Bug 14709: Do not access to C4::Context->userenv from the commandline
[koha.git] / tools / export.pl
blob80a15b4cf9c3b6da02580e382379d92fa185d51a
1 #!/usr/bin/perl
4 # This file is part of Koha.
6 # Koha is free software; you can redistribute it and/or modify it
7 # under the terms of the GNU General Public License as published by
8 # the Free Software Foundation; either version 3 of the License, or
9 # (at your option) any later version.
11 # Koha is distributed in the hope that it will be useful, but
12 # WITHOUT ANY WARRANTY; without even the implied warranty of
13 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
14 # GNU General Public License for more details.
16 # You should have received a copy of the GNU General Public License
17 # along with Koha; if not, see <http://www.gnu.org/licenses>.
19 use Modern::Perl;
20 use MARC::File::XML;
21 use List::MoreUtils qw(uniq);
22 use Getopt::Long;
23 use CGI qw ( -utf8 );
24 use C4::Auth;
25 use C4::AuthoritiesMarc; # GetAuthority
26 use C4::Biblio; # GetMarcBiblio
27 use C4::Branch; # GetBranches
28 use C4::Csv;
29 use C4::Koha; # GetItemTypes
30 use C4::Output;
31 use C4::Record;
33 my $query = new CGI;
35 my $clean;
36 my $dont_export_items;
37 my $deleted_barcodes;
38 my $timestamp;
39 my $record_type;
40 my $id_list_file;
41 my $help;
42 my $op = $query->param("op") || '';
43 my $filename = $query->param("filename") || 'koha.mrc';
44 my $dbh = C4::Context->dbh;
45 my $marcflavour = C4::Context->preference("marcflavour");
46 my $output_format = $query->param("format") || $query->param("output_format") || 'iso2709';
48 # Checks if the script is called from commandline
49 my $commandline = not defined $ENV{GATEWAY_INTERFACE};
52 # @biblionumbers is only use for csv export from circulation.pl
53 my @biblionumbers = uniq $query->param("biblionumbers");
55 if ( $commandline ) {
57 # Getting parameters
58 $op = 'export';
59 GetOptions(
60 'format=s' => \$output_format,
61 'date=s' => \$timestamp,
62 'dont_export_items' => \$dont_export_items,
63 'deleted_barcodes' => \$deleted_barcodes,
64 'clean' => \$clean,
65 'filename=s' => \$filename,
66 'record-type=s' => \$record_type,
67 'id_list_file=s' => \$id_list_file,
68 'help|?' => \$help
71 if ($help) {
72 print <<_USAGE_;
73 export.pl [--format=format] [--date=date] [--record-type=TYPE] [--dont_export_items] [--deleted_barcodes] [--clean] [--id_list_file=PATH] --filename=outputfile
76 --format=FORMAT FORMAT is either 'xml' or 'marc' (default)
78 --date=DATE DATE should be entered as the 'dateformat' syspref is
79 set (dd/mm/yyyy for metric, yyyy-mm-dd for iso,
80 mm/dd/yyyy for us) records exported are the ones that
81 have been modified since DATE
83 --record-type=TYPE TYPE is 'bibs' or 'auths'
85 --deleted_barcodes If used, a list of barcodes of items deleted since DATE
86 is produced (or from all deleted items if no date is
87 specified). Used only if TYPE is 'bibs'
89 --clean removes NSE/NSB
91 --id_list_file=PATH PATH is a path to a file containing a list of
92 IDs (biblionumber or authid) with one ID per line.
93 This list works as a filter; it is compatible with
94 other parameters for selecting records
95 _USAGE_
96 exit;
99 # Default parameters values :
100 $timestamp ||= '';
101 $dont_export_items ||= 0;
102 $deleted_barcodes ||= 0;
103 $clean ||= 0;
104 $record_type ||= "bibs";
105 $id_list_file ||= 0;
107 # Redirect stdout
108 open STDOUT, '>', $filename if $filename;
111 else {
113 $op = $query->param("op") || '';
114 $filename = $query->param("filename") || 'koha.mrc';
115 $filename =~ s/(\r|\n)//;
119 # Default value for output_format is 'iso2709'
120 $output_format ||= 'iso2709';
121 # Retrocompatibility for the format parameter
122 $output_format = 'iso2709' if $output_format eq 'marc';
124 my ( $template, $loggedinuser, $cookie, $flags ) = get_template_and_user(
126 template_name => "tools/export.tt",
127 query => $query,
128 type => "intranet",
129 authnotrequired => $commandline,
130 flagsrequired => { tools => 'export_catalog' },
131 debug => 1,
135 my $limit_ind_branch =
136 ( C4::Context->preference('IndependentBranches')
137 && C4::Context->userenv
138 && !C4::Context->IsSuperLibrarian()
139 && C4::Context->userenv->{branch} ) ? 1 : 0;
141 my @branch = $query->param("branch");
142 if ( C4::Context->preference("IndependentBranches")
143 && C4::Context->userenv
144 && !C4::Context->IsSuperLibrarian() )
146 @branch = ( C4::Context->userenv->{'branch'} );
148 # if stripping nonlocal items, use loggedinuser's branch
149 my $localbranch = C4::Context->userenv ? C4::Context->userenv->{'branch'} : undef;
151 my %branchmap = map { $_ => 1 } @branch; # for quick lookups
153 my $backupdir = C4::Context->config('backupdir');
155 if ( $op eq "export" ) {
156 if (
157 $output_format eq "iso2709"
158 or $output_format eq "xml"
159 or (
160 $output_format eq 'csv'
161 and not @biblionumbers
164 my $charset = 'utf-8';
165 my $mimetype = 'application/octet-stream';
167 binmode STDOUT, ':encoding(UTF-8)'
168 if $filename =~ m/\.gz$/
169 or $filename =~ m/\.bz2$/
170 or $output_format ne 'csv';
172 if ( $filename =~ m/\.gz$/ ) {
173 $mimetype = 'application/x-gzip';
174 $charset = '';
175 binmode STDOUT;
177 elsif ( $filename =~ m/\.bz2$/ ) {
178 $mimetype = 'application/x-bzip2';
179 binmode STDOUT;
180 $charset = '';
182 print $query->header(
183 -type => $mimetype,
184 -charset => $charset,
185 -attachment => $filename,
186 ) unless ($commandline);
188 $record_type = $query->param("record_type") unless ($commandline);
189 my $export_remove_fields = $query->param("export_remove_fields");
190 my @biblionumbers = $query->param("biblionumbers");
191 my @itemnumbers = $query->param("itemnumbers");
192 my @sql_params;
193 my $sql_query;
194 my @recordids;
196 my $StartingBiblionumber = $query->param("StartingBiblionumber");
197 my $EndingBiblionumber = $query->param("EndingBiblionumber");
198 my $itemtype = $query->param("itemtype");
199 my $start_callnumber = $query->param("start_callnumber");
200 my $end_callnumber = $query->param("end_callnumber");
201 $timestamp = ($timestamp) ? C4::Dates->new($timestamp) : ''
202 if ($commandline);
203 my $start_accession =
204 ( $query->param("start_accession") )
205 ? C4::Dates->new( $query->param("start_accession") )
206 : '';
207 my $end_accession =
208 ( $query->param("end_accession") )
209 ? C4::Dates->new( $query->param("end_accession") )
210 : '';
211 $dont_export_items = $query->param("dont_export_item")
212 unless ($commandline);
214 my $strip_nonlocal_items = $query->param("strip_nonlocal_items");
216 my $biblioitemstable =
217 ( $commandline and $deleted_barcodes )
218 ? 'deletedbiblioitems'
219 : 'biblioitems';
220 my $itemstable =
221 ( $commandline and $deleted_barcodes )
222 ? 'deleteditems'
223 : 'items';
225 my $starting_authid = $query->param('starting_authid');
226 my $ending_authid = $query->param('ending_authid');
227 my $authtype = $query->param('authtype');
228 my $filefh;
229 if ($commandline) {
230 open $filefh,"<", $id_list_file or die "cannot open $id_list_file: $!" if $id_list_file;
231 } else {
232 $filefh = $query->upload("id_list_file");
234 my %id_filter;
235 if ($filefh) {
236 while (my $number=<$filefh>){
237 $number=~s/[\r\n]*$//;
238 $id_filter{$number}=1 if $number=~/^\d+$/;
242 if ( $record_type eq 'bibs' and not @biblionumbers ) {
243 if ($timestamp) {
245 # Specific query when timestamp is used
246 # Actually it's used only with CLI and so all previous filters
247 # are not used.
248 # If one day timestamp is used via the web interface, this part will
249 # certainly have to be rewrited
250 my ( $query, $params ) = construct_query(
252 recordtype => $record_type,
253 timestamp => $timestamp,
254 biblioitemstable => $biblioitemstable,
257 $sql_query = $query;
258 @sql_params = @$params;
261 else {
262 my ( $query, $params ) = construct_query(
264 recordtype => $record_type,
265 biblioitemstable => $biblioitemstable,
266 itemstable => $itemstable,
267 StartingBiblionumber => $StartingBiblionumber,
268 EndingBiblionumber => $EndingBiblionumber,
269 branch => \@branch,
270 start_callnumber => $start_callnumber,
271 end_callnumber => $end_callnumber,
272 start_accession => $start_accession,
273 end_accession => $end_accession,
274 itemtype => $itemtype,
277 $sql_query = $query;
278 @sql_params = @$params;
281 elsif ( $record_type eq 'auths' ) {
282 my ( $query, $params ) = construct_query(
284 recordtype => $record_type,
285 starting_authid => $starting_authid,
286 ending_authid => $ending_authid,
287 authtype => $authtype,
290 $sql_query = $query;
291 @sql_params = @$params;
294 elsif ( $record_type eq 'db' ) {
295 my $successful_export;
296 if ( $flags->{superlibrarian}
297 && C4::Context->config('backup_db_via_tools') )
299 $successful_export = download_backup(
301 directory => "$backupdir",
302 extension => 'sql',
303 filename => "$filename"
307 unless ($successful_export) {
308 my $remotehost = $query->remote_host();
309 $remotehost =~ s/(\n|\r)//;
310 warn
311 "A suspicious attempt was made to download the db at '$filename' by someone at "
312 . $remotehost . "\n";
314 exit;
316 elsif ( $record_type eq 'conf' ) {
317 my $successful_export;
318 if ( $flags->{superlibrarian}
319 && C4::Context->config('backup_conf_via_tools') )
321 $successful_export = download_backup(
323 directory => "$backupdir",
324 extension => 'tar',
325 filename => "$filename"
329 unless ($successful_export) {
330 my $remotehost = $query->remote_host();
331 $remotehost =~ s/(\n|\r)//;
332 warn
333 "A suspicious attempt was made to download the configuration at '$filename' by someone at "
334 . $remotehost . "\n";
336 exit;
338 elsif (@biblionumbers) {
339 push @recordids, (@biblionumbers);
341 else {
343 # Someone is trying to mess us up
344 exit;
346 unless (@biblionumbers) {
347 my $sth = $dbh->prepare($sql_query);
348 $sth->execute(@sql_params);
349 push @recordids, map {
350 map { $$_[0] } $_
351 } @{ $sth->fetchall_arrayref };
352 @recordids = grep { exists($id_filter{$_}) } @recordids if scalar(%id_filter);
355 my $xml_header_written = 0;
356 for my $recordid ( uniq @recordids ) {
357 if ($deleted_barcodes) {
358 my $q = "
359 SELECT DISTINCT barcode
360 FROM deleteditems
361 WHERE deleteditems.biblionumber = ?
363 my $sth = $dbh->prepare($q);
364 $sth->execute($recordid);
365 while ( my $row = $sth->fetchrow_array ) {
366 print "$row\n";
369 else {
370 my $record;
371 if ( $record_type eq 'bibs' ) {
372 $record = eval { GetMarcBiblio($recordid); };
374 next if $@;
375 next if not defined $record;
376 C4::Biblio::EmbedItemsInMarcBiblio( $record, $recordid,
377 \@itemnumbers )
378 unless $dont_export_items;
379 if ( $strip_nonlocal_items
380 || $limit_ind_branch
381 || $dont_export_items )
383 my ( $homebranchfield, $homebranchsubfield ) =
384 GetMarcFromKohaField( 'items.homebranch', '' );
385 for my $itemfield ( $record->field($homebranchfield) ) {
386 $record->delete_field($itemfield)
387 if ( $dont_export_items
388 || $localbranch ne $itemfield->subfield(
389 $homebranchsubfield) );
393 elsif ( $record_type eq 'auths' ) {
394 $record = C4::AuthoritiesMarc::GetAuthority($recordid);
395 next if not defined $record;
398 if ($export_remove_fields) {
399 for my $f ( split / /, $export_remove_fields ) {
400 if ( $f =~ m/^(\d{3})(.)?$/ ) {
401 my ( $field, $subfield ) = ( $1, $2 );
403 # skip if this record doesn't have this field
404 if ( defined $record->field($field) ) {
405 if ( defined $subfield ) {
406 my @tags = $record->field($field);
407 foreach my $t (@tags) {
408 $t->delete_subfields($subfield);
411 else {
412 $record->delete_fields($record->field($field));
418 RemoveAllNsb($record) if ($clean);
419 if ( $output_format eq "xml" ) {
420 unless ($xml_header_written) {
421 MARC::File::XML->default_record_format(
423 $marcflavour eq 'UNIMARC'
424 && $record_type eq 'auths'
425 ) ? 'UNIMARCAUTH' : $marcflavour
427 print MARC::File::XML::header();
428 print "\n";
429 $xml_header_written = 1;
431 print MARC::File::XML::record($record);
432 print "\n";
434 elsif ( $output_format eq 'iso2709' ) {
435 my $errorcount_on_decode = eval { scalar(MARC::File::USMARC->decode( $record->as_usmarc )->warnings()) };
436 if ($errorcount_on_decode or $@){
437 warn $@ if $@;
438 warn "record (number $recordid) is invalid and therefore not exported because its reopening generates warnings above";
439 next;
441 print $record->as_usmarc();
445 if ($xml_header_written) {
446 print MARC::File::XML::footer();
447 print "\n";
449 if ( $output_format eq 'csv' ) {
450 my $csv_profile_id = $query->param('csv_profile')
451 || GetCsvProfileId( C4::Context->preference('ExportWithCsvProfile') );
452 my $output =
453 marc2csv( \@recordids,
454 $csv_profile_id );
456 print $output;
459 exit;
461 elsif ( $output_format eq "csv" ) {
462 my @biblionumbers = uniq $query->param("biblionumbers");
463 my @itemnumbers = $query->param("itemnumbers");
464 my $csv_profile_id = $query->param('csv_profile') || GetCsvProfileId( C4::Context->preference('ExportWithCsvProfile') );
465 my $output =
466 marc2csv( \@biblionumbers,
467 $csv_profile_id,
468 \@itemnumbers, );
469 print $query->header(
470 -type => 'application/octet-stream',
471 -'Content-Transfer-Encoding' => 'binary',
472 -attachment => "export.csv"
474 print $output;
475 exit;
477 } # if export
479 else {
481 my $itemtypes = GetItemTypes;
482 my @itemtypesloop;
483 foreach my $thisitemtype ( sort keys %$itemtypes ) {
484 my %row = (
485 value => $thisitemtype,
486 description => $itemtypes->{$thisitemtype}->{'description'},
488 push @itemtypesloop, \%row;
490 my $branches = GetBranches($limit_ind_branch);
491 my @branchloop;
492 for my $thisbranch (
493 sort { $branches->{$a}->{branchname} cmp $branches->{$b}->{branchname} }
494 keys %{$branches}
497 push @branchloop,
499 value => $thisbranch,
500 selected => %branchmap ? $branchmap{$thisbranch} : 1,
501 branchname => $branches->{$thisbranch}->{'branchname'},
505 my $authtypes = getauthtypes;
506 my @authtypesloop;
507 foreach my $thisauthtype ( sort keys %$authtypes ) {
508 next unless $thisauthtype;
509 my %row = (
510 value => $thisauthtype,
511 description => $authtypes->{$thisauthtype}->{'authtypetext'},
513 push @authtypesloop, \%row;
516 if ( $flags->{superlibrarian}
517 && C4::Context->config('backup_db_via_tools')
518 && $backupdir
519 && -d $backupdir )
521 $template->{VARS}->{'allow_db_export'} = 1;
522 $template->{VARS}->{'dbfiles'} = getbackupfilelist(
523 { directory => "$backupdir", extension => 'sql' } );
526 if ( $flags->{superlibrarian}
527 && C4::Context->config('backup_conf_via_tools')
528 && $backupdir
529 && -d $backupdir )
531 $template->{VARS}->{'allow_conf_export'} = 1;
532 $template->{VARS}->{'conffiles'} = getbackupfilelist(
533 { directory => "$backupdir", extension => 'tar' } );
536 $template->param(
537 branchloop => \@branchloop,
538 itemtypeloop => \@itemtypesloop,
539 authtypeloop => \@authtypesloop,
540 export_remove_fields => C4::Context->preference("ExportRemoveFields"),
541 csv_profiles => C4::Csv::GetCsvProfiles('marc'),
544 output_html_with_http_headers $query, $cookie, $template->output;
547 sub construct_query {
548 my ($params) = @_;
550 my ( $sql_query, @sql_params );
552 if ( $params->{recordtype} eq "bibs" ) {
553 if ( $params->{timestamp} ) {
554 my $biblioitemstable = $params->{biblioitemstable};
555 $sql_query = " (
556 SELECT biblionumber
557 FROM $biblioitemstable
558 LEFT JOIN items USING(biblionumber)
559 WHERE $biblioitemstable.timestamp >= ?
560 OR items.timestamp >= ?
561 ) UNION (
562 SELECT biblionumber
563 FROM $biblioitemstable
564 LEFT JOIN deleteditems USING(biblionumber)
565 WHERE $biblioitemstable.timestamp >= ?
566 OR deleteditems.timestamp >= ?
567 ) ";
568 my $ts = $timestamp->output('iso');
569 @sql_params = ( $ts, $ts, $ts, $ts );
571 else {
572 my $biblioitemstable = $params->{biblioitemstable};
573 my $itemstable = $params->{itemstable};
574 my $StartingBiblionumber = $params->{StartingBiblionumber};
575 my $EndingBiblionumber = $params->{EndingBiblionumber};
576 my @branch = @{ $params->{branch} };
577 my $start_callnumber = $params->{start_callnumber};
578 my $end_callnumber = $params->{end_callnumber};
579 my $start_accession = $params->{start_accession};
580 my $end_accession = $params->{end_accession};
581 my $itemtype = $params->{itemtype};
582 my $items_filter =
583 @branch
584 || $start_callnumber
585 || $end_callnumber
586 || $start_accession
587 || $end_accession
588 || ( $itemtype && C4::Context->preference('item-level_itypes') );
589 $sql_query = $items_filter
590 ? "SELECT DISTINCT $biblioitemstable.biblionumber
591 FROM $biblioitemstable JOIN $itemstable
592 USING (biblionumber) WHERE 1"
593 : "SELECT $biblioitemstable.biblionumber FROM $biblioitemstable WHERE biblionumber >0 ";
595 if ($StartingBiblionumber) {
596 $sql_query .= " AND $biblioitemstable.biblionumber >= ? ";
597 push @sql_params, $StartingBiblionumber;
600 if ($EndingBiblionumber) {
601 $sql_query .= " AND $biblioitemstable.biblionumber <= ? ";
602 push @sql_params, $EndingBiblionumber;
605 if (@branch) {
606 $sql_query .= " AND homebranch IN (".join(',',map({'?'} @branch)).")";
607 push @sql_params, @branch;
610 if ($start_callnumber) {
611 $sql_query .= " AND itemcallnumber >= ? ";
612 push @sql_params, $start_callnumber;
615 if ($end_callnumber) {
616 $sql_query .= " AND itemcallnumber <= ? ";
617 push @sql_params, $end_callnumber;
619 if ($start_accession) {
620 $sql_query .= " AND dateaccessioned >= ? ";
621 push @sql_params, $start_accession->output('iso');
624 if ($end_accession) {
625 $sql_query .= " AND dateaccessioned <= ? ";
626 push @sql_params, $end_accession->output('iso');
629 if ($itemtype) {
630 $sql_query .=
631 ( C4::Context->preference('item-level_itypes') )
632 ? " AND items.itype = ? "
633 : " AND biblioitems.itemtype = ?";
634 push @sql_params, $itemtype;
638 elsif ( $params->{recordtype} eq "auths" ) {
639 if ( $params->{timestamp} ) {
641 #TODO
643 else {
644 my $starting_authid = $params->{starting_authid};
645 my $ending_authid = $params->{ending_authid};
646 my $authtype = $params->{authtype};
647 $sql_query =
648 "SELECT DISTINCT auth_header.authid FROM auth_header WHERE 1";
650 if ($starting_authid) {
651 $sql_query .= " AND auth_header.authid >= ? ";
652 push @sql_params, $starting_authid;
655 if ($ending_authid) {
656 $sql_query .= " AND auth_header.authid <= ? ";
657 push @sql_params, $ending_authid;
660 if ($authtype) {
661 $sql_query .= " AND auth_header.authtypecode = ? ";
662 push @sql_params, $authtype;
666 return ( $sql_query, \@sql_params );
669 sub getbackupfilelist {
670 my $args = shift;
671 my $directory = $args->{directory};
672 my $extension = $args->{extension};
673 my @files;
675 if ( opendir( my $dir, $directory ) ) {
676 while ( my $file = readdir($dir) ) {
677 next unless ( $file =~ m/\.$extension(\.(gz|bz2|xz))?/ );
678 push @files, $file
679 if ( -f "$directory/$file" && -r "$directory/$file" );
681 closedir($dir);
683 return \@files;
686 sub download_backup {
687 my $args = shift;
688 my $directory = $args->{directory};
689 my $extension = $args->{extension};
690 my $filename = $args->{filename};
692 return unless ( $directory && -d $directory );
693 return unless ( $filename =~ m/\.$extension(\.(gz|bz2|xz))?$/ );
694 return if ( $filename =~ m#/# );
695 $filename = "$directory/$filename";
696 return unless ( -f $filename && -r $filename );
697 return unless ( open( my $dump, '<', $filename ) );
698 binmode $dump;
700 while ( read( $dump, my $data, 64 * 1024 ) ) {
701 print $data;
703 close($dump);
704 return 1;