--- trunk/lib/WebPAC/Normalize.pm 2006/12/10 12:45:11 786 +++ trunk/lib/WebPAC/Normalize.pm 2007/11/19 16:33:09 1048 @@ -1,19 +1,16 @@ 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 - tag search display - marc marc_indicators marc_repeatable_subfield - marc_compose marc_leader - marc_duplicate marc_remove - marc_original_order + search_display search display sorted rec1 rec2 rec + frec frec_eq frec_ne regex prefix suffix surround first lookup join_with save_into_lookup @@ -21,6 +18,8 @@ split_rec_on get set + count + /; use warnings; @@ -28,24 +27,32 @@ #use base qw/WebPAC::Common/; use Data::Dump qw/dump/; -use Storable qw/dclone/; use Carp qw/confess/; # debugging warn(s) my $debug = 0; +_debug( $debug ); + +# FIXME +use WebPAC::Normalize::ISBN; +push @EXPORT, ( 'isbn_10', 'isbn_13' ); +use WebPAC::Normalize::MARC; +push @EXPORT, ( qw/ + marc marc_indicators marc_repeatable_subfield + marc_compose marc_leader marc_fixed + marc_duplicate marc_remove marc_count + marc_original_order + marc_template +/); =head1 NAME WebPAC::Normalize - describe normalisaton rules using sets -=head1 VERSION - -Version 0.25 - =cut -our $VERSION = '0.25'; +our $VERSION = '0.35'; =head1 SYNOPSIS @@ -58,7 +65,7 @@ C. Normalisation can generate multiple output normalized data. For now, supported output -types (on the left side of definition) are: C, C, C and +types (on the left side of definition) are: C, C, C and C. =head1 FUNCTIONS @@ -77,7 +84,7 @@ marc_encoding => 'utf-8', config => $config, load_row_coderef => sub { - my ($database,$input,$mfn) = shift; + my ($database,$input,$mfn) = @_; $store->load_row( database => $database, input => $input, id => $mfn ); }, ); @@ -103,33 +110,43 @@ die "need row argument" unless ($arg->{row}); die "need normalisation argument" unless ($arg->{rules}); - no strict 'subs'; _set_lookup( $arg->{lookup} ) if defined($arg->{lookup}); - _set_rec( $arg->{row} ); + _set_ds( $arg->{row} ); _set_config( $arg->{config} ) if defined($arg->{config}); _clean_ds( %{ $arg } ); $load_row_coderef = $arg->{load_row_coderef}; - eval "$arg->{rules}"; + 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"; + $WebPAC::Normalize::MARC::rec = $rec; } +=head2 + + my $rec = _get_rec(); + +=cut + +sub _get_rec { $rec }; + =head2 _set_config Set current config hash @@ -166,10 +183,10 @@ =cut -my ($out, $marc_record, $marc_encoding, $marc_repeatable_subfield, $marc_indicators, $leader); -my ($marc_record_offset, $marc_fetch_offset) = (0, 0); +my $out; sub _get_ds { +#warn "## out = ",dump($out); return $out; } @@ -183,9 +200,8 @@ sub _clean_ds { my $a = {@_}; - ($out,$marc_record, $marc_encoding, $marc_repeatable_subfield, $marc_indicators, $leader) = (); - ($marc_record_offset, $marc_fetch_offset) = (0,0); - $marc_encoding = $a->{marc_encoding}; + $out = undef; + WebPAC::Normalize::MARC::_clean(); } =head2 _set_lookup @@ -233,180 +249,6 @@ $load_row_coderef = $coderef; } -=head2 _get_marc_fields - -Get all fields defined by calls to C - - $marc->add_fields( WebPAC::Normalize:_get_marc_fields() ); - -We are using I which detect repeatable fields only from -sequence of field/subfield data generated by normalization. - -Repeatable field is created when there is second occurence of same subfield or -if any of indicators are different. - -This is sane for most cases. Something like: - - 900a-1 900b-1 900c-1 - 900a-2 900b-2 - 900a-3 - -will be created from any combination of: - - 900a-1 900a-2 900a-3 900b-1 900b-2 900c-1 - -and following rules: - - marc('900','a', rec('200','a') ); - marc('900','b', rec('200','b') ); - marc('900','c', rec('200','c') ); - -which might not be what you have in mind. If you need repeatable subfield, -define it using C 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 - -sub _get_marc_fields { - - my $arg = {@_}; - warn "### _get_marc_fields arg: ", dump($arg), $/ if ($debug > 2); - my $offset = $marc_fetch_offset; - if ($arg->{offset}) { - $offset = $arg->{offset}; - } elsif($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->[ $offset ]; - - warn "## _get_marc_fields (at offset: $offset) -- marc_record = ", dump( @$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_rec }; - - @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] . ( $_->[3] || '') }++ } @sorted_marc_record; - - if ($debug) { - warn "## marc_repeatable_subfield = ", dump( $marc_repeatable_subfield ), $/ if ( $marc_repeatable_subfield ); - warn "## marc_record[$offset] = ", dump( $marc_rec ), $/; - warn "## sorted_marc_record = ", dump( \@sorted_marc_record ), $/; - warn "## subfield count = ", dump( $u ), $/; - } - - my $len = $#sorted_marc_record; - my $visited; - my $i = 0; - my $field; - - foreach ( 0 .. $len ) { - - # find next element which isn't visited - while ($visited->{$i}) { - $i = ($i + 1) % ($len + 1); - } - - # mark it visited - $visited->{$i}++; - - my $row = dclone( $sorted_marc_record[$i] ); - - # field and subfield which is key for - # marc_repeatable_subfield and u - my $fsf = $row->[0] . ( $row->[3] || '' ); - - if ($debug > 1) { - - print "### field so far [", $#$field, "] : ", dump( $field ), " ", $field ? 'T' : 'F', $/; - print "### this [$i]: ", dump( $row ),$/; - print "### sf: ", $row->[3], " vs ", $field->[3], - $marc_repeatable_subfield->{ $row->[0] . $row->[3] } ? ' (repeatable)' : '', $/, - if ($#$field >= 0); - - } - - # if field exists - if ( $#$field >= 0 ) { - if ( - $row->[0] ne $field->[0] || # field - $row->[1] ne $field->[1] || # i1 - $row->[2] ne $field->[2] # i2 - ) { - push @m, $field; - warn "## saved/1 ", dump( $field ),$/ if ($debug); - $field = $row; - - } elsif ( - ( $row->[3] lt $field->[-2] ) # subfield which is not next (e.g. a after c) - || - ( $row->[3] eq $field->[-2] && # same subfield, but not repeatable - ! $marc_repeatable_subfield->{ $fsf } - ) - ) { - push @m, $field; - warn "## saved/2 ", dump( $field ),$/ if ($debug); - $field = $row; - - } else { - # append new subfields to existing field - push @$field, ( $row->[3], $row->[4] ); - } - } else { - # insert first field - $field = $row; - } - - if (! $marc_repeatable_subfield->{ $fsf }) { - # make step to next subfield - $i = ($i + $u->{ $fsf } ) % ($len + 1); - } - } - - if ($#$field >= 0) { - push @m, $field; - warn "## saved/3 ", dump( $field ),$/ if ($debug); - } - - return \@m; -} - =head2 _debug Change level of debug warnings @@ -420,358 +262,78 @@ return $debug unless defined($l); warn "debug level $l",$/ if ($l > 0); $debug = $l; + $WebPAC::Normalize::MARC::debug = $debug; } =head1 Functions to create C Those functions generally have to first in your normalization file. -=head2 tag +=head2 search_display -Define new tag for I and I. +Define output for L and L 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 display +=head2 tag -Define tag just for I - - @v = display('Title', rec('200','a') ); +Old name for L, but supported =cut -sub display { - my $name = shift or die "display needs name as first argument"; - my @o = grep { defined($_) && $_ ne '' } @_; - return unless (@o); - $out->{$name}->{tag} = $name; - $out->{$name}->{display} = \@o; +sub tag { + search_display( @_ ); } -=head2 search +=head2 display -Prepare values just for I +Define output just for I - @v = search('Title', rec('200','a') ); + @v = display('Title', rec('200','a') ); =cut -sub search { - my $name = shift or die "search 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}->{search} = \@o; -} - -=head2 marc_leader - -Setup fields within MARC leader or get leader - - marc_leader('05','c'); - my $leader = marc_leader(); - -=cut - -sub marc_leader { - my ($offset,$value) = @_; - - if ($offset) { - $leader->{ $offset } = $value; - } else { - return $leader; - } -} - -=head2 marc - -Save value for MARC field - - marc('900','a', rec('200','a') ); - marc('001', rec('000') ); - -=cut - -sub marc { - my $f = shift or die "marc needs field"; - die "marc field must be numer" unless ($f =~ /^\d+$/); - - 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) && $v !~ /^\s*$/); - my ($i1,$i2) = defined($marc_indicators->{$f}) ? @{ $marc_indicators->{$f} } : (' ',' '); - if (defined $sf) { - push @{ $marc_record->[ $marc_record_offset ] }, [ $f, $i1, $i2, $sf => $v ]; - } else { - push @{ $marc_record->[ $marc_record_offset ] }, [ $f, $v ]; - } - } -} - -=head2 marc_repeatable_subfield - -Save values for MARC repetable subfield - - marc_repeatable_subfield('910', 'z', rec('909') ); - -=cut - -sub marc_repeatable_subfield { - my ($f,$sf) = @_; - die "marc_repeatable_subfield need field and subfield!\n" unless ($f && $sf); - $marc_repeatable_subfield->{ $f . $sf }++; - marc(@_); -} - -=head2 marc_indicators - -Set both indicators for MARC field - - marc_indicators('900', ' ', 1); - -Any indicator value other than C<0-9> will be treated as undefined. - -=cut - -sub marc_indicators { - my $f = shift || die "marc_indicators need field!\n"; - my ($i1,$i2) = @_; - die "marc_indicators($f, ...) need i1!\n" unless(defined($i1)); - die "marc_indicators($f, $i1, ...) need i2!\n" unless(defined($i2)); - - $i1 = ' ' if ($i1 !~ /^\d$/); - $i2 = ' ' if ($i2 !~ /^\d$/); - @{ $marc_indicators->{$f} } = ($i1,$i2); -} - -=head2 marc_compose - -Save values for each MARC subfield explicitly - - marc_compose('900', - 'a', rec('200','a') - 'b', rec('201','a') - 'a', rec('200','b') - 'c', rec('200','c') - ); - -If you specify C<+> for subfield, value will be appended -to previous defined subfield. - -=cut - -sub marc_compose { - my $f = shift or die "marc_compose needs field"; - die "marc_compose field must be numer" unless ($f =~ /^\d+$/); - - my ($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; - my $v = shift; - - next unless (defined($v) && $v !~ /^\s*$/); - 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 current marc = ", dump( $m ),$/ if ($debug > 1); - - push @{ $marc_record->[ $marc_record_offset ] }, $m if ($#{$m} > 2); + $out->{$name}->{$type} = \@o; } -=head2 marc_duplicate +sub display { _field( 'display', @_ ) } -Generate copy of current MARC record and continue working on copy +=head2 search - marc_duplicate(); +Prepare values just for I -Copies can be accessed using C<< _get_marc_fields( fetch_next => 1 ) >> or -C<< _get_marc_fields( offset => 42 ) >>. + @v = search('Title', rec('200','a') ); =cut -sub marc_duplicate { - my $m = $marc_record->[ -1 ]; - die "can't duplicate record which isn't defined" unless ($m); - push @{ $marc_record }, dclone( $m ); - warn "## marc_duplicate = ", dump(@$marc_record), $/ if ($debug > 1); - $marc_record_offset = $#{ $marc_record }; - warn "## marc_record_offset = $marc_record_offset", $/ if ($debug > 1); -} - -=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. +sub search { _field( 'search', @_ ) } - marc_remove('*'); +=head2 sorted -Will remove all fields in current MARC record. +Insert into lists which will be automatically sorted -This is useful after calling C or on it's own (but, you -should probably just remove that subfield definition if you are not -using C). - -FIXME: support fields < 10. + sorted('Title', rec('200','a') ); =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 ] ); - - } 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. +sub sorted { _field( 'sorted', @_ ) } - 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 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 ($debug > 1); -} =head1 Functions to extract data from input @@ -795,6 +357,9 @@ my ($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; @@ -884,7 +449,7 @@ } else { $_->{$sf}; } - } grep { ref($_) eq 'HASH' && $_->{$sf} } @{ $rec->{$f} }; + } grep { ref($_) eq 'HASH' && defined $_->{$sf} } @{ $rec->{$f} }; } =head2 rec @@ -915,6 +480,54 @@ } } +=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 frec_eq + +=head2 frec_ne + +Check if first values from two fields are same or different + + if ( frec_eq( 900 => 'a', 910 => 'c' ) ) { + # values are same + } else { + # values are different + } + +Strictly speaking C and C wouldn't be needed if you +could write something like: + + if ( frec( '900','a' ) eq frec( '910','c' ) ) { + # yada tada + } + +but you can't since our parser L will remove all whitespaces +in order to parse text and create invalid function C. + +=cut + +sub frec_eq { + my ( $f1,$sf1, $f2, $sf2 ) = @_; + return (rec( $f1, $sf1 ))[0] eq (rec( $f2, $sf2 ))[0]; +} + +sub frec_ne { + return ! frec_eq( @_ ); +} + =head2 regex Apply regex to some or all values @@ -944,7 +557,8 @@ =cut sub prefix { - my $p = shift or return; + my $p = shift; + return @_ unless defined( $p ); return map { $p . $_ } grep { defined($_) } @_; } @@ -957,7 +571,8 @@ =cut sub suffix { - my $s = shift or die "suffix needs string as first argument"; + my $s = shift; + return @_ unless defined( $s ); return map { $_ . $s } grep { defined($_) } @_; } @@ -970,8 +585,10 @@ =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 . $_ . $s } grep { defined($_) } @_; } @@ -1146,7 +763,6 @@ $database_code = config(); # use _ from hash $database_name = config('name'); $database_input_name = config('input name'); - $tag = config('input normalize tag'); Up to three levels are supported. @@ -1207,7 +823,7 @@ sub id { my $mfn = $config->{_mfn} || die "no _mfn in config data"; - return $mfn . $#{$marc_record} ? $#{$marc_record} + 1 : ''; + return $mfn . ( WebPAC::Normalize::MARC::_created_marc_records() || '' ); } =head2 join_with @@ -1272,7 +888,7 @@ sub set { my ($k,$v) = @_; - warn "## set ( $k => ", dump($v), " )", $/; + warn "## set ( $k => ", dump($v), " )", $/ if ( $debug ); $hash->{$k} = $v; }; @@ -1285,10 +901,22 @@ sub get { my $k = shift || return; my $v = $hash->{$k}; - warn "## get $k = ", dump( $v ), $/; + 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;