misc/bulkauthimport.pl

   1 #!/usr/bin/perl
   2 # small script that import an iso2709 file into koha 2.0
   3
   4 use strict;
   5 BEGIN {
   6     # find Koha's Perl modules
   7     # test carefully before changing this
   8     use FindBin;
   9     eval { require "$FindBin::Bin/kohalib.pl" };
  10 }
  11
  12 # Koha modules used
  13 use MARC::File::USMARC;
  14 use MARC::Record;
  15 use MARC::Batch;
  16 use C4::Context;
  17 use C4::AuthoritiesMarc;
  18 use Time::HiRes qw(gettimeofday);
  19
  20 use Getopt::Long;
  21 my ( $input_marc_file, $number) = ('',0);
  22 my ($version, $delete, $test_parameter,$char_encoding, $verbose);
  23 GetOptions(
  24     'file:s'    => \$input_marc_file,
  25     'n' => \$number,
  26     'h' => \$version,
  27     'd' => \$delete,
  28     't' => \$test_parameter,
  29     'c:s' => \$char_encoding,
  30     'v:s' => \$verbose,
  31 );
  32
  33 if ($version || ($input_marc_file eq '')) {
  34         print <<EOF
  35 small script to import an iso2709 file into Koha.
  36 parameters :
  37 \th : this version/help screen
  38 \tfile /path/to/file/to/dump : the file to dump
  39 \tv : verbose mode. 1 means "some infos", 2 means "MARC dumping"
  40 \tn : the number of the record to import. If missing, all the file is imported
  41 \tt : test mode : parses the file, saying what he would do, but doing nothing.
  42 \tc : the char encoding. At the moment, only MARC21 and UNIMARC supported. MARC21 by default.
  43 \td : delete EVERYTHING related to biblio in koha-DB before import  :tables :
  44 \t\tbiblio, \t\tbiblioitems, \t\tsubjects,\titems
  45 \tmarc_biblio,
  46 \t\tmarc_subfield_table, \tmarc_word, \t\tmarc_blob_subfield
  47 IMPORTANT : don't use this script before you've entered and checked twice (or more) your  MARC parameters tables.
  48 If you fail this, the import won't work correctly and you will get invalid datas.
  49
  50 SAMPLE : ./bulkmarcimport.pl -file /home/paul/koha.dev/local/npl -n 1
  51 EOF
  52 ;#'
  53 die;
  54 }
  55
  56 my $dbh = C4::Context->dbh;
  57
  58 if ($delete) {
  59         print "deleting authorities\n";
  60         $dbh->do("delete from auth_header");
  61 }
  62 if ($test_parameter) {
  63         print "TESTING MODE ONLY\n    DOING NOTHING\n===============\n";
  64 }
  65
  66 $char_encoding = 'MARC21' unless ($char_encoding);
  67 print "CHAR : $char_encoding\n" if $verbose;
  68 my $starttime = gettimeofday;
  69 my $batch = MARC::Batch->new( 'USMARC', $input_marc_file );
  70 $batch->warnings_off();
  71 $batch->strict_off();
  72 my $i=0;
  73 while ( my $record = $batch->next() ) {
  74         $i++;
  75         #now, parse the record, extract the item fields, and store them in somewhere else.
  76
  77     ## create an empty record object to populate
  78     my $newRecord = MARC::Record->new();
  79         $newRecord->leader($record->leader);
  80     # go through each field in the existing record
  81     foreach my $oldField ( $record->fields() ) {
  82                 # just reproduce tags < 010 in our new record
  83                 if ( $oldField->tag() < 10 ) {
  84                         $newRecord->append_fields( $oldField );
  85                         next();
  86                 }
  87                 # store our new subfield data in this list
  88                 my @newSubfields = ();
  89
  90                 # go through each subfield code/data pair
  91                 foreach my $pair ( $oldField->subfields() ) {
  92                         $pair->[1] =~ s/\<//g;
  93                         $pair->[1] =~ s/\>//g;
  94                         push( @newSubfields, $pair->[0], char_decode($pair->[1],$char_encoding) );
  95                 }
  96
  97                 # add the new field to our new record
  98                 my $newField = MARC::Field->new(
  99                         $oldField->tag(),
 100                         $oldField->indicator(1),
 101                         $oldField->indicator(2),
 102                         @newSubfields
 103                 );
 104                 $newRecord->append_fields( $newField );
 105     }
 106         warn "$i ==>".$newRecord->as_formatted() if $verbose eq 2;
 107         my $authtypecode=substr($newRecord->leader(),9,1);
 108         $authtypecode="NP" if ($authtypecode eq 'a'); # personnes
 109         $authtypecode="CO" if ($authtypecode eq 'b'); # collectivit�
 110         $authtypecode="NG" if ($authtypecode eq 'c'); # g�graphique
 111         $authtypecode="NM" if ($authtypecode eq 'd'); # marque
 112         $authtypecode="NF" if ($authtypecode eq 'e'); # famille
 113         $authtypecode="TI" if ($authtypecode eq 'f'); # Titre uniforme
 114         $authtypecode="TI" if ($authtypecode eq 'h'); # auteur/titre
 115         $authtypecode="MM" if ($authtypecode eq 'j'); # mot mati�e
 116         warn "XX => $authtypecode";
 117         # now, create biblio and items with NEWnewXX call.
 118         unless ($test_parameter) {
 119                 my ($authid) = AddAuthority($newRecord,0,$authtypecode);
 120                 warn "ADDED authority NB $authid in DB\n" if $verbose;
 121         }
 122 }
 123 # $dbh->do("unlock tables");
 124 my $timeneeded = gettimeofday - $starttime;
 125 print "$i MARC record done in $timeneeded seconds";