} = GetAuthorisedValueDesc('','',$item->{notforloan},'','',$notforloan_authorised_value) if $notforloan_authorised_value;
1400                                         $other_items->{$key}->{count}++ if $item->{$hbranch};
1401                                         $other_items->{$key}->{location} = $shelflocations->{ $item->{location} };
1402                                         $other_items->{$key}->{imageurl} = getitemtypeimagelocation( 'opac', $itemtypes{ $item->{itype} }->{imageurl} );
1403                 }
1404                 # item is available
1405                 else {
1406                     $can_place_holds = 1;
1407                     $available_count++;
1408                                         $available_items->{$prefix}->{count}++ if $item->{$hbranch};
1409                                         foreach (qw(branchname itemcallnumber)) {
1410                         $available_items->{$prefix}->{$_} = $item->{$_};
1411                                         }
1412                                         $available_items->{$prefix}->{location} = $shelflocations->{ $item->{location} };
1413                                         $available_items->{$prefix}->{imageurl} = getitemtypeimagelocation( 'opac', $itemtypes{ $item->{itype} }->{imageurl} );
1414                 }
1415             }
1416         }    # notforloan, item level and biblioitem level
1417         my ( $availableitemscount, $onloanitemscount, $otheritemscount );
1418         $maxitems =
1419           ( C4::Context->preference('maxItemsinSearchResults') )
1420           ? C4::Context->preference('maxItemsinSearchResults') - 1
1421           : 1;
1422         for my $key ( sort keys %$onloan_items ) {
1423             (++$onloanitemscount > $maxitems) and last;
1424             push @onloan_items_loop, $onloan_items->{$key};
1425         }
1426         for my $key ( sort keys %$other_items ) {
1427             (++$otheritemscount > $maxitems) and last;
1428             push @other_items_loop, $other_items->{$key};
1429         }
1430         for my $key ( sort keys %$available_items ) {
1431             (++$availableitemscount > $maxitems) and last;
1432             push @available_items_loop, $available_items->{$key}
1433         }
1434
1435         # XSLT processing of some stuff
1436         if (C4::Context->preference("XSLTResultsDisplay") && !$scan) {
1437             $oldbiblio->{XSLTResultsRecord} = XSLTParse4Display(
1438                 $oldbiblio->{biblionumber}, $marcrecord, 'Results' );
1439         }
1440
1441         # last check for norequest : if itemtype is notforloan, it can't be reserved either, whatever the items
1442         $can_place_holds = 0
1443           if $itemtypes{ $oldbiblio->{itemtype} }->{notforloan};
1444         $oldbiblio->{norequests} = 1 unless $can_place_holds;
1445         $oldbiblio->{itemsplural}          = 1 if $items_count > 1;
1446         $oldbiblio->{items_count}          = $items_count;
1447         $oldbiblio->{available_items_loop} = \@available_items_loop;
1448         $oldbiblio->{onloan_items_loop}    = \@onloan_items_loop;
1449         $oldbiblio->{other_items_loop}     = \@other_items_loop;
1450         $oldbiblio->{availablecount}       = $available_count;
1451         $oldbiblio->{availableplural}      = 1 if $available_count > 1;
1452         $oldbiblio->{onloancount}          = $onloan_count;
1453         $oldbiblio->{onloanplural}         = 1 if $onloan_count > 1;
1454         $oldbiblio->{othercount}           = $other_count;
1455         $oldbiblio->{otherplural}          = 1 if $other_count > 1;
1456         $oldbiblio->{wthdrawncount}        = $wthdrawn_count;
1457         $oldbiblio->{itemlostcount}        = $itemlost_count;
1458         $oldbiblio->{damagedcount}         = $itemdamaged_count;
1459         $oldbiblio->{intransitcount}       = $item_in_transit_count;
1460         $oldbiblio->{orderedcount}         = $ordered_count;
1461         $oldbiblio->{isbn} =~
1462           s/-//g;    # deleting - in isbn to enable amazon content
1463         push( @newresults, $oldbiblio )
1464             if(not $hidelostitems
1465                or (($items_count > $itemlost_count )
1466                     && $hidelostitems));
1467     }
1468
1469     return @newresults;
1470 }
1471
1472 =head2 SearchAcquisitions
1473     Search for acquisitions
1474 =cut
1475
1476 sub SearchAcquisitions{
1477     my ($datebegin, $dateend, $itemtypes,$criteria, $orderby) = @_;
1478
1479     my $dbh=C4::Context->dbh;
1480     # Variable initialization
1481     my $str=qq|
1482     SELECT marcxml
1483     FROM biblio
1484     LEFT JOIN biblioitems ON biblioitems.biblionumber=biblio.biblionumber
1485     LEFT JOIN items ON items.biblionumber=biblio.biblionumber
1486     WHERE dateaccessioned BETWEEN ? AND ?
1487     |;
1488
1489     my (@params,@loopcriteria);
1490
1491     push @params, $datebegin->output("iso");
1492     push @params, $dateend->output("iso");
1493
1494     if (scalar(@$itemtypes)>0 and $criteria ne "itemtype" ){
1495         if(C4::Context->preference("item-level_itypes")){
1496             $str .= "AND items.itype IN (?".( ',?' x scalar @$itemtypes - 1 ).") ";
1497         }else{
1498             $str .= "AND biblioitems.itemtype IN (?".( ',?' x scalar @$itemtypes - 1 ).") ";
1499         }
1500         push @params, @$itemtypes;
1501     }
1502
1503     if ($criteria =~/itemtype/){
1504         if(C4::Context->preference("item-level_itypes")){
1505             $str .= "AND items.itype=? ";
1506         }else{
1507             $str .= "AND biblioitems.itemtype=? ";
1508         }
1509
1510         if(scalar(@$itemtypes) == 0){
1511             my $itypes = GetItemTypes();
1512             for my $key (keys %$itypes){
1513                 push @$itemtypes, $key;
1514             }
1515         }
1516
1517         @loopcriteria= @$itemtypes;
1518     }elsif ($criteria=~/itemcallnumber/){
1519         $str .= "AND (items.itemcallnumber LIKE CONCAT(?,'%')
1520                  OR items.itemcallnumber is NULL
1521                  OR items.itemcallnumber = '')";
1522
1523         @loopcriteria = ("AA".."ZZ", "") unless (scalar(@loopcriteria)>0);
1524     }else {
1525         $str .= "AND biblio.title LIKE CONCAT(?,'%') ";
1526         @loopcriteria = ("A".."z") unless (scalar(@loopcriteria)>0);
1527     }
1528
1529     if ($orderby =~ /date_desc/){
1530         $str.=" ORDER BY dateaccessioned DESC";
1531     } else {
1532         $str.=" ORDER BY title";
1533     }
1534
1535     my $qdataacquisitions=$dbh->prepare($str);
1536
1537     my @loopacquisitions;
1538     foreach my $value(@loopcriteria){
1539         push @params,$value;
1540         my %cell;
1541         $cell{"title"}=$value;
1542         $cell{"titlecode"}=$value;
1543
1544         eval{$qdataacquisitions->execute(@params);};
1545
1546         if ($@){ warn "recentacquisitions Error :$@";}
1547         else {
1548             my @loopdata;
1549             while (my $data=$qdataacquisitions->fetchrow_hashref){
1550                 push @loopdata, {"summary"=>GetBiblioSummary( $data->{'marcxml'} ) };
1551             }
1552             $cell{"loopdata"}=\@loopdata;
1553         }
1554         push @loopacquisitions,\%cell if (scalar(@{$cell{loopdata}})>0);
1555         pop @params;
1556     }
1557     $qdataacquisitions->finish;
1558     return \@loopacquisitions;
1559 }
1560 #----------------------------------------------------------------------
1561 #
1562 # Non-Zebra GetRecords#
1563 #----------------------------------------------------------------------
1564
1565 =head2 NZgetRecords
1566
1567   NZgetRecords has the same API as zera getRecords, even if some parameters are not managed
1568
1569 =cut
1570
1571 sub NZgetRecords {
1572     my (
1573         $query,            $simple_query, $sort_by_ref,    $servers_ref,
1574         $results_per_page, $offset,       $expanded_facet, $branches,
1575         $query_type,       $scan
1576     ) = @_;
1577     warn "query =$query" if $DEBUG;
1578     my $result = NZanalyse($query);
1579     warn "results =$result" if $DEBUG;
1580     return ( undef,
1581         NZorder( $result, @$sort_by_ref[0], $results_per_page, $offset ),
1582         undef );
1583 }
1584
1585 =head2 NZanalyse
1586
1587   NZanalyse : get a CQL string as parameter, and returns a list of biblionumber;title,biblionumber;title,...
1588   the list is built from an inverted index in the nozebra SQL table
1589   note that title is here only for convenience : the sorting will be very fast when requested on title
1590   if the sorting is requested on something else, we will have to reread all results, and that may be longer.
1591
1592 =cut
1593
1594 sub NZanalyse {
1595     my ( $string, $server ) = @_;
1596 #     warn "---------"       if $DEBUG;
1597     warn " NZanalyse" if $DEBUG;
1598 #     warn "---------"       if $DEBUG;
1599
1600  # $server contains biblioserver or authorities, depending on what we search on.
1601  #warn "querying : $string on $server";
1602     $server = 'biblioserver' unless $server;
1603
1604 # if we have a ", replace the content to discard temporarily any and/or/not inside
1605     my $commacontent;
1606     if ( $string =~ /"/ ) {
1607         $string =~ s/"(.*?)"/__X__/;
1608         $commacontent = $1;
1609         warn "commacontent : $commacontent" if $DEBUG;
1610     }
1611
1612 # split the query string in 3 parts : X AND Y means : $left="X", $operand="AND" and $right="Y"
1613 # then, call again NZanalyse with $left and $right
1614 # (recursive until we find a leaf (=> something without and/or/not)
1615 # delete repeated operator... Would then go in infinite loop
1616     while ( $string =~ s/( and| or| not| AND| OR| NOT)\1/$1/g ) {
1617     }
1618
1619     #process parenthesis before.
1620     if ( $string =~ /^\s*\((.*)\)(( and | or | not | AND | OR | NOT )(.*))?/ ) {
1621         my $left     = $1;
1622         my $right    = $4;
1623         my $operator = lc($3);   # FIXME: and/or/not are operators, not operands
1624         warn
1625 "dealing w/parenthesis before recursive sub call. left :$left operator:$operator right:$right"
1626           if $DEBUG;
1627         my $leftresult = NZanalyse( $left, $server );
1628         if ($operator) {
1629             my $rightresult = NZanalyse( $right, $server );
1630
1631             # OK, we have the results for right and left part of the query
1632             # depending of operand, intersect, union or exclude both lists
1633             # to get a result list
1634             if ( $operator eq ' and ' ) {
1635                 return NZoperatorAND($leftresult,$rightresult);
1636             }
1637             elsif ( $operator eq ' or ' ) {
1638
1639                 # just merge the 2 strings
1640                 return $leftresult . $rightresult;
1641             }
1642             elsif ( $operator eq ' not ' ) {
1643                 return NZoperatorNOT($leftresult,$rightresult);
1644             }
1645         }
1646         else {
1647 # this error is impossible, because of the regexp that isolate the operand, but just in case...
1648             return $leftresult;
1649         }
1650     }
1651     warn "string :" . $string if $DEBUG;
1652     my $left = "";
1653     my $right = "";
1654     my $operator = "";
1655     if ($string =~ /(.*?)( and | or | not | AND | OR | NOT )(.*)/) {
1656         $left     = $1;
1657         $right    = $3;
1658         $operator = lc($2);    # FIXME: and/or/not are operators, not operands
1659     }
1660     warn "no parenthesis. left : $left operator: $operator right: $right"
1661       if $DEBUG;
1662
1663     # it's not a leaf, we have a and/or/not
1664     if ($operator) {
1665
1666         # reintroduce comma content if needed
1667         $right =~ s/__X__/"$commacontent"/ if $commacontent;
1668         $left  =~ s/__X__/"$commacontent"/ if $commacontent;
1669         warn "node : $left / $operator / $right\n" if $DEBUG;
1670         my $leftresult  = NZanalyse( $left,  $server );
1671         my $rightresult = NZanalyse( $right, $server );
1672         warn " leftresult : $leftresult" if $DEBUG;
1673         warn " rightresult : $rightresult" if $DEBUG;
1674         # OK, we have the results for right and left part of the query
1675         # depending of operand, intersect, union or exclude both lists
1676         # to get a result list
1677         if ( $operator eq ' and ' ) {
1678             warn "NZAND";
1679             return NZoperatorAND($leftresult,$rightresult);
1680         }
1681         elsif ( $operator eq ' or ' ) {
1682
1683             # just merge the 2 strings
1684             return $leftresult . $rightresult;
1685         }
1686         elsif ( $operator eq ' not ' ) {
1687             return NZoperatorNOT($leftresult,$rightresult);
1688         }
1689         else {
1690
1691 # this error is impossible, because of the regexp that isolate the operand, but just in case...
1692             die "error : operand unknown : $operator for $string";
1693         }
1694
1695         # it's a leaf, do the real SQL query and return the result
1696     }
1697     else {
1698         $string =~ s/__X__/"$commacontent"/ if $commacontent;
1699         $string =~ s/-|\.|\?|,|;|!|'|\(|\)|\[|\]|{|}|"|&|\+|\*|\// /g;
1700         #remove trailing blank at the beginning
1701         $string =~ s/^ //g;
1702         warn "leaf:$string" if $DEBUG;
1703
1704         # parse the string in in operator/operand/value again
1705         my $left = "";
1706         my $operator = "";
1707         my $right = "";
1708         if ($string =~ /(.*)(>=|<=)(.*)/) {
1709             $left     = $1;
1710             $operator = $2;
1711             $right    = $3;
1712         } else {
1713             $left = $string;
1714         }
1715 #         warn "handling leaf... left:$left operator:$operator right:$right"
1716 #           if $DEBUG;
1717         unless ($operator) {
1718             if ($string =~ /(.*)(>|<|=)(.*)/) {
1719                 $left     = $1;
1720                 $operator = $2;
1721                 $right    = $3;
1722                 warn
1723     "handling unless (operator)... left:$left operator:$operator right:$right"
1724                 if $DEBUG;
1725             } else {
1726                 $left = $string;
1727             }
1728         }
1729         my $results;
1730
1731 # strip adv, zebra keywords, currently not handled in nozebra: wrdl, ext, phr...
1732         $left =~ s/ .*$//;
1733
1734         # automatic replace for short operators
1735         $left = 'title'            if $left =~ '^ti$';
1736         $left = 'author'           if $left =~ '^au$';
1737         $left = 'publisher'        if $left =~ '^pb$';
1738         $left = 'subject'          if $left =~ '^su$';
1739         $left = 'koha-Auth-Number' if $left =~ '^an$';
1740         $left = 'keyword'          if $left =~ '^kw$';
1741         $left = 'itemtype'         if $left =~ '^mc$'; # Fix for Bug 2599 - Search limits not working for NoZebra
1742         warn "handling leaf... left:$left operator:$operator right:$right" if $DEBUG;
1743         my $dbh = C4::Context->dbh;
1744         if ( $operator && $left ne 'keyword' ) {
1745             #do a specific search
1746             $operator = 'LIKE' if $operator eq '=' and $right =~ /%/;
1747             my $sth = $dbh->prepare(
1748 "SELECT biblionumbers,value FROM nozebra WHERE server=? AND indexname=? AND value $operator ?"
1749             );
1750             warn "$left / $operator / $right\n" if $DEBUG;
1751
1752             # split each word, query the DB and build the biblionumbers result
1753             #sanitizing leftpart
1754             $left =~ s/^\s+|\s+$//;
1755             foreach ( split / /, $right ) {
1756                 my $biblionumbers;
1757                 $_ =~ s/^\s+|\s+$//;
1758                 next unless $_;
1759                 warn "EXECUTE : $server, $left, $_" if $DEBUG;
1760                 $sth->execute( $server, $left, $_ )
1761                   or warn "execute failed: $!";
1762                 while ( my ( $line, $value ) = $sth->fetchrow ) {
1763
1764 # if we are dealing with a numeric value, use only numeric results (in case of >=, <=, > or <)
1765 # otherwise, fill the result
1766                     $biblionumbers .= $line
1767                       unless ( $right =~ /^\d+$/ && $value =~ /\D/ );
1768                     warn "result : $value "
1769                       . ( $right  =~ /\d/ ) . "=="
1770                       . ( $value =~ /\D/?$line:"" ) if $DEBUG;         #= $line";
1771                 }
1772
1773 # do a AND with existing list if there is one, otherwise, use the biblionumbers list as 1st result list
1774                 if ($results) {
1775                     warn "NZAND" if $DEBUG;
1776                     $results = NZoperatorAND($biblionumbers,$results);
1777                 } else {
1778                     $results = $biblionumbers;
1779                 }
1780             }
1781         }
1782         else {
1783       #do a complete search (all indexes), if index='kw' do complete search too.
1784             my $sth = $dbh->prepare(
1785 "SELECT biblionumbers FROM nozebra WHERE server=? AND value LIKE ?"
1786             );
1787
1788             # split each word, query the DB and build the biblionumbers result
1789             foreach ( split / /, $string ) {
1790                 next if C4::Context->stopwords->{ uc($_) };   # skip if stopword
1791                 warn "search on all indexes on $_" if $DEBUG;
1792                 my $biblionumbers;
1793                 next unless $_;
1794                 $sth->execute( $server, $_ );
1795                 while ( my $line = $sth->fetchrow ) {
1796                     $biblionumbers .= $line;
1797                 }
1798
1799 # do a AND with existing list if there is one, otherwise, use the biblionumbers list as 1st result list
1800                 if ($results) {
1801                     $results = NZoperatorAND($biblionumbers,$results);
1802                 }
1803                 else {
1804                     warn "NEW RES for $_ = $biblionumbers" if $DEBUG;
1805                     $results = $biblionumbers;
1806                 }
1807             }
1808         }
1809         warn "return : $results for LEAF : $string" if $DEBUG;
1810         return $results;
1811     }
1812     warn "---------\nLeave NZanalyse\n---------" if $DEBUG;
1813 }
1814
1815 sub NZoperatorAND{
1816     my ($rightresult, $leftresult)=@_;
1817
1818     my @leftresult = split /;/, $leftresult;
1819     warn " @leftresult / $rightresult \n" if $DEBUG;
1820
1821     #             my @rightresult = split /;/,$leftresult;
1822     my $finalresult;
1823
1824 # parse the left results, and if the biblionumber exist in the right result, save it in finalresult
1825 # the result is stored twice, to have the same weight for AND than OR.
1826 # example : TWO : 61,61,64,121 (two is twice in the biblio #61) / TOWER : 61,64,130
1827 # result : 61,61,61,61,64,64 for two AND tower : 61 has more weight than 64
1828     foreach (@leftresult) {
1829         my $value = $_;
1830         my $countvalue;
1831         ( $value, $countvalue ) = ( $1, $2 ) if ($value=~/(.*)-(\d+)$/);
1832         if ( $rightresult =~ /\Q$value\E-(\d+);/ ) {
1833             $countvalue = ( $1 > $countvalue ? $countvalue : $1 );
1834             $finalresult .=
1835                 "$value-$countvalue;$value-$countvalue;";
1836         }
1837     }
1838     warn "NZAND DONE : $finalresult \n" if $DEBUG;
1839     return $finalresult;
1840 }
1841
1842 sub NZoperatorOR{
1843     my ($rightresult, $leftresult)=@_;
1844     return $rightresult.$leftresult;
1845 }
1846
1847 sub NZoperatorNOT{
1848     my ($leftresult, $rightresult)=@_;
1849
1850     my @leftresult = split /;/, $leftresult;
1851
1852     #             my @rightresult = split /;/,$leftresult;
1853     my $finalresult;
1854     foreach (@leftresult) {
1855         my $value=$_;
1856         $value=$1 if $value=~m/(.*)-\d+$/;
1857         unless ($rightresult =~ "$value-") {
1858             $finalresult .= "$_;";
1859         }
1860     }
1861     return $finalresult;
1862 }
1863
1864 =head2 NZorder
1865
1866   $finalresult = NZorder($biblionumbers, $ordering,$results_per_page,$offset);
1867
1868   TODO :: Description
1869
1870 =cut
1871
1872 sub NZorder {
1873     my ( $biblionumbers, $ordering, $results_per_page, $offset ) = @_;
1874     warn "biblionumbers = $biblionumbers and ordering = $ordering\n" if $DEBUG;
1875
1876     # order title asc by default
1877     #     $ordering = '1=36 <i' unless $ordering;
1878     $results_per_page = 20 unless $results_per_page;
1879     $offset           = 0  unless $offset;
1880     my $dbh = C4::Context->dbh;
1881
1882     #
1883     # order by POPULARITY
1884     #
1885     if ( $ordering =~ /popularity/ ) {
1886         my %result;
1887         my %popularity;
1888
1889         # popularity is not in MARC record, it's builded from a specific query
1890         my $sth =
1891           $dbh->prepare("select sum(issues) from items where biblionumber=?");
1892         foreach ( split /;/, $biblionumbers ) {
1893             my ( $biblionumber, $title ) = split /,/, $_;
1894             $result{$biblionumber} = GetMarcBiblio($biblionumber);
1895             $sth->execute($biblionumber);
1896             my $popularity = $sth->fetchrow || 0;
1897
1898 # hint : the key is popularity.title because we can have
1899 # many results with the same popularity. In this case, sub-ordering is done by title
1900 # we also have biblionumber to avoid bug for 2 biblios with the same title & popularity
1901 # (un-frequent, I agree, but we won't forget anything that way ;-)
1902             $popularity{ sprintf( "%10d", $popularity ) . $title
1903                   . $biblionumber } = $biblionumber;
1904         }
1905
1906     # sort the hash and return the same structure as GetRecords (Zebra querying)
1907         my $result_hash;
1908         my $numbers = 0;
1909         if ( $ordering eq 'popularity_dsc' ) {    # sort popularity DESC
1910             foreach my $key ( sort { $b cmp $a } ( keys %popularity ) ) {
1911                 $result_hash->{'RECORDS'}[ $numbers++ ] =
1912                   $result{ $popularity{$key} }->as_usmarc();
1913             }
1914         }
1915         else {                                    # sort popularity ASC
1916             foreach my $key ( sort ( keys %popularity ) ) {
1917                 $result_hash->{'RECORDS'}[ $numbers++ ] =
1918                   $result{ $popularity{$key} }->as_usmarc();
1919             }
1920         }
1921         my $finalresult = ();
1922         $result_hash->{'hits'}         = $numbers;
1923         $finalresult->{'biblioserver'} = $result_hash;
1924         return $finalresult;
1925
1926         #
1927         # ORDER BY author
1928         #
1929     }
1930     elsif ( $ordering =~ /author/ ) {
1931         my %result;
1932         foreach ( split /;/, $biblionumbers ) {
1933             my ( $biblionumber, $title ) = split /,/, $_;
1934             my $record = GetMarcBiblio($biblionumber);
1935             my $author;
1936             if ( C4::Context->preference('marcflavour') eq 'UNIMARC' ) {
1937                 $author = $record->subfield( '200', 'f' );
1938                 $author = $record->subfield( '700', 'a' ) unless $author;
1939             }
1940             else {
1941                 $author = $record->subfield( '100', 'a' );
1942             }
1943
1944 # hint : the result is sorted by title.biblionumber because we can have X biblios with the same title
1945 # and we don't want to get only 1 result for each of them !!!
1946             $result{ $author . $biblionumber } = $record;
1947         }
1948
1949     # sort the hash and return the same structure as GetRecords (Zebra querying)
1950         my $result_hash;
1951         my $numbers = 0;
1952         if ( $ordering eq 'author_za' ) {    # sort by author desc
1953             foreach my $key ( sort { $b cmp $a } ( keys %result ) ) {
1954                 $result_hash->{'RECORDS'}[ $numbers++ ] =
1955                   $result{$key}->as_usmarc();
1956             }
1957         }
1958         else {                               # sort by author ASC
1959             foreach my $key ( sort ( keys %result ) ) {
1960                 $result_hash->{'RECORDS'}[ $numbers++ ] =
1961                   $result{$key}->as_usmarc();
1962             }
1963         }
1964         my $finalresult = ();
1965         $result_hash->{'hits'}         = $numbers;
1966         $finalresult->{'biblioserver'} = $result_hash;
1967         return $finalresult;
1968
1969         #
1970         # ORDER BY callnumber
1971         #
1972     }
1973     elsif ( $ordering =~ /callnumber/ ) {
1974         my %result;
1975         foreach ( split /;/, $biblionumbers ) {
1976             my ( $biblionumber, $title ) = split /,/, $_;
1977             my $record = GetMarcBiblio($biblionumber);
1978             my $callnumber;
1979             my $frameworkcode = GetFrameworkCode($biblionumber);
1980             my ( $callnumber_tag, $callnumber_subfield ) = GetMarcFromKohaField(  'items.itemcallnumber', $frameworkcode);
1981                ( $callnumber_tag, $callnumber_subfield ) = GetMarcFromKohaField('biblioitems.callnumber', $frameworkcode)
1982                 unless $callnumber_tag;
1983             if ( C4::Context->preference('marcflavour') eq 'UNIMARC' ) {
1984                 $callnumber = $record->subfield( '200', 'f' );
1985             } else {
1986                 $callnumber = $record->subfield( '100', 'a' );
1987             }
1988
1989 # hint : the result is sorted by title.biblionumber because we can have X biblios with the same title
1990 # and we don't want to get only 1 result for each of them !!!
1991             $result{ $callnumber . $biblionumber } = $record;
1992         }
1993
1994     # sort the hash and return the same structure as GetRecords (Zebra querying)
1995         my $result_hash;
1996         my $numbers = 0;
1997         if ( $ordering eq 'call_number_dsc' ) {    # sort by title desc
1998             foreach my $key ( sort { $b cmp $a } ( keys %result ) ) {
1999                 $result_hash->{'RECORDS'}[ $numbers++ ] =
2000                   $result{$key}->as_usmarc();
2001             }
2002         }
2003         else {                                     # sort by title ASC
2004             foreach my $key ( sort { $a cmp $b } ( keys %result ) ) {
2005                 $result_hash->{'RECORDS'}[ $numbers++ ] =
2006                   $result{$key}->as_usmarc();
2007             }
2008         }
2009         my $finalresult = ();
2010         $result_hash->{'hits'}         = $numbers;
2011         $finalresult->{'biblioserver'} = $result_hash;
2012         return $finalresult;
2013     }
2014     elsif ( $ordering =~ /pubdate/ ) {             #pub year
2015         my %result;
2016         foreach ( split /;/, $biblionumbers ) {
2017             my ( $biblionumber, $title ) = split /,/, $_;
2018             my $record = GetMarcBiblio($biblionumber);
2019             my ( $publicationyear_tag, $publicationyear_subfield ) =
2020               GetMarcFromKohaField( 'biblioitems.publicationyear', '' );
2021             my $publicationyear =
2022               $record->subfield( $publicationyear_tag,
2023                 $publicationyear_subfield );
2024
2025 # hint : the result is sorted by title.biblionumber because we can have X biblios with the same title
2026 # and we don't want to get only 1 result for each of them !!!
2027             $result{ $publicationyear . $biblionumber } = $record;
2028         }
2029
2030     # sort the hash and return the same structure as GetRecords (Zebra querying)
2031         my $result_hash;
2032         my $numbers = 0;
2033         if ( $ordering eq 'pubdate_dsc' ) {    # sort by pubyear desc
2034             foreach my $key ( sort { $b cmp $a } ( keys %result ) ) {
2035                 $result_hash->{'RECORDS'}[ $numbers++ ] =
2036                   $result{$key}->as_usmarc();
2037             }
2038         }
2039         else {                                 # sort by pub year ASC
2040             foreach my $key ( sort ( keys %result ) ) {
2041                 $result_hash->{'RECORDS'}[ $numbers++ ] =
2042                   $result{$key}->as_usmarc();
2043             }
2044         }
2045         my $finalresult = ();
2046         $result_hash->{'hits'}         = $numbers;
2047         $finalresult->{'biblioserver'} = $result_hash;
2048         return $finalresult;
2049
2050         #
2051         # ORDER BY title
2052         #
2053     }
2054     elsif ( $ordering =~ /title/ ) {
2055
2056 # the title is in the biblionumbers string, so we just need to build a hash, sort it and return
2057         my %result;
2058         foreach ( split /;/, $biblionumbers ) {
2059             my ( $biblionumber, $title ) = split /,/, $_;
2060
2061 # hint : the result is sorted by title.biblionumber because we can have X biblios with the same title
2062 # and we don't want to get only 1 result for each of them !!!
2063 # hint & speed improvement : we can order without reading the record
2064 # so order, and read records only for the requested page !
2065             $result{ $title . $biblionumber } = $biblionumber;
2066         }
2067
2068     # sort the hash and return the same structure as GetRecords (Zebra querying)
2069         my $result_hash;
2070         my $numbers = 0;
2071         if ( $ordering eq 'title_az' ) {    # sort by title desc
2072             foreach my $key ( sort ( keys %result ) ) {
2073                 $result_hash->{'RECORDS'}[ $numbers++ ] = $result{$key};
2074             }
2075         }
2076         else {                              # sort by title ASC
2077             foreach my $key ( sort { $b cmp $a } ( keys %result ) ) {
2078                 $result_hash->{'RECORDS'}[ $numbers++ ] = $result{$key};
2079             }
2080         }
2081
2082         # limit the $results_per_page to result size if it's more
2083         $results_per_page = $numbers - 1 if $numbers < $results_per_page;
2084
2085         # for the requested page, replace biblionumber by the complete record
2086         # speed improvement : avoid reading too much things
2087         for (
2088             my $counter = $offset ;
2089             $counter <= $offset + $results_per_page ;
2090             $counter++
2091           )
2092         {
2093             $result_hash->{'RECORDS'}[$counter] =
2094               GetMarcBiblio( $result_hash->{'RECORDS'}[$counter] )->as_usmarc;
2095         }
2096         my $finalresult = ();
2097         $result_hash->{'hits'}         = $numbers;
2098         $finalresult->{'biblioserver'} = $result_hash;
2099         return $finalresult;
2100     }
2101     else {
2102
2103 #
2104 # order by ranking
2105 #
2106 # we need 2 hashes to order by ranking : the 1st one to count the ranking, the 2nd to order by ranking
2107         my %result;
2108         my %count_ranking;
2109         foreach ( split /;/, $biblionumbers ) {
2110             my ( $biblionumber, $title ) = split /,/, $_;
2111             $title =~ /(.*)-(\d)/;
2112
2113             # get weight
2114             my $ranking = $2;
2115
2116 # note that we + the ranking because ranking is calculated on weight of EACH term requested.
2117 # if we ask for "two towers", and "two" has weight 2 in biblio N, and "towers" has weight 4 in biblio N
2118 # biblio N has ranking = 6
2119             $count_ranking{$biblionumber} += $ranking;
2120         }
2121
2122 # build the result by "inverting" the count_ranking hash
2123 # 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
2124 #         warn "counting";
2125         foreach ( keys %count_ranking ) {
2126             $result{ sprintf( "%10d", $count_ranking{$_} ) . '-' . $_ } = $_;
2127         }
2128
2129     # sort the hash and return the same structure as GetRecords (Zebra querying)
2130         my $result_hash;
2131         my $numbers = 0;
2132         foreach my $key ( sort { $b cmp $a } ( keys %result ) ) {
2133             $result_hash->{'RECORDS'}[ $numbers++ ] = $result{$key};
2134         }
2135
2136         # limit the $results_per_page to result size if it's more
2137         $results_per_page = $numbers - 1 if $numbers < $results_per_page;
2138
2139         # for the requested page, replace biblionumber by the complete record
2140         # speed improvement : avoid reading too much things
2141         for (
2142             my $counter = $offset ;
2143             $counter <= $offset + $results_per_page ;
2144             $counter++
2145           )
2146         {
2147             $result_hash->{'RECORDS'}[$counter] =
2148               GetMarcBiblio( $result_hash->{'RECORDS'}[$counter] )->as_usmarc
2149               if $result_hash->{'RECORDS'}[$counter];
2150         }
2151         my $finalresult = ();
2152         $result_hash->{'hits'}         = $numbers;
2153         $finalresult->{'biblioserver'} = $result_hash;
2154         return $finalresult;
2155     }
2156 }
2157
2158 =head2 enabled_staff_search_views
2159
2160 %hash = enabled_staff_search_views()
2161
2162 This function returns a hash that contains three flags obtained from the system
2163 preferences, used to determine whether a particular staff search results view
2164 is enabled.
2165
2166 =over 2
2167
2168 =item C<Output arg:>
2169
2170     * $hash{can_view_MARC} is true only if the MARC view is enabled
2171     * $hash{can_view_ISBD} is true only if the ISBD view is enabled
2172     * $hash{can_view_labeledMARC} is true only if the Labeled MARC view is enabled
2173
2174 =item C<usage in the script:>
2175
2176 =back
2177
2178 $template->param ( C4::Search::enabled_staff_search_views );
2179
2180 =cut
2181
2182 sub enabled_staff_search_views
2183 {
2184         return (
2185                 can_view_MARC                   => C4::Context->preference('viewMARC'),                 # 1 if the staff search allows the MARC view
2186                 can_view_ISBD                   => C4::Context->preference('viewISBD'),                 # 1 if the staff search allows the ISBD view
2187                 can_view_labeledMARC    => C4::Context->preference('viewLabeledMARC'),  # 1 if the staff search allows the Labeled MARC view
2188         );
2189 }
2190
2191 sub AddSearchHistory{
2192         my ($borrowernumber,$session,$query_desc,$query_cgi, $total)=@_;
2193     my $dbh = C4::Context->dbh;
2194
2195     # Add the request the user just made
2196     my $sql = "INSERT INTO search_history(userid, sessionid, query_desc, query_cgi, total, time) VALUES(?, ?, ?, ?, ?, NOW())";
2197     my $sth   = $dbh->prepare($sql);
2198     $sth->execute($borrowernumber, $session, $query_desc, $query_cgi, $total);
2199         return $dbh->last_insert_id(undef, 'search_history', undef,undef,undef);
2200 }
2201
2202 sub GetSearchHistory{
2203         my ($borrowernumber,$session)=@_;
2204     my $dbh = C4::Context->dbh;
2205
2206     # Add the request the user just made
2207     my $query = "SELECT FROM search_history WHERE (userid=? OR sessionid=?)";
2208     my $sth   = $dbh->prepare($query);
2209         $sth->execute($borrowernumber, $session);
2210     return  $sth->fetchall_hashref({});
2211 }
2212
2213 =head2 z3950_search_args
2214
2215 $arrayref = z3950_search_args($matchpoints)
2216
2217 This function returns an array reference that contains the search parameters to be
2218 passed to the Z39.50 search script (z3950_search.pl). The array elements
2219 are hash refs whose keys are name, value and encvalue, and whose values are the
2220 name of a search parameter, the value of that search parameter and the URL encoded
2221 value of that parameter.
2222
2223 The search parameter names are lccn, isbn, issn, title, author, dewey and subject.
2224
2225 The search parameter values are obtained from the bibliographic record whose
2226 data is in a hash reference in $matchpoints, as returned by Biblio::GetBiblioData().
2227
2228 If $matchpoints is a scalar, it is assumed to be an unnamed query descriptor, e.g.
2229 a general purpose search argument. In this case, the returned array contains only
2230 entry: the key is 'title' and the value and encvalue are derived from $matchpoints.
2231
2232 If a search parameter value is undefined or empty, it is not included in the returned
2233 array.
2234
2235 The returned array reference may be passed directly to the template parameters.
2236
2237 =over 2
2238
2239 =item C<Output arg:>
2240
2241     * $array containing hash refs as described above
2242
2243 =item C<usage in the script:>
2244
2245 =back
2246
2247 $data = Biblio::GetBiblioData($bibno);
2248 $template->param ( MYLOOP => C4::Search::z3950_search_args($data) )
2249
2250 *OR*
2251
2252 $template->param ( MYLOOP => C4::Search::z3950_search_args($searchscalar) )
2253
2254 =cut
2255
2256 sub z3950_search_args {
2257     my $bibrec = shift;
2258     $bibrec = { title => $bibrec } if !ref $bibrec;
2259     my $array = [];
2260     for my $field (qw/ lccn isbn issn title author dewey subject /)
2261     {
2262         my $encvalue = URI::Escape::uri_escape_utf8($bibrec->{$field});
2263         push @$array, { name=>$field, value=>$bibrec->{$field}, encvalue=>$encvalue } if defined $bibrec->{$field};
2264     }
2265     return $array;
2266 }
2267
2268 =head2 BiblioAddAuthorities
2269
2270 ( $countlinked, $countcreated ) = BiblioAddAuthorities($record, $frameworkcode);
2271
2272 this function finds the authorities linked to the biblio
2273     * search in the authority DB for the same authid (in $9 of the biblio)
2274     * search in the authority DB for the same 001 (in $3 of the biblio in UNIMARC)
2275     * search in the authority DB for the same values (exactly) (in all subfields of the biblio)
2276 OR adds a new authority record
2277
2278 =over 2
2279
2280 =item C<input arg:>
2281
2282     * $record is the MARC record in question (marc blob)
2283     * $frameworkcode is the bibliographic framework to use (if it is "" it uses the default framework)
2284
2285 =item C<Output arg:>
2286
2287     * $countlinked is the number of authorities records that are linked to this authority
2288     * $countcreated
2289
2290 =item C<BUGS>
2291     * I had to add this to Search.pm (instead of the logical Biblio.pm) because of a circular dependency (this sub uses SimpleSearch, and Search.pm uses Biblio.pm)
2292 =back
2293
2294 =cut
2295
2296
2297 sub BiblioAddAuthorities{
2298   my ( $record, $frameworkcode ) = @_;
2299   my $dbh=C4::Context->dbh;
2300   my $query=$dbh->prepare(qq|
2301 SELECT authtypecode,tagfield
2302 FROM marc_subfield_structure
2303 WHERE frameworkcode=?
2304 AND (authtypecode IS NOT NULL AND authtypecode<>\"\")|);
2305 # SELECT authtypecode,tagfield
2306 # FROM marc_subfield_structure
2307 # WHERE frameworkcode=?
2308 # AND (authtypecode IS NOT NULL OR authtypecode<>\"\")|);
2309   $query->execute($frameworkcode);
2310   my ($countcreated,$countlinked);
2311   while (my $data=$query->fetchrow_hashref){
2312     foreach my $field ($record->field($data->{tagfield})){
2313       next if ($field->subfield('3')||$field->subfield('9'));
2314       # No authorities id in the tag.
2315       # Search if there is any authorities to link to.
2316       my $query='at='.$data->{authtypecode}.' ';
2317       map {$query.= ' and he,ext="'.$_->[1].'"' if ($_->[0]=~/[A-z]/)}  $field->subfields();
2318       my ($error, $results, $total_hits)=SimpleSearch( $query, undef, undef, [ "authorityserver" ] );
2319     # there is only 1 result
2320           if ( $error ) {
2321         warn "BIBLIOADDSAUTHORITIES: $error";
2322             return (0,0) ;
2323           }
2324       if ($results && scalar(@$results)==1) {
2325         my $marcrecord = MARC::File::USMARC::decode($results->[0]);
2326         $field->add_subfields('9'=>$marcrecord->field('001')->data);
2327         $countlinked++;
2328       } elsif (scalar(@$results)>1) {
2329    #More than One result
2330    #This can comes out of a lack of a subfield.
2331 #         my $marcrecord = MARC::File::USMARC::decode($results->[0]);
2332 #         $record->field($data->{tagfield})->add_subfields('9'=>$marcrecord->field('001')->data);
2333   $countlinked++;
2334       } else {
2335   #There are no results, build authority record, add it to Authorities, get authid and add it to 9
2336   ###NOTICE : This is only valid if a subfield is linked to one and only one authtypecode
2337   ###NOTICE : This can be a problem. We should also look into other types and rejected forms.
2338          my $authtypedata=C4::AuthoritiesMarc->GetAuthType($data->{authtypecode});
2339          next unless $authtypedata;
2340          my $marcrecordauth=MARC::Record->new();
2341          my $authfield=MARC::Field->new($authtypedata->{auth_tag_to_report},'','',"a"=>"".$field->subfield('a'));
2342          map { $authfield->add_subfields($_->[0]=>$_->[1]) if ($_->[0]=~/[A-z]/ && $_->[0] ne "a" )}  $field->subfields();
2343          $marcrecordauth->insert_fields_ordered($authfield);
2344
2345          # bug 2317: ensure new authority knows it's using UTF-8; currently
2346          # only need to do this for MARC21, as MARC::Record->as_xml_record() handles
2347          # automatically for UNIMARC (by not transcoding)
2348          # FIXME: AddAuthority() instead should simply explicitly require that the MARC::Record
2349          # use UTF-8, but as of 2008-08-05, did not want to introduce that kind
2350          # of change to a core API just before the 3.0 release.
2351          if (C4::Context->preference('marcflavour') eq 'MARC21') {
2352             SetMarcUnicodeFlag($marcrecordauth, 'MARC21');
2353          }
2354
2355 #          warn "AUTH RECORD ADDED : ".$marcrecordauth->as_formatted;
2356
2357          my $authid=AddAuthority($marcrecordauth,'',$data->{authtypecode});
2358          $countcreated++;
2359          $field->add_subfields('9'=>$authid);
2360       }
2361     }
2362   }
2363   return ($countlinked,$countcreated);
2364 }
2365
2366 =head2 GetDistinctValues($field);
2367
2368 C<$field> is a reference to the fields array
2369
2370 =cut
2371
2372 sub GetDistinctValues {
2373     my ($fieldname,$string)=@_;
2374     # returns a reference to a hash of references to branches...
2375     if ($fieldname=~/\./){
2376                         my ($table,$column)=split /\./, $fieldname;
2377                         my $dbh = C4::Context->dbh;
2378                         warn "select DISTINCT($column) as value, count(*) as cnt from $table group by lib order by $column ";
2379                         my $sth = $dbh->prepare("select DISTINCT($column) as value, count(*) as cnt from $table ".($string?" where $column like \"$string%\"":"")."group by value order by $column ");
2380                         $sth->execute;
2381                         my $elements=$sth->fetchall_arrayref({});
2382                         return $elements;
2383    }
2384    else {
2385                 $string||= qq("");
2386                 my @servers=qw<biblioserver authorityserver>;
2387                 my (@zconns,@results);
2388         for ( my $i = 0 ; $i < @servers ; $i++ ) {
2389                 $zconns[$i] = C4::Context->Zconn( $servers[$i], 1 );
2390                         $results[$i] =
2391                       $zconns[$i]->scan(
2392                         ZOOM::Query::CCL2RPN->new( qq"$fieldname $string", $zconns[$i])
2393                       );
2394                 }
2395                 # The big moment: asynchronously retrieve results from all servers
2396                 my @elements;
2397                 while ( ( my $i = ZOOM::event( \@zconns ) ) != 0 ) {
2398                         my $ev = $zconns[ $i - 1 ]->last_event();
2399                         if ( $ev == ZOOM::Event::ZEND ) {
2400                                 next unless $results[ $i - 1 ];
2401                                 my $size = $results[ $i - 1 ]->size();
2402                                 if ( $size > 0 ) {
2403                       for (my $j=0;$j<$size;$j++){
2404                                                 my %hashscan;
2405                                                 @hashscan{qw(value cnt)}=$results[ $i - 1 ]->display_term($j);
2406                                                 push @elements, \%hashscan;
2407                                           }
2408                                 }
2409                         }
2410                 }
2411                 return \@elements;
2412    }
2413 }
2414
2415
2416 END { }    # module clean-up code here (global destructor)
2417
2418 1;
2419 __END__
2420
2421 =head1 AUTHOR
2422
2423 Koha Developement team <info@koha.org>
2424
2425 =cut