/[webpac2]/trunk/lib/WebPAC/Normalize.pm
This is repository of my old source code which isn't updated any more. Go to git.rot13.org for current projects!
ViewVC logotype

Diff of /trunk/lib/WebPAC/Normalize.pm

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 786 by dpavlin, Sun Dec 10 12:45:11 2006 UTC revision 973 by dpavlin, Fri Nov 2 14:59:12 2007 UTC
# Line 7  use Exporter 'import'; Line 7  use Exporter 'import';
7          _debug          _debug
8          _pack_subfields_hash          _pack_subfields_hash
9    
10          tag search display          search_display search display sorted
11    
12          marc marc_indicators marc_repeatable_subfield          marc marc_indicators marc_repeatable_subfield
13          marc_compose marc_leader          marc_compose marc_leader marc_fixed
14          marc_duplicate marc_remove          marc_duplicate marc_remove marc_count
15          marc_original_order          marc_original_order
16    
17          rec1 rec2 rec          rec1 rec2 rec
# Line 21  use Exporter 'import'; Line 22  use Exporter 'import';
22          split_rec_on          split_rec_on
23    
24          get set          get set
25            count
26  /;  /;
27    
28  use warnings;  use warnings;
# Line 39  my $debug = 0; Line 41  my $debug = 0;
41    
42  WebPAC::Normalize - describe normalisaton rules using sets  WebPAC::Normalize - describe normalisaton rules using sets
43    
 =head1 VERSION  
   
 Version 0.25  
   
44  =cut  =cut
45    
46  our $VERSION = '0.25';  our $VERSION = '0.31';
47    
48  =head1 SYNOPSIS  =head1 SYNOPSIS
49    
# Line 58  means that you check it's validity befor Line 56  means that you check it's validity befor
56  C<perl -c normalize.pl>.  C<perl -c normalize.pl>.
57    
58  Normalisation can generate multiple output normalized data. For now, supported output  Normalisation can generate multiple output normalized data. For now, supported output
59  types (on the left side of definition) are: C<tag>, C<display>, C<search> and  types (on the left side of definition) are: C<search_display>, C<display>, C<search> and
60  C<marc>.  C<marc>.
61    
62  =head1 FUNCTIONS  =head1 FUNCTIONS
# Line 110  sub data_structure { Line 108  sub data_structure {
108          _clean_ds( %{ $arg } );          _clean_ds( %{ $arg } );
109          $load_row_coderef = $arg->{load_row_coderef};          $load_row_coderef = $arg->{load_row_coderef};
110    
111            # FIXME load this conditionally
112            use WebPAC::Normalize::ISBN;
113    
114          eval "$arg->{rules}";          eval "$arg->{rules}";
115          die "error evaling $arg->{rules}: $@\n" if ($@);          die "error evaling $arg->{rules}: $@\n" if ($@);
116    
# Line 166  Return hash formatted as data structure Line 167  Return hash formatted as data structure
167    
168  =cut  =cut
169    
170  my ($out, $marc_record, $marc_encoding, $marc_repeatable_subfield, $marc_indicators, $leader);  my ($out, $marc_record, $marc_encoding, $marc_repeatable_subfield, $marc_indicators, $marc_leader);
171  my ($marc_record_offset, $marc_fetch_offset) = (0, 0);  my ($marc_record_offset, $marc_fetch_offset) = (0, 0);
172    
173  sub _get_ds {  sub _get_ds {
# Line 183  Clean data structure hash for next recor Line 184  Clean data structure hash for next recor
184    
185  sub _clean_ds {  sub _clean_ds {
186          my $a = {@_};          my $a = {@_};
187          ($out,$marc_record, $marc_encoding, $marc_repeatable_subfield, $marc_indicators, $leader) = ();          ($out,$marc_record, $marc_encoding, $marc_repeatable_subfield, $marc_indicators, $marc_leader) = ();
188          ($marc_record_offset, $marc_fetch_offset) = (0,0);          ($marc_record_offset, $marc_fetch_offset) = (0,0);
189          $marc_encoding = $a->{marc_encoding};          $marc_encoding = $a->{marc_encoding};
190  }  }
# Line 288  will return 42th copy record (if it exis Line 289  will return 42th copy record (if it exis
289    
290  =cut  =cut
291    
292    my $fetch_pos;
293    
294  sub _get_marc_fields {  sub _get_marc_fields {
295    
296          my $arg = {@_};          my $arg = {@_};
297          warn "### _get_marc_fields arg: ", dump($arg), $/ if ($debug > 2);          warn "### _get_marc_fields arg: ", dump($arg), $/ if ($debug > 2);
298          my $offset = $marc_fetch_offset;          $fetch_pos = $marc_fetch_offset;
299          if ($arg->{offset}) {          if ($arg->{offset}) {
300                  $offset = $arg->{offset};                  $fetch_pos = $arg->{offset};
301          } elsif($arg->{fetch_next}) {          } elsif($arg->{fetch_next}) {
302                  $marc_fetch_offset++;                  $marc_fetch_offset++;
303          }          }
# Line 303  sub _get_marc_fields { Line 306  sub _get_marc_fields {
306    
307          warn "### full marc_record = ", dump( @{ $marc_record }), $/ if ($debug > 2);          warn "### full marc_record = ", dump( @{ $marc_record }), $/ if ($debug > 2);
308    
309          my $marc_rec = $marc_record->[ $offset ];          my $marc_rec = $marc_record->[ $fetch_pos ];
310    
311          warn "## _get_marc_fields (at offset: $offset) -- marc_record = ", dump( @$marc_rec ), $/ if ($debug > 1);          warn "## _get_marc_fields (at offset: $fetch_pos) -- marc_record = ", dump( @$marc_rec ), $/ if ($debug > 1);
312    
313          return if (! $marc_rec || ref($marc_rec) ne 'ARRAY' || $#{ $marc_rec } < 0);          return if (! $marc_rec || ref($marc_rec) ne 'ARRAY' || $#{ $marc_rec } < 0);
314    
# Line 326  sub _get_marc_fields { Line 329  sub _get_marc_fields {
329    
330          if ($debug) {          if ($debug) {
331                  warn "## marc_repeatable_subfield = ", dump( $marc_repeatable_subfield ), $/ if ( $marc_repeatable_subfield );                  warn "## marc_repeatable_subfield = ", dump( $marc_repeatable_subfield ), $/ if ( $marc_repeatable_subfield );
332                  warn "## marc_record[$offset] = ", dump( $marc_rec ), $/;                  warn "## marc_record[$fetch_pos] = ", dump( $marc_rec ), $/;
333                  warn "## sorted_marc_record = ", dump( \@sorted_marc_record ), $/;                  warn "## sorted_marc_record = ", dump( \@sorted_marc_record ), $/;
334                  warn "## subfield count = ", dump( $u ), $/;                  warn "## subfield count = ", dump( $u ), $/;
335          }          }
# Line 407  sub _get_marc_fields { Line 410  sub _get_marc_fields {
410          return \@m;          return \@m;
411  }  }
412    
413    =head2 _get_marc_leader
414    
415    Return leader from currently fetched record by L</_get_marc_fields>
416    
417      print WebPAC::Normalize::_get_marc_leader();
418    
419    =cut
420    
421    sub _get_marc_leader {
422            die "no fetch_pos, did you called _get_marc_fields first?" unless ( defined( $fetch_pos ) );
423            return $marc_leader->[ $fetch_pos ];
424    }
425    
426  =head2 _debug  =head2 _debug
427    
428  Change level of debug warnings  Change level of debug warnings
# Line 426  sub _debug { Line 442  sub _debug {
442    
443  Those functions generally have to first in your normalization file.  Those functions generally have to first in your normalization file.
444    
445  =head2 tag  =head2 search_display
446    
447  Define new tag for I<search> and I<display>.  Define output for L<search> and L<display> at the same time
448    
449    tag('Title', rec('200','a') );    search_display('Title', rec('200','a') );
450    
451    
452  =cut  =cut
453    
454  sub tag {  sub search_display {
455          my $name = shift or die "tag needs name as first argument";          my $name = shift or die "search_display needs name as first argument";
456          my @o = grep { defined($_) && $_ ne '' } @_;          my @o = grep { defined($_) && $_ ne '' } @_;
457          return unless (@o);          return unless (@o);
         $out->{$name}->{tag} = $name;  
458          $out->{$name}->{search} = \@o;          $out->{$name}->{search} = \@o;
459          $out->{$name}->{display} = \@o;          $out->{$name}->{display} = \@o;
460  }  }
461    
462    =head2 tag
463    
464    Old name for L<search_display>, but supported
465    
466    =cut
467    
468    sub tag {
469            search_display( @_ );
470    }
471    
472  =head2 display  =head2 display
473    
474  Define tag just for I<display>  Define output just for I<display>
475    
476    @v = display('Title', rec('200','a') );    @v = display('Title', rec('200','a') );
477    
478  =cut  =cut
479    
480  sub display {  sub _field {
481          my $name = shift or die "display needs name as first argument";          my $type = shift or confess "need type -- BUG?";
482            my $name = shift or confess "needs name as first argument";
483          my @o = grep { defined($_) && $_ ne '' } @_;          my @o = grep { defined($_) && $_ ne '' } @_;
484          return unless (@o);          return unless (@o);
485          $out->{$name}->{tag} = $name;          $out->{$name}->{$type} = \@o;
         $out->{$name}->{display} = \@o;  
486  }  }
487    
488    sub display { _field( 'display', @_ ) }
489    
490  =head2 search  =head2 search
491    
492  Prepare values just for I<search>  Prepare values just for I<search>
# Line 468  Prepare values just for I<search> Line 495  Prepare values just for I<search>
495    
496  =cut  =cut
497    
498  sub search {  sub search { _field( 'search', @_ ) }
499          my $name = shift or die "search needs name as first argument";  
500          my @o = grep { defined($_) && $_ ne '' } @_;  =head2 sorted
501          return unless (@o);  
502          $out->{$name}->{tag} = $name;  Insert into lists which will be automatically sorted
503          $out->{$name}->{search} = \@o;  
504  }   sorted('Title', rec('200','a') );
505    
506    =cut
507    
508    sub sorted { _field( 'sorted', @_ ) }
509    
510    
511  =head2 marc_leader  =head2 marc_leader
512    
# Line 489  sub marc_leader { Line 521  sub marc_leader {
521          my ($offset,$value) = @_;          my ($offset,$value) = @_;
522    
523          if ($offset) {          if ($offset) {
524                  $leader->{ $offset } = $value;                  $marc_leader->[ $marc_record_offset ]->{ $offset } = $value;
525          } else {          } else {
526                  return $leader;                  
527                    if (defined($marc_leader)) {
528                            die "marc_leader not array = ", dump( $marc_leader ) unless (ref($marc_leader) eq 'ARRAY');
529                            return $marc_leader->[ $marc_record_offset ];
530                    } else {
531                            return;
532                    }
533            }
534    }
535    
536    =head2 marc_fixed
537    
538    Create control/indentifier fields with values in fixed positions
539    
540      marc_fixed('008', 00, '070402');
541      marc_fixed('008', 39, '|');
542    
543    Positions not specified will be filled with spaces (C<0x20>).
544    
545    There will be no effort to extend last specified value to full length of
546    field in standard.
547    
548    =cut
549    
550    sub marc_fixed {
551            my ($f, $pos, $val) = @_;
552            die "need marc(field, position, value)" unless defined($f) && defined($pos);
553    
554            confess "need val" unless defined $val;
555    
556            my $update = 0;
557    
558            map {
559                    if ($_->[0] eq $f) {
560                            my $old = $_->[1];
561                            if (length($old) <= $pos) {
562                                    $_->[1] .= ' ' x ( $pos - length($old) ) . $val;
563                                    warn "## marc_fixed($f,$pos,'$val') append '$old' -> '$_->[1]'\n" if ($debug > 1);
564                            } else {
565                                    $_->[1] = substr($old, 0, $pos) . $val . substr($old, $pos + length($val));
566                                    warn "## marc_fixed($f,$pos,'$val') update '$old' -> '$_->[1]'\n" if ($debug > 1);
567                            }
568                            $update++;
569                    }
570            } @{ $marc_record->[ $marc_record_offset ] };
571    
572            if (! $update) {
573                    my $v = ' ' x $pos . $val;
574                    push @{ $marc_record->[ $marc_record_offset ] }, [ $f, $v ];
575                    warn "## marc_fixed($f,$pos,'val') created '$v'\n" if ($debug > 1);
576          }          }
577  }  }
578    
# Line 623  sub marc_duplicate { Line 704  sub marc_duplicate {
704           my $m = $marc_record->[ -1 ];           my $m = $marc_record->[ -1 ];
705           die "can't duplicate record which isn't defined" unless ($m);           die "can't duplicate record which isn't defined" unless ($m);
706           push @{ $marc_record }, dclone( $m );           push @{ $marc_record }, dclone( $m );
707           warn "## marc_duplicate = ", dump(@$marc_record), $/ if ($debug > 1);           push @{ $marc_leader }, dclone( marc_leader() );
708             warn "## marc_duplicate = ", dump(@$marc_leader, @$marc_record), $/ if ($debug > 1);
709           $marc_record_offset = $#{ $marc_record };           $marc_record_offset = $#{ $marc_record };
710           warn "## marc_record_offset = $marc_record_offset", $/ if ($debug > 1);           warn "## marc_record_offset = $marc_record_offset", $/ if ($debug > 1);
711    
712  }  }
713    
714  =head2 marc_remove  =head2 marc_remove
# Line 661  sub marc_remove { Line 744  sub marc_remove {
744          if ($f eq '*') {          if ($f eq '*') {
745    
746                  delete( $marc_record->[ $marc_record_offset ] );                  delete( $marc_record->[ $marc_record_offset ] );
747                    warn "## full marc_record = ", dump( @{ $marc_record }), $/ if ($debug > 1);
748    
749          } else {          } else {
750    
# Line 699  sub marc_remove { Line 783  sub marc_remove {
783                  $marc_record->[ $marc_record_offset ] = $marc;                  $marc_record->[ $marc_record_offset ] = $marc;
784          }          }
785    
   
786          warn "## full marc_record = ", dump( @{ $marc_record }), $/ if ($debug > 1);          warn "## full marc_record = ", dump( @{ $marc_record }), $/ if ($debug > 1);
787  }  }
788    
# Line 773  sub marc_original_order { Line 856  sub marc_original_order {
856          warn "## marc_record = ", dump( $marc_record ),$/ if ($debug > 1);          warn "## marc_record = ", dump( $marc_record ),$/ if ($debug > 1);
857  }  }
858    
859    =head2 marc_count
860    
861    Return number of MARC records created using L</marc_duplicate>.
862    
863      print "created ", marc_count(), " records";
864    
865    =cut
866    
867    sub marc_count {
868            return $#{ $marc_record };
869    }
870    
871    
872  =head1 Functions to extract data from input  =head1 Functions to extract data from input
873    
# Line 795  sub _pack_subfields_hash { Line 890  sub _pack_subfields_hash {
890    
891          my ($h,$include_subfields) = @_;          my ($h,$include_subfields) = @_;
892    
893            # sanity and ease of use
894            return $h if (ref($h) ne 'HASH');
895    
896          if ( defined($h->{subfields}) ) {          if ( defined($h->{subfields}) ) {
897                  my $sfs = delete $h->{subfields} || die "no subfields?";                  my $sfs = delete $h->{subfields} || die "no subfields?";
898                  my @out;                  my @out;
# Line 944  Prefix all values with a string Line 1042  Prefix all values with a string
1042  =cut  =cut
1043    
1044  sub prefix {  sub prefix {
1045          my $p = shift or return;          my $p = shift;
1046            return @_ unless defined( $p );
1047          return map { $p . $_ } grep { defined($_) } @_;          return map { $p . $_ } grep { defined($_) } @_;
1048  }  }
1049    
# Line 957  suffix all values with a string Line 1056  suffix all values with a string
1056  =cut  =cut
1057    
1058  sub suffix {  sub suffix {
1059          my $s = shift or die "suffix needs string as first argument";          my $s = shift;
1060            return @_ unless defined( $s );
1061          return map { $_ . $s } grep { defined($_) } @_;          return map { $_ . $s } grep { defined($_) } @_;
1062  }  }
1063    
# Line 970  surround all values with a two strings Line 1070  surround all values with a two strings
1070  =cut  =cut
1071    
1072  sub surround {  sub surround {
1073          my $p = shift or die "surround need prefix as first argument";          my $p = shift;
1074          my $s = shift or die "surround needs suffix as second argument";          my $s = shift;
1075            $p = '' unless defined( $p );
1076            $s = '' unless defined( $s );
1077          return map { $p . $_ . $s } grep { defined($_) } @_;          return map { $p . $_ . $s } grep { defined($_) } @_;
1078  }  }
1079    
# Line 1146  Consult config values stored in C<config Line 1248  Consult config values stored in C<config
1248    $database_code = config();    # use _ from hash    $database_code = config();    # use _ from hash
1249    $database_name = config('name');    $database_name = config('name');
1250    $database_input_name = config('input name');    $database_input_name = config('input name');
   $tag = config('input normalize tag');  
1251    
1252  Up to three levels are supported.  Up to three levels are supported.
1253    
# Line 1272  my $hash; Line 1373  my $hash;
1373    
1374  sub set {  sub set {
1375          my ($k,$v) = @_;          my ($k,$v) = @_;
1376          warn "## set ( $k => ", dump($v), " )", $/;          warn "## set ( $k => ", dump($v), " )", $/ if ( $debug );
1377          $hash->{$k} = $v;          $hash->{$k} = $v;
1378  };  };
1379    
# Line 1285  sub set { Line 1386  sub set {
1386  sub get {  sub get {
1387          my $k = shift || return;          my $k = shift || return;
1388          my $v = $hash->{$k};          my $v = $hash->{$k};
1389          warn "## get $k = ", dump( $v ), $/;          warn "## get $k = ", dump( $v ), $/ if ( $debug );
1390          return $v;          return $v;
1391  }  }
1392    
1393    =head2 count
1394    
1395      if ( count( @result ) == 1 ) {
1396            # do something if only 1 result is there
1397      }
1398    
1399    =cut
1400    
1401    sub count {
1402            warn "## count ",dump(@_),$/ if ( $debug );
1403            return @_ . '';
1404    }
1405    
1406  # END  # END
1407  1;  1;

Legend:
Removed from v.786  
changed lines
  Added in v.973

  ViewVC Help
Powered by ViewVC 1.1.26