3 # Copyright 2000-2002 Katipo Communications
4 # Parts Copyright 2013 Prosentient Systems
6 # This file is part of Koha.
8 # Koha is free software; you can redistribute it and/or modify it
9 # under the terms of the GNU General Public License as published by
10 # the Free Software Foundation; either version 3 of the License, or
11 # (at your option) any later version.
13 # Koha is distributed in the hope that it will be useful, but
14 # WITHOUT ANY WARRANTY; without even the implied warranty of
15 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
16 # GNU General Public License for more details.
18 # You should have received a copy of the GNU General Public License
19 # along with Koha; if not, see <http://www.gnu.org/licenses>.
27 use MARC::File::USMARC;
30 use C4::AuthoritiesMarc; #GuessAuthTypeCode, FindDuplicateAuthority
33 use Koha::XSLT_Handler;
35 use vars qw(@ISA @EXPORT @EXPORT_OK %EXPORT_TAGS);
40 @EXPORT = qw(&BreedingSearch &Z3950Search &Z3950SearchAuth);
45 C4::Breeding : module to add biblios to import_records via
46 the breeding/reservoir API.
50 Z3950Search($pars, $template);
51 ($count, @results) = &BreedingSearch($title,$isbn,$random);
55 This module contains routines related to Koha's Z39.50 search into
56 cataloguing reservoir features.
60 ($count, @results) = &BreedingSearch($title,$isbn,$random);
61 C<$title> contains the title,
62 C<$isbn> contains isbn or issn,
63 C<$random> contains the random seed from a z3950 search.
65 C<$count> is the number of items in C<@results>. C<@results> is an
66 array of references-to-hash; the keys are the items from the C<import_records> and
67 C<import_biblios> tables of the Koha database.
72 my ($search,$isbn,$z3950random) = @_;
73 my $dbh = C4::Context->dbh;
79 # normalise ISBN like at import
80 $isbn = C4::Koha::GetNormalizedISBN($isbn);
82 $query = "SELECT import_record_id, file_name, isbn, title, author
84 JOIN import_records USING (import_record_id)
85 JOIN import_batches USING (import_batch_id)
88 $query .= "z3950random = ?";
92 if (defined($search) && length($search)>0) {
93 $search =~ s/(\s+)/\%/g;
94 $query .= "title like ? OR author like ?";
95 push(@bind,"%$search%", "%$search%");
97 if ($#bind!=-1 && defined($isbn) && length($isbn)>0) {
100 if (defined($isbn) && length($isbn)>0) {
101 $query .= "isbn like ?";
102 push(@bind,"$isbn%");
105 $sth = $dbh->prepare($query);
106 $sth->execute(@bind);
107 while (my $data = $sth->fetchrow_hashref) {
108 $results[$count] = $data;
109 # FIXME - hack to reflect difference in name
110 # of columns in old marc_breeding and import_records
111 # There needs to be more separation between column names and
112 # field names used in the templates </soapbox>
113 $data->{'file'} = $data->{'file_name'};
114 $data->{'id'} = $data->{'import_record_id'};
119 return($count, @results);
120 } # sub breedingsearch
125 Z3950Search($pars, $template);
127 Parameters for Z3950 search are all passed via the $pars hash. It may contain isbn, title, author, dewey, subject, lccall, controlnumber, stdid, srchany.
128 Also it should contain an arrayref id that points to a list of id's of the z3950 targets to be queried (see z3950servers table).
129 This code is used in acqui/z3950_search and cataloging/z3950_search.
130 The second parameter $template is a Template object. The routine uses this parameter to store the found values into the template.
135 my ($pars, $template)= @_;
137 my @id= @{$pars->{id}};
138 my $page= $pars->{page};
139 my $biblionumber= $pars->{biblionumber};
144 my @breeding_loop = ();
151 my ( $zquery, $squery ) = _bib_build_query( $pars );
153 my $schema = Koha::Database->new()->schema();
154 my $rs = $schema->resultset('Z3950server')->search(
156 { result_class => 'DBIx::Class::ResultClass::HashRefInflator' },
158 my @servers = $rs->all;
159 foreach my $server ( @servers ) {
160 $oConnection[$s] = _create_connection( $server );
162 $server->{servertype} eq 'zed'?
163 $oConnection[$s]->search_pqf( $zquery ):
164 $oConnection[$s]->search(new ZOOM::Query::CQL(
165 _translate_query( $server, $squery )));
168 my $xslh = Koha::XSLT_Handler->new;
171 while ( $nremaining-- ) {
174 while ( ( $k = ZOOM::event( \@oConnection ) ) != 0 ) {
175 $event = $oConnection[ $k - 1 ]->last_event();
176 last if $event == ZOOM::Event::ZEND;
181 my ($error)= $oConnection[$k]->error_x(); #ignores errmsg, addinfo, diagset
183 if ($error =~ m/^(10000|10007)$/ ) {
184 push(@errconn, { server => $servers[$k]->{host}, error => $error } );
188 my $numresults = $oResult[$k]->size();
191 if ( $numresults > 0 and $numresults >= (($page-1)*20)) {
192 $show_next = 1 if $numresults >= ($page*20);
193 $total_pages = int($numresults/20)+1 if $total_pages < ($numresults/20);
194 for ($i = ($page-1)*20; $i < (($numresults < ($page*20)) ? $numresults : ($page*20)); $i++) {
195 if ( $oResult[$k]->record($i) ) {
197 ( $res, $error ) = _handle_one_result( $oResult[$k]->record($i), $servers[$k], ++$imported, $biblionumber, $xslh ); #ignores error in sequence numbering
198 push @breeding_loop, $res if $res;
199 push @errconn, { server => $servers[$k]->{servername}, error => $error, seq => $i+1 } if $error;
202 push @errconn, { 'server' => $servers[$k]->{servername}, error => ( ( $oConnection[$k]->error_x() )[0] ), seq => $i+1 };
210 numberpending => $nremaining,
211 current_page => $page,
212 total_pages => $total_pages,
213 show_nextbutton => $show_next?1:0,
214 show_prevbutton => $page!=1,
218 #close result sets and connections
220 $oResult[$_]->destroy();
221 $oConnection[$_]->destroy();
225 breeding_loop => \@breeding_loop,
226 servers => \@servers,
231 sub _auth_build_query {
235 nameany => '@attr 1=1002 "#term" ',
236 authorany => '@attr 1=1003 "#term" ',
237 authorcorp => '@attr 1=2 "#term" ',
238 authorpersonal => '@attr 1=1 "#term" ',
239 authormeetingcon => '@attr 1=3 "#term" ',
240 subject => '@attr 1=21 "#term" ',
241 subjectsubdiv => '@attr 1=47 "#term" ',
242 title => '@attr 1=4 "#term" ',
243 uniformtitle => '@attr 1=6 "#term" ',
244 srchany => '@attr 1=1016 "#term" ',
245 controlnumber => '@attr 1=12 "#term" ',
248 return _build_query( $pars, $qry_build );
251 sub _bib_build_query {
256 isbn => '@attr 1=7 @attr 5=1 "#term" ',
257 issn => '@attr 1=8 @attr 5=1 "#term" ',
258 title => '@attr 1=4 "#term" ',
259 author => '@attr 1=1003 "#term" ',
260 dewey => '@attr 1=16 "#term" ',
261 subject => '@attr 1=21 "#term" ',
262 lccall => '@attr 1=16 @attr 2=3 @attr 3=1 @attr 4=1 @attr 5=1 '.
263 '@attr 6=1 "#term" ',
264 controlnumber => '@attr 1=12 "#term" ',
265 srchany => '@attr 1=1016 "#term" ',
266 stdid => '@attr 1=1007 "#term" ',
269 return _build_query( $pars, $qry_build );
274 my ( $pars, $qry_build ) = @_;
279 foreach my $k ( sort keys %$pars ) {
280 #note that the sort keys forces an identical result under Perl 5.18
281 #one of the unit tests is based on that assumption
282 if( ( my $val=$pars->{$k} ) && $qry_build->{$k} ) {
283 $qry_build->{$k} =~ s/#term/$val/g;
284 $zquery .= $qry_build->{$k};
285 $squery .= "[$k]=\"$val\" and ";
289 $zquery = "\@and " . $zquery for 2..$nterms;
290 $squery =~ s/ and $//;
291 return ( $zquery, $squery );
294 sub _handle_one_result {
295 my ( $zoomrec, $servhref, $seq, $bib, $xslh )= @_;
297 my $raw= $zoomrec->raw();
299 if( $servhref->{servertype} eq 'sru' ) {
300 $marcrecord= MARC::Record->new_from_xml( $raw, 'UTF-8',
301 $servhref->{syntax} );
303 ($marcrecord) = MarcToUTF8Record($raw, C4::Context->preference('marcflavour'), $servhref->{encoding} // "iso-5426" ); #ignores charset return values
305 SetUTF8Flag($marcrecord);
307 ( $marcrecord, $error ) = _do_xslt_proc($marcrecord, $servhref, $xslh);
309 my $batch_id = GetZ3950BatchId($servhref->{servername});
310 my $breedingid = AddBiblioToBatch($batch_id, $seq, $marcrecord, 'UTF-8', 0, 0);
311 #FIXME passing 0 for z3950random
312 #Will eliminate this unused field in a followup report
313 #Last zero indicates: no update for batch record counts
316 #call to TransformMarcToKoha replaced by next call
317 #we only need six fields from the marc record
321 biblionumber => $bib,
322 server => $servhref->{servername},
323 breedingid => $breedingid,
324 }, $marcrecord) if $breedingid;
325 return ( $row, $error );
329 my ( $marc, $server, $xslh ) = @_;
330 return $marc if !$server->{add_xslt};
332 my $htdocs = C4::Context->config('intrahtdocs');
333 my $theme = C4::Context->preference("template"); #staff
334 my $lang = C4::Languages::getlanguage() || 'en';
336 my @files= split ',', $server->{add_xslt};
337 my $xml = $marc->as_xml;
338 foreach my $f ( @files ) {
339 $f =~ s/^\s+//; $f =~ s/\s+$//; next if !$f;
340 $f = C4::XSLT::_get_best_default_xslt_filename(
341 $htdocs, $theme, $lang, $f ) unless $f =~ /^\//;
342 $xml = $xslh->transform( $xml, $f );
343 last if $xslh->err; #skip other files
346 return MARC::Record->new_from_xml($xml, 'UTF-8');
348 return ( $marc, $xslh->err ); #original record in case of errors
353 my ($row, $record)=@_;
355 title => 'biblio.title',
356 author => 'biblio.author',
357 isbn =>'biblioitems.isbn',
358 lccn =>'biblioitems.lccn', #LC control number (not call number)
359 edition =>'biblioitems.editionstatement'
361 $fetch{date} = C4::Context->preference('marcflavour') eq "MARC21" ? 'biblio.copyrightdate' : 'biblioitems.publicationyear';
363 foreach my $k (keys %fetch) {
364 $row->{$k} = C4::Biblio::TransformMarcToKohaOneField( $fetch{$k}, $record );
366 $row->{date}//= $row->{date2};
367 $row->{isbn}=_isbn_replace($row->{isbn});
369 $row = _add_custom_field_rowdata($row, $record);
374 sub _add_custom_field_rowdata
376 my ( $row, $record ) = @_;
377 my $pref_newtags = C4::Context->preference('AdditionalFieldsInZ3950ResultSearch');
379 $pref_newtags =~ s/^\s+|\s+$//g;
380 $pref_newtags =~ s/\h+/ /g;
384 foreach my $field (split /\,/, $pref_newtags) {
385 $field =~ s/^\s+|\s+$//g ; # trim whitespace
386 my ($tag, $subtags) = split(/\$/, $field);
388 if ( $record->field($tag) ) {
391 for my $marcfield ($record->field($tag)) {
394 for my $code (split //, $subtags) {
395 if ( $marcfield->subfield($code) ) {
396 $str .= $marcfield->subfield($code) . ' ';
399 if ( not $str eq '') {
402 } elsif ( $tag <= 10 ) {
403 push @content, $marcfield->data();
405 push @content, $marcfield->as_string();
410 $row->{$field} = \@content;
411 push( @addnumberfields, $field );
416 $row->{'addnumberfields'} = \@addnumberfields;
423 return unless defined $isbn;
424 $isbn =~ s/ |-|\.//g;
425 $isbn =~ s/\|/ \| /g;
430 sub _create_connection {
432 my $option1= new ZOOM::Options();
433 $option1->option( 'async' => 1 );
434 $option1->option( 'elementSetName', 'F' );
435 $option1->option( 'preferredRecordSyntax', $server->{syntax} );
436 $option1->option( 'timeout', $server->{timeout} ) if $server->{timeout};
438 if( $server->{servertype} eq 'sru' ) {
439 foreach( split ',', $server->{sru_options}//'' ) {
440 #first remove surrounding spaces at comma and equals-sign
442 my @temp= split '=', $_, 2;
443 @temp= map { my $c=$_; $c=~s/^\s+|\s+$//g; $c; } @temp;
444 $option1->option( $temp[0] => $temp[1] ) if @temp;
446 } elsif( $server->{servertype} eq 'zed' ) {
447 $option1->option( 'databaseName', $server->{db} );
448 $option1->option( 'user', $server->{userid} ) if $server->{userid};
449 $option1->option( 'password', $server->{password} ) if $server->{password};
451 my $obj= ZOOM::Connection->create($option1);
452 if( $server->{servertype} eq 'sru' ) {
453 my $host= $server->{host};
454 if( $host !~ /^https?:\/\// ) {
455 #Normally, host will not be prefixed by protocol.
456 #In that case we can (safely) assume http.
457 #In case someone prefixed with https, give it a try..
458 $host = 'http://' . $host;
460 $obj->connect( $host.':'.$server->{port}.'/'.$server->{db} );
462 $obj->connect( $server->{host}, $server->{port} );
467 sub _translate_query { #SRU query adjusted per server cf. srufields column
468 my ($server, $query) = @_;
470 #sru_fields is in format title=field,isbn=field,...
471 #if a field doesn't exist, try anywhere or remove [field]=
472 my @parts= split(',', $server->{sru_fields} );
473 my %trans= map { if( /=/ ) { ( $`,$' ) } else { () } } @parts;
474 my $any= $trans{srchany}?$trans{srchany}.'=':'';
477 foreach my $key (keys %trans) {
482 $q=~s/\[$key\]=/$any/g;
485 $q=~s/\[\w+\]=/$any/g; # remove remaining fields (not found in field list)
489 =head2 ImportBreedingAuth
491 ImportBreedingAuth($marcrecords,$overwrite_auth,$filename,$encoding,$z3950random);
493 ImportBreedingAuth imports MARC records in the reservoir (import_records table).
494 ImportBreedingAuth is based on the ImportBreeding subroutine.
498 sub ImportBreedingAuth {
499 my ($marcrecord,$overwrite_auth,$filename,$encoding,$z3950random) = @_;
500 my $dbh = C4::Context->dbh;
502 my $batch_id = GetZ3950BatchId($filename);
503 my $searchbreeding = $dbh->prepare("select import_record_id from import_auths where control_number=? and authorized_heading=?");
505 my $marcflavour = C4::Context->preference('marcflavour');
506 my $marc_type = $marcflavour eq 'UNIMARC' ? 'UNIMARCAUTH' : $marcflavour;
508 # fields used for import results
511 my $alreadyinfarm = 0;
512 my $notmarcrecord = 0;
515 # Normalize the record so it doesn't have separated diacritics
516 SetUTF8Flag($marcrecord);
518 if (scalar($marcrecord->fields()) == 0) {
522 $heading = C4::AuthoritiesMarc::GetAuthorizedHeading({ record => $marcrecord });
524 my $heading_authtype_code;
525 $heading_authtype_code = GuessAuthTypeCode($marcrecord);
528 $controlnumber = $marcrecord->field('001')->data;
530 #Check if the authority record already exists in the database...
531 my ($duplicateauthid,$duplicateauthvalue);
532 if ($marcrecord && $heading_authtype_code) {
533 ($duplicateauthid,$duplicateauthvalue) = FindDuplicateAuthority( $marcrecord, $heading_authtype_code);
536 if ($duplicateauthid && $overwrite_auth ne 2) {
537 #If the authority record exists and $overwrite_auth doesn't equal 2, then mark it as already in the DB
540 if ($controlnumber && $heading) {
541 $searchbreeding->execute($controlnumber,$heading);
542 ($breedingid) = $searchbreeding->fetchrow;
544 if ($breedingid && $overwrite_auth eq '0') {
547 if ($breedingid && $overwrite_auth eq '1') {
548 ModAuthorityInBatch($breedingid, $marcrecord);
550 my $import_id = AddAuthToBatch($batch_id, $imported, $marcrecord, $encoding, $z3950random);
551 $breedingid = $import_id;
557 return ($notmarcrecord,$alreadyindb,$alreadyinfarm,$imported,$breedingid);
560 =head2 Z3950SearchAuth
562 Z3950SearchAuth($pars, $template);
564 Parameters for Z3950 search are all passed via the $pars hash. It may contain nameany, namepersonal, namecorp, namemeetingcon,
565 title, uniform title, subject, subjectsubdiv, srchany.
566 Also it should contain an arrayref id that points to a list of IDs of the z3950 targets to be queried (see z3950servers table).
567 This code is used in cataloging/z3950_auth_search.
568 The second parameter $template is a Template object. The routine uses this parameter to store the found values into the template.
572 sub Z3950SearchAuth {
573 my ($pars, $template)= @_;
575 my $dbh = C4::Context->dbh;
576 my @id= @{$pars->{id}};
577 my $random= $pars->{random};
578 my $page= $pars->{page};
594 my @breeding_loop = ();
604 my $marcflavour = C4::Context->preference('marcflavour');
605 my $marc_type = $marcflavour eq 'UNIMARC' ? 'UNIMARCAUTH' : $marcflavour;
606 my $authid= $pars->{authid};
607 my ( $zquery, $squery ) = _auth_build_query( $pars );
608 foreach my $servid (@id) {
609 my $sth = $dbh->prepare("select * from z3950servers where id=?");
610 $sth->execute($servid);
611 while ( $server = $sth->fetchrow_hashref ) {
612 $oConnection[$s] = _create_connection( $server );
615 $server->{servertype} eq 'zed'?
616 $oConnection[$s]->search_pqf( $zquery ):
617 $oConnection[$s]->search(new ZOOM::Query::CQL(
618 _translate_query( $server, $squery )));
619 $encoding[$s] = ($server->{encoding}?$server->{encoding}:"iso-5426");
620 $servers[$s] = $server;
626 while ( $nremaining-- ) {
629 while ( ( $k = ZOOM::event( \@oConnection ) ) != 0 ) {
630 $event = $oConnection[ $k - 1 ]->last_event();
631 last if $event == ZOOM::Event::ZEND;
636 my ($error, $errmsg, $addinfo, $diagset)= $oConnection[$k]->error_x();
638 if ($error =~ m/^(10000|10007)$/ ) {
639 push(@errconn, {'server' => $serverhost[$k]});
643 my $numresults = $oResult[$k]->size();
646 if ( $numresults > 0 and $numresults >= (($page-1)*20)) {
647 $show_next = 1 if $numresults >= ($page*20);
648 $total_pages = int($numresults/20)+1 if $total_pages < ($numresults/20);
649 for ($i = ($page-1)*20; $i < (($numresults < ($page*20)) ? $numresults : ($page*20)); $i++) {
650 my $rec = $oResult[$k]->record($i);
654 $marcdata = $rec->raw();
656 my ($charset_result, $charset_errors);
657 if( $servers[$k]->{servertype} eq 'sru' ) {
658 $marcrecord = MARC::Record->new_from_xml( $marcdata, 'UTF-8', $servers[$k]->{syntax} );
660 ( $marcrecord, $charset_result, $charset_errors ) = MarcToUTF8Record( $marcdata, $marc_type, $encoding[$k] );
663 my $heading_authtype_code;
664 $heading_authtype_code = GuessAuthTypeCode($marcrecord);
665 $heading = C4::AuthoritiesMarc::GetAuthorizedHeading({ record => $marcrecord });
667 my ($notmarcrecord, $alreadyindb, $alreadyinfarm, $imported, $breedingid)= ImportBreedingAuth( $marcrecord, 2, $serverhost[$k], $encoding[$k], $random);
669 $row_data{server} = $servers[$k]->{'servername'};
670 $row_data{breedingid} = $breedingid;
671 $row_data{heading} = $heading;
672 $row_data{authid} = $authid;
673 $row_data{heading_code} = $heading_authtype_code;
674 push( @breeding_loop, \%row_data );
677 push(@breeding_loop,{'server'=>$servers[$k]->{'servername'},'title'=>join(': ',$oConnection[$k]->error_x()),'breedingid'=>-1,'authid'=>-1});
685 numberpending => $nremaining,
686 current_page => $page,
687 total_pages => $total_pages,
688 show_nextbutton => $show_next?1:0,
689 show_prevbutton => $page!=1,
693 #close result sets and connections
695 $oResult[$_]->destroy();
696 $oConnection[$_]->destroy();
700 foreach my $id (@id) {
701 push @servers, {id => $id};
704 breeding_loop => \@breeding_loop,
705 servers => \@servers,