/[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 721 by dpavlin, Fri Sep 29 12:27:47 2006 UTC revision 752 by dpavlin, Sun Oct 8 18:21:26 2006 UTC
# Line 2  package WebPAC::Normalize; Line 2  package WebPAC::Normalize;
2  use Exporter 'import';  use Exporter 'import';
3  @EXPORT = qw/  @EXPORT = qw/
4          _set_rec _set_lookup          _set_rec _set_lookup
5            _set_load_row
6          _get_ds _clean_ds          _get_ds _clean_ds
7          _debug          _debug
8          _pack_subfields_hash          _pack_subfields_hash
# Line 26  use strict; Line 27  use strict;
27  #use base qw/WebPAC::Common/;  #use base qw/WebPAC::Common/;
28  use Data::Dump qw/dump/;  use Data::Dump qw/dump/;
29  use Storable qw/dclone/;  use Storable qw/dclone/;
30    use Carp qw/confess/;
31    
32  # debugging warn(s)  # debugging warn(s)
33  my $debug = 0;  my $debug = 0;
# Line 37  WebPAC::Normalize - describe normalisato Line 39  WebPAC::Normalize - describe normalisato
39    
40  =head1 VERSION  =head1 VERSION
41    
42  Version 0.20  Version 0.23
43    
44  =cut  =cut
45    
46  our $VERSION = '0.20';  our $VERSION = '0.23';
47    
48  =head1 SYNOPSIS  =head1 SYNOPSIS
49    
# Line 67  All other functions are available for us Line 69  All other functions are available for us
69  Return data structure  Return data structure
70    
71    my $ds = WebPAC::Normalize::data_structure(    my $ds = WebPAC::Normalize::data_structure(
72          lookup => $lookup_variable,          lookup => $lookup_hash,
73          row => $row,          row => $row,
74          rules => $normalize_pl_config,          rules => $normalize_pl_config,
75          marc_encoding => 'utf-8',          marc_encoding => 'utf-8',
76          config => $config,          config => $config,
77            load_row_coderef => sub {
78                    my ($database,$input,$mfn) = shift;
79                    $store->load_row( database => $database, input => $input, id => $mfn );
80            },
81    );    );
82    
83  Options C<row>, C<rules> and C<log> are mandatory while all  Options C<row>, C<rules> and C<log> are mandatory while all
84  other are optional.  other are optional.
85    
86    C<load_row_coderef> is closure only used when executing lookups, so they will
87    die if it's not defined.
88    
89  This function will B<die> if normalizastion can't be evaled.  This function will B<die> if normalizastion can't be evaled.
90    
91  Since this function isn't exported you have to call it with  Since this function isn't exported you have to call it with
# Line 84  C<WebPAC::Normalize::data_structure>. Line 93  C<WebPAC::Normalize::data_structure>.
93    
94  =cut  =cut
95    
96    my $load_row_coderef;
97    
98  sub data_structure {  sub data_structure {
99          my $arg = {@_};          my $arg = {@_};
100    
# Line 91  sub data_structure { Line 102  sub data_structure {
102          die "need normalisation argument" unless ($arg->{rules});          die "need normalisation argument" unless ($arg->{rules});
103    
104          no strict 'subs';          no strict 'subs';
105          _set_lookup( $arg->{lookup} ) if (defined( $arg->{lookup} ));          _set_lookup( $arg->{lookup} ) if defined($arg->{lookup});
106          _set_rec( $arg->{row} );          _set_rec( $arg->{row} );
107          _set_config( $arg->{config} ) if (defined( $arg->{config} ));          _set_config( $arg->{config} ) if defined($arg->{config});
108          _clean_ds( %{ $arg } );          _clean_ds( %{ $arg } );
109            $load_row_coderef = $arg->{load_row_coderef};
110    
111          eval "$arg->{rules}";          eval "$arg->{rules}";
112          die "error evaling $arg->{rules}: $@\n" if ($@);          die "error evaling $arg->{rules}: $@\n" if ($@);
113    
# Line 151  Return hash formatted as data structure Line 164  Return hash formatted as data structure
164    
165  =cut  =cut
166    
167  my ($out, $marc_record, $marc_encoding, $marc_repeatable_subfield, $marc_indicators);  my ($out, $marc_record, $marc_encoding, $marc_repeatable_subfield, $marc_indicators, $leader);
168  my ($marc_record_offset, $marc_fetch_offset) = (0, 0);  my ($marc_record_offset, $marc_fetch_offset) = (0, 0);
169    
170  sub _get_ds {  sub _get_ds {
# Line 168  Clean data structure hash for next recor Line 181  Clean data structure hash for next recor
181    
182  sub _clean_ds {  sub _clean_ds {
183          my $a = {@_};          my $a = {@_};
184          ($out,$marc_record, $marc_encoding, $marc_repeatable_subfield, $marc_indicators) = ();          ($out,$marc_record, $marc_encoding, $marc_repeatable_subfield, $marc_indicators, $leader) = ();
185          ($marc_record_offset, $marc_fetch_offset) = (0,0);          ($marc_record_offset, $marc_fetch_offset) = (0,0);
186          $marc_encoding = $a->{marc_encoding};          $marc_encoding = $a->{marc_encoding};
187  }  }
# Line 199  sub _get_lookup { Line 212  sub _get_lookup {
212          return $lookup;          return $lookup;
213  }  }
214    
215    =head2 _set_load_row
216    
217    Setup code reference which will return L<data_structure> from
218    L<WebPAC::Store>
219    
220      _set_load_row(sub {
221                    my ($database,$input,$mfn) = @_;
222                    $store->load_row( database => $database, input => $input, id => $mfn );
223      });
224    
225    =cut
226    
227    sub _set_load_row {
228            my $coderef = shift;
229            confess "argument isn't CODE" unless ref($coderef) eq 'CODE';
230    
231            $load_row_coderef = $coderef;
232    }
233    
234  =head2 _get_marc_fields  =head2 _get_marc_fields
235    
236  Get all fields defined by calls to C<marc>  Get all fields defined by calls to C<marc>
# Line 455  sub marc_leader { Line 487  sub marc_leader {
487          my ($offset,$value) = @_;          my ($offset,$value) = @_;
488    
489          if ($offset) {          if ($offset) {
490                  $out->{' leader'}->{ $offset } = $value;                  $leader->{ $offset } = $value;
491          } else {          } else {
492                  return $out->{' leader'};                  return $leader;
493          }          }
494  }  }
495    
# Line 848  syntaxtic sugar for Line 880  syntaxtic sugar for
880    @v = rec('200')    @v = rec('200')
881    @v = rec('200','a')    @v = rec('200','a')
882    
883    If rec() returns just single value, it will
884    return scalar, not array.
885    
886  =cut  =cut
887    
888  sub rec {  sub rec {
# Line 857  sub rec { Line 892  sub rec {
892          } elsif ($#_ == 1) {          } elsif ($#_ == 1) {
893                  @out = rec2(@_);                  @out = rec2(@_);
894          }          }
895          if (@out) {          if ($#out == 0 && ! wantarray) {
896                    return $out[0];
897            } elsif (@out) {
898                  return @out;                  return @out;
899          } else {          } else {
900                  return '';                  return '';
# Line 941  sub first { Line 978  sub first {
978    
979  Consult lookup hashes for some value  Consult lookup hashes for some value
980    
981    @v = lookup( $v );    @v = lookup(
982    @v = lookup( @v );          sub {
983                    'ffkk/peri/mfn'.rec('000')
984            },
985            'ffkk','peri','200-a-200-e',
986            sub {
987                    first(rec(200,'a')).' '.first(rec('200','e'))
988            }
989      );
990    
991    Code like above will be B<automatically generated> using L<WebPAC::Parse> from
992    normal lookup definition in C<conf/lookup/something.pl> which looks like:
993    
994      lookup(
995            # which results to return from record recorded in lookup
996            sub { 'ffkk/peri/mfn' . rec('000') },
997            # from which database and input
998            'ffkk','peri',
999            # such that following values match
1000            sub { first(rec(200,'a')) . ' ' . first(rec('200','e')) },
1001            # if this part is missing, we will try to match same fields
1002            # from lookup record and current one, or you can override
1003            # which records to use from current record using
1004            sub { rec('900','x') . ' ' . rec('900','y') },
1005      )
1006    
1007    You can think about this lookup as SQL (if that helps):
1008    
1009      select
1010            sub { what }
1011      from
1012            database, input
1013      where
1014        sub { filter from lookuped record }
1015      having
1016        sub { optional filter on current record }
1017    
1018  FIXME B<currently this one is broken!>  Easy as pie, right?
1019    
1020  =cut  =cut
1021    
1022  sub lookup {  sub lookup {
1023          my $k = shift or return;          my ($what, $database, $input, $key, $having) = @_;
1024          return unless (defined($lookup->{$k}));  
1025          if (ref($lookup->{$k}) eq 'ARRAY') {          confess "lookup needs 5 arguments: what, database, input, key, having" unless ($#_ == 4);
1026                  return @{ $lookup->{$k} };  
1027            warn "## lookup ($database, $input, $key)", $/ if ($debug > 1);
1028            return unless (defined($lookup->{$database}->{$input}->{$key}));
1029    
1030            confess "lookup really need load_row_coderef added to data_structure\n" unless ($load_row_coderef);
1031    
1032            my $mfns;
1033            my @having = $having->();
1034    
1035            warn "## having = ", dump( @having ) if ($debug > 2);
1036    
1037            foreach my $h ( @having ) {
1038                    if (defined($lookup->{$database}->{$input}->{$key}->{$h})) {
1039                            warn "lookup for $database/$input/$key/$h return ",dump($lookup->{$database}->{$input}->{$key}->{$h}),"\n" if ($debug);
1040                            $mfns->{$_}++ foreach keys %{ $lookup->{$database}->{$input}->{$key}->{$h} };
1041                    }
1042            }
1043    
1044            return unless ($mfns);
1045    
1046            my @mfns = sort keys %$mfns;
1047    
1048            warn "# lookup loading $database/$input/$key mfn ", join(",",@mfns)," having ",dump(@having),"\n" if ($debug);
1049    
1050            my $old_rec = $rec;
1051            my @out;
1052    
1053            foreach my $mfn (@mfns) {
1054                    $rec = $load_row_coderef->( $database, $input, $mfn );
1055    
1056                    warn "got $database/$input/$mfn = ", dump($rec), $/ if ($debug);
1057    
1058                    my @vals = $what->();
1059    
1060                    push @out, ( @vals );
1061    
1062                    warn "lookup for mfn $mfn returned ", dump(@vals), $/ if ($debug);
1063            }
1064    
1065    #       if (ref($lookup->{$k}) eq 'ARRAY') {
1066    #               return @{ $lookup->{$k} };
1067    #       } else {
1068    #               return $lookup->{$k};
1069    #       }
1070    
1071            $rec = $old_rec;
1072    
1073            warn "## lookup returns = ", dump(@out), $/ if ($debug);
1074    
1075            if ($#out == 0) {
1076                    return $out[0];
1077          } else {          } else {
1078                  return $lookup->{$k};                  return @out;
1079          }          }
1080  }  }
1081    
1082  =head2 save_into_lookup  =head2 save_into_lookup
1083    
1084  Save value into lookup.  Save value into lookup. It associates current database, input
1085    and specific keys with one or more values which will be
1086    associated over MFN.
1087    
1088    MFN will be extracted from first occurence current of field 000
1089    in current record, or if it doesn't exist from L<_set_config> C<_mfn>.
1090    
1091    save_into_lookup($database,$input,$key,sub {    my $nr = save_into_lookup($database,$input,$key,sub {
1092          # code which produce one or more values          # code which produce one or more values
1093    });    });
1094    
1095  This function shouldn't be called directly, it's called from code created by L<WebPAC::Parser>.  It returns number of items saved.
1096    
1097    This function shouldn't be called directly, it's called from code created by
1098    L<WebPAC::Parser>.
1099    
1100  =cut  =cut
1101    
# Line 976  sub save_into_lookup { Line 1105  sub save_into_lookup {
1105          die "save_into_lookup needs input" unless defined($input);          die "save_into_lookup needs input" unless defined($input);
1106          die "save_into_lookup needs key" unless defined($key);          die "save_into_lookup needs key" unless defined($key);
1107          die "save_into_lookup needs CODE" unless ( defined($coderef) && ref($coderef) eq 'CODE' );          die "save_into_lookup needs CODE" unless ( defined($coderef) && ref($coderef) eq 'CODE' );
1108          my $mfn = $rec->{'000'}->[0] || die "mfn not defined or zero";  
1109            warn "## save_into_lookup rec = ", dump($rec), " config = ", dump($config), $/ if ($debug > 2);
1110    
1111            my $mfn =
1112                    defined($rec->{'000'}->[0])     ?       $rec->{'000'}->[0]      :
1113                    defined($config->{_mfn})        ?       $config->{_mfn}         :
1114                                                                                    die "mfn not defined or zero";
1115    
1116            my $nr = 0;
1117    
1118          foreach my $v ( $coderef->() ) {          foreach my $v ( $coderef->() ) {
1119                  $lookup->{$database}->{$input}->{$key}->{$v}->{$mfn}++;                  $lookup->{$database}->{$input}->{$key}->{$v}->{$mfn}++;
1120                  warn "# saved lookup $database/$input/$key [$v] $mfn\n" if ($debug > 1);                  warn "# saved lookup $database/$input/$key [$v] $mfn\n" if ($debug > 1);
1121                    $nr++;
1122          }          }
1123    
1124            return $nr;
1125  }  }
1126    
1127  =head2 config  =head2 config

Legend:
Removed from v.721  
changed lines
  Added in v.752

  ViewVC Help
Powered by ViewVC 1.1.26