package WebPAC::Normalize;
use Exporter 'import';
-@EXPORT = qw/
- _set_rec _set_lookup
+our @EXPORT = qw/
+ _set_ds _set_lookup
+ _set_load_row
_get_ds _clean_ds
_debug
+ _pack_subfields_hash
+
+ search_display search display sorted
- tag search display
marc marc_indicators marc_repeatable_subfield
- marc_compose marc_leader
+ marc_compose marc_leader marc_fixed
+ marc_duplicate marc_remove marc_count
+ marc_original_order
+ marc_template
rec1 rec2 rec
+ frec
regex prefix suffix surround
first lookup join_with
+ save_into_lookup
split_rec_on
+
+ get set
+ count
+
/;
use warnings;
#use base qw/WebPAC::Common/;
use Data::Dump qw/dump/;
-use Encode qw/from_to/;
+use Storable qw/dclone/;
+use Carp qw/confess/;
# debugging warn(s)
my $debug = 0;
+use WebPAC::Normalize::ISBN;
+push @EXPORT, ( 'isbn_10', 'isbn_13' );
=head1 NAME
WebPAC::Normalize - describe normalisaton rules using sets
-=head1 VERSION
-
-Version 0.09
-
=cut
-our $VERSION = '0.09';
+our $VERSION = '0.33';
=head1 SYNOPSIS
C<perl -c normalize.pl>.
Normalisation can generate multiple output normalized data. For now, supported output
-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
C<marc>.
=head1 FUNCTIONS
Return data structure
my $ds = WebPAC::Normalize::data_structure(
- lookup => $lookup->lookup_hash,
+ lookup => $lookup_hash,
row => $row,
rules => $normalize_pl_config,
marc_encoding => 'utf-8',
+ config => $config,
+ load_row_coderef => sub {
+ my ($database,$input,$mfn) = @_;
+ $store->load_row( database => $database, input => $input, id => $mfn );
+ },
);
-Options C<lookup>, C<row>, C<rules> and C<log> are mandatory while all
+Options C<row>, C<rules> and C<log> are mandatory while all
other are optional.
+C<load_row_coderef> is closure only used when executing lookups, so they will
+die if it's not defined.
+
This function will B<die> if normalizastion can't be evaled.
Since this function isn't exported you have to call it with
=cut
+my $load_row_coderef;
+
sub data_structure {
my $arg = {@_};
die "need row argument" unless ($arg->{row});
die "need normalisation argument" unless ($arg->{rules});
- no strict 'subs';
- _set_lookup( $arg->{lookup} );
- _set_rec( $arg->{row} );
+ _set_lookup( $arg->{lookup} ) if defined($arg->{lookup});
+ _set_ds( $arg->{row} );
+ _set_config( $arg->{config} ) if defined($arg->{config});
_clean_ds( %{ $arg } );
- eval "$arg->{rules}";
+ $load_row_coderef = $arg->{load_row_coderef};
+
+ no strict 'subs';
+ no warnings 'redefine';
+ eval "$arg->{rules};";
die "error evaling $arg->{rules}: $@\n" if ($@);
return _get_ds();
}
-=head2 _set_rec
+=head2 _set_ds
Set current record hash
- _set_rec( $rec );
+ _set_ds( $rec );
=cut
my $rec;
-sub _set_rec {
+sub _set_ds {
$rec = shift or die "no record hash";
}
+=head2 _set_config
+
+Set current config hash
+
+ _set_config( $config );
+
+Magic keys are:
+
+=over 4
+
+=item _
+
+Code of current database
+
+=item _mfn
+
+Current MFN
+
+=back
+
+=cut
+
+my $config;
+
+sub _set_config {
+ $config = shift;
+}
+
=head2 _get_ds
Return hash formatted as data structure
=cut
-my ($out,$marc_record, $marc_encoding, $marc_repeatable_subfield, $marc_indicators);
+my ($out, $marc_record, $marc_encoding, $marc_repeatable_subfield, $marc_indicators, $marc_leader);
+my ($marc_record_offset, $marc_fetch_offset) = (0, 0);
sub _get_ds {
+#warn "## out = ",dump($out);
return $out;
}
sub _clean_ds {
my $a = {@_};
- ($out,$marc_record, $marc_encoding, $marc_repeatable_subfield, $marc_indicators) = ();
+ ($out,$marc_record, $marc_encoding, $marc_repeatable_subfield, $marc_indicators, $marc_leader) = ();
+ ($marc_record_offset, $marc_fetch_offset) = (0,0);
$marc_encoding = $a->{marc_encoding};
}
$lookup = shift;
}
+=head2 _get_lookup
+
+Get current lookup hash
+
+ my $lookup = _get_lookup();
+
+=cut
+
+sub _get_lookup {
+ return $lookup;
+}
+
+=head2 _set_load_row
+
+Setup code reference which will return L<data_structure> from
+L<WebPAC::Store>
+
+ _set_load_row(sub {
+ my ($database,$input,$mfn) = @_;
+ $store->load_row( database => $database, input => $input, id => $mfn );
+ });
+
+=cut
+
+sub _set_load_row {
+ my $coderef = shift;
+ confess "argument isn't CODE" unless ref($coderef) eq 'CODE';
+
+ $load_row_coderef = $coderef;
+}
+
=head2 _get_marc_fields
Get all fields defined by calls to C<marc>
which might not be what you have in mind. If you need repeatable subfield,
define it using C<marc_repeatable_subfield> like this:
-....
+ marc_repeatable_subfield('900','a');
+ marc('900','a', rec('200','a') );
+ marc('900','b', rec('200','b') );
+ marc('900','c', rec('200','c') );
+
+will create:
+
+ 900a-1 900a-2 900a-3 900b-1 900c-1
+ 900b-2
+
+There is also support for returning next or specific using:
+
+ while (my $mf = WebPAC::Normalize:_get_marc_fields( fetch_next => 1 ) ) {
+ # do something with $mf
+ }
+
+will always return fields from next MARC record or
+
+ my $mf = WebPAC::Normalize::_get_marc_fields( offset => 42 );
+
+will return 42th copy record (if it exists).
=cut
+my $fetch_pos;
+
sub _get_marc_fields {
- return if (! $marc_record || ref($marc_record) ne 'ARRAY' || $#{ $marc_record } < 0);
+ my $arg = {@_};
+ warn "### _get_marc_fields arg: ", dump($arg), $/ if ($debug > 2);
+ $fetch_pos = $marc_fetch_offset;
+ if ($arg->{offset}) {
+ $fetch_pos = $arg->{offset};
+ } elsif($arg->{fetch_next}) {
+ $marc_fetch_offset++;
+ }
+
+ return if (! $marc_record || ref($marc_record) ne 'ARRAY');
+
+ warn "### full marc_record = ", dump( @{ $marc_record }), $/ if ($debug > 2);
+
+ my $marc_rec = $marc_record->[ $fetch_pos ];
+
+ warn "## _get_marc_fields (at offset: $fetch_pos) -- marc_record = ", dump( @$marc_rec ), $/ if ($debug > 1);
+
+ return if (! $marc_rec || ref($marc_rec) ne 'ARRAY' || $#{ $marc_rec } < 0);
# first, sort all existing fields
# XXX might not be needed, but modern perl might randomize elements in hash
my @sorted_marc_record = sort {
- $a->[0] . $a->[3] cmp $b->[0] . $b->[3]
- } @{ $marc_record };
+ $a->[0] . ( $a->[3] || '' ) cmp $b->[0] . ( $b->[3] || '')
+ } @{ $marc_rec };
- @sorted_marc_record = @{ $marc_record }; ### FIXME disable sorting
+ @sorted_marc_record = @{ $marc_rec }; ### FIXME disable sorting
# output marc fields
my @m;
# count unique field-subfields (used for offset when walking to next subfield)
my $u;
- map { $u->{ $_->[0] . $_->[3] }++ } @sorted_marc_record;
+ map { $u->{ $_->[0] . ( $_->[3] || '') }++ } @sorted_marc_record;
if ($debug) {
- warn "## marc_repeatable_subfield ", dump( $marc_repeatable_subfield ), $/;
- warn "## marc_record ", dump( $marc_record ), $/;
- warn "## sorted_marc_record ", dump( \@sorted_marc_record ), $/;
- warn "## subfield count ", dump( $u ), $/;
+ warn "## marc_repeatable_subfield = ", dump( $marc_repeatable_subfield ), $/ if ( $marc_repeatable_subfield );
+ warn "## marc_record[$fetch_pos] = ", dump( $marc_rec ), $/;
+ warn "## sorted_marc_record = ", dump( \@sorted_marc_record ), $/;
+ warn "## subfield count = ", dump( $u ), $/;
}
my $len = $#sorted_marc_record;
# mark it visited
$visited->{$i}++;
- my $row = $sorted_marc_record[$i];
+ my $row = dclone( $sorted_marc_record[$i] );
# field and subfield which is key for
# marc_repeatable_subfield and u
- my $fsf = $row->[0] . $row->[3];
+ my $fsf = $row->[0] . ( $row->[3] || '' );
if ($debug > 1) {
warn "## saved/3 ", dump( $field ),$/ if ($debug);
}
- return @m;
+ return \@m;
+}
+
+=head2 _get_marc_leader
+
+Return leader from currently fetched record by L</_get_marc_fields>
+
+ print WebPAC::Normalize::_get_marc_leader();
+
+=cut
+
+sub _get_marc_leader {
+ die "no fetch_pos, did you called _get_marc_fields first?" unless ( defined( $fetch_pos ) );
+ return $marc_leader->[ $fetch_pos ];
}
=head2 _debug
Those functions generally have to first in your normalization file.
-=head2 tag
+=head2 search_display
-Define new tag for I<search> and I<display>.
+Define output for L<search> and L<display> at the same time
- tag('Title', rec('200','a') );
+ search_display('Title', rec('200','a') );
=cut
-sub tag {
- my $name = shift or die "tag needs name as first argument";
+sub search_display {
+ my $name = shift or die "search_display needs name as first argument";
my @o = grep { defined($_) && $_ ne '' } @_;
return unless (@o);
- $out->{$name}->{tag} = $name;
$out->{$name}->{search} = \@o;
$out->{$name}->{display} = \@o;
}
+=head2 tag
+
+Old name for L<search_display>, but supported
+
+=cut
+
+sub tag {
+ search_display( @_ );
+}
+
=head2 display
-Define tag just for I<display>
+Define output just for I<display>
@v = display('Title', rec('200','a') );
=cut
-sub display {
- my $name = shift or die "display needs name as first argument";
+sub _field {
+ my $type = shift or confess "need type -- BUG?";
+ my $name = shift or confess "needs name as first argument";
my @o = grep { defined($_) && $_ ne '' } @_;
return unless (@o);
- $out->{$name}->{tag} = $name;
- $out->{$name}->{display} = \@o;
+ $out->{$name}->{$type} = \@o;
}
+sub display { _field( 'display', @_ ) }
+
=head2 search
Prepare values just for I<search>
=cut
-sub search {
- my $name = shift or die "search needs name as first argument";
- my @o = grep { defined($_) && $_ ne '' } @_;
- return unless (@o);
- $out->{$name}->{tag} = $name;
- $out->{$name}->{search} = \@o;
-}
+sub search { _field( 'search', @_ ) }
+
+=head2 sorted
+
+Insert into lists which will be automatically sorted
+
+ sorted('Title', rec('200','a') );
+
+=cut
+
+sub sorted { _field( 'sorted', @_ ) }
+
=head2 marc_leader
my ($offset,$value) = @_;
if ($offset) {
- $out->{' leader'}->{ $offset } = $value;
+ $marc_leader->[ $marc_record_offset ]->{ $offset } = $value;
} else {
- return $out->{' leader'};
+
+ if (defined($marc_leader)) {
+ die "marc_leader not array = ", dump( $marc_leader ) unless (ref($marc_leader) eq 'ARRAY');
+ return $marc_leader->[ $marc_record_offset ];
+ } else {
+ return;
+ }
+ }
+}
+
+=head2 marc_fixed
+
+Create control/indentifier fields with values in fixed positions
+
+ marc_fixed('008', 00, '070402');
+ marc_fixed('008', 39, '|');
+
+Positions not specified will be filled with spaces (C<0x20>).
+
+There will be no effort to extend last specified value to full length of
+field in standard.
+
+=cut
+
+sub marc_fixed {
+ my ($f, $pos, $val) = @_;
+ die "need marc(field, position, value)" unless defined($f) && defined($pos);
+
+ confess "need val" unless defined $val;
+
+ my $update = 0;
+
+ map {
+ if ($_->[0] eq $f) {
+ my $old = $_->[1];
+ if (length($old) <= $pos) {
+ $_->[1] .= ' ' x ( $pos - length($old) ) . $val;
+ warn "## marc_fixed($f,$pos,'$val') append '$old' -> '$_->[1]'\n" if ($debug > 1);
+ } else {
+ $_->[1] = substr($old, 0, $pos) . $val . substr($old, $pos + length($val));
+ warn "## marc_fixed($f,$pos,'$val') update '$old' -> '$_->[1]'\n" if ($debug > 1);
+ }
+ $update++;
+ }
+ } @{ $marc_record->[ $marc_record_offset ] };
+
+ if (! $update) {
+ my $v = ' ' x $pos . $val;
+ push @{ $marc_record->[ $marc_record_offset ] }, [ $f, $v ];
+ warn "## marc_fixed($f,$pos,'val') created '$v'\n" if ($debug > 1);
}
}
Save value for MARC field
marc('900','a', rec('200','a') );
+ marc('001', rec('000') );
=cut
my $f = shift or die "marc needs field";
die "marc field must be numer" unless ($f =~ /^\d+$/);
- my $sf = shift or die "marc needs subfield";
+ my $sf;
+ if ($f >= 10) {
+ $sf = shift or die "marc needs subfield";
+ }
foreach (@_) {
my $v = $_; # make var read-write for Encode
next unless (defined($v) && $v !~ /^\s*$/);
- from_to($v, 'iso-8859-2', $marc_encoding) if ($marc_encoding);
my ($i1,$i2) = defined($marc_indicators->{$f}) ? @{ $marc_indicators->{$f} } : (' ',' ');
- push @{ $marc_record }, [ $f, $i1, $i2, $sf => $v ];
+ if (defined $sf) {
+ push @{ $marc_record->[ $marc_record_offset ] }, [ $f, $i1, $i2, $sf => $v ];
+ } else {
+ push @{ $marc_record->[ $marc_record_offset ] }, [ $f, $v ];
+ }
}
}
'c', rec('200','c')
);
+If you specify C<+> for subfield, value will be appended
+to previous defined subfield.
+
=cut
sub marc_compose {
my ($i1,$i2) = defined($marc_indicators->{$f}) ? @{ $marc_indicators->{$f} } : (' ',' ');
my $m = [ $f, $i1, $i2 ];
+ warn "### marc_compose input subfields = ", dump(@_),$/ if ($debug > 2);
+
+ if ($#_ % 2 != 1) {
+ die "ERROR: marc_compose",dump($f,@_)," not valid (must be even).\nDo you need to add first() or join() around some argument?\n";
+ }
+
while (@_) {
- my $sf = shift or die "marc_compose $f needs subfield";
+ my $sf = shift;
my $v = shift;
next unless (defined($v) && $v !~ /^\s*$/);
- from_to($v, 'iso-8859-2', $marc_encoding) if ($marc_encoding);
- push @$m, ( $sf, $v );
warn "## ++ marc_compose($f,$sf,$v) ", dump( $m ),$/ if ($debug > 1);
+ if ($sf ne '+') {
+ push @$m, ( $sf, $v );
+ } else {
+ $m->[ $#$m ] .= $v;
+ }
}
- warn "## marc_compose(d) ", dump( $m ),$/ if ($debug > 1);
+ warn "## marc_compose current marc = ", dump( $m ),$/ if ($debug > 1);
+
+ push @{ $marc_record->[ $marc_record_offset ] }, $m if ($#{$m} > 2);
+}
+
+=head2 marc_duplicate
+
+Generate copy of current MARC record and continue working on copy
+
+ marc_duplicate();
+
+Copies can be accessed using C<< _get_marc_fields( fetch_next => 1 ) >> or
+C<< _get_marc_fields( offset => 42 ) >>.
+
+=cut
+
+sub marc_duplicate {
+ my $m = $marc_record->[ -1 ];
+ die "can't duplicate record which isn't defined" unless ($m);
+ push @{ $marc_record }, dclone( $m );
+ push @{ $marc_leader }, dclone( marc_leader() );
+ warn "## marc_duplicate = ", dump(@$marc_leader, @$marc_record), $/ if ($debug > 1);
+ $marc_record_offset = $#{ $marc_record };
+ warn "## marc_record_offset = $marc_record_offset", $/ if ($debug > 1);
- push @{ $marc_record }, $m if ($#{$m} > 2);
+}
+
+=head2 marc_remove
+
+Remove some field or subfield from MARC record.
+
+ marc_remove('200');
+ marc_remove('200','a');
+
+This will erase field C<200> or C<200^a> from current MARC record.
+
+ marc_remove('*');
+
+Will remove all fields in current MARC record.
+
+This is useful after calling C<marc_duplicate> or on it's own (but, you
+should probably just remove that subfield definition if you are not
+using C<marc_duplicate>).
+
+FIXME: support fields < 10.
+
+=cut
+
+sub marc_remove {
+ my ($f, $sf) = @_;
+
+ die "marc_remove needs record number" unless defined($f);
+
+ my $marc = $marc_record->[ $marc_record_offset ];
+
+ warn "### marc_remove before = ", dump( $marc ), $/ if ($debug > 2);
+
+ if ($f eq '*') {
+
+ delete( $marc_record->[ $marc_record_offset ] );
+ warn "## full marc_record = ", dump( @{ $marc_record }), $/ if ($debug > 1);
+
+ } else {
+
+ my $i = 0;
+ foreach ( 0 .. $#{ $marc } ) {
+ last unless (defined $marc->[$i]);
+ warn "#### working on ",dump( @{ $marc->[$i] }), $/ if ($debug > 3);
+ if ($marc->[$i]->[0] eq $f) {
+ if (! defined $sf) {
+ # remove whole field
+ splice @$marc, $i, 1;
+ warn "#### slice \@\$marc, $i, 1 = ",dump( @{ $marc }), $/ if ($debug > 3);
+ $i--;
+ } else {
+ foreach my $j ( 0 .. (( $#{ $marc->[$i] } - 3 ) / 2) ) {
+ my $o = ($j * 2) + 3;
+ if ($marc->[$i]->[$o] eq $sf) {
+ # remove subfield
+ splice @{$marc->[$i]}, $o, 2;
+ warn "#### slice \@{\$marc->[$i]}, $o, 2 = ", dump( @{ $marc }), $/ if ($debug > 3);
+ # is record now empty?
+ if ($#{ $marc->[$i] } == 2) {
+ splice @$marc, $i, 1;
+ warn "#### slice \@\$marc, $i, 1 = ", dump( @{ $marc }), $/ if ($debug > 3);
+ $i--;
+ };
+ }
+ }
+ }
+ }
+ $i++;
+ }
+
+ warn "### marc_remove($f", $sf ? ",$sf" : "", ") after = ", dump( $marc ), $/ if ($debug > 2);
+
+ $marc_record->[ $marc_record_offset ] = $marc;
+ }
+
+ warn "## full marc_record = ", dump( @{ $marc_record }), $/ if ($debug > 1);
+}
+
+=head2 marc_original_order
+
+Copy all subfields preserving original order to marc field.
+
+ marc_original_order( marc_field_number, original_input_field_number );
+
+Please note that field numbers are consistent with other commands (marc
+field number first), but somewhat counter-intuitive (destination and then
+source).
+
+You might want to use this command if you are just renaming subfields or
+using pre-processing modify_record in C<config.yml> and don't need any
+post-processing or want to preserve order of original subfields.
+
+
+=cut
+
+sub marc_original_order {
+
+ my ($to, $from) = @_;
+ die "marc_original_order needs from and to fields\n" unless ($from && $to);
+
+ return unless defined($rec->{$from});
+
+ my $r = $rec->{$from};
+ die "record field $from isn't array\n" unless (ref($r) eq 'ARRAY');
+
+ my ($i1,$i2) = defined($marc_indicators->{$to}) ? @{ $marc_indicators->{$to} } : (' ',' ');
+ warn "## marc_original_order($to,$from) source = ", dump( $r ),$/ if ($debug > 1);
+
+ foreach my $d (@$r) {
+
+ if (! defined($d->{subfields}) && ref($d->{subfields}) ne 'ARRAY') {
+ warn "# marc_original_order($to,$from): field $from doesn't have subfields specification\n";
+ next;
+ }
+
+ my @sfs = @{ $d->{subfields} };
+
+ die "field $from doesn't have even number of subfields specifications\n" unless($#sfs % 2 == 1);
+
+ warn "#--> d: ",dump($d), "\n#--> sfs: ",dump(@sfs),$/ if ($debug > 2);
+
+ my $m = [ $to, $i1, $i2 ];
+
+ while (my $sf = shift @sfs) {
+
+ warn "#--> sf: ",dump($sf), $/ if ($debug > 2);
+ my $offset = shift @sfs;
+ die "corrupted sufields specification for field $from\n" unless defined($offset);
+
+ my $v;
+ if (ref($d->{$sf}) eq 'ARRAY') {
+ $v = $d->{$sf}->[$offset] if (defined($d->{$sf}->[$offset]));
+ } elsif ($offset == 0) {
+ $v = $d->{$sf};
+ } else {
+ die "field $from subfield '$sf' need occurence $offset which doesn't exist", dump($d->{$sf});
+ }
+ push @$m, ( $sf, $v ) if (defined($v));
+ }
+
+ if ($#{$m} > 2) {
+ push @{ $marc_record->[ $marc_record_offset ] }, $m;
+ }
+ }
+
+ warn "## marc_record = ", dump( $marc_record ),$/ if ($debug > 1);
+}
+
+=head2 marc_template
+
+=cut
+
+sub marc_template {
+ my $args = {@_};
+ warn "## marc_template(",dump($args),")";
+
+ foreach ( qw/subfields_rename marc_template/ ) {
+# warn "ref($_) = ",ref($args->{$_});
+ die "$_ not ARRAY" if ref($args->{$_}) ne 'ARRAY';
+ }
+
+ my $r = $rec->{ $args->{from} }; # || return;
+ die "record field ", $args->{from}, " isn't array\n" unless (ref($r) eq 'ARRAY');
+
+ my @subfields_rename = @{ $args->{subfields_rename} };
+# warn "### subfields_rename [$#subfields_rename] = ",dump( @subfields_rename );
+
+ confess "need mapping in pairs for subfields_rename"
+ if $#subfields_rename % 2 != 1;
+
+ my ( $subfields_rename, $from_subfields, $to_subfields );
+ while ( my ( $from, $to ) = splice(@subfields_rename, 0, 2) ) {
+ my ( $f, $t ) = (
+ $from_subfields->{ $from }++,
+ $to_subfields->{ $to }++
+ );
+ $subfields_rename->{ $from }->[ $f ] = [ $to => $t ];
+ }
+ warn "### subfields_rename = ",dump( $subfields_rename ),$/;
+ warn "### from_subfields = ", dump( $from_subfields ),$/;
+ warn "### to_subfields = ", dump( $to_subfields ),$/;
+
+ my $fields_re = join('|', keys %$to_subfields );
+
+ my $pos_templates;
+ my $count;
+ my @marc_order;
+ my $marc_template_order;
+ my $fill_in;
+ my @marc_out;
+
+ foreach my $template ( @{ $args->{marc_template} } ) {
+ $count = {};
+ @marc_order = ();
+ sub my_count {
+ my $sf = shift;
+ my $nr = $count->{$sf}++;
+ push @marc_order, [ $sf, $nr ];
+ return $sf . $nr;
+ }
+ my $pos_template = $template;
+ $pos_template =~ s/($fields_re)/my_count($1)/ge;
+ my $count_key = dump( $count );
+ warn "### template: |$template| -> |$pos_template| count = $count_key marc_order = ",dump( @marc_order ),$/;
+ $pos_templates->{ $count_key } = $pos_template;
+ $marc_template_order->{ $pos_template } = [ @marc_order ];
+ }
+ warn "### from ",dump( $args->{marc_template} ), " created ", dump( $pos_templates ), " and ", dump( $marc_template_order );
+
+ my $m;
+
+ foreach my $r ( @{ $rec->{ $args->{from} } } ) {
+
+ my $i1 = $r->{i1} || ' ';
+ my $i2 = $r->{i2} || ' ';
+ $m = [ $args->{to}, $i1, $i2 ];
+
+ warn "### r = ",dump( $r );
+
+ my ( $new_r, $from_count, $to_count );
+ foreach my $sf ( keys %{$r} ) {
+ my $nr = $from_count->{$sf}++;
+ my $rename_to = $subfields_rename->{ $sf }; # ||
+# die "can't find subfield rename for $sf/$nr in ", dump( $subfields_rename );
+ warn "### rename $sf/$nr to ", dump( $rename_to->[$nr] ), $/;
+ my ( $to_sf, $to_nr ) = @{ $rename_to->[$nr] };
+ $new_r->{ $to_sf }->[ $to_nr ] = [ $sf => $nr ];
+
+ $to_count->{ $to_sf }++;
+ }
+
+ warn "### new_r = ",dump( $new_r );
+
+ my $from_count_key = dump( $to_count );
+
+ warn "### from_count = ",dump( $from_count ), $/;
+ warn "### to_count = ",dump( $to_count ), $/;
+
+ my $template = $pos_templates->{ $from_count_key } ||
+ die "I don't have template for:\n$from_count_key\n## available templates\n", dump( $pos_templates );
+
+ warn "### selected template: |$template|\n";
+
+ $fill_in = {};
+
+ foreach my $sf ( split(/\|/, $template ) ) {
+ sub fill_in {
+ my ( $r, $sf, $nr ) = @_;
+ my ( $from_sf, $from_nr ) = @{ $new_r->{ $sf }->[ $nr ] };
+ my $v = $r->{ $from_sf }; # || die "no $from_sf/$from_nr";
+ warn "#### fill_in( $sf, $nr ) = $from_sf/$from_nr >>>> ",dump( $v ), $/;
+ if ( ref( $v ) eq 'ARRAY' ) {
+ $fill_in->{$sf}->[$nr] = $v->[$from_nr];
+ return $v->[$from_nr];
+ } elsif ( $from_nr == 0 ) {
+ $fill_in->{$sf}->[$nr] = $v;
+ return $v;
+ } else {
+ die "requested subfield $from_sf/$from_nr but it's ",dump( $v );
+ }
+ }
+ warn "#### $sf <<<< $fields_re\n";
+ $sf =~ s/($fields_re)(\d+)/fill_in($r,$1,$2)/ge;
+ warn "#### >>>> $sf with fill_in = ",dump( $fill_in ),$/;
+ }
+
+ warn "## template: |$template|\n## marc_template_order = ",dump( $marc_template_order );
+
+ foreach my $sf ( @{ $marc_template_order->{$template} } ) {
+ my ( $sf, $nr ) = @$sf;
+ my $v = $fill_in->{$sf}->[$nr] || die "can't find fill_in $sf/$nr";
+ warn "++ $sf/$nr |$v|\n";
+ push @$m, [ $sf, $v ];
+ }
+
+ warn "#### >>>> created marc: ", dump( $m );
+
+ push @marc_out, $m;
+ }
+
+ warn "### marc_template produced: ",dump( @marc_out );
+ return @marc_out;
+}
+
+=head2 marc_count
+
+Return number of MARC records created using L</marc_duplicate>.
+
+ print "created ", marc_count(), " records";
+
+=cut
+
+sub marc_count {
+ return $#{ $marc_record };
}
This function should be used inside functions to create C<data_structure> described
above.
+=head2 _pack_subfields_hash
+
+ @subfields = _pack_subfields_hash( $h );
+ $subfields = _pack_subfields_hash( $h, 1 );
+
+Return each subfield value in array or pack them all together and return scalar
+with subfields (denoted by C<^>) and values.
+
+=cut
+
+sub _pack_subfields_hash {
+
+ warn "## _pack_subfields_hash( ",dump(@_), " )\n" if ($debug > 1);
+
+ my ($h,$include_subfields) = @_;
+
+ # sanity and ease of use
+ return $h if (ref($h) ne 'HASH');
+
+ if ( defined($h->{subfields}) ) {
+ my $sfs = delete $h->{subfields} || die "no subfields?";
+ my @out;
+ while (@$sfs) {
+ my $sf = shift @$sfs;
+ push @out, '^' . $sf if ($include_subfields);
+ my $o = shift @$sfs;
+ if ($o == 0 && ref( $h->{$sf} ) ne 'ARRAY' ) {
+ # single element subfields are not arrays
+#warn "====> $sf $o / $#$sfs ", dump( $sfs, $h->{$sf} ), "\n";
+
+ push @out, $h->{$sf};
+ } else {
+#warn "====> $sf $o / $#$sfs ", dump( $sfs, $h->{$sf} ), "\n";
+ push @out, $h->{$sf}->[$o];
+ }
+ }
+ if ($include_subfields) {
+ return join('', @out);
+ } else {
+ return @out;
+ }
+ } else {
+ if ($include_subfields) {
+ my $out = '';
+ foreach my $sf (sort keys %$h) {
+ if (ref($h->{$sf}) eq 'ARRAY') {
+ $out .= '^' . $sf . join('^' . $sf, @{ $h->{$sf} });
+ } else {
+ $out .= '^' . $sf . $h->{$sf};
+ }
+ }
+ return $out;
+ } else {
+ # FIXME this should probably be in alphabetical order instead of hash order
+ values %{$h};
+ }
+ }
+}
+
=head2 rec1
Return all values in some field
sub rec1 {
my $f = shift;
+ warn "rec1($f) = ", dump( $rec->{$f} ), $/ if ($debug > 1);
return unless (defined($rec) && defined($rec->{$f}));
+ warn "rec1($f) = ", dump( $rec->{$f} ), $/ if ($debug > 1);
if (ref($rec->{$f}) eq 'ARRAY') {
- return map {
- if (ref($_) eq 'HASH') {
- values %{$_};
+ my @out;
+ foreach my $h ( @{ $rec->{$f} } ) {
+ if (ref($h) eq 'HASH') {
+ push @out, ( _pack_subfields_hash( $h ) );
} else {
- $_;
+ push @out, $h;
}
- } @{ $rec->{$f} };
+ }
+ return @out;
} elsif( defined($rec->{$f}) ) {
return $rec->{$f};
}
my $f = shift;
return unless (defined($rec && $rec->{$f}));
my $sf = shift;
- return map { $_->{$sf} } grep { ref($_) eq 'HASH' && $_->{$sf} } @{ $rec->{$f} };
+ warn "rec2($f,$sf) = ", dump( $rec->{$f} ), $/ if ($debug > 1);
+ return map {
+ if (ref($_->{$sf}) eq 'ARRAY') {
+ @{ $_->{$sf} };
+ } else {
+ $_->{$sf};
+ }
+ } grep { ref($_) eq 'HASH' && $_->{$sf} } @{ $rec->{$f} };
}
=head2 rec
@v = rec('200')
@v = rec('200','a')
+If rec() returns just single value, it will
+return scalar, not array.
+
=cut
sub rec {
+ my @out;
if ($#_ == 0) {
- return rec1(@_);
+ @out = rec1(@_);
} elsif ($#_ == 1) {
- return rec2(@_);
+ @out = rec2(@_);
}
+ if ($#out == 0 && ! wantarray) {
+ return $out[0];
+ } elsif (@out) {
+ return @out;
+ } else {
+ return '';
+ }
+}
+
+=head2 frec
+
+Returns first value from field
+
+ $v = frec('200');
+ $v = frec('200','a');
+
+=cut
+
+sub frec {
+ my @out = rec(@_);
+ warn "rec(",dump(@_),") has more than one return value, ignoring\n" if $#out > 0;
+ return shift @out;
}
=head2 regex
=cut
sub prefix {
- my $p = shift or die "prefix needs string as first argument";
+ my $p = shift;
+ return @_ unless defined( $p );
return map { $p . $_ } grep { defined($_) } @_;
}
=cut
sub suffix {
- my $s = shift or die "suffix needs string as first argument";
+ my $s = shift;
+ return @_ unless defined( $s );
return map { $_ . $s } grep { defined($_) } @_;
}
=cut
sub surround {
- my $p = shift or die "surround need prefix as first argument";
- my $s = shift or die "surround needs suffix as second argument";
+ my $p = shift;
+ my $s = shift;
+ $p = '' unless defined( $p );
+ $s = '' unless defined( $s );
return map { $p . $_ . $s } grep { defined($_) } @_;
}
Consult lookup hashes for some value
- @v = lookup( $v );
- @v = lookup( @v );
+ @v = lookup(
+ sub {
+ 'ffkk/peri/mfn'.rec('000')
+ },
+ 'ffkk','peri','200-a-200-e',
+ sub {
+ first(rec(200,'a')).' '.first(rec('200','e'))
+ }
+ );
+
+Code like above will be B<automatically generated> using L<WebPAC::Parse> from
+normal lookup definition in C<conf/lookup/something.pl> which looks like:
+
+ lookup(
+ # which results to return from record recorded in lookup
+ sub { 'ffkk/peri/mfn' . rec('000') },
+ # from which database and input
+ 'ffkk','peri',
+ # such that following values match
+ sub { first(rec(200,'a')) . ' ' . first(rec('200','e')) },
+ # if this part is missing, we will try to match same fields
+ # from lookup record and current one, or you can override
+ # which records to use from current record using
+ sub { rec('900','x') . ' ' . rec('900','y') },
+ )
+
+You can think about this lookup as SQL (if that helps):
+
+ select
+ sub { what }
+ from
+ database, input
+ where
+ sub { filter from lookuped record }
+ having
+ sub { optional filter on current record }
+
+Easy as pie, right?
=cut
sub lookup {
- my $k = shift or return;
- return unless (defined($lookup->{$k}));
- if (ref($lookup->{$k}) eq 'ARRAY') {
- return @{ $lookup->{$k} };
+ my ($what, $database, $input, $key, $having) = @_;
+
+ confess "lookup needs 5 arguments: what, database, input, key, having\n" unless ($#_ == 4);
+
+ warn "## lookup ($database, $input, $key)", $/ if ($debug > 1);
+ return unless (defined($lookup->{$database}->{$input}->{$key}));
+
+ confess "lookup really need load_row_coderef added to data_structure\n" unless ($load_row_coderef);
+
+ my $mfns;
+ my @having = $having->();
+
+ warn "## having = ", dump( @having ) if ($debug > 2);
+
+ foreach my $h ( @having ) {
+ if (defined($lookup->{$database}->{$input}->{$key}->{$h})) {
+ warn "lookup for $database/$input/$key/$h return ",dump($lookup->{$database}->{$input}->{$key}->{$h}),"\n" if ($debug);
+ $mfns->{$_}++ foreach keys %{ $lookup->{$database}->{$input}->{$key}->{$h} };
+ }
+ }
+
+ return unless ($mfns);
+
+ my @mfns = sort keys %$mfns;
+
+ warn "# lookup loading $database/$input/$key mfn ", join(",",@mfns)," having ",dump(@having),"\n" if ($debug);
+
+ my $old_rec = $rec;
+ my @out;
+
+ foreach my $mfn (@mfns) {
+ $rec = $load_row_coderef->( $database, $input, $mfn );
+
+ warn "got $database/$input/$mfn = ", dump($rec), $/ if ($debug);
+
+ my @vals = $what->();
+
+ push @out, ( @vals );
+
+ warn "lookup for mfn $mfn returned ", dump(@vals), $/ if ($debug);
+ }
+
+# if (ref($lookup->{$k}) eq 'ARRAY') {
+# return @{ $lookup->{$k} };
+# } else {
+# return $lookup->{$k};
+# }
+
+ $rec = $old_rec;
+
+ warn "## lookup returns = ", dump(@out), $/ if ($debug);
+
+ if ($#out == 0) {
+ return $out[0];
+ } else {
+ return @out;
+ }
+}
+
+=head2 save_into_lookup
+
+Save value into lookup. It associates current database, input
+and specific keys with one or more values which will be
+associated over MFN.
+
+MFN will be extracted from first occurence current of field 000
+in current record, or if it doesn't exist from L<_set_config> C<_mfn>.
+
+ my $nr = save_into_lookup($database,$input,$key,sub {
+ # code which produce one or more values
+ });
+
+It returns number of items saved.
+
+This function shouldn't be called directly, it's called from code created by
+L<WebPAC::Parser>.
+
+=cut
+
+sub save_into_lookup {
+ my ($database,$input,$key,$coderef) = @_;
+ die "save_into_lookup needs database" unless defined($database);
+ die "save_into_lookup needs input" unless defined($input);
+ die "save_into_lookup needs key" unless defined($key);
+ die "save_into_lookup needs CODE" unless ( defined($coderef) && ref($coderef) eq 'CODE' );
+
+ warn "## save_into_lookup rec = ", dump($rec), " config = ", dump($config), $/ if ($debug > 2);
+
+ my $mfn =
+ defined($rec->{'000'}->[0]) ? $rec->{'000'}->[0] :
+ defined($config->{_mfn}) ? $config->{_mfn} :
+ die "mfn not defined or zero";
+
+ my $nr = 0;
+
+ foreach my $v ( $coderef->() ) {
+ $lookup->{$database}->{$input}->{$key}->{$v}->{$mfn}++;
+ warn "# saved lookup $database/$input/$key [$v] $mfn\n" if ($debug > 1);
+ $nr++;
+ }
+
+ return $nr;
+}
+
+=head2 config
+
+Consult config values stored in C<config.yml>
+
+ # return database code (key under databases in yaml)
+ $database_code = config(); # use _ from hash
+ $database_name = config('name');
+ $database_input_name = config('input name');
+
+Up to three levels are supported.
+
+=cut
+
+sub config {
+ return unless ($config);
+
+ my $p = shift;
+
+ $p ||= '';
+
+ my $v;
+
+ warn "### getting config($p)\n" if ($debug > 1);
+
+ my @p = split(/\s+/,$p);
+ if ($#p < 0) {
+ $v = $config->{ '_' }; # special, database code
} else {
- return $lookup->{$k};
+
+ my $c = dclone( $config );
+
+ foreach my $k (@p) {
+ warn "### k: $k c = ",dump($c),$/ if ($debug > 1);
+ if (ref($c) eq 'ARRAY') {
+ $c = shift @$c;
+ warn "config($p) taking first occurence of '$k', probably not what you wanted!\n";
+ last;
+ }
+
+ if (! defined($c->{$k}) ) {
+ $c = undef;
+ last;
+ } else {
+ $c = $c->{$k};
+ }
+ }
+ $v = $c if ($c);
+
}
+
+ warn "## config( '$p' ) = ",dump( $v ),$/ if ($v && $debug);
+ warn "config( '$p' ) is empty\n" if (! $v);
+
+ return $v;
+}
+
+=head2 id
+
+Returns unique id of this record
+
+ $id = id();
+
+Returns C<42/2> for 2nd occurence of MFN 42.
+
+=cut
+
+sub id {
+ my $mfn = $config->{_mfn} || die "no _mfn in config data";
+ return $mfn . $#{$marc_record} ? $#{$marc_record} + 1 : '';
}
=head2 join_with
sub join_with {
my $d = shift;
- return join($d, grep { defined($_) && $_ ne '' } @_);
+ warn "### join_with('$d',",dump(@_),")\n" if ($debug > 2);
+ my $v = join($d, grep { defined($_) && $_ ne '' } @_);
+ return '' unless defined($v);
+ return $v;
}
=head2 split_rec_on
my $v = shift @r;
warn "### first rec($fld,$sf) = ",dump($v),$/ if ($debug > 2);
- return '' if( ! defined($v) || $v =~ /^\s*$/);
+ return '' if ( ! defined($v) || $v =~ /^\s*$/);
my @s = split( $regex, $v );
warn "## split_rec_on($fld,$sf,$regex,$part) = ",dump(@s),$/ if ($debug > 1);
if ($part && $part > 0) {
return $s[ $part - 1 ];
} else {
- return [ @s ];
+ return @s;
}
}
+my $hash;
+
+=head2 set
+
+ set( key => 'value' );
+
+=cut
+
+sub set {
+ my ($k,$v) = @_;
+ warn "## set ( $k => ", dump($v), " )", $/ if ( $debug );
+ $hash->{$k} = $v;
+};
+
+=head2 get
+
+ get( 'key' );
+
+=cut
+
+sub get {
+ my $k = shift || return;
+ my $v = $hash->{$k};
+ warn "## get $k = ", dump( $v ), $/ if ( $debug );
+ return $v;
+}
+
+=head2 count
+
+ if ( count( @result ) == 1 ) {
+ # do something if only 1 result is there
+ }
+
+=cut
+
+sub count {
+ warn "## count ",dump(@_),$/ if ( $debug );
+ return @_ . '';
+}
+
# END
1;
projects / webpac2 / blobdiff