/[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 550 by dpavlin, Fri Jun 30 18:48:33 2006 UTC revision 562 by dpavlin, Sun Jul 2 16:14:41 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
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 0.08
37    
38  =cut  =cut
39    
40  our $VERSION = '0.06';  our $VERSION = '0.08';
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      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  You can change behaviour of that using C<marc_repeatable_subfield>.    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    
189  sub _get_marc_fields {  sub _get_marc_fields {
190    
191            return if (! $marc_record || ref($marc_record) ne 'ARRAY' || $#{ $marc_record } < 0);
192    
193          # first, sort all existing fields          # first, sort all existing fields
194          # XXX might not be needed, but modern perl might randomize elements in hash          # XXX might not be needed, but modern perl might randomize elements in hash
# Line 175  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 265  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 378  sub marc_indicators { Line 416  sub marc_indicators {
416          @{ $marc_indicators->{$f} } = ($i1,$i2);          @{ $marc_indicators->{$f} } = ($i1,$i2);
417  }  }
418    
419    =head2 marc_compose
420    
421    Save values for each MARC subfield explicitly
422    
423      marc_compose('900',
424            'a', rec('200','a')
425            'b', rec('201','a')
426            'a', rec('200','b')
427            'c', rec('200','c')
428      );
429    
430    =cut
431    
432    sub marc_compose {
433            my $f = shift or die "marc_compose needs field";
434            die "marc_compose field must be numer" unless ($f =~ /^\d+$/);
435    
436            my ($i1,$i2) = defined($marc_indicators->{$f}) ? @{ $marc_indicators->{$f} } : (' ',' ');
437            my $m = [ $f, $i1, $i2 ];
438    
439            while (@_) {
440                    my $sf = shift or die "marc_compose $f needs subfield";
441                    my $v = shift or die "marc_compose $f needs value for subfield $sf";
442    
443                    next unless (defined($v) && $v !~ /^\s*$/);
444                    from_to($v, 'iso-8859-2', $marc_encoding) if ($marc_encoding);
445                    push @$m, ( $sf, $v );
446                    warn "## ++ marc_compose($f,$sf,$v) ", dump( $m ) if ($debug > 1);
447            }
448    
449            warn "## marc_compose(d) ", dump( $m ) if ($debug > 1);
450    
451            push @{ $marc_record }, $m;
452    }
453    
454    
455  =head1 Functions to extract data from input  =head1 Functions to extract data from input
456    
# Line 547  sub join_with { Line 620  sub join_with {
620          return join($d, grep { defined($_) && $_ ne '' } @_);          return join($d, grep { defined($_) && $_ ne '' } @_);
621  }  }
622    
623    =head2 split_rec_on
624    
625    Split record subfield on some regex and take one of parts out
626    
627      $a_before_semi_column =
628            split_rec_on('200','a', /\s*;\s*/, $part);
629    
630    C<$part> is optional number of element. First element is
631    B<1>, not 0!
632    
633    If there is no C<$part> parameter or C<$part> is 0, this function will
634    return all values produced by splitting.
635    
636    =cut
637    
638    sub split_rec_on {
639            die "split_rec_on need (fld,sf,regex[,part]" if ($#_ < 2);
640    
641            my ($fld, $sf, $regex, $part) = @_;
642            warn "### regex ", ref($regex), $regex if ($debug > 2);
643    
644            my @r = rec( $fld, $sf );
645            my $v = shift @r;
646            warn "### first rec($fld,$sf) = ",dump($v) if ($debug > 2);
647    
648            my @s = split( $regex, $v );
649            warn "## split_rec_on($fld,$sf,$regex,$part) = ",dump(@s) if ($debug > 1);
650            if ($part > 0) {
651                    return $s[ $part - 1 ];
652            } else {
653                    return @s;
654            }
655    }
656    
657  # END  # END
658  1;  1;

Legend:
Removed from v.550  
changed lines
  Added in v.562

  ViewVC Help
Powered by ViewVC 1.1.26