misc/bulkauthimport.pl

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