1 |
package WebPAC::Normalize; |
package WebPAC::Normalize; |
2 |
use Exporter 'import'; |
use Exporter 'import'; |
3 |
@EXPORT = qw/ |
our @EXPORT = qw/ |
4 |
_set_rec _set_lookup |
_set_ds _set_lookup |
5 |
_set_load_row |
_set_load_row |
6 |
_get_ds _clean_ds |
_get_ds _clean_ds |
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_fixed |
marc_compose marc_leader marc_fixed |
14 |
marc_duplicate marc_remove marc_count |
marc_duplicate marc_remove marc_count |
15 |
marc_original_order |
marc_original_order |
16 |
|
marc_template |
17 |
|
|
18 |
rec1 rec2 rec |
rec1 rec2 rec |
19 |
|
frec frec_eq frec_ne |
20 |
regex prefix suffix surround |
regex prefix suffix surround |
21 |
first lookup join_with |
first lookup join_with |
22 |
save_into_lookup |
save_into_lookup |
25 |
|
|
26 |
get set |
get set |
27 |
count |
count |
28 |
|
|
29 |
/; |
/; |
30 |
|
|
31 |
use warnings; |
use warnings; |
39 |
# debugging warn(s) |
# debugging warn(s) |
40 |
my $debug = 0; |
my $debug = 0; |
41 |
|
|
42 |
|
use WebPAC::Normalize::ISBN; |
43 |
|
push @EXPORT, ( 'isbn_10', 'isbn_13' ); |
44 |
|
|
45 |
=head1 NAME |
=head1 NAME |
46 |
|
|
47 |
WebPAC::Normalize - describe normalisaton rules using sets |
WebPAC::Normalize - describe normalisaton rules using sets |
48 |
|
|
|
=head1 VERSION |
|
|
|
|
|
Version 0.29 |
|
|
|
|
49 |
=cut |
=cut |
50 |
|
|
51 |
our $VERSION = '0.29'; |
our $VERSION = '0.34'; |
52 |
|
|
53 |
=head1 SYNOPSIS |
=head1 SYNOPSIS |
54 |
|
|
61 |
C<perl -c normalize.pl>. |
C<perl -c normalize.pl>. |
62 |
|
|
63 |
Normalisation can generate multiple output normalized data. For now, supported output |
Normalisation can generate multiple output normalized data. For now, supported output |
64 |
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 |
65 |
C<marc>. |
C<marc>. |
66 |
|
|
67 |
=head1 FUNCTIONS |
=head1 FUNCTIONS |
80 |
marc_encoding => 'utf-8', |
marc_encoding => 'utf-8', |
81 |
config => $config, |
config => $config, |
82 |
load_row_coderef => sub { |
load_row_coderef => sub { |
83 |
my ($database,$input,$mfn) = shift; |
my ($database,$input,$mfn) = @_; |
84 |
$store->load_row( database => $database, input => $input, id => $mfn ); |
$store->load_row( database => $database, input => $input, id => $mfn ); |
85 |
}, |
}, |
86 |
); |
); |
106 |
die "need row argument" unless ($arg->{row}); |
die "need row argument" unless ($arg->{row}); |
107 |
die "need normalisation argument" unless ($arg->{rules}); |
die "need normalisation argument" unless ($arg->{rules}); |
108 |
|
|
|
no strict 'subs'; |
|
109 |
_set_lookup( $arg->{lookup} ) if defined($arg->{lookup}); |
_set_lookup( $arg->{lookup} ) if defined($arg->{lookup}); |
110 |
_set_rec( $arg->{row} ); |
_set_ds( $arg->{row} ); |
111 |
_set_config( $arg->{config} ) if defined($arg->{config}); |
_set_config( $arg->{config} ) if defined($arg->{config}); |
112 |
_clean_ds( %{ $arg } ); |
_clean_ds( %{ $arg } ); |
113 |
$load_row_coderef = $arg->{load_row_coderef}; |
$load_row_coderef = $arg->{load_row_coderef}; |
114 |
|
|
115 |
eval "$arg->{rules}"; |
no strict 'subs'; |
116 |
|
no warnings 'redefine'; |
117 |
|
eval "$arg->{rules};"; |
118 |
die "error evaling $arg->{rules}: $@\n" if ($@); |
die "error evaling $arg->{rules}: $@\n" if ($@); |
119 |
|
|
120 |
return _get_ds(); |
return _get_ds(); |
121 |
} |
} |
122 |
|
|
123 |
=head2 _set_rec |
=head2 _set_ds |
124 |
|
|
125 |
Set current record hash |
Set current record hash |
126 |
|
|
127 |
_set_rec( $rec ); |
_set_ds( $rec ); |
128 |
|
|
129 |
=cut |
=cut |
130 |
|
|
131 |
my $rec; |
my $rec; |
132 |
|
|
133 |
sub _set_rec { |
sub _set_ds { |
134 |
$rec = shift or die "no record hash"; |
$rec = shift or die "no record hash"; |
135 |
} |
} |
136 |
|
|
174 |
my ($marc_record_offset, $marc_fetch_offset) = (0, 0); |
my ($marc_record_offset, $marc_fetch_offset) = (0, 0); |
175 |
|
|
176 |
sub _get_ds { |
sub _get_ds { |
177 |
|
#warn "## out = ",dump($out); |
178 |
return $out; |
return $out; |
179 |
} |
} |
180 |
|
|
446 |
|
|
447 |
Those functions generally have to first in your normalization file. |
Those functions generally have to first in your normalization file. |
448 |
|
|
449 |
=head2 tag |
=head2 search_display |
450 |
|
|
451 |
Define new tag for I<search> and I<display>. |
Define output for L<search> and L<display> at the same time |
452 |
|
|
453 |
tag('Title', rec('200','a') ); |
search_display('Title', rec('200','a') ); |
454 |
|
|
455 |
|
|
456 |
=cut |
=cut |
457 |
|
|
458 |
sub tag { |
sub search_display { |
459 |
my $name = shift or die "tag needs name as first argument"; |
my $name = shift or die "search_display needs name as first argument"; |
460 |
my @o = grep { defined($_) && $_ ne '' } @_; |
my @o = grep { defined($_) && $_ ne '' } @_; |
461 |
return unless (@o); |
return unless (@o); |
|
$out->{$name}->{tag} = $name; |
|
462 |
$out->{$name}->{search} = \@o; |
$out->{$name}->{search} = \@o; |
463 |
$out->{$name}->{display} = \@o; |
$out->{$name}->{display} = \@o; |
464 |
} |
} |
465 |
|
|
466 |
|
=head2 tag |
467 |
|
|
468 |
|
Old name for L<search_display>, but supported |
469 |
|
|
470 |
|
=cut |
471 |
|
|
472 |
|
sub tag { |
473 |
|
search_display( @_ ); |
474 |
|
} |
475 |
|
|
476 |
=head2 display |
=head2 display |
477 |
|
|
478 |
Define tag just for I<display> |
Define output just for I<display> |
479 |
|
|
480 |
@v = display('Title', rec('200','a') ); |
@v = display('Title', rec('200','a') ); |
481 |
|
|
482 |
=cut |
=cut |
483 |
|
|
484 |
sub display { |
sub _field { |
485 |
my $name = shift or die "display needs name as first argument"; |
my $type = shift or confess "need type -- BUG?"; |
486 |
|
my $name = shift or confess "needs name as first argument"; |
487 |
my @o = grep { defined($_) && $_ ne '' } @_; |
my @o = grep { defined($_) && $_ ne '' } @_; |
488 |
return unless (@o); |
return unless (@o); |
489 |
$out->{$name}->{tag} = $name; |
$out->{$name}->{$type} = \@o; |
|
$out->{$name}->{display} = \@o; |
|
490 |
} |
} |
491 |
|
|
492 |
|
sub display { _field( 'display', @_ ) } |
493 |
|
|
494 |
=head2 search |
=head2 search |
495 |
|
|
496 |
Prepare values just for I<search> |
Prepare values just for I<search> |
499 |
|
|
500 |
=cut |
=cut |
501 |
|
|
502 |
sub search { |
sub search { _field( 'search', @_ ) } |
503 |
my $name = shift or die "search needs name as first argument"; |
|
504 |
my @o = grep { defined($_) && $_ ne '' } @_; |
=head2 sorted |
505 |
return unless (@o); |
|
506 |
$out->{$name}->{tag} = $name; |
Insert into lists which will be automatically sorted |
507 |
$out->{$name}->{search} = \@o; |
|
508 |
} |
sorted('Title', rec('200','a') ); |
509 |
|
|
510 |
|
=cut |
511 |
|
|
512 |
|
sub sorted { _field( 'sorted', @_ ) } |
513 |
|
|
514 |
|
|
515 |
=head2 marc_leader |
=head2 marc_leader |
516 |
|
|
815 |
return unless defined($rec->{$from}); |
return unless defined($rec->{$from}); |
816 |
|
|
817 |
my $r = $rec->{$from}; |
my $r = $rec->{$from}; |
818 |
die "record field $from isn't array\n" unless (ref($r) eq 'ARRAY'); |
die "record field $from isn't array ",dump( $rec ) unless (ref($r) eq 'ARRAY'); |
819 |
|
|
820 |
my ($i1,$i2) = defined($marc_indicators->{$to}) ? @{ $marc_indicators->{$to} } : (' ',' '); |
my ($i1,$i2) = defined($marc_indicators->{$to}) ? @{ $marc_indicators->{$to} } : (' ',' '); |
821 |
warn "## marc_original_order($to,$from) source = ", dump( $r ),$/ if ($debug > 1); |
warn "## marc_original_order($to,$from) source = ", dump( $r ),$/ if ($debug > 1); |
860 |
warn "## marc_record = ", dump( $marc_record ),$/ if ($debug > 1); |
warn "## marc_record = ", dump( $marc_record ),$/ if ($debug > 1); |
861 |
} |
} |
862 |
|
|
863 |
|
=head2 marc_template |
864 |
|
|
865 |
|
=cut |
866 |
|
|
867 |
|
sub marc_template { |
868 |
|
my $args = {@_}; |
869 |
|
warn "## marc_template(",dump($args),")"; |
870 |
|
|
871 |
|
foreach ( qw/subfields_rename marc_template/ ) { |
872 |
|
# warn "ref($_) = ",ref($args->{$_}); |
873 |
|
die "$_ not ARRAY" if ref($args->{$_}) ne 'ARRAY'; |
874 |
|
} |
875 |
|
|
876 |
|
my $r = $rec->{ $args->{from} } || return; |
877 |
|
die "record field ", $args->{from}, " isn't array ",dump( $rec ) unless (ref($r) eq 'ARRAY'); |
878 |
|
|
879 |
|
my @subfields_rename = @{ $args->{subfields_rename} }; |
880 |
|
# warn "### subfields_rename [$#subfields_rename] = ",dump( @subfields_rename ); |
881 |
|
|
882 |
|
confess "need mapping in pairs for subfields_rename" |
883 |
|
if $#subfields_rename % 2 != 1; |
884 |
|
|
885 |
|
my ( $subfields_rename, $from_subfields, $to_subfields ); |
886 |
|
while ( my ( $from, $to ) = splice(@subfields_rename, 0, 2) ) { |
887 |
|
my ( $f, $t ) = ( |
888 |
|
$from_subfields->{ $from }++, |
889 |
|
$to_subfields->{ $to }++ |
890 |
|
); |
891 |
|
$subfields_rename->{ $from }->[ $f ] = [ $to => $t ]; |
892 |
|
} |
893 |
|
warn "### subfields_rename = ",dump( $subfields_rename ),$/; |
894 |
|
warn "### from_subfields = ", dump( $from_subfields ),$/; |
895 |
|
warn "### to_subfields = ", dump( $to_subfields ),$/; |
896 |
|
|
897 |
|
my $fields_re = join('|', keys %$to_subfields ); |
898 |
|
|
899 |
|
my $pos_templates; |
900 |
|
my $count; |
901 |
|
my @marc_order; |
902 |
|
my $marc_template_order; |
903 |
|
my $fill_in; |
904 |
|
my @marc_out; |
905 |
|
|
906 |
|
foreach my $template ( @{ $args->{marc_template} } ) { |
907 |
|
$count = {}; |
908 |
|
@marc_order = (); |
909 |
|
sub my_count { |
910 |
|
my $sf = shift; |
911 |
|
my $nr = $count->{$sf}++; |
912 |
|
push @marc_order, [ $sf, $nr ]; |
913 |
|
return $sf . $nr; |
914 |
|
} |
915 |
|
my $pos_template = $template; |
916 |
|
$pos_template =~ s/($fields_re)/my_count($1)/ge; |
917 |
|
my $count_key = dump( $count ); |
918 |
|
warn "### template: |$template| -> |$pos_template| count = $count_key marc_order = ",dump( @marc_order ),$/; |
919 |
|
$pos_templates->{ $count_key } = $pos_template; |
920 |
|
$marc_template_order->{ $pos_template } = [ @marc_order ]; |
921 |
|
} |
922 |
|
warn "### from ",dump( $args->{marc_template} ), " created ", dump( $pos_templates ), " and ", dump( $marc_template_order ); |
923 |
|
|
924 |
|
my $m; |
925 |
|
|
926 |
|
foreach my $r ( @{ $rec->{ $args->{from} } } ) { |
927 |
|
|
928 |
|
my $i1 = $r->{i1} || ' '; |
929 |
|
my $i2 = $r->{i2} || ' '; |
930 |
|
$m = [ $args->{to}, $i1, $i2 ]; |
931 |
|
|
932 |
|
warn "### r = ",dump( $r ); |
933 |
|
|
934 |
|
my ( $new_r, $from_count, $to_count ); |
935 |
|
foreach my $sf ( keys %{$r} ) { |
936 |
|
# skip everything which isn't one char subfield (e.g. 'subfields') |
937 |
|
next unless $sf =~ m/^\w$/; |
938 |
|
my $nr = $from_count->{$sf}++; |
939 |
|
my $rename_to = $subfields_rename->{ $sf } || |
940 |
|
die "can't find subfield rename for $sf/$nr in ", dump( $subfields_rename ); |
941 |
|
warn "### rename $sf/$nr to ", dump( $rename_to->[$nr] ), $/; |
942 |
|
my ( $to_sf, $to_nr ) = @{ $rename_to->[$nr] }; |
943 |
|
$new_r->{ $to_sf }->[ $to_nr ] = [ $sf => $nr ]; |
944 |
|
|
945 |
|
$to_count->{ $to_sf }++; |
946 |
|
} |
947 |
|
|
948 |
|
warn "### new_r = ",dump( $new_r ); |
949 |
|
|
950 |
|
my $from_count_key = dump( $to_count ); |
951 |
|
|
952 |
|
warn "### from_count = ",dump( $from_count ), $/; |
953 |
|
warn "### to_count = ",dump( $to_count ), $/; |
954 |
|
|
955 |
|
my $template = $pos_templates->{ $from_count_key } || |
956 |
|
die "I don't have template for:\n$from_count_key\n## available templates\n", dump( $pos_templates ); |
957 |
|
|
958 |
|
warn "### selected template: |$template|\n"; |
959 |
|
|
960 |
|
$fill_in = {}; |
961 |
|
|
962 |
|
my @templates = split(/\|/, $template ); |
963 |
|
@templates = ( $template ); |
964 |
|
|
965 |
|
foreach my $sf ( @templates ) { |
966 |
|
sub fill_in { |
967 |
|
my ( $r, $sf, $nr ) = @_; |
968 |
|
my ( $from_sf, $from_nr ) = @{ $new_r->{ $sf }->[ $nr ] }; |
969 |
|
my $v = $r->{ $from_sf }; # || die "no $from_sf/$from_nr"; |
970 |
|
warn "#### fill_in( $sf, $nr ) = $from_sf/$from_nr >>>> ",dump( $v ), $/; |
971 |
|
if ( ref( $v ) eq 'ARRAY' ) { |
972 |
|
$fill_in->{$sf}->[$nr] = $v->[$from_nr]; |
973 |
|
return $v->[$from_nr]; |
974 |
|
} elsif ( $from_nr == 0 ) { |
975 |
|
$fill_in->{$sf}->[$nr] = $v; |
976 |
|
return $v; |
977 |
|
} else { |
978 |
|
die "requested subfield $from_sf/$from_nr but it's ",dump( $v ); |
979 |
|
} |
980 |
|
} |
981 |
|
warn "#### $sf <<<< $fields_re\n"; |
982 |
|
$sf =~ s/($fields_re)(\d+)/fill_in($r,$1,$2)/ge; |
983 |
|
warn "#### >>>> $sf with fill_in = ",dump( $fill_in ),$/; |
984 |
|
} |
985 |
|
|
986 |
|
warn "## template: |$template|\n## marc_template_order = ",dump( $marc_template_order ); |
987 |
|
|
988 |
|
foreach my $sf ( @{ $marc_template_order->{$template} } ) { |
989 |
|
my ( $sf, $nr ) = @$sf; |
990 |
|
my $v = $fill_in->{$sf}->[$nr] || die "can't find fill_in $sf/$nr"; |
991 |
|
warn "++ $sf/$nr |$v|\n"; |
992 |
|
push @$m, ( $sf, $v ); |
993 |
|
} |
994 |
|
|
995 |
|
warn "#### >>>> created marc: ", dump( $m ); |
996 |
|
|
997 |
|
push @marc_out, $m; |
998 |
|
} |
999 |
|
|
1000 |
|
warn "### marc_template produced: ",dump( @marc_out ); |
1001 |
|
|
1002 |
|
foreach my $marc ( @marc_out ) { |
1003 |
|
warn "+++ ",dump( $marc ); |
1004 |
|
push @{ $marc_record->[ $marc_record_offset ] }, $marc; |
1005 |
|
} |
1006 |
|
} |
1007 |
|
|
1008 |
=head2 marc_count |
=head2 marc_count |
1009 |
|
|
1010 |
Return number of MARC records created using L</marc_duplicate>. |
Return number of MARC records created using L</marc_duplicate>. |
1162 |
} |
} |
1163 |
} |
} |
1164 |
|
|
1165 |
|
=head2 frec |
1166 |
|
|
1167 |
|
Returns first value from field |
1168 |
|
|
1169 |
|
$v = frec('200'); |
1170 |
|
$v = frec('200','a'); |
1171 |
|
|
1172 |
|
=cut |
1173 |
|
|
1174 |
|
sub frec { |
1175 |
|
my @out = rec(@_); |
1176 |
|
warn "rec(",dump(@_),") has more than one return value, ignoring\n" if $#out > 0; |
1177 |
|
return shift @out; |
1178 |
|
} |
1179 |
|
|
1180 |
|
=head2 frec_eq |
1181 |
|
|
1182 |
|
=head2 frec_ne |
1183 |
|
|
1184 |
|
Check if first values from two fields are same or different |
1185 |
|
|
1186 |
|
if ( frec_eq( 900 => 'a', 910 => 'c' ) ) { |
1187 |
|
# values are same |
1188 |
|
} else { |
1189 |
|
# values are different |
1190 |
|
} |
1191 |
|
|
1192 |
|
Strictly speaking C<frec_eq> and C<frec_ne> wouldn't be needed if you |
1193 |
|
could write something like: |
1194 |
|
|
1195 |
|
if ( frec( '900','a' ) eq frec( '910','c' ) ) { |
1196 |
|
# yada tada |
1197 |
|
} |
1198 |
|
|
1199 |
|
but you can't since our parser L<WebPAC::Parser> will remove all whitespaces |
1200 |
|
in order to parse text and create invalid function C<eqfrec>. |
1201 |
|
|
1202 |
|
=cut |
1203 |
|
|
1204 |
|
sub frec_eq { |
1205 |
|
my ( $f1,$sf1, $f2, $sf2 ) = @_; |
1206 |
|
return (rec( $f1, $sf1 ))[0] eq (rec( $f2, $sf2 ))[0]; |
1207 |
|
} |
1208 |
|
|
1209 |
|
sub frec_ne { |
1210 |
|
return ! frec_eq( @_ ); |
1211 |
|
} |
1212 |
|
|
1213 |
=head2 regex |
=head2 regex |
1214 |
|
|
1215 |
Apply regex to some or all values |
Apply regex to some or all values |
1445 |
$database_code = config(); # use _ from hash |
$database_code = config(); # use _ from hash |
1446 |
$database_name = config('name'); |
$database_name = config('name'); |
1447 |
$database_input_name = config('input name'); |
$database_input_name = config('input name'); |
|
$tag = config('input normalize tag'); |
|
1448 |
|
|
1449 |
Up to three levels are supported. |
Up to three levels are supported. |
1450 |
|
|