/[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 551 by dpavlin, Fri Jun 30 20:43:09 2006 UTC revision 571 by dpavlin, Mon Jul 3 14:30:22 2006 UTC
# Line 3  use Exporter 'import'; Line 3  use Exporter 'import';
3  @EXPORT = qw/  @EXPORT = qw/
4          _set_rec _set_lookup          _set_rec _set_lookup
5          _get_ds _clean_ds          _get_ds _clean_ds
6            _debug
7    
8          tag search display          tag search display
9          marc marc_indicators marc_repeatable_subfield          marc marc_indicators marc_repeatable_subfield
10            marc_compose marc_leader
11    
12          rec1 rec2 rec          rec1 rec2 rec
13          regex prefix suffix surround          regex prefix suffix surround
14          first lookup join_with          first lookup join_with
15    
16            split_rec_on
17  /;  /;
18    
19  use warnings;  use warnings;
# Line 29  WebPAC::Normalize - describe normalisato Line 33  WebPAC::Normalize - describe normalisato
33    
34  =head1 VERSION  =head1 VERSION
35    
36  Version 0.06  Version 0.09
37    
38  =cut  =cut
39    
40  our $VERSION = '0.06';  our $VERSION = '0.09';
41    
42  =head1 SYNOPSIS  =head1 SYNOPSIS
43    
# Line 153  Get all fields defined by calls to C<mar Line 157  Get all fields defined by calls to C<mar
157    
158          $marc->add_fields( WebPAC::Normalize:_get_marc_fields() );          $marc->add_fields( WebPAC::Normalize:_get_marc_fields() );
159    
   
   
160  We are using I<magic> which detect repeatable fields only from  We are using I<magic> which detect repeatable fields only from
161  sequence of field/subfield data generated by normalization.  sequence of field/subfield data generated by normalization.
162    
163  Repeatable field is created if there is second occurence of same subfield or  Repeatable field is created when there is second occurence of same subfield or
164  if any of indicators are different. This is sane for most cases except for  if any of indicators are different.
165  non-repeatable fields with repeatable subfields.  
166    This is sane for most cases. Something like:
167    
168  You can change behaviour of that using C<marc_repeatable_subfield>.    900a-1 900b-1 900c-1
169      900a-2 900b-2
170      900a-3
171    
172    will be created from any combination of:
173    
174      900a-1 900a-2 900a-3 900b-1 900b-2 900c-1
175    
176    and following rules:
177    
178      marc('900','a', rec('200','a') );
179      marc('900','b', rec('200','b') );
180      marc('900','c', rec('200','c') );
181    
182    which might not be what you have in mind. If you need repeatable subfield,
183    define it using C<marc_repeatable_subfield> like this:
184    
185    ....
186    
187  =cut  =cut
188    
# Line 176  sub _get_marc_fields { Line 196  sub _get_marc_fields {
196                  $a->[0] . $a->[3] cmp $b->[0] . $b->[3]                  $a->[0] . $a->[3] cmp $b->[0] . $b->[3]
197          } @{ $marc_record };          } @{ $marc_record };
198    
199            @sorted_marc_record = @{ $marc_record };        ### FIXME disable sorting
200            
201          # output marc fields          # output marc fields
202          my @m;          my @m;
203    
# Line 266  sub _get_marc_fields { Line 288  sub _get_marc_fields {
288          return @m;          return @m;
289  }  }
290    
291    =head2 _debug
292    
293    Change level of debug warnings
294    
295      _debug( 2 );
296    
297    =cut
298    
299    sub _debug {
300            my $l = shift;
301            return $debug unless defined($l);
302            warn "debug level $l",$/ if ($l > 0);
303            $debug = $l;
304    }
305    
306  =head1 Functions to create C<data_structure>  =head1 Functions to create C<data_structure>
307    
308  Those functions generally have to first in your normalization file.  Those functions generally have to first in your normalization file.
# Line 320  sub search { Line 357  sub search {
357          $out->{$name}->{search} = \@o;          $out->{$name}->{search} = \@o;
358  }  }
359    
360    =head2 marc_leader
361    
362    Setup fields within MARC leader or get leader
363    
364      marc_leader('05','c');
365      my $leader = marc_leader();
366    
367    =cut
368    
369    sub marc_leader {
370            my ($offset,$value) = @_;
371    
372            if ($offset) {
373                    $out->{' leader'}->{ $offset } = $value;
374            } else {
375                    return $out->{' leader'};
376            }
377    }
378    
379  =head2 marc  =head2 marc
380    
381  Save value for MARC field  Save value for MARC field
382    
383    marc('900','a', rec('200','a') );    marc('900','a', rec('200','a') );
384      marc('001', rec('000') );
385    
386  =cut  =cut
387    
# Line 332  sub marc { Line 389  sub marc {
389          my $f = shift or die "marc needs field";          my $f = shift or die "marc needs field";
390          die "marc field must be numer" unless ($f =~ /^\d+$/);          die "marc field must be numer" unless ($f =~ /^\d+$/);
391    
392          my $sf = shift or die "marc needs subfield";          my $sf;
393            if ($f >= 10) {
394                    $sf = shift or die "marc needs subfield";
395            }
396    
397          foreach (@_) {          foreach (@_) {
398                  my $v = $_;             # make var read-write for Encode                  my $v = $_;             # make var read-write for Encode
399                  next unless (defined($v) && $v !~ /^\s*$/);                  next unless (defined($v) && $v !~ /^\s*$/);
400                  from_to($v, 'iso-8859-2', $marc_encoding) if ($marc_encoding);                  from_to($v, 'iso-8859-2', $marc_encoding) if ($marc_encoding);
401                  my ($i1,$i2) = defined($marc_indicators->{$f}) ? @{ $marc_indicators->{$f} } : (' ',' ');                  my ($i1,$i2) = defined($marc_indicators->{$f}) ? @{ $marc_indicators->{$f} } : (' ',' ');
402                  push @{ $marc_record }, [ $f, $i1, $i2, $sf => $v ];                  if (defined $sf) {
403                            push @{ $marc_record }, [ $f, $i1, $i2, $sf => $v ];
404                    } else {
405                            push @{ $marc_record }, [ $f, $v ];
406                    }
407          }          }
408  }  }
409    
# Line 379  sub marc_indicators { Line 443  sub marc_indicators {
443          @{ $marc_indicators->{$f} } = ($i1,$i2);          @{ $marc_indicators->{$f} } = ($i1,$i2);
444  }  }
445    
446    =head2 marc_compose
447    
448    Save values for each MARC subfield explicitly
449    
450      marc_compose('900',
451            'a', rec('200','a')
452            'b', rec('201','a')
453            'a', rec('200','b')
454            'c', rec('200','c')
455      );
456    
457    =cut
458    
459    sub marc_compose {
460            my $f = shift or die "marc_compose needs field";
461            die "marc_compose field must be numer" unless ($f =~ /^\d+$/);
462    
463            my ($i1,$i2) = defined($marc_indicators->{$f}) ? @{ $marc_indicators->{$f} } : (' ',' ');
464            my $m = [ $f, $i1, $i2 ];
465    
466            while (@_) {
467                    my $sf = shift or die "marc_compose $f needs subfield";
468                    my $v = shift;
469    
470                    next unless (defined($v) && $v !~ /^\s*$/);
471                    from_to($v, 'iso-8859-2', $marc_encoding) if ($marc_encoding);
472                    push @$m, ( $sf, $v );
473                    warn "## ++ marc_compose($f,$sf,$v) ", dump( $m ),$/ if ($debug > 1);
474            }
475    
476            warn "## marc_compose(d) ", dump( $m ),$/ if ($debug > 1);
477    
478            push @{ $marc_record }, $m if ($#{$m} > 2);
479    }
480    
481    
482  =head1 Functions to extract data from input  =head1 Functions to extract data from input
483    
# Line 397  TODO: order of values is probably same a Line 496  TODO: order of values is probably same a
496    
497  sub rec1 {  sub rec1 {
498          my $f = shift;          my $f = shift;
499            warn "rec1($f) = ", dump( $rec->{$f} ), $/ if ($debug > 1);
500          return unless (defined($rec) && defined($rec->{$f}));          return unless (defined($rec) && defined($rec->{$f}));
501            warn "rec1($f) = ", dump( $rec->{$f} ), $/ if ($debug > 1);
502          if (ref($rec->{$f}) eq 'ARRAY') {          if (ref($rec->{$f}) eq 'ARRAY') {
503                  return map {                  return map {
504                          if (ref($_) eq 'HASH') {                          if (ref($_) eq 'HASH') {
# Line 548  sub join_with { Line 649  sub join_with {
649          return join($d, grep { defined($_) && $_ ne '' } @_);          return join($d, grep { defined($_) && $_ ne '' } @_);
650  }  }
651    
652    =head2 split_rec_on
653    
654    Split record subfield on some regex and take one of parts out
655    
656      $a_before_semi_column =
657            split_rec_on('200','a', /\s*;\s*/, $part);
658    
659    C<$part> is optional number of element. First element is
660    B<1>, not 0!
661    
662    If there is no C<$part> parameter or C<$part> is 0, this function will
663    return all values produced by splitting.
664    
665    =cut
666    
667    sub split_rec_on {
668            die "split_rec_on need (fld,sf,regex[,part]" if ($#_ < 2);
669    
670            my ($fld, $sf, $regex, $part) = @_;
671            warn "### regex ", ref($regex), $regex, $/ if ($debug > 2);
672    
673            my @r = rec( $fld, $sf );
674            my $v = shift @r;
675            warn "### first rec($fld,$sf) = ",dump($v),$/ if ($debug > 2);
676    
677            return '' if( ! defined($v) || $v =~ /^\s*$/);
678    
679            my @s = split( $regex, $v );
680            warn "## split_rec_on($fld,$sf,$regex,$part) = ",dump(@s),$/ if ($debug > 1);
681            if ($part && $part > 0) {
682                    return $s[ $part - 1 ];
683            } else {
684                    return @s;
685            }
686    }
687    
688  # END  # END
689  1;  1;
Legend:
Removed from v.551  
changed lines
  Added in v.571
  ViewVC Help
Powered by ViewVC 1.1.26