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 - Bug 2505
  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::Reserves;    # CheckReserves
  31 use C4::Debug;
  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 ( $_ =~ /(and|or|not)/ );    # don't remove operators
 653                         if ( my ($matched) = ($operand =~
 654                                 /([^\X\p{isAlnum}]\Q$_\E[^\X\p{isAlnum}]|[^\X\p{isAlnum}]\Q$_\E$|^\Q$_\E[^\X\p{isAlnum}])/gi))
 655                         {
 656                                 $operand =~ s/\Q$matched\E/ /gi;
 657                                 push @stopwords_removed, $_;
 658                         }
 659                 }
 660         }
 661     return ( $operand, \@stopwords_removed );
 662 }
 663
 664 # TRUNCATION
 665 sub _detect_truncation {
 666     my ( $operand, $index ) = @_;
 667     my ( @nontruncated, @righttruncated, @lefttruncated, @rightlefttruncated,
 668         @regexpr );
 669     $operand =~ s/^ //g;
 670     my @wordlist = split( /\s/, $operand );
 671     foreach my $word (@wordlist) {
 672         if ( $word =~ s/^\*([^\*]+)\*$/$1/ ) {
 673             push @rightlefttruncated, $word;
 674         }
 675         elsif ( $word =~ s/^\*([^\*]+)$/$1/ ) {
 676             push @lefttruncated, $word;
 677         }
 678         elsif ( $word =~ s/^([^\*]+)\*$/$1/ ) {
 679             push @righttruncated, $word;
 680         }
 681         elsif ( index( $word, "*" ) < 0 ) {
 682             push @nontruncated, $word;
 683         }
 684         else {
 685             push @regexpr, $word;
 686         }
 687     }
 688     return (
 689         \@nontruncated,       \@righttruncated, \@lefttruncated,
 690         \@rightlefttruncated, \@regexpr
 691     );
 692 }
 693
 694 # STEMMING
 695 sub _build_stemmed_operand {
 696     my ($operand,$lang) = @_;
 697     require Lingua::Stem::Snowball ;
 698     my $stemmed_operand;
 699
 700     # If operand contains a digit, it is almost certainly an identifier, and should
 701     # not be stemmed.  This is particularly relevant for ISBNs and ISSNs, which
 702     # can contain the letter "X" - for example, _build_stemmend_operand would reduce
 703     # "014100018X" to "x ", which for a MARC21 database would bring up irrelevant
 704     # results (e.g., "23 x 29 cm." from the 300$c).  Bug 2098.
 705     return $operand if $operand =~ /\d/;
 706
 707 # FIXME: the locale should be set based on the user's language and/or search choice
 708     #warn "$lang";
 709     my $stemmer = Lingua::Stem::Snowball->new( lang => $lang,
 710                                                encoding => "UTF-8" );
 711
 712     my @words = split( / /, $operand );
 713     my @stems = $stemmer->stem(\@words);
 714     for my $stem (@stems) {
 715         $stemmed_operand .= "$stem";
 716         $stemmed_operand .= "?"
 717           unless ( $stem =~ /(and$|or$|not$)/ ) || ( length($stem) < 3 );
 718         $stemmed_operand .= " ";
 719     }
 720     warn "STEMMED OPERAND: $stemmed_operand" if $DEBUG;
 721     return $stemmed_operand;
 722 }
 723
 724 # FIELD WEIGHTING
 725 sub _build_weighted_query {
 726
 727 # FIELD WEIGHTING - This is largely experimental stuff. What I'm committing works
 728 # pretty well but could work much better if we had a smarter query parser
 729     my ( $operand, $stemmed_operand, $index ) = @_;
 730     my $stemming      = C4::Context->preference("QueryStemming")     || 0;
 731     my $weight_fields = C4::Context->preference("QueryWeightFields") || 0;
 732     my $fuzzy_enabled = C4::Context->preference("QueryFuzzy")        || 0;
 733
 734     my $weighted_query .= "(rk=(";    # Specifies that we're applying rank
 735
 736     # Keyword, or, no index specified
 737     if ( ( $index eq 'kw' ) || ( !$index ) ) {
 738         $weighted_query .=
 739           "Title-cover,ext,r1=\"$operand\"";    # exact title-cover
 740         $weighted_query .= " or ti,ext,r2=\"$operand\"";    # exact title
 741         $weighted_query .= " or ti,phr,r3=\"$operand\"";    # phrase title
 742           #$weighted_query .= " or any,ext,r4=$operand";               # exact any
 743           #$weighted_query .=" or kw,wrdl,r5=\"$operand\"";            # word list any
 744         $weighted_query .= " or wrdl,fuzzy,r8=\"$operand\""
 745           if $fuzzy_enabled;    # add fuzzy, word list
 746         $weighted_query .= " or wrdl,right-Truncation,r9=\"$stemmed_operand\""
 747           if ( $stemming and $stemmed_operand )
 748           ;                     # add stemming, right truncation
 749         $weighted_query .= " or wrdl,r9=\"$operand\"";
 750
 751         # embedded sorting: 0 a-z; 1 z-a
 752         # $weighted_query .= ") or (sort1,aut=1";
 753     }
 754
 755     # Barcode searches should skip this process
 756     elsif ( $index eq 'bc' ) {
 757         $weighted_query .= "bc=\"$operand\"";
 758     }
 759
 760     # Authority-number searches should skip this process
 761     elsif ( $index eq 'an' ) {
 762         $weighted_query .= "an=\"$operand\"";
 763     }
 764
 765     # If the index already has more than one qualifier, wrap the operand
 766     # in quotes and pass it back (assumption is that the user knows what they
 767     # are doing and won't appreciate us mucking up their query
 768     elsif ( $index =~ ',' ) {
 769         $weighted_query .= " $index=\"$operand\"";
 770     }
 771
 772     #TODO: build better cases based on specific search indexes
 773     else {
 774         $weighted_query .= " $index,ext,r1=\"$operand\"";    # exact index
 775           #$weighted_query .= " or (title-sort-az=0 or $index,startswithnt,st-word,r3=$operand #)";
 776         $weighted_query .= " or $index,phr,r3=\"$operand\"";    # phrase index
 777         $weighted_query .=
 778           " or $index,rt,wrdl,r3=\"$operand\"";    # word list index
 779     }
 780
 781     $weighted_query .= "))";                       # close rank specification
 782     return $weighted_query;
 783 }
 784
 785 =head2 getIndexes
 786
 787 Return an array with available indexes.
 788
 789 =cut
 790
 791 sub getIndexes{
 792     my @indexes = (
 793                     # biblio indexes
 794                     'ab',
 795                     'Abstract',
 796                     'acqdate',
 797                     'allrecords',
 798                     'an',
 799                     'Any',
 800                     'at',
 801                     'au',
 802                     'aub',
 803                     'aud',
 804                     'audience',
 805                     'auo',
 806                     'aut',
 807                     'Author',
 808                     'Author-in-order ',
 809                     'Author-personal-bibliography',
 810                     'Authority-Number',
 811                     'authtype',
 812                     'bc',
 813                     'biblionumber',
 814                     'bio',
 815                     'biography',
 816                     'callnum',
 817                     'cfn',
 818                     'Chronological-subdivision',
 819                     'cn-bib-source',
 820                     'cn-bib-sort',
 821                     'cn-class',
 822                     'cn-item',
 823                     'cn-prefix',
 824                     'cn-suffix',
 825                     'cpn',
 826                     'Code-institution',
 827                     'Conference-name',
 828                     'Conference-name-heading',
 829                     'Conference-name-see',
 830                     'Conference-name-seealso',
 831                     'Content-type',
 832                     'Control-number',
 833                     'copydate',
 834                     'Corporate-name',
 835                     'Corporate-name-heading',
 836                     'Corporate-name-see',
 837                     'Corporate-name-seealso',
 838                     'ctype',
 839                     'date-entered-on-file',
 840                     'Date-of-acquisition',
 841                     'Date-of-publication',
 842                     'Dewey-classification',
 843                     'extent',
 844                     'fic',
 845                     'fiction',
 846                     'Form-subdivision',
 847                     'format',
 848                     'Geographic-subdivision',
 849                     'he',
 850                     'Heading',
 851                     'Heading-use-main-or-added-entry',
 852                     'Heading-use-series-added-entry ',
 853                     'Heading-use-subject-added-entry',
 854                     'Host-item',
 855                     'id-other',
 856                     'Illustration-code',
 857                     'ISBN',
 858                     'ISSN',
 859                     'itemtype',
 860                     'kw',
 861                     'Koha-Auth-Number',
 862                     'l-format',
 863                     'language',
 864                     'lc-card',
 865                     'LC-card-number',
 866                     'lcn',
 867                     'llength',
 868                     'ln',
 869                     'Local-classification',
 870                     'Local-number',
 871                     'Match-heading',
 872                     'Match-heading-see-from',
 873                     'Material-type',
 874                     'mc-itemtype',
 875                     'mc-rtype',
 876                     'mus',
 877                     'Name-geographic',
 878                     'Name-geographic-heading',
 879                     'Name-geographic-see',
 880                     'Name-geographic-seealso',
 881                     'nb',
 882                     'Note',
 883                     'ns',
 884                     'nt',
 885                     'pb',
 886                     'Personal-name',
 887                     'Personal-name-heading',
 888                     'Personal-name-see',
 889                     'Personal-name-seealso',
 890                     'pl',
 891                     'Place-publication',
 892                     'pn',
 893                     'popularity',
 894                     'pubdate',
 895                     'Publisher',
 896                     'Record-control-number',
 897                     'rcn',
 898                     'Record-type',
 899                     'rtype',
 900                     'se',
 901                     'See',
 902                     'See-also',
 903                     'sn',
 904                     'Stock-number',
 905                     'su',
 906                     'Subject',
 907                     'Subject-heading-thesaurus',
 908                     'Subject-name-personal',
 909                     'Subject-subdivision',
 910                     'Summary',
 911                     'Suppress',
 912                     'su-geo',
 913                     'su-na',
 914                     'su-to',
 915                     'su-ut',
 916                     'ut',
 917                     'Term-genre-form',
 918                     'Term-genre-form-heading',
 919                     'Term-genre-form-see',
 920                     'Term-genre-form-seealso',
 921                     'ti',
 922                     'Title',
 923                     'Title-cover',
 924                     'Title-series',
 925                     'Title-uniform',
 926                     'Title-uniform-heading',
 927                     'Title-uniform-see',
 928                     'Title-uniform-seealso',
 929                     'totalissues',
 930                     'yr',
 931
 932                     # items indexes
 933                     'acqsource',
 934                     'barcode',
 935                     'bc',
 936                     'branch',
 937                     'ccode',
 938                     'classification-source',
 939                     'cn-sort',
 940                     'coded-location-qualifier',
 941                     'copynumber',
 942                     'damaged',
 943                     'datelastborrowed',
 944                     'datelastseen',
 945                     'holdingbranch',
 946                     'homebranch',
 947                     'issues',
 948                     'item',
 949                     'itemnumber',
 950                     'itype',
 951                     'Local-classification',
 952                     'location',
 953                     'lost',
 954                     'materials-specified',
 955                     'mc-ccode',
 956                     'mc-itype',
 957                     'mc-loc',
 958                     'notforloan',
 959                     'onloan',
 960                     'price',
 961                     'renewals',
 962                     'replacementprice',
 963                     'replacementpricedate',
 964                     'reserves',
 965                     'restricted',
 966                     'stack',
 967                     'uri',
 968                     'withdrawn',
 969
 970                     # subject related
 971                   );
 972
 973     return \@indexes;
 974 }
 975
 976 =head2 buildQuery
 977
 978 ( $error, $query,
 979 $simple_query, $query_cgi,
 980 $query_desc, $limit,
 981 $limit_cgi, $limit_desc,
 982 $stopwords_removed, $query_type ) = buildQuery ( $operators, $operands, $indexes, $limits, $sort_by, $scan, $lang);
 983
 984 Build queries and limits in CCL, CGI, Human,
 985 handle truncation, stemming, field weighting, stopwords, fuzziness, etc.
 986
 987 See verbose embedded documentation.
 988
 989
 990 =cut
 991
 992 sub buildQuery {
 993     my ( $operators, $operands, $indexes, $limits, $sort_by, $scan, $lang) = @_;
 994
 995     warn "---------\nEnter buildQuery\n---------" if $DEBUG;
 996
 997     # dereference
 998     my @operators = $operators ? @$operators : ();
 999     my @indexes   = $indexes   ? @$indexes   : ();
1000     my @operands  = $operands  ? @$operands  : ();
1001     my @limits    = $limits    ? @$limits    : ();
1002     my @sort_by   = $sort_by   ? @$sort_by   : ();
1003
1004     my $stemming         = C4::Context->preference("QueryStemming")        || 0;
1005     my $auto_truncation  = C4::Context->preference("QueryAutoTruncate")    || 0;
1006     my $weight_fields    = C4::Context->preference("QueryWeightFields")    || 0;
1007     my $fuzzy_enabled    = C4::Context->preference("QueryFuzzy")           || 0;
1008     my $remove_stopwords = C4::Context->preference("QueryRemoveStopwords") || 0;
1009
1010     # no stemming/weight/fuzzy in NoZebra
1011     if ( C4::Context->preference("NoZebra") ) {
1012         $stemming         = 0;
1013         $weight_fields    = 0;
1014         $fuzzy_enabled    = 0;
1015         $auto_truncation  = 0;
1016     }
1017
1018     my $query        = $operands[0];
1019     my $simple_query = $operands[0];
1020
1021     # initialize the variables we're passing back
1022     my $query_cgi;
1023     my $query_desc;
1024     my $query_type;
1025
1026     my $limit;
1027     my $limit_cgi;
1028     my $limit_desc;
1029
1030     my $stopwords_removed;    # flag to determine if stopwords have been removed
1031
1032     my $cclq;
1033     my $cclindexes = getIndexes();
1034     if( $query !~ /\s*ccl=/ ){
1035         for my $index (@$cclindexes){
1036             if($query =~ /($index)(,?\w)*[:=]/){
1037                 $cclq = 1;
1038             }
1039         }
1040         $query = "ccl=$query" if($cclq);
1041     }
1042
1043 # for handling ccl, cql, pqf queries in diagnostic mode, skip the rest of the steps
1044 # DIAGNOSTIC ONLY!!
1045     if ( $query =~ /^ccl=/ ) {
1046         my $q=$';
1047         # This is needed otherwise ccl= and &limit won't work together, and
1048         # this happens when selecting a subject on the opac-detail page
1049         if (@limits) {
1050             $q .= ' and '.join(' and ', @limits);
1051         }
1052         return ( undef, $q, $q, "q=ccl=$q", $q, '', '', '', '', 'ccl' );
1053     }
1054     if ( $query =~ /^cql=/ ) {
1055         return ( undef, $', $', "q=cql=$'", $', '', '', '', '', 'cql' );
1056     }
1057     if ( $query =~ /^pqf=/ ) {
1058         return ( undef, $', $', "q=pqf=$'", $', '', '', '', '', 'pqf' );
1059     }
1060
1061     # pass nested queries directly
1062     # FIXME: need better handling of some of these variables in this case
1063     # Nested queries aren't handled well and this implementation is flawed and causes users to be
1064     # unable to search for anything containing () commenting out, will be rewritten for 3.4.0
1065 #    if ( $query =~ /(\(|\))/ ) {
1066 #        return (
1067 #            undef,              $query, $simple_query, $query_cgi,
1068 #            $query,             $limit, $limit_cgi,    $limit_desc,
1069 #            $stopwords_removed, 'ccl'
1070 #        );
1071 #    }
1072
1073 # Form-based queries are non-nested and fixed depth, so we can easily modify the incoming
1074 # query operands and indexes and add stemming, truncation, field weighting, etc.
1075 # Once we do so, we'll end up with a value in $query, just like if we had an
1076 # incoming $query from the user
1077     else {
1078         $query = ""
1079           ; # clear it out so we can populate properly with field-weighted, stemmed, etc. query
1080         my $previous_operand
1081           ;    # a flag used to keep track if there was a previous query
1082                # if there was, we can apply the current operator
1083                # for every operand
1084         for ( my $i = 0 ; $i <= @operands ; $i++ ) {
1085
1086             # COMBINE OPERANDS, INDEXES AND OPERATORS
1087             if ( $operands[$i] ) {
1088                 $operands[$i]=~s/^\s+//;
1089
1090               # A flag to determine whether or not to add the index to the query
1091                 my $indexes_set;
1092
1093 # If the user is sophisticated enough to specify an index, turn off field weighting, stemming, and stopword handling
1094                 if ( $operands[$i] =~ /(:|=)/ || $scan ) {
1095                     $weight_fields    = 0;
1096                     $stemming         = 0;
1097                     $remove_stopwords = 0;
1098                 }
1099                 my $operand = $operands[$i];
1100                 my $index   = $indexes[$i];
1101
1102                 # Add index-specific attributes
1103                 # Date of Publication
1104                 if ( $index eq 'yr' ) {
1105                     $index .= ",st-numeric";
1106                     $indexes_set++;
1107                                         $stemming = $auto_truncation = $weight_fields = $fuzzy_enabled = $remove_stopwords = 0;
1108                 }
1109
1110                 # Date of Acquisition
1111                 elsif ( $index eq 'acqdate' ) {
1112                     $index .= ",st-date-normalized";
1113                     $indexes_set++;
1114                                         $stemming = $auto_truncation = $weight_fields = $fuzzy_enabled = $remove_stopwords = 0;
1115                 }
1116                 # ISBN,ISSN,Standard Number, don't need special treatment
1117                 elsif ( $index eq 'nb' || $index eq 'ns' ) {
1118                     $indexes_set++;
1119                     (
1120                         $stemming,      $auto_truncation,
1121                         $weight_fields, $fuzzy_enabled,
1122                         $remove_stopwords
1123                     ) = ( 0, 0, 0, 0, 0 );
1124
1125                 }
1126
1127                 if(not $index){
1128                     $index = 'kw';
1129                 }
1130
1131                 # Set default structure attribute (word list)
1132                 my $struct_attr = q{};
1133                 unless ( $indexes_set || !$index || $index =~ /(st-|phr|ext|wrdl)/ ) {
1134                     $struct_attr = ",wrdl";
1135                 }
1136
1137                 # Some helpful index variants
1138                 my $index_plus       = $index . $struct_attr . ':';
1139                 my $index_plus_comma = $index . $struct_attr . ',';
1140
1141                 # Remove Stopwords
1142                 if ($remove_stopwords) {
1143                     ( $operand, $stopwords_removed ) =
1144                       _remove_stopwords( $operand, $index );
1145                     warn "OPERAND w/out STOPWORDS: >$operand<" if $DEBUG;
1146                     warn "REMOVED STOPWORDS: @$stopwords_removed"
1147                       if ( $stopwords_removed && $DEBUG );
1148                 }
1149
1150                 if ($auto_truncation){
1151                                         unless ( $index =~ /(st-|phr|ext)/ ) {
1152                                                 #FIXME only valid with LTR scripts
1153                                                 $operand=join(" ",map{
1154                                                                                         (index($_,"*")>0?"$_":"$_*")
1155                                                                                          }split (/\s+/,$operand));
1156                                                 warn $operand if $DEBUG;
1157                                         }
1158                                 }
1159
1160                 # Detect Truncation
1161                 my $truncated_operand;
1162                 my( $nontruncated, $righttruncated, $lefttruncated,
1163                     $rightlefttruncated, $regexpr
1164                 ) = _detect_truncation( $operand, $index );
1165                 warn
1166 "TRUNCATION: NON:>@$nontruncated< RIGHT:>@$righttruncated< LEFT:>@$lefttruncated< RIGHTLEFT:>@$rightlefttruncated< REGEX:>@$regexpr<"
1167                   if $DEBUG;
1168
1169                 # Apply Truncation
1170                 if (
1171                     scalar(@$righttruncated) + scalar(@$lefttruncated) +
1172                     scalar(@$rightlefttruncated) > 0 )
1173                 {
1174
1175                # Don't field weight or add the index to the query, we do it here
1176                     $indexes_set = 1;
1177                     undef $weight_fields;
1178                     my $previous_truncation_operand;
1179                     if (scalar @$nontruncated) {
1180                         $truncated_operand .= "$index_plus @$nontruncated ";
1181                         $previous_truncation_operand = 1;
1182                     }
1183                     if (scalar @$righttruncated) {
1184                         $truncated_operand .= "and " if $previous_truncation_operand;
1185                         $truncated_operand .= $index_plus_comma . "rtrn:@$righttruncated ";
1186                         $previous_truncation_operand = 1;
1187                     }
1188                     if (scalar @$lefttruncated) {
1189                         $truncated_operand .= "and " if $previous_truncation_operand;
1190                         $truncated_operand .= $index_plus_comma . "ltrn:@$lefttruncated ";
1191                         $previous_truncation_operand = 1;
1192                     }
1193                     if (scalar @$rightlefttruncated) {
1194                         $truncated_operand .= "and " if $previous_truncation_operand;
1195                         $truncated_operand .= $index_plus_comma . "rltrn:@$rightlefttruncated ";
1196                         $previous_truncation_operand = 1;
1197                     }
1198                 }
1199                 $operand = $truncated_operand if $truncated_operand;
1200                 warn "TRUNCATED OPERAND: >$truncated_operand<" if $DEBUG;
1201
1202                 # Handle Stemming
1203                 my $stemmed_operand;
1204                 $stemmed_operand = _build_stemmed_operand($operand, $lang)
1205                                                                                 if $stemming;
1206
1207                 warn "STEMMED OPERAND: >$stemmed_operand<" if $DEBUG;
1208
1209                 # Handle Field Weighting
1210                 my $weighted_operand;
1211                 if ($weight_fields) {
1212                     $weighted_operand = _build_weighted_query( $operand, $stemmed_operand, $index );
1213                     $operand = $weighted_operand;
1214                     $indexes_set = 1;
1215                 }
1216
1217                 warn "FIELD WEIGHTED OPERAND: >$weighted_operand<" if $DEBUG;
1218
1219                 # If there's a previous operand, we need to add an operator
1220                 if ($previous_operand) {
1221
1222                     # User-specified operator
1223                     if ( $operators[ $i - 1 ] ) {
1224                         $query     .= " $operators[$i-1] ";
1225                         $query     .= " $index_plus " unless $indexes_set;
1226                         $query     .= " $operand";
1227                         $query_cgi .= "&op=$operators[$i-1]";
1228                         $query_cgi .= "&idx=$index" if $index;
1229                         $query_cgi .= "&q=$operands[$i]" if $operands[$i];
1230                         $query_desc .=
1231                           " $operators[$i-1] $index_plus $operands[$i]";
1232                     }
1233
1234                     # Default operator is and
1235                     else {
1236                         $query      .= " and ";
1237                         $query      .= "$index_plus " unless $indexes_set;
1238                         $query      .= "$operand";
1239                         $query_cgi  .= "&op=and&idx=$index" if $index;
1240                         $query_cgi  .= "&q=$operands[$i]" if $operands[$i];
1241                         $query_desc .= " and $index_plus $operands[$i]";
1242                     }
1243                 }
1244
1245                 # There isn't a pervious operand, don't need an operator
1246                 else {
1247
1248                     # Field-weighted queries already have indexes set
1249                     $query .= " $index_plus " unless $indexes_set;
1250                     $query .= $operand;
1251                     $query_desc .= " $index_plus $operands[$i]";
1252                     $query_cgi  .= "&idx=$index" if $index;
1253                     $query_cgi  .= "&q=$operands[$i]" if $operands[$i];
1254                     $previous_operand = 1;
1255                 }
1256             }    #/if $operands
1257         }    # /for
1258     }
1259     warn "QUERY BEFORE LIMITS: >$query<" if $DEBUG;
1260
1261     # add limits
1262     my $group_OR_limits;
1263     my $availability_limit;
1264     foreach my $this_limit (@limits) {
1265         if ( $this_limit =~ /available/ ) {
1266 #
1267 ## 'available' is defined as (items.onloan is NULL) and (items.itemlost = 0)
1268 ## In English:
1269 ## all records not indexed in the onloan register (zebra) and all records with a value of lost equal to 0
1270             $availability_limit .=
1271 "( ( allrecords,AlwaysMatches='' not onloan,AlwaysMatches='') and (lost,st-numeric=0) )"; #or ( allrecords,AlwaysMatches='' not lost,AlwaysMatches='')) )";
1272             $limit_cgi  .= "&limit=available";
1273             $limit_desc .= "";
1274         }
1275
1276         # group_OR_limits, prefixed by mc-
1277         # OR every member of the group
1278         elsif ( $this_limit =~ /mc/ ) {
1279
1280             if ( $this_limit =~ /mc-ccode:/ ) {
1281                 # in case the mc-ccode value has complicating chars like ()'s inside it we wrap in quotes
1282                 $this_limit =~ tr/"//d;
1283                 my ($k,$v) = split(/:/, $this_limit,2);
1284                 $this_limit = $k.":\"".$v."\"";
1285             }
1286
1287             $group_OR_limits .= " or " if $group_OR_limits;
1288             $limit_desc      .= " or " if $group_OR_limits;
1289             $group_OR_limits .= "$this_limit";
1290             $limit_cgi       .= "&limit=$this_limit";
1291             $limit_desc      .= " $this_limit";
1292         }
1293
1294         # Regular old limits
1295         else {
1296             $limit .= " and " if $limit || $query;
1297             $limit      .= "$this_limit";
1298             $limit_cgi  .= "&limit=$this_limit";
1299             if ($this_limit =~ /^branch:(.+)/) {
1300                 my $branchcode = $1;
1301                 my $branchname = GetBranchName($branchcode);
1302                 if (defined $branchname) {
1303                     $limit_desc .= " branch:$branchname";
1304                 } else {
1305                     $limit_desc .= " $this_limit";
1306                 }
1307             } else {
1308                 $limit_desc .= " $this_limit";
1309             }
1310         }
1311     }
1312     if ($group_OR_limits) {
1313         $limit .= " and " if ( $query || $limit );
1314         $limit .= "($group_OR_limits)";
1315     }
1316     if ($availability_limit) {
1317         $limit .= " and " if ( $query || $limit );
1318         $limit .= "($availability_limit)";
1319     }
1320
1321     # Normalize the query and limit strings
1322     # This is flawed , means we can't search anything with : in it
1323     # if user wants to do ccl or cql, start the query with that
1324 #    $query =~ s/:/=/g;
1325     $query =~ s/(?<=(ti|au|pb|su|an|kw|mc)):/=/g;
1326     $query =~ s/(?<=(wrdl)):/=/g;
1327     $query =~ s/(?<=(trn|phr)):/=/g;
1328     $limit =~ s/:/=/g;
1329     for ( $query, $query_desc, $limit, $limit_desc ) {
1330         s/  +/ /g;    # remove extra spaces
1331         s/^ //g;     # remove any beginning spaces
1332         s/ $//g;     # remove any ending spaces
1333         s/==/=/g;    # remove double == from query
1334     }
1335     $query_cgi =~ s/^&//; # remove unnecessary & from beginning of the query cgi
1336
1337     for ($query_cgi,$simple_query) {
1338         s/"//g;
1339     }
1340     # append the limit to the query
1341     $query .= " " . $limit;
1342
1343     # Warnings if DEBUG
1344     if ($DEBUG) {
1345         warn "QUERY:" . $query;
1346         warn "QUERY CGI:" . $query_cgi;
1347         warn "QUERY DESC:" . $query_desc;
1348         warn "LIMIT:" . $limit;
1349         warn "LIMIT CGI:" . $limit_cgi;
1350         warn "LIMIT DESC:" . $limit_desc;
1351         warn "---------\nLeave buildQuery\n---------";
1352     }
1353     return (
1354         undef,              $query, $simple_query, $query_cgi,
1355         $query_desc,        $limit, $limit_cgi,    $limit_desc,
1356         $stopwords_removed, $query_type
1357     );
1358 }
1359
1360 =head2 searchResults
1361
1362   my @search_results = searchResults($search_context, $searchdesc, $hits,
1363                                      $results_per_page, $offset, $scan,
1364                                      @marcresults, $hidelostitems);
1365
1366 Format results in a form suitable for passing to the template
1367
1368 =cut
1369
1370 # IMO this subroutine is pretty messy still -- it's responsible for
1371 # building the HTML output for the template
1372 sub searchResults {
1373     my ( $search_context, $searchdesc, $hits, $results_per_page, $offset, $scan, @marcresults, $hidelostitems ) = @_;
1374     my $dbh = C4::Context->dbh;
1375     my @newresults;
1376
1377     $search_context = 'opac' unless $search_context eq 'opac' or $search_context eq 'intranet';
1378
1379     #Build branchnames hash
1380     #find branchname
1381     #get branch information.....
1382     my %branches;
1383     my $bsth =$dbh->prepare("SELECT branchcode,branchname FROM branches"); # FIXME : use C4::Branch::GetBranches
1384     $bsth->execute();
1385     while ( my $bdata = $bsth->fetchrow_hashref ) {
1386         $branches{ $bdata->{'branchcode'} } = $bdata->{'branchname'};
1387     }
1388 # FIXME - We build an authorised values hash here, using the default framework
1389 # though it is possible to have different authvals for different fws.
1390
1391     my $shelflocations =GetKohaAuthorisedValues('items.location','');
1392
1393     # get notforloan authorised value list (see $shelflocations  FIXME)
1394     my $notforloan_authorised_value = GetAuthValCode('items.notforloan','');
1395
1396     #Build itemtype hash
1397     #find itemtype & itemtype image
1398     my %itemtypes;
1399     $bsth =
1400       $dbh->prepare(
1401         "SELECT itemtype,description,imageurl,summary,notforloan FROM itemtypes"
1402       );
1403     $bsth->execute();
1404     while ( my $bdata = $bsth->fetchrow_hashref ) {
1405                 foreach (qw(description imageurl summary notforloan)) {
1406                 $itemtypes{ $bdata->{'itemtype'} }->{$_} = $bdata->{$_};
1407                 }
1408     }
1409
1410     #search item field code
1411     my $sth =
1412       $dbh->prepare(
1413 "SELECT tagfield FROM marc_subfield_structure WHERE kohafield LIKE 'items.itemnumber'"
1414       );
1415     $sth->execute;
1416     my ($itemtag) = $sth->fetchrow;
1417
1418     ## find column names of items related to MARC
1419     my $sth2 = $dbh->prepare("SHOW COLUMNS FROM items");
1420     $sth2->execute;
1421     my %subfieldstosearch;
1422     while ( ( my $column ) = $sth2->fetchrow ) {
1423         my ( $tagfield, $tagsubfield ) =
1424           &GetMarcFromKohaField( "items." . $column, "" );
1425         $subfieldstosearch{$column} = $tagsubfield;
1426     }
1427
1428     # handle which records to actually retrieve
1429     my $times;
1430     if ( $hits && $offset + $results_per_page <= $hits ) {
1431         $times = $offset + $results_per_page;
1432     }
1433     else {
1434         $times = $hits;  # FIXME: if $hits is undefined, why do we want to equal it?
1435     }
1436
1437         my $marcflavour = C4::Context->preference("marcflavour");
1438     # We get the biblionumber position in MARC
1439     my ($bibliotag,$bibliosubf)=GetMarcFromKohaField('biblio.biblionumber','');
1440     my $fw;
1441
1442     # loop through all of the records we've retrieved
1443     for ( my $i = $offset ; $i <= $times - 1 ; $i++ ) {
1444         my $marcrecord = MARC::File::USMARC::decode( $marcresults[$i] );
1445         $fw = $scan
1446              ? undef
1447              : $bibliotag < 10
1448                ? GetFrameworkCode($marcrecord->field($bibliotag)->data)
1449                : GetFrameworkCode($marcrecord->subfield($bibliotag,$bibliosubf));
1450         my $oldbiblio = TransformMarcToKoha( $dbh, $marcrecord, $fw );
1451         $oldbiblio->{subtitle} = GetRecordValue('subtitle', $marcrecord, $fw);
1452         $oldbiblio->{result_number} = $i + 1;
1453
1454         # add imageurl to itemtype if there is one
1455         $oldbiblio->{imageurl} = getitemtypeimagelocation( 'opac', $itemtypes{ $oldbiblio->{itemtype} }->{imageurl} );
1456
1457         $oldbiblio->{'authorised_value_images'}  = C4::Items::get_authorised_value_images( C4::Biblio::get_biblio_authorised_values( $oldbiblio->{'biblionumber'}, $marcrecord ) );
1458                 $oldbiblio->{normalized_upc}  = GetNormalizedUPC(       $marcrecord,$marcflavour);
1459                 $oldbiblio->{normalized_ean}  = GetNormalizedEAN(       $marcrecord,$marcflavour);
1460                 $oldbiblio->{normalized_oclc} = GetNormalizedOCLCNumber($marcrecord,$marcflavour);
1461                 $oldbiblio->{normalized_isbn} = GetNormalizedISBN(undef,$marcrecord,$marcflavour);
1462                 $oldbiblio->{content_identifier_exists} = 1 if ($oldbiblio->{normalized_isbn} or $oldbiblio->{normalized_oclc} or $oldbiblio->{normalized_ean} or $oldbiblio->{normalized_upc});
1463
1464                 # edition information, if any
1465         $oldbiblio->{edition} = $oldbiblio->{editionstatement};
1466                 $oldbiblio->{description} = $itemtypes{ $oldbiblio->{itemtype} }->{description};
1467  # Build summary if there is one (the summary is defined in the itemtypes table)
1468  # FIXME: is this used anywhere, I think it can be commented out? -- JF
1469         if ( $itemtypes{ $oldbiblio->{itemtype} }->{summary} ) {
1470             my $summary = $itemtypes{ $oldbiblio->{itemtype} }->{summary};
1471             my @fields  = $marcrecord->fields();
1472
1473             my $newsummary;
1474             foreach my $line ( "$summary\n" =~ /(.*)\n/g ){
1475                 my $tags = {};
1476                 foreach my $tag ( $line =~ /\[(\d{3}[\w|\d])\]/ ) {
1477                     $tag =~ /(.{3})(.)/;
1478                     if($marcrecord->field($1)){
1479                         my @abc = $marcrecord->field($1)->subfield($2);
1480                         $tags->{$tag} = $#abc + 1 ;
1481                     }
1482                 }
1483
1484                 # We catch how many times to repeat this line
1485                 my $max = 0;
1486                 foreach my $tag (keys(%$tags)){
1487                     $max = $tags->{$tag} if($tags->{$tag} > $max);
1488                  }
1489
1490                 # we replace, and repeat each line
1491                 for (my $i = 0 ; $i < $max ; $i++){
1492                     my $newline = $line;
1493
1494                     foreach my $tag ( $newline =~ /\[(\d{3}[\w|\d])\]/g ) {
1495                         $tag =~ /(.{3})(.)/;
1496
1497                         if($marcrecord->field($1)){
1498                             my @repl = $marcrecord->field($1)->subfield($2);
1499                             my $subfieldvalue = $repl[$i];
1500
1501                             if (! utf8::is_utf8($subfieldvalue)) {
1502                                 utf8::decode($subfieldvalue);
1503                             }
1504
1505                              $newline =~ s/\[$tag\]/$subfieldvalue/g;
1506                         }
1507                     }
1508                     $newsummary .= "$newline\n";
1509                 }
1510             }
1511
1512             $newsummary =~ s/\[(.*?)]//g;
1513             $newsummary =~ s/\n/<br\/>/g;
1514             $oldbiblio->{summary} = $newsummary;
1515         }
1516
1517         # Pull out the items fields
1518         my @fields = $marcrecord->field($itemtag);
1519
1520         # Setting item statuses for display
1521         my @available_items_loop;
1522         my @onloan_items_loop;
1523         my @other_items_loop;
1524
1525         my $available_items;
1526         my $onloan_items;
1527         my $other_items;
1528
1529         my $ordered_count         = 0;
1530         my $available_count       = 0;
1531         my $onloan_count          = 0;
1532         my $longoverdue_count     = 0;
1533         my $other_count           = 0;
1534         my $wthdrawn_count        = 0;
1535         my $itemlost_count        = 0;
1536         my $itembinding_count     = 0;
1537         my $itemdamaged_count     = 0;
1538         my $item_in_transit_count = 0;
1539         my $can_place_holds       = 0;
1540         my $item_onhold_count     = 0;
1541         my $items_count           = scalar(@fields);
1542         my $maxitems =
1543           ( C4::Context->preference('maxItemsinSearchResults') )
1544           ? C4::Context->preference('maxItemsinSearchResults') - 1
1545           : 1;
1546
1547         # loop through every item
1548         foreach my $field (@fields) {
1549             my $item;
1550
1551             # populate the items hash
1552             foreach my $code ( keys %subfieldstosearch ) {
1553                 $item->{$code} = $field->subfield( $subfieldstosearch{$code} );
1554             }
1555
1556                         my $hbranch     = C4::Context->preference('HomeOrHoldingBranch') eq 'homebranch' ? 'homebranch'    : 'holdingbranch';
1557                         my $otherbranch = C4::Context->preference('HomeOrHoldingBranch') eq 'homebranch' ? 'holdingbranch' : 'homebranch';
1558             # set item's branch name, use HomeOrHoldingBranch syspref first, fall back to the other one
1559             if ($item->{$hbranch}) {
1560                 $item->{'branchname'} = $branches{$item->{$hbranch}};
1561             }
1562             elsif ($item->{$otherbranch}) {     # Last resort
1563                 $item->{'branchname'} = $branches{$item->{$otherbranch}};
1564             }
1565
1566                         my $prefix = $item->{$hbranch} . '--' . $item->{location} . $item->{itype} . $item->{itemcallnumber};
1567 # For each grouping of items (onloan, available, unavailable), we build a key to store relevant info about that item
1568             if ( $item->{onloan} ) {
1569                 $onloan_count++;
1570                                 my $key = $prefix . $item->{onloan} . $item->{barcode};
1571                                 $onloan_items->{$key}->{due_date} = format_date($item->{onloan});
1572                                 $onloan_items->{$key}->{count}++ if $item->{$hbranch};
1573                                 $onloan_items->{$key}->{branchname} = $item->{branchname};
1574                                 $onloan_items->{$key}->{location} = $shelflocations->{ $item->{location} };
1575                                 $onloan_items->{$key}->{itemcallnumber} = $item->{itemcallnumber};
1576                                 $onloan_items->{$key}->{imageurl} = getitemtypeimagelocation( 'opac', $itemtypes{ $item->{itype} }->{imageurl} );
1577                 # if something's checked out and lost, mark it as 'long overdue'
1578                 if ( $item->{itemlost} ) {
1579                     $onloan_items->{$prefix}->{longoverdue}++;
1580                     $longoverdue_count++;
1581                 } else {        # can place holds as long as item isn't lost
1582                     $can_place_holds = 1;
1583                 }
1584             }
1585
1586          # items not on loan, but still unavailable ( lost, withdrawn, damaged )
1587             else {
1588
1589                 # item is on order
1590                 if ( $item->{notforloan} == -1 ) {
1591                     $ordered_count++;
1592                 }
1593
1594                 # is item in transit?
1595                 my $transfertwhen = '';
1596                 my ($transfertfrom, $transfertto);
1597
1598                 # is item on the reserve shelf?
1599                 my $reservestatus = 0;
1600                 my $reserveitem;
1601
1602                 unless ($item->{wthdrawn}
1603                         || $item->{itemlost}
1604                         || $item->{damaged}
1605                         || $item->{notforloan}
1606                         || $items_count > 20) {
1607
1608                     # A couple heuristics to limit how many times
1609                     # we query the database for item transfer information, sacrificing
1610                     # accuracy in some cases for speed;
1611                     #
1612                     # 1. don't query if item has one of the other statuses
1613                     # 2. don't check transit status if the bib has
1614                     #    more than 20 items
1615                     #
1616                     # FIXME: to avoid having the query the database like this, and to make
1617                     #        the in transit status count as unavailable for search limiting,
1618                     #        should map transit status to record indexed in Zebra.
1619                     #
1620                     ($transfertwhen, $transfertfrom, $transfertto) = C4::Circulation::GetTransfers($item->{itemnumber});
1621                     ($reservestatus, $reserveitem) = C4::Reserves::CheckReserves($item->{itemnumber});
1622                 }
1623
1624                 # item is withdrawn, lost or damaged
1625                 if (   $item->{wthdrawn}
1626                     || $item->{itemlost}
1627                     || $item->{damaged}
1628                     || $item->{notforloan} > 0
1629                     || $reservestatus eq 'Waiting'
1630                     || ($transfertwhen ne ''))
1631                 {
1632                     $wthdrawn_count++        if $item->{wthdrawn};
1633                     $itemlost_count++        if $item->{itemlost};
1634                     $itemdamaged_count++     if $item->{damaged};
1635                     $item_in_transit_count++ if $transfertwhen ne '';
1636                     $item_onhold_count++     if $reservestatus eq 'Waiting';
1637                     $item->{status} = $item->{wthdrawn} . "-" . $item->{itemlost} . "-" . $item->{damaged} . "-" . $item->{notforloan};
1638                     $other_count++;
1639
1640                                         my $key = $prefix . $item->{status};
1641                                         foreach (qw(wthdrawn itemlost damaged branchname itemcallnumber)) {
1642                         $other_items->{$key}->{$_} = $item->{$_};
1643                                         }
1644                     $other_items->{$key}->{intransit} = ($transfertwhen ne '') ? 1 : 0;
1645                     $other_items->{$key}->{onhold} = ($reservestatus) ? 1 : 0;
1646                                         $other_items->{$key}->{notforloan} = GetAuthorisedValueDesc('','',$item->{notforloan},'','',$notforloan_authorised_value) if $notforloan_authorised_value;
1647                                         $other_items->{$key}->{count}++ if $item->{$hbranch};
1648                                         $other_items->{$key}->{location} = $shelflocations->{ $item->{location} };
1649                                         $other_items->{$key}->{imageurl} = getitemtypeimagelocation( 'opac', $itemtypes{ $item->{itype} }->{imageurl} );
1650                 }
1651                 # item is available
1652                 else {
1653                     $can_place_holds = 1;
1654                     $available_count++;
1655                                         $available_items->{$prefix}->{count}++ if $item->{$hbranch};
1656                                         foreach (qw(branchname itemcallnumber)) {
1657                         $available_items->{$prefix}->{$_} = $item->{$_};
1658                                         }
1659                                         $available_items->{$prefix}->{location} = $shelflocations->{ $item->{location} };
1660                                         $available_items->{$prefix}->{imageurl} = getitemtypeimagelocation( 'opac', $itemtypes{ $item->{itype} }->{imageurl} );
1661                 }
1662             }
1663         }    # notforloan, item level and biblioitem level
1664         my ( $availableitemscount, $onloanitemscount, $otheritemscount );
1665         $maxitems =
1666           ( C4::Context->preference('maxItemsinSearchResults') )
1667           ? C4::Context->preference('maxItemsinSearchResults') - 1
1668           : 1;
1669         for my $key ( sort keys %$onloan_items ) {
1670             (++$onloanitemscount > $maxitems) and last;
1671             push @onloan_items_loop, $onloan_items->{$key};
1672         }
1673         for my $key ( sort keys %$other_items ) {
1674             (++$otheritemscount > $maxitems) and last;
1675             push @other_items_loop, $other_items->{$key};
1676         }
1677         for my $key ( sort keys %$available_items ) {
1678             (++$availableitemscount > $maxitems) and last;
1679             push @available_items_loop, $available_items->{$key}
1680         }
1681
1682         # XSLT processing of some stuff
1683         use C4::Charset;
1684         SetUTF8Flag($marcrecord);
1685         $debug && warn $marcrecord->as_formatted;
1686         if (!$scan && $search_context eq 'opac' && C4::Context->preference("OPACXSLTResultsDisplay")) {
1687             # FIXME note that XSLTResultsDisplay (use of XSLT to format staff interface bib search results)
1688             # is not implemented yet
1689             $oldbiblio->{XSLTResultsRecord} = XSLTParse4Display($oldbiblio->{biblionumber}, $marcrecord, 'Results',
1690                                                                 $search_context, 1);
1691                 # the last parameter tells Koha to clean up the problematic ampersand entities that Zebra outputs
1692
1693         }
1694
1695         # last check for norequest : if itemtype is notforloan, it can't be reserved either, whatever the items
1696         $can_place_holds = 0
1697           if $itemtypes{ $oldbiblio->{itemtype} }->{notforloan};
1698         $oldbiblio->{norequests} = 1 unless $can_place_holds;
1699         $oldbiblio->{itemsplural}          = 1 if $items_count > 1;
1700         $oldbiblio->{items_count}          = $items_count;
1701         $oldbiblio->{available_items_loop} = \@available_items_loop;
1702         $oldbiblio->{onloan_items_loop}    = \@onloan_items_loop;
1703         $oldbiblio->{other_items_loop}     = \@other_items_loop;
1704         $oldbiblio->{availablecount}       = $available_count;
1705         $oldbiblio->{availableplural}      = 1 if $available_count > 1;
1706         $oldbiblio->{onloancount}          = $onloan_count;
1707         $oldbiblio->{onloanplural}         = 1 if $onloan_count > 1;
1708         $oldbiblio->{othercount}           = $other_count;
1709         $oldbiblio->{otherplural}          = 1 if $other_count > 1;
1710         $oldbiblio->{wthdrawncount}        = $wthdrawn_count;
1711         $oldbiblio->{itemlostcount}        = $itemlost_count;
1712         $oldbiblio->{damagedcount}         = $itemdamaged_count;
1713         $oldbiblio->{intransitcount}       = $item_in_transit_count;
1714         $oldbiblio->{onholdcount}          = $item_onhold_count;
1715         $oldbiblio->{orderedcount}         = $ordered_count;
1716         $oldbiblio->{isbn} =~
1717           s/-//g;    # deleting - in isbn to enable amazon content
1718         push( @newresults, $oldbiblio )
1719             if(not $hidelostitems
1720                or (($items_count > $itemlost_count )
1721                     && $hidelostitems));
1722     }
1723
1724     return @newresults;
1725 }
1726
1727 =head2 SearchAcquisitions
1728     Search for acquisitions
1729 =cut
1730
1731 sub SearchAcquisitions{
1732     my ($datebegin, $dateend, $itemtypes,$criteria, $orderby) = @_;
1733
1734     my $dbh=C4::Context->dbh;
1735     # Variable initialization
1736     my $str=qq|
1737     SELECT marcxml
1738     FROM biblio
1739     LEFT JOIN biblioitems ON biblioitems.biblionumber=biblio.biblionumber
1740     LEFT JOIN items ON items.biblionumber=biblio.biblionumber
1741     WHERE dateaccessioned BETWEEN ? AND ?
1742     |;
1743
1744     my (@params,@loopcriteria);
1745
1746     push @params, $datebegin->output("iso");
1747     push @params, $dateend->output("iso");
1748
1749     if (scalar(@$itemtypes)>0 and $criteria ne "itemtype" ){
1750         if(C4::Context->preference("item-level_itypes")){
1751             $str .= "AND items.itype IN (?".( ',?' x scalar @$itemtypes - 1 ).") ";
1752         }else{
1753             $str .= "AND biblioitems.itemtype IN (?".( ',?' x scalar @$itemtypes - 1 ).") ";
1754         }
1755         push @params, @$itemtypes;
1756     }
1757
1758     if ($criteria =~/itemtype/){
1759         if(C4::Context->preference("item-level_itypes")){
1760             $str .= "AND items.itype=? ";
1761         }else{
1762             $str .= "AND biblioitems.itemtype=? ";
1763         }
1764
1765         if(scalar(@$itemtypes) == 0){
1766             my $itypes = GetItemTypes();
1767             for my $key (keys %$itypes){
1768                 push @$itemtypes, $key;
1769             }
1770         }
1771
1772         @loopcriteria= @$itemtypes;
1773     }elsif ($criteria=~/itemcallnumber/){
1774         $str .= "AND (items.itemcallnumber LIKE CONCAT(?,'%')
1775                  OR items.itemcallnumber is NULL
1776                  OR items.itemcallnumber = '')";
1777
1778         @loopcriteria = ("AA".."ZZ", "") unless (scalar(@loopcriteria)>0);
1779     }else {
1780         $str .= "AND biblio.title LIKE CONCAT(?,'%') ";
1781         @loopcriteria = ("A".."z") unless (scalar(@loopcriteria)>0);
1782     }
1783
1784     if ($orderby =~ /date_desc/){
1785         $str.=" ORDER BY dateaccessioned DESC";
1786     } else {
1787         $str.=" ORDER BY title";
1788     }
1789
1790     my $qdataacquisitions=$dbh->prepare($str);
1791
1792     my @loopacquisitions;
1793     foreach my $value(@loopcriteria){
1794         push @params,$value;
1795         my %cell;
1796         $cell{"title"}=$value;
1797         $cell{"titlecode"}=$value;
1798
1799         eval{$qdataacquisitions->execute(@params);};
1800
1801         if ($@){ warn "recentacquisitions Error :$@";}
1802         else {
1803             my @loopdata;
1804             while (my $data=$qdataacquisitions->fetchrow_hashref){
1805                 push @loopdata, {"summary"=>GetBiblioSummary( $data->{'marcxml'} ) };
1806             }
1807             $cell{"loopdata"}=\@loopdata;
1808         }
1809         push @loopacquisitions,\%cell if (scalar(@{$cell{loopdata}})>0);
1810         pop @params;
1811     }
1812     $qdataacquisitions->finish;
1813     return \@loopacquisitions;
1814 }
1815 #----------------------------------------------------------------------
1816 #
1817 # Non-Zebra GetRecords#
1818 #----------------------------------------------------------------------
1819
1820 =head2 NZgetRecords
1821
1822   NZgetRecords has the same API as zera getRecords, even if some parameters are not managed
1823
1824 =cut
1825
1826 sub NZgetRecords {
1827     my (
1828         $query,            $simple_query, $sort_by_ref,    $servers_ref,
1829         $results_per_page, $offset,       $expanded_facet, $branches,
1830         $query_type,       $scan
1831     ) = @_;
1832     warn "query =$query" if $DEBUG;
1833     my $result = NZanalyse($query);
1834     warn "results =$result" if $DEBUG;
1835     return ( undef,
1836         NZorder( $result, @$sort_by_ref[0], $results_per_page, $offset ),
1837         undef );
1838 }
1839
1840 =head2 NZanalyse
1841
1842   NZanalyse : get a CQL string as parameter, and returns a list of biblionumber;title,biblionumber;title,...
1843   the list is built from an inverted index in the nozebra SQL table
1844   note that title is here only for convenience : the sorting will be very fast when requested on title
1845   if the sorting is requested on something else, we will have to reread all results, and that may be longer.
1846
1847 =cut
1848
1849 sub NZanalyse {
1850     my ( $string, $server ) = @_;
1851 #     warn "---------"       if $DEBUG;
1852     warn " NZanalyse" if $DEBUG;
1853 #     warn "---------"       if $DEBUG;
1854
1855  # $server contains biblioserver or authorities, depending on what we search on.
1856  #warn "querying : $string on $server";
1857     $server = 'biblioserver' unless $server;
1858
1859 # if we have a ", replace the content to discard temporarily any and/or/not inside
1860     my $commacontent;
1861     if ( $string =~ /"/ ) {
1862         $string =~ s/"(.*?)"/__X__/;
1863         $commacontent = $1;
1864         warn "commacontent : $commacontent" if $DEBUG;
1865     }
1866
1867 # split the query string in 3 parts : X AND Y means : $left="X", $operand="AND" and $right="Y"
1868 # then, call again NZanalyse with $left and $right
1869 # (recursive until we find a leaf (=> something without and/or/not)
1870 # delete repeated operator... Would then go in infinite loop
1871     while ( $string =~ s/( and| or| not| AND| OR| NOT)\1/$1/g ) {
1872     }
1873
1874     #process parenthesis before.
1875     if ( $string =~ /^\s*\((.*)\)(( and | or | not | AND | OR | NOT )(.*))?/ ) {
1876         my $left     = $1;
1877         my $right    = $4;
1878         my $operator = lc($3);   # FIXME: and/or/not are operators, not operands
1879         warn
1880 "dealing w/parenthesis before recursive sub call. left :$left operator:$operator right:$right"
1881           if $DEBUG;
1882         my $leftresult = NZanalyse( $left, $server );
1883         if ($operator) {
1884             my $rightresult = NZanalyse( $right, $server );
1885
1886             # OK, we have the results for right and left part of the query
1887             # depending of operand, intersect, union or exclude both lists
1888             # to get a result list
1889             if ( $operator eq ' and ' ) {
1890                 return NZoperatorAND($leftresult,$rightresult);
1891             }
1892             elsif ( $operator eq ' or ' ) {
1893
1894                 # just merge the 2 strings
1895                 return $leftresult . $rightresult;
1896             }
1897             elsif ( $operator eq ' not ' ) {
1898                 return NZoperatorNOT($leftresult,$rightresult);
1899             }
1900         }
1901         else {
1902 # this error is impossible, because of the regexp that isolate the operand, but just in case...
1903             return $leftresult;
1904         }
1905     }
1906     warn "string :" . $string if $DEBUG;
1907     my $left = "";
1908     my $right = "";
1909     my $operator = "";
1910     if ($string =~ /(.*?)( and | or | not | AND | OR | NOT )(.*)/) {
1911         $left     = $1;
1912         $right    = $3;
1913         $operator = lc($2);    # FIXME: and/or/not are operators, not operands
1914     }
1915     warn "no parenthesis. left : $left operator: $operator right: $right"
1916       if $DEBUG;
1917
1918     # it's not a leaf, we have a and/or/not
1919     if ($operator) {
1920
1921         # reintroduce comma content if needed
1922         $right =~ s/__X__/"$commacontent"/ if $commacontent;
1923         $left  =~ s/__X__/"$commacontent"/ if $commacontent;
1924         warn "node : $left / $operator / $right\n" if $DEBUG;
1925         my $leftresult  = NZanalyse( $left,  $server );
1926         my $rightresult = NZanalyse( $right, $server );
1927         warn " leftresult : $leftresult" if $DEBUG;
1928         warn " rightresult : $rightresult" if $DEBUG;
1929         # OK, we have the results for right and left part of the query
1930         # depending of operand, intersect, union or exclude both lists
1931         # to get a result list
1932         if ( $operator eq ' and ' ) {
1933             return NZoperatorAND($leftresult,$rightresult);
1934         }
1935         elsif ( $operator eq ' or ' ) {
1936
1937             # just merge the 2 strings
1938             return $leftresult . $rightresult;
1939         }
1940         elsif ( $operator eq ' not ' ) {
1941             return NZoperatorNOT($leftresult,$rightresult);
1942         }
1943         else {
1944
1945 # this error is impossible, because of the regexp that isolate the operand, but just in case...
1946             die "error : operand unknown : $operator for $string";
1947         }
1948
1949         # it's a leaf, do the real SQL query and return the result
1950     }
1951     else {
1952         $string =~ s/__X__/"$commacontent"/ if $commacontent;
1953         $string =~ s/-|\.|\?|,|;|!|'|\(|\)|\[|\]|{|}|"|&|\+|\*|\// /g;
1954         #remove trailing blank at the beginning
1955         $string =~ s/^ //g;
1956         warn "leaf:$string" if $DEBUG;
1957
1958         # parse the string in in operator/operand/value again
1959         my $left = "";
1960         my $operator = "";
1961         my $right = "";
1962         if ($string =~ /(.*)(>=|<=)(.*)/) {
1963             $left     = $1;
1964             $operator = $2;
1965             $right    = $3;
1966         } else {
1967             $left = $string;
1968         }
1969 #         warn "handling leaf... left:$left operator:$operator right:$right"
1970 #           if $DEBUG;
1971         unless ($operator) {
1972             if ($string =~ /(.*)(>|<|=)(.*)/) {
1973                 $left     = $1;
1974                 $operator = $2;
1975                 $right    = $3;
1976                 warn
1977     "handling unless (operator)... left:$left operator:$operator right:$right"
1978                 if $DEBUG;
1979             } else {
1980                 $left = $string;
1981             }
1982         }
1983         my $results;
1984
1985 # strip adv, zebra keywords, currently not handled in nozebra: wrdl, ext, phr...
1986         $left =~ s/ .*$//;
1987
1988         # automatic replace for short operators
1989         $left = 'title'            if $left =~ '^ti$';
1990         $left = 'author'           if $left =~ '^au$';
1991         $left = 'publisher'        if $left =~ '^pb$';
1992         $left = 'subject'          if $left =~ '^su$';
1993         $left = 'koha-Auth-Number' if $left =~ '^an$';
1994         $left = 'keyword'          if $left =~ '^kw$';
1995         $left = 'itemtype'         if $left =~ '^mc$'; # Fix for Bug 2599 - Search limits not working for NoZebra
1996         warn "handling leaf... left:$left operator:$operator right:$right" if $DEBUG;
1997         my $dbh = C4::Context->dbh;
1998         if ( $operator && $left ne 'keyword' ) {
1999             #do a specific search
2000             $operator = 'LIKE' if $operator eq '=' and $right =~ /%/;
2001             my $sth = $dbh->prepare(
2002 "SELECT biblionumbers,value FROM nozebra WHERE server=? AND indexname=? AND value $operator ?"
2003             );
2004             warn "$left / $operator / $right\n" if $DEBUG;
2005
2006             # split each word, query the DB and build the biblionumbers result
2007             #sanitizing leftpart
2008             $left =~ s/^\s+|\s+$//;
2009             foreach ( split / /, $right ) {
2010                 my $biblionumbers;
2011                 $_ =~ s/^\s+|\s+$//;
2012                 next unless $_;
2013                 warn "EXECUTE : $server, $left, $_" if $DEBUG;
2014                 $sth->execute( $server, $left, $_ )
2015                   or warn "execute failed: $!";
2016                 while ( my ( $line, $value ) = $sth->fetchrow ) {
2017
2018 # if we are dealing with a numeric value, use only numeric results (in case of >=, <=, > or <)
2019 # otherwise, fill the result
2020                     $biblionumbers .= $line
2021                       unless ( $right =~ /^\d+$/ && $value =~ /\D/ );
2022                     warn "result : $value "
2023                       . ( $right  =~ /\d/ ) . "=="
2024                       . ( $value =~ /\D/?$line:"" ) if $DEBUG;         #= $line";
2025                 }
2026
2027 # do a AND with existing list if there is one, otherwise, use the biblionumbers list as 1st result list
2028                 if ($results) {
2029                     warn "NZAND" if $DEBUG;
2030                     $results = NZoperatorAND($biblionumbers,$results);
2031                 } else {
2032                     $results = $biblionumbers;
2033                 }
2034             }
2035         }
2036         else {
2037       #do a complete search (all indexes), if index='kw' do complete search too.
2038             my $sth = $dbh->prepare(
2039 "SELECT biblionumbers FROM nozebra WHERE server=? AND value LIKE ?"
2040             );
2041
2042             # split each word, query the DB and build the biblionumbers result
2043             foreach ( split / /, $string ) {
2044                 next if C4::Context->stopwords->{ uc($_) };   # skip if stopword
2045                 warn "search on all indexes on $_" if $DEBUG;
2046                 my $biblionumbers;
2047                 next unless $_;
2048                 $sth->execute( $server, $_ );
2049                 while ( my $line = $sth->fetchrow ) {
2050                     $biblionumbers .= $line;
2051                 }
2052
2053 # do a AND with existing list if there is one, otherwise, use the biblionumbers list as 1st result list
2054                 if ($results) {
2055                     $results = NZoperatorAND($biblionumbers,$results);
2056                 }
2057                 else {
2058                     warn "NEW RES for $_ = $biblionumbers" if $DEBUG;
2059                     $results = $biblionumbers;
2060                 }
2061             }
2062         }
2063         warn "return : $results for LEAF : $string" if $DEBUG;
2064         return $results;
2065     }
2066     warn "---------\nLeave NZanalyse\n---------" if $DEBUG;
2067 }
2068
2069 sub NZoperatorAND{
2070     my ($rightresult, $leftresult)=@_;
2071
2072     my @leftresult = split /;/, $leftresult;
2073     warn " @leftresult / $rightresult \n" if $DEBUG;
2074
2075     #             my @rightresult = split /;/,$leftresult;
2076     my $finalresult;
2077
2078 # parse the left results, and if the biblionumber exist in the right result, save it in finalresult
2079 # the result is stored twice, to have the same weight for AND than OR.
2080 # example : TWO : 61,61,64,121 (two is twice in the biblio #61) / TOWER : 61,64,130
2081 # result : 61,61,61,61,64,64 for two AND tower : 61 has more weight than 64
2082     foreach (@leftresult) {
2083         my $value = $_;
2084         my $countvalue;
2085         ( $value, $countvalue ) = ( $1, $2 ) if ($value=~/(.*)-(\d+)$/);
2086         if ( $rightresult =~ /\Q$value\E-(\d+);/ ) {
2087             $countvalue = ( $1 > $countvalue ? $countvalue : $1 );
2088             $finalresult .=
2089                 "$value-$countvalue;$value-$countvalue;";
2090         }
2091     }
2092     warn "NZAND DONE : $finalresult \n" if $DEBUG;
2093     return $finalresult;
2094 }
2095
2096 sub NZoperatorOR{
2097     my ($rightresult, $leftresult)=@_;
2098     return $rightresult.$leftresult;
2099 }
2100
2101 sub NZoperatorNOT{
2102     my ($leftresult, $rightresult)=@_;
2103
2104     my @leftresult = split /;/, $leftresult;
2105
2106     #             my @rightresult = split /;/,$leftresult;
2107     my $finalresult;
2108     foreach (@leftresult) {
2109         my $value=$_;
2110         $value=$1 if $value=~m/(.*)-\d+$/;
2111         unless ($rightresult =~ "$value-") {
2112             $finalresult .= "$_;";
2113         }
2114     }
2115     return $finalresult;
2116 }
2117
2118 =head2 NZorder
2119
2120   $finalresult = NZorder($biblionumbers, $ordering,$results_per_page,$offset);
2121
2122   TODO :: Description
2123
2124 =cut
2125
2126 sub NZorder {
2127     my ( $biblionumbers, $ordering, $results_per_page, $offset ) = @_;
2128     warn "biblionumbers = $biblionumbers and ordering = $ordering\n" if $DEBUG;
2129
2130     # order title asc by default
2131     #     $ordering = '1=36 <i' unless $ordering;
2132     $results_per_page = 20 unless $results_per_page;
2133     $offset           = 0  unless $offset;
2134     my $dbh = C4::Context->dbh;
2135
2136     #
2137     # order by POPULARITY
2138     #
2139     if ( $ordering =~ /popularity/ ) {
2140         my %result;
2141         my %popularity;
2142
2143         # popularity is not in MARC record, it's builded from a specific query
2144         my $sth =
2145           $dbh->prepare("select sum(issues) from items where biblionumber=?");
2146         foreach ( split /;/, $biblionumbers ) {
2147             my ( $biblionumber, $title ) = split /,/, $_;
2148             $result{$biblionumber} = GetMarcBiblio($biblionumber);
2149             $sth->execute($biblionumber);
2150             my $popularity = $sth->fetchrow || 0;
2151
2152 # hint : the key is popularity.title because we can have
2153 # many results with the same popularity. In this case, sub-ordering is done by title
2154 # we also have biblionumber to avoid bug for 2 biblios with the same title & popularity
2155 # (un-frequent, I agree, but we won't forget anything that way ;-)
2156             $popularity{ sprintf( "%10d", $popularity ) . $title
2157                   . $biblionumber } = $biblionumber;
2158         }
2159
2160     # sort the hash and return the same structure as GetRecords (Zebra querying)
2161         my $result_hash;
2162         my $numbers = 0;
2163         if ( $ordering eq 'popularity_dsc' ) {    # sort popularity DESC
2164             foreach my $key ( sort { $b cmp $a } ( keys %popularity ) ) {
2165                 $result_hash->{'RECORDS'}[ $numbers++ ] =
2166                   $result{ $popularity{$key} }->as_usmarc();
2167             }
2168         }
2169         else {                                    # sort popularity ASC
2170             foreach my $key ( sort ( keys %popularity ) ) {
2171                 $result_hash->{'RECORDS'}[ $numbers++ ] =
2172                   $result{ $popularity{$key} }->as_usmarc();
2173             }
2174         }
2175         my $finalresult = ();
2176         $result_hash->{'hits'}         = $numbers;
2177         $finalresult->{'biblioserver'} = $result_hash;
2178         return $finalresult;
2179
2180         #
2181         # ORDER BY author
2182         #
2183     }
2184     elsif ( $ordering =~ /author/ ) {
2185         my %result;
2186         foreach ( split /;/, $biblionumbers ) {
2187             my ( $biblionumber, $title ) = split /,/, $_;
2188             my $record = GetMarcBiblio($biblionumber);
2189             my $author;
2190             if ( C4::Context->preference('marcflavour') eq 'UNIMARC' ) {
2191                 $author = $record->subfield( '200', 'f' );
2192                 $author = $record->subfield( '700', 'a' ) unless $author;
2193             }
2194             else {
2195                 $author = $record->subfield( '100', 'a' );
2196             }
2197
2198 # hint : the result is sorted by title.biblionumber because we can have X biblios with the same title
2199 # and we don't want to get only 1 result for each of them !!!
2200             $result{ $author . $biblionumber } = $record;
2201         }
2202
2203     # sort the hash and return the same structure as GetRecords (Zebra querying)
2204         my $result_hash;
2205         my $numbers = 0;
2206         if ( $ordering eq 'author_za' ) {    # sort by author desc
2207             foreach my $key ( sort { $b cmp $a } ( keys %result ) ) {
2208                 $result_hash->{'RECORDS'}[ $numbers++ ] =
2209                   $result{$key}->as_usmarc();
2210             }
2211         }
2212         else {                               # sort by author ASC
2213             foreach my $key ( sort ( keys %result ) ) {
2214                 $result_hash->{'RECORDS'}[ $numbers++ ] =
2215                   $result{$key}->as_usmarc();
2216             }
2217         }
2218         my $finalresult = ();
2219         $result_hash->{'hits'}         = $numbers;
2220         $finalresult->{'biblioserver'} = $result_hash;
2221         return $finalresult;
2222
2223         #
2224         # ORDER BY callnumber
2225         #
2226     }
2227     elsif ( $ordering =~ /callnumber/ ) {
2228         my %result;
2229         foreach ( split /;/, $biblionumbers ) {
2230             my ( $biblionumber, $title ) = split /,/, $_;
2231             my $record = GetMarcBiblio($biblionumber);
2232             my $callnumber;
2233             my $frameworkcode = GetFrameworkCode($biblionumber);
2234             my ( $callnumber_tag, $callnumber_subfield ) = GetMarcFromKohaField(  'items.itemcallnumber', $frameworkcode);
2235                ( $callnumber_tag, $callnumber_subfield ) = GetMarcFromKohaField('biblioitems.callnumber', $frameworkcode)
2236                 unless $callnumber_tag;
2237             if ( C4::Context->preference('marcflavour') eq 'UNIMARC' ) {
2238                 $callnumber = $record->subfield( '200', 'f' );
2239             } else {
2240                 $callnumber = $record->subfield( '100', 'a' );
2241             }
2242
2243 # hint : the result is sorted by title.biblionumber because we can have X biblios with the same title
2244 # and we don't want to get only 1 result for each of them !!!
2245             $result{ $callnumber . $biblionumber } = $record;
2246         }
2247
2248     # sort the hash and return the same structure as GetRecords (Zebra querying)
2249         my $result_hash;
2250         my $numbers = 0;
2251         if ( $ordering eq 'call_number_dsc' ) {    # sort by title desc
2252             foreach my $key ( sort { $b cmp $a } ( keys %result ) ) {
2253                 $result_hash->{'RECORDS'}[ $numbers++ ] =
2254                   $result{$key}->as_usmarc();
2255             }
2256         }
2257         else {                                     # sort by title ASC
2258             foreach my $key ( sort { $a cmp $b } ( keys %result ) ) {
2259                 $result_hash->{'RECORDS'}[ $numbers++ ] =
2260                   $result{$key}->as_usmarc();
2261             }
2262         }
2263         my $finalresult = ();
2264         $result_hash->{'hits'}         = $numbers;
2265         $finalresult->{'biblioserver'} = $result_hash;
2266         return $finalresult;
2267     }
2268     elsif ( $ordering =~ /pubdate/ ) {             #pub year
2269         my %result;
2270         foreach ( split /;/, $biblionumbers ) {
2271             my ( $biblionumber, $title ) = split /,/, $_;
2272             my $record = GetMarcBiblio($biblionumber);
2273             my ( $publicationyear_tag, $publicationyear_subfield ) =
2274               GetMarcFromKohaField( 'biblioitems.publicationyear', '' );
2275             my $publicationyear =
2276               $record->subfield( $publicationyear_tag,
2277                 $publicationyear_subfield );
2278
2279 # hint : the result is sorted by title.biblionumber because we can have X biblios with the same title
2280 # and we don't want to get only 1 result for each of them !!!
2281             $result{ $publicationyear . $biblionumber } = $record;
2282         }
2283
2284     # sort the hash and return the same structure as GetRecords (Zebra querying)
2285         my $result_hash;
2286         my $numbers = 0;
2287         if ( $ordering eq 'pubdate_dsc' ) {    # sort by pubyear desc
2288             foreach my $key ( sort { $b cmp $a } ( keys %result ) ) {
2289                 $result_hash->{'RECORDS'}[ $numbers++ ] =
2290                   $result{$key}->as_usmarc();
2291             }
2292         }
2293         else {                                 # sort by pub year ASC
2294             foreach my $key ( sort ( keys %result ) ) {
2295                 $result_hash->{'RECORDS'}[ $numbers++ ] =
2296                   $result{$key}->as_usmarc();
2297             }
2298         }
2299         my $finalresult = ();
2300         $result_hash->{'hits'}         = $numbers;
2301         $finalresult->{'biblioserver'} = $result_hash;
2302         return $finalresult;
2303
2304         #
2305         # ORDER BY title
2306         #
2307     }
2308     elsif ( $ordering =~ /title/ ) {
2309
2310 # the title is in the biblionumbers string, so we just need to build a hash, sort it and return
2311         my %result;
2312         foreach ( split /;/, $biblionumbers ) {
2313             my ( $biblionumber, $title ) = split /,/, $_;
2314
2315 # hint : the result is sorted by title.biblionumber because we can have X biblios with the same title
2316 # and we don't want to get only 1 result for each of them !!!
2317 # hint & speed improvement : we can order without reading the record
2318 # so order, and read records only for the requested page !
2319             $result{ $title . $biblionumber } = $biblionumber;
2320         }
2321
2322     # sort the hash and return the same structure as GetRecords (Zebra querying)
2323         my $result_hash;
2324         my $numbers = 0;
2325         if ( $ordering eq 'title_az' ) {    # sort by title desc
2326             foreach my $key ( sort ( keys %result ) ) {
2327                 $result_hash->{'RECORDS'}[ $numbers++ ] = $result{$key};
2328             }
2329         }
2330         else {                              # sort by title ASC
2331             foreach my $key ( sort { $b cmp $a } ( keys %result ) ) {
2332                 $result_hash->{'RECORDS'}[ $numbers++ ] = $result{$key};
2333             }
2334         }
2335
2336         # limit the $results_per_page to result size if it's more
2337         $results_per_page = $numbers - 1 if $numbers < $results_per_page;
2338
2339         # for the requested page, replace biblionumber by the complete record
2340         # speed improvement : avoid reading too much things
2341         for (
2342             my $counter = $offset ;
2343             $counter <= $offset + $results_per_page ;
2344             $counter++
2345           )
2346         {
2347             $result_hash->{'RECORDS'}[$counter] =
2348               GetMarcBiblio( $result_hash->{'RECORDS'}[$counter] )->as_usmarc;
2349         }
2350         my $finalresult = ();
2351         $result_hash->{'hits'}         = $numbers;
2352         $finalresult->{'biblioserver'} = $result_hash;
2353         return $finalresult;
2354     }
2355     else {
2356
2357 #
2358 # order by ranking
2359 #
2360 # we need 2 hashes to order by ranking : the 1st one to count the ranking, the 2nd to order by ranking
2361         my %result;
2362         my %count_ranking;
2363         foreach ( split /;/, $biblionumbers ) {
2364             my ( $biblionumber, $title ) = split /,/, $_;
2365             $title =~ /(.*)-(\d)/;
2366
2367             # get weight
2368             my $ranking = $2;
2369
2370 # note that we + the ranking because ranking is calculated on weight of EACH term requested.
2371 # if we ask for "two towers", and "two" has weight 2 in biblio N, and "towers" has weight 4 in biblio N
2372 # biblio N has ranking = 6
2373             $count_ranking{$biblionumber} += $ranking;
2374         }
2375
2376 # build the result by "inverting" the count_ranking hash
2377 # 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
2378 #         warn "counting";
2379         foreach ( keys %count_ranking ) {
2380             $result{ sprintf( "%10d", $count_ranking{$_} ) . '-' . $_ } = $_;
2381         }
2382
2383     # sort the hash and return the same structure as GetRecords (Zebra querying)
2384         my $result_hash;
2385         my $numbers = 0;
2386         foreach my $key ( sort { $b cmp $a } ( keys %result ) ) {
2387             $result_hash->{'RECORDS'}[ $numbers++ ] = $result{$key};
2388         }
2389
2390         # limit the $results_per_page to result size if it's more
2391         $results_per_page = $numbers - 1 if $numbers < $results_per_page;
2392
2393         # for the requested page, replace biblionumber by the complete record
2394         # speed improvement : avoid reading too much things
2395         for (
2396             my $counter = $offset ;
2397             $counter <= $offset + $results_per_page ;
2398             $counter++
2399           )
2400         {
2401             $result_hash->{'RECORDS'}[$counter] =
2402               GetMarcBiblio( $result_hash->{'RECORDS'}[$counter] )->as_usmarc
2403               if $result_hash->{'RECORDS'}[$counter];
2404         }
2405         my $finalresult = ();
2406         $result_hash->{'hits'}         = $numbers;
2407         $finalresult->{'biblioserver'} = $result_hash;
2408         return $finalresult;
2409     }
2410 }
2411
2412 =head2 enabled_staff_search_views
2413
2414 %hash = enabled_staff_search_views()
2415
2416 This function returns a hash that contains three flags obtained from the system
2417 preferences, used to determine whether a particular staff search results view
2418 is enabled.
2419
2420 =over 2
2421
2422 =item C<Output arg:>
2423
2424     * $hash{can_view_MARC} is true only if the MARC view is enabled
2425     * $hash{can_view_ISBD} is true only if the ISBD view is enabled
2426     * $hash{can_view_labeledMARC} is true only if the Labeled MARC view is enabled
2427
2428 =item C<usage in the script:>
2429
2430 =back
2431
2432 $template->param ( C4::Search::enabled_staff_search_views );
2433
2434 =cut
2435
2436 sub enabled_staff_search_views
2437 {
2438         return (
2439                 can_view_MARC                   => C4::Context->preference('viewMARC'),                 # 1 if the staff search allows the MARC view
2440                 can_view_ISBD                   => C4::Context->preference('viewISBD'),                 # 1 if the staff search allows the ISBD view
2441                 can_view_labeledMARC    => C4::Context->preference('viewLabeledMARC'),  # 1 if the staff search allows the Labeled MARC view
2442         );
2443 }
2444
2445 sub AddSearchHistory{
2446         my ($borrowernumber,$session,$query_desc,$query_cgi, $total)=@_;
2447     my $dbh = C4::Context->dbh;
2448
2449     # Add the request the user just made
2450     my $sql = "INSERT INTO search_history(userid, sessionid, query_desc, query_cgi, total, time) VALUES(?, ?, ?, ?, ?, NOW())";
2451     my $sth   = $dbh->prepare($sql);
2452     $sth->execute($borrowernumber, $session, $query_desc, $query_cgi, $total);
2453         return $dbh->last_insert_id(undef, 'search_history', undef,undef,undef);
2454 }
2455
2456 sub GetSearchHistory{
2457         my ($borrowernumber,$session)=@_;
2458     my $dbh = C4::Context->dbh;
2459
2460     # Add the request the user just made
2461     my $query = "SELECT FROM search_history WHERE (userid=? OR sessionid=?)";
2462     my $sth   = $dbh->prepare($query);
2463         $sth->execute($borrowernumber, $session);
2464     return  $sth->fetchall_hashref({});
2465 }
2466
2467 =head2 z3950_search_args
2468
2469 $arrayref = z3950_search_args($matchpoints)
2470
2471 This function returns an array reference that contains the search parameters to be
2472 passed to the Z39.50 search script (z3950_search.pl). The array elements
2473 are hash refs whose keys are name, value and encvalue, and whose values are the
2474 name of a search parameter, the value of that search parameter and the URL encoded
2475 value of that parameter.
2476
2477 The search parameter names are lccn, isbn, issn, title, author, dewey and subject.
2478
2479 The search parameter values are obtained from the bibliographic record whose
2480 data is in a hash reference in $matchpoints, as returned by Biblio::GetBiblioData().
2481
2482 If $matchpoints is a scalar, it is assumed to be an unnamed query descriptor, e.g.
2483 a general purpose search argument. In this case, the returned array contains only
2484 entry: the key is 'title' and the value and encvalue are derived from $matchpoints.
2485
2486 If a search parameter value is undefined or empty, it is not included in the returned
2487 array.
2488
2489 The returned array reference may be passed directly to the template parameters.
2490
2491 =over 2
2492
2493 =item C<Output arg:>
2494
2495     * $array containing hash refs as described above
2496
2497 =item C<usage in the script:>
2498
2499 =back
2500
2501 $data = Biblio::GetBiblioData($bibno);
2502 $template->param ( MYLOOP => C4::Search::z3950_search_args($data) )
2503
2504 *OR*
2505
2506 $template->param ( MYLOOP => C4::Search::z3950_search_args($searchscalar) )
2507
2508 =cut
2509
2510 sub z3950_search_args {
2511     my $bibrec = shift;
2512     $bibrec = { title => $bibrec } if !ref $bibrec;
2513     my $array = [];
2514     for my $field (qw/ lccn isbn issn title author dewey subject /)
2515     {
2516         my $encvalue = URI::Escape::uri_escape_utf8($bibrec->{$field});
2517         push @$array, { name=>$field, value=>$bibrec->{$field}, encvalue=>$encvalue } if defined $bibrec->{$field};
2518     }
2519     return $array;
2520 }
2521
2522 =head2 BiblioAddAuthorities
2523
2524 ( $countlinked, $countcreated ) = BiblioAddAuthorities($record, $frameworkcode);
2525
2526 this function finds the authorities linked to the biblio
2527     * search in the authority DB for the same authid (in $9 of the biblio)
2528     * search in the authority DB for the same 001 (in $3 of the biblio in UNIMARC)
2529     * search in the authority DB for the same values (exactly) (in all subfields of the biblio)
2530 OR adds a new authority record
2531
2532 =over 2
2533
2534 =item C<input arg:>
2535
2536     * $record is the MARC record in question (marc blob)
2537     * $frameworkcode is the bibliographic framework to use (if it is "" it uses the default framework)
2538
2539 =item C<Output arg:>
2540
2541     * $countlinked is the number of authorities records that are linked to this authority
2542     * $countcreated
2543
2544 =item C<BUGS>
2545     * 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)
2546
2547 =back
2548
2549 =cut
2550
2551
2552 sub BiblioAddAuthorities{
2553   my ( $record, $frameworkcode ) = @_;
2554   my $dbh=C4::Context->dbh;
2555   my $query=$dbh->prepare(qq|
2556 SELECT authtypecode,tagfield
2557 FROM marc_subfield_structure
2558 WHERE frameworkcode=?
2559 AND (authtypecode IS NOT NULL AND authtypecode<>\"\")|);
2560 # SELECT authtypecode,tagfield
2561 # FROM marc_subfield_structure
2562 # WHERE frameworkcode=?
2563 # AND (authtypecode IS NOT NULL OR authtypecode<>\"\")|);
2564   $query->execute($frameworkcode);
2565   my ($countcreated,$countlinked);
2566   while (my $data=$query->fetchrow_hashref){
2567     foreach my $field ($record->field($data->{tagfield})){
2568       next if ($field->subfield('3')||$field->subfield('9'));
2569       # No authorities id in the tag.
2570       # Search if there is any authorities to link to.
2571       my $query='at='.$data->{authtypecode}.' ';
2572       map {$query.= ' and he,ext="'.$_->[1].'"' if ($_->[0]=~/[A-z]/)}  $field->subfields();
2573       my ($error, $results, $total_hits)=SimpleSearch( $query, undef, undef, [ "authorityserver" ] );
2574     # there is only 1 result
2575           if ( $error ) {
2576         warn "BIBLIOADDSAUTHORITIES: $error";
2577             return (0,0) ;
2578           }
2579       if ($results && scalar(@$results)==1) {
2580         my $marcrecord = MARC::File::USMARC::decode($results->[0]);
2581         $field->add_subfields('9'=>$marcrecord->field('001')->data);
2582         $countlinked++;
2583       } elsif (scalar(@$results)>1) {
2584    #More than One result
2585    #This can comes out of a lack of a subfield.
2586 #         my $marcrecord = MARC::File::USMARC::decode($results->[0]);
2587 #         $record->field($data->{tagfield})->add_subfields('9'=>$marcrecord->field('001')->data);
2588   $countlinked++;
2589       } else {
2590   #There are no results, build authority record, add it to Authorities, get authid and add it to 9
2591   ###NOTICE : This is only valid if a subfield is linked to one and only one authtypecode
2592   ###NOTICE : This can be a problem. We should also look into other types and rejected forms.
2593          my $authtypedata=C4::AuthoritiesMarc::GetAuthType($data->{authtypecode});
2594          next unless $authtypedata;
2595          my $marcrecordauth=MARC::Record->new();
2596          my $authfield=MARC::Field->new($authtypedata->{auth_tag_to_report},'','',"a"=>"".$field->subfield('a'));
2597          map { $authfield->add_subfields($_->[0]=>$_->[1]) if ($_->[0]=~/[A-z]/ && $_->[0] ne "a" )}  $field->subfields();
2598          $marcrecordauth->insert_fields_ordered($authfield);
2599
2600          # bug 2317: ensure new authority knows it's using UTF-8; currently
2601          # only need to do this for MARC21, as MARC::Record->as_xml_record() handles
2602          # automatically for UNIMARC (by not transcoding)
2603          # FIXME: AddAuthority() instead should simply explicitly require that the MARC::Record
2604          # use UTF-8, but as of 2008-08-05, did not want to introduce that kind
2605          # of change to a core API just before the 3.0 release.
2606          if (C4::Context->preference('marcflavour') eq 'MARC21') {
2607             SetMarcUnicodeFlag($marcrecordauth, 'MARC21');
2608          }
2609
2610 #          warn "AUTH RECORD ADDED : ".$marcrecordauth->as_formatted;
2611
2612          my $authid=AddAuthority($marcrecordauth,'',$data->{authtypecode});
2613          $countcreated++;
2614          $field->add_subfields('9'=>$authid);
2615       }
2616     }
2617   }
2618   return ($countlinked,$countcreated);
2619 }
2620
2621 =head2 GetDistinctValues($field);
2622
2623 C<$field> is a reference to the fields array
2624
2625 =cut
2626
2627 sub GetDistinctValues {
2628     my ($fieldname,$string)=@_;
2629     # returns a reference to a hash of references to branches...
2630     if ($fieldname=~/\./){
2631                         my ($table,$column)=split /\./, $fieldname;
2632                         my $dbh = C4::Context->dbh;
2633                         warn "select DISTINCT($column) as value, count(*) as cnt from $table group by lib order by $column " if $DEBUG;
2634                         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 ");
2635                         $sth->execute;
2636                         my $elements=$sth->fetchall_arrayref({});
2637                         return $elements;
2638    }
2639    else {
2640                 $string||= qq("");
2641                 my @servers=qw<biblioserver authorityserver>;
2642                 my (@zconns,@results);
2643         for ( my $i = 0 ; $i < @servers ; $i++ ) {
2644                 $zconns[$i] = C4::Context->Zconn( $servers[$i], 1 );
2645                         $results[$i] =
2646                       $zconns[$i]->scan(
2647                         ZOOM::Query::CCL2RPN->new( qq"$fieldname $string", $zconns[$i])
2648                       );
2649                 }
2650                 # The big moment: asynchronously retrieve results from all servers
2651                 my @elements;
2652                 while ( ( my $i = ZOOM::event( \@zconns ) ) != 0 ) {
2653                         my $ev = $zconns[ $i - 1 ]->last_event();
2654                         if ( $ev == ZOOM::Event::ZEND ) {
2655                                 next unless $results[ $i - 1 ];
2656                                 my $size = $results[ $i - 1 ]->size();
2657                                 if ( $size > 0 ) {
2658                       for (my $j=0;$j<$size;$j++){
2659                                                 my %hashscan;
2660                                                 @hashscan{qw(value cnt)}=$results[ $i - 1 ]->display_term($j);
2661                                                 push @elements, \%hashscan;
2662                                           }
2663                                 }
2664                         }
2665                 }
2666                 return \@elements;
2667    }
2668 }
2669
2670
2671 END { }    # module clean-up code here (global destructor)
2672
2673 1;
2674 __END__
2675
2676 =head1 AUTHOR
2677
2678 Koha Development Team <http://koha-community.org/>
2679
2680 =cut