--- trunk/lib/WebPAC/Normalize.pm 2005/12/20 23:31:37 312 +++ trunk/lib/WebPAC/Normalize.pm 2007/09/06 19:12:15 889 @@ -1,712 +1,1392 @@ package WebPAC::Normalize; +use Exporter 'import'; +@EXPORT = qw/ + _set_rec _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_fixed + marc_duplicate marc_remove marc_count + marc_original_order + + rec1 rec2 rec + regex prefix suffix surround + first lookup join_with + save_into_lookup + + split_rec_on + + get set + count +/; use warnings; use strict; -use base 'WebPAC::Common'; -use Data::Dumper; + +#use base qw/WebPAC::Common/; +use Data::Dump qw/dump/; +use Storable qw/dclone/; +use Carp qw/confess/; + +# debugging warn(s) +my $debug = 0; + =head1 NAME -WebPAC::Normalize - data mungling for normalisation +WebPAC::Normalize - describe normalisaton rules using sets =head1 VERSION -Version 0.07 +Version 0.29 =cut -our $VERSION = '0.07'; +our $VERSION = '0.29'; =head1 SYNOPSIS -This package contains code that mungle data to produce normalized format. +This module uses C files to perform normalisation +from input records using perl functions which are specialized for set +processing. + +Sets are implemented as arrays, and normalisation file is valid perl, which +means that you check it's validity before running WebPAC using +C. + +Normalisation can generate multiple output normalized data. For now, supported output +types (on the left side of definition) are: C, C, C and +C. -It contains several assumptions: +=head1 FUNCTIONS -=over +Functions which start with C<_> are private and used by WebPAC internally. +All other functions are available for use within normalisation rules. -=item * +=head2 data_structure -format of fields is defined using C notation for repeatable fields -or C for single (or first) value, where C<123> is field number and -C is subfield. +Return data structure -=item * + my $ds = WebPAC::Normalize::data_structure( + lookup => $lookup_hash, + row => $row, + rules => $normalize_pl_config, + marc_encoding => 'utf-8', + config => $config, + load_row_coderef => sub { + my ($database,$input,$mfn) = shift; + $store->load_row( database => $database, input => $input, id => $mfn ); + }, + ); -source data records (C<$rec>) have unique identifiers in field C<000> +Options C, C and C are mandatory while all +other are optional. -=item * +C is closure only used when executing lookups, so they will +die if it's not defined. -optional C tag at B will be -perl code that is evaluated before producing output (value of field will be -interpolated before that) +This function will B if normalizastion can't be evaled. -=item * +Since this function isn't exported you have to call it with +C. -optional C at B will apply perl -code defined as code ref on format after field substitution to producing -output +=cut -There is one built-in filter called C which can be use like this: +my $load_row_coderef; - filter{regex(s/foo/bar/)} +sub data_structure { + my $arg = {@_}; -=item * + die "need row argument" unless ($arg->{row}); + die "need normalisation argument" unless ($arg->{rules}); -optional C will be then performed. See C. + no strict 'subs'; + _set_lookup( $arg->{lookup} ) if defined($arg->{lookup}); + _set_rec( $arg->{row} ); + _set_config( $arg->{config} ) if defined($arg->{config}); + _clean_ds( %{ $arg } ); + $load_row_coderef = $arg->{load_row_coderef}; -=item * + eval "$arg->{rules}"; + die "error evaling $arg->{rules}: $@\n" if ($@); -at end, optional Cs rules are resolved. Format rules are similar to -C and can also contain C which is performed after -values are inserted in format. + return _get_ds(); +} -=back +=head2 _set_rec -This also describes order in which transformations are applied (eval, -filter, lookup, format) which is important to undestand when deciding how to -solve your data mungling and normalisation process. +Set current record hash + _set_rec( $rec ); +=cut +my $rec; -=head1 FUNCTIONS +sub _set_rec { + $rec = shift or die "no record hash"; +} -=head2 new +=head2 _set_config -Create new normalisation object +Set current config hash - my $n = new WebPAC::Normalize::Something( - filter => { - 'filter_name_1' => sub { - # filter code - return length($_); - }, ... - }, - db => $db_obj, - lookup_regex => $lookup->regex, - lookup => $lookup_obj, - prefix => 'foobar', - ); + _set_config( $config ); -Parametar C defines user supplied snippets of perl code which can -be use with C notation. +Magic keys are: -C is used to form filename for database record (to support multiple -source files which are joined in one database). +=over 4 -Recommended parametar C is used to enable parsing of lookups -in structures. If you pass this parametar, you must also pass C -which is C object. +=item _ + +Code of current database + +=item _mfn + +Current MFN + +=back =cut -sub new { - my $class = shift; - my $self = {@_}; - bless($self, $class); +my $config; - my $r = $self->{'lookup_regex'} ? 1 : 0; - my $l = $self->{'lookup'} ? 1 : 0; +sub _set_config { + $config = shift; +} - my $log = $self->_get_logger(); +=head2 _get_ds - # those two must be in pair - if ( ($r & $l) != ($r || $l) ) { - my $log = $self->_get_logger(); - $log->logdie("lookup_regex and lookup must be in pair"); - } +Return hash formatted as data structure - $log->logdie("lookup must be WebPAC::Lookup object") if ($self->{'lookup'} && ! $self->{'lookup'}->isa('WebPAC::Lookup')); + my $ds = _get_ds(); - $log->warn("no prefix defined. please check that!") unless ($self->{'prefix'}); +=cut - $log->debug("using lookup regex: ", $self->{lookup_regex}) if ($r && $l); +my ($out, $marc_record, $marc_encoding, $marc_repeatable_subfield, $marc_indicators, $marc_leader); +my ($marc_record_offset, $marc_fetch_offset) = (0, 0); - if (! $self->{filter} || ! $self->{filter}->{regex}) { - $log->debug("adding built-in filter regex"); - $self->{filter}->{regex} = sub { - my ($val, $regex) = @_; - eval "\$val =~ $regex"; - return $val; - }; - } +sub _get_ds { + return $out; +} + +=head2 _clean_ds + +Clean data structure hash for next record + + _clean_ds(); + +=cut - $self ? return $self : return undef; +sub _clean_ds { + my $a = {@_}; + ($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}; } +=head2 _set_lookup -=head2 data_structure +Set current lookup hash + + _set_lookup( $lookup ); + +=cut -Create in-memory data structure which represents normalized layout from -C. +my $lookup; -This structures are used to produce output. +sub _set_lookup { + $lookup = shift; +} + +=head2 _get_lookup + +Get current lookup hash - my $ds = $webpac->data_structure($rec); + my $lookup = _get_lookup(); =cut -sub data_structure { - my $self = shift; +sub _get_lookup { + return $lookup; +} - my $log = $self->_get_logger(); +=head2 _set_load_row - my $rec = shift; - $log->logconfess("need HASH as first argument!") if ($rec !~ /HASH/o); +Setup code reference which will return L from +L - $log->debug("data_structure rec = ", sub { Dumper($rec) }); + _set_load_row(sub { + my ($database,$input,$mfn) = @_; + $store->load_row( database => $database, input => $input, id => $mfn ); + }); - $log->logdie("need unique ID (mfn) in field 000 of record " . Dumper($rec) ) unless (defined($rec->{'000'})); +=cut - my $id = $rec->{'000'}->[0] || $log->logdie("field 000 isn't array!"); +sub _set_load_row { + my $coderef = shift; + confess "argument isn't CODE" unless ref($coderef) eq 'CODE'; - my $cache_file; + $load_row_coderef = $coderef; +} - if ($self->{'db'}) { - my $ds = $self->{'db'}->load_ds( id => $id, prefix => $self->{prefix} ); - $log->debug("load_ds( rec = ", sub { Dumper($rec) }, ") = ", sub { Dumper($ds) }); - return $ds if ($ds); - $log->debug("cache miss, creating"); - } +=head2 _get_marc_fields - my @sorted_tags; - if ($self->{tags_by_order}) { - @sorted_tags = @{$self->{tags_by_order}}; - } else { - @sorted_tags = sort { $self->_sort_by_order } keys %{$self->{'import_xml'}->{'indexer'}}; - $self->{tags_by_order} = \@sorted_tags; - } +Get all fields defined by calls to C - my $ds; + $marc->add_fields( WebPAC::Normalize:_get_marc_fields() ); - $log->debug("tags: ",sub { join(", ",@sorted_tags) }); +We are using I which detect repeatable fields only from +sequence of field/subfield data generated by normalization. - foreach my $field (@sorted_tags) { +Repeatable field is created when there is second occurence of same subfield or +if any of indicators are different. - my $row; +This is sane for most cases. Something like: -#print "field $field [",$self->{'tag'},"] = ",Dumper($self->{'import_xml'}->{'indexer'}->{$field}->{$self->{'tag'}}); + 900a-1 900b-1 900c-1 + 900a-2 900b-2 + 900a-3 - foreach my $tag (@{$self->{'import_xml'}->{'indexer'}->{$field}->{$self->{'tag'}}}) { - my $format; +will be created from any combination of: - $log->logdie("expected tag HASH and got $tag") unless (ref($tag) eq 'HASH'); - $format = $tag->{'value'} || $tag->{'content'}; + 900a-1 900a-2 900a-3 900b-1 900b-2 900c-1 - $log->debug("format: $format"); +and following rules: - my @v; - if ($self->{'lookup_regex'} && $format =~ $self->{'lookup_regex'}) { - @v = $self->fill_in_to_arr($rec,$format); - } else { - @v = $self->parse_to_arr($rec,$format); - } - next if (! @v); + marc('900','a', rec('200','a') ); + marc('900','b', rec('200','b') ); + marc('900','c', rec('200','c') ); - if ($tag->{'sort'}) { - @v = $self->sort_arr(@v); - } +which might not be what you have in mind. If you need repeatable subfield, +define it using C like this: - # use format? - if ($tag->{'format_name'}) { - @v = map { $self->apply_format($tag->{'format_name'},$tag->{'format_delimiter'},$_) } @v; - } + marc_repeatable_subfield('900','a'); + marc('900','a', rec('200','a') ); + marc('900','b', rec('200','b') ); + marc('900','c', rec('200','c') ); - # delimiter will join repeatable fields - if ($tag->{'delimiter'}) { - @v = ( join($tag->{'delimiter'}, @v) ); - } +will create: - # default types - my @types = qw(display search); - # override by type attribute - @types = ( $tag->{'type'} ) if ($tag->{'type'}); - - foreach my $type (@types) { - # append to previous line? - $log->debug("type: $type ",sub { join(" ",@v) }, " ", $row->{'append'} || 'no append'); - if ($tag->{'append'}) { - - # I will delimit appended part with - # delimiter (or ,) - my $d = $tag->{'delimiter'}; - # default delimiter - $d ||= " "; - - my $last = pop @{$row->{$type}}; - $d = "" if (! $last); - $last .= $d . join($d, @v); - push @{$row->{$type}}, $last; + 900a-1 900a-2 900a-3 900b-1 900c-1 + 900b-2 - } else { - push @{$row->{$type}}, @v; - } - } +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 { + + 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_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 ( $marc_repeatable_subfield ); + warn "## marc_record[$fetch_pos] = ", 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) % ($len + 1); } - if ($row) { - $row->{'tag'} = $field; + # mark it visited + $visited->{$i}++; - # TODO: name_sigular, name_plural - my $name = $self->{'import_xml'}->{'indexer'}->{$field}->{'name'}; - my $row_name = $name ? $self->_x($name) : $field; - - # post-sort all values in field - if ($self->{'import_xml'}->{'indexer'}->{$field}->{'sort'}) { - $log->warn("sort at field tag not implemented"); - } + my $row = dclone( $sorted_marc_record[$i] ); - $ds->{$row_name} = $row; + # 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; + } - $log->debug("row $field: ",sub { Dumper($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); } - $self->{'db'}->save_ds( - id => $id, - ds => $ds, - prefix => $self->{prefix}, - ) if ($self->{'db'}); + return \@m; +} - $log->debug("ds: ", sub { Dumper($ds) }); +=head2 _get_marc_leader - $log->logconfess("data structure returned is not array any more!") if wantarray; +Return leader from currently fetched record by L - return $ds; + 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 parse +=head2 _debug -Perform smart parsing of string, skipping delimiters for fields which aren't -defined. It can also eval code in format starting with C and -return output or nothing depending on eval code. +Change level of debug warnings - my $text = $webpac->parse($rec,'eval{"v901^a" eq "Deskriptor"}descriptor: v250^a', $i); + _debug( 2 ); -Filters are implemented here. While simple form of filters looks like this: +=cut - filter{name_of_filter} +sub _debug { + my $l = shift; + return $debug unless defined($l); + warn "debug level $l",$/ if ($l > 0); + $debug = $l; +} -but, filters can also have variable number of parametars like this: +=head1 Functions to create C - filter{name_of_filter(param,param,param)} +Those functions generally have to first in your normalization file. -=cut +=head2 tag -my $warn_once; +Define new tag for I and I. -sub parse { - my $self = shift; + tag('Title', rec('200','a') ); - my ($rec, $format_utf8, $i) = @_; - return if (! $format_utf8); +=cut - my $log = $self->_get_logger(); +sub tag { + my $name = shift or die "tag needs name as first argument"; + my @o = grep { defined($_) && $_ ne '' } @_; + return unless (@o); + $out->{$name}->{tag} = $name; + $out->{$name}->{search} = \@o; + $out->{$name}->{display} = \@o; +} - $log->logconfess("need HASH as first argument!") if ($rec !~ /HASH/o); +=head2 display - $i = 0 if (! $i); +Define tag just for I - my $format = $self->_x($format_utf8) || $log->logconfess("can't convert '$format_utf8' from UTF-8 to ",$self->{'code_page'}); + @v = display('Title', rec('200','a') ); - my @out; +=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; +} - $log->debug("format: $format"); +=head2 search - my $eval_code; - # remove eval{...} from beginning - $eval_code = $1 if ($format =~ s/^eval{([^}]+)}//s); +Prepare values just for I - my $filter_name; - # remove filter{...} from beginning - $filter_name = $1 if ($format =~ s/^filter{([^}]+)}//s); + @v = search('Title', rec('200','a') ); - my $prefix; - my $all_found=0; +=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; +} - while ($format =~ s/^(.*?)(v|s)(\d+)(?:\^(\w))?//s) { +=head2 marc_leader - my $del = $1 || ''; - $prefix ||= $del if ($all_found == 0); +Setup fields within MARC leader or get leader - # repeatable index - my $r = $i; - $r = 0 if (lc("$2") eq 's'); + marc_leader('05','c'); + my $leader = marc_leader(); + +=cut - my $found = 0; - my $tmp = $self->get_data(\$rec,$3,$4,$r,\$found); +sub marc_leader { + my ($offset,$value) = @_; - if ($found) { - push @out, $del; - push @out, $tmp; - $all_found += $found; + if ($offset) { + $marc_leader->[ $marc_record_offset ]->{ $offset } = $value; + } else { + + if (defined($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 - return if (! $all_found); + marc_fixed('008', 00, '070402'); + marc_fixed('008', 39, '|'); - my $out = join('',@out); +Positions not specified will be filled with spaces (C<0x20>). - if ($out) { - # add rest of format (suffix) - $out .= $format; +There will be no effort to extend last specified value to full length of +field in standard. - # add prefix if not there - $out = $prefix . $out if ($out !~ m/^\Q$prefix\E/); +=cut + +sub marc_fixed { + my ($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 ] }; - $log->debug("result: $out"); + 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 ($debug > 1); } +} + +=head2 marc + +Save value for MARC field + + marc('900','a', rec('200','a') ); + marc('001', rec('000') ); + +=cut - if ($eval_code) { - my $eval = $self->fill_in($rec,$eval_code,$i) || return; - $log->debug("about to eval{$eval} format: $out"); - return if (! $self->_eval($eval)); +sub marc { + my $f = shift or die "marc needs field"; + die "marc field must be numer" unless ($f =~ /^\d+$/); + + my $sf; + if ($f >= 10) { + $sf = shift or die "marc needs subfield"; } - - if ($filter_name) { - my @filter_args; - if ($filter_name =~ s/(\w+)\((.*)\)/$1/) { - @filter_args = split(/,/, $2); - } - if ($self->{'filter'}->{$filter_name}) { - $log->debug("about to filter{$filter_name} format: $out with arguments: ", join(",", @filter_args)); - unshift @filter_args, $out; - $out = $self->{'filter'}->{$filter_name}->(@filter_args); - return unless(defined($out)); - $log->debug("filter result: $out"); - } elsif (! $warn_once->{$filter_name}) { - $log->warn("trying to use undefined filter $filter_name"); - $warn_once->{$filter_name}++; + + 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 ]; } } +} - return $out; +=head2 marc_repeatable_subfield + +Save values for MARC repetable subfield + + marc_repeatable_subfield('910', 'z', rec('909') ); + +=cut + +sub marc_repeatable_subfield { + my ($f,$sf) = @_; + die "marc_repeatable_subfield need field and subfield!\n" unless ($f && $sf); + $marc_repeatable_subfield->{ $f . $sf }++; + marc(@_); } -=head2 parse_to_arr +=head2 marc_indicators -Similar to C, but returns array of all repeatable fields +Set both indicators for MARC field - my @arr = $webpac->parse_to_arr($rec,'v250^a'); + marc_indicators('900', ' ', 1); + +Any indicator value other than C<0-9> will be treated as undefined. =cut -sub parse_to_arr { - my $self = shift; +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); +} - my ($rec, $format_utf8) = @_; +=head2 marc_compose - my $log = $self->_get_logger(); +Save values for each MARC subfield explicitly - $log->logconfess("need HASH as first argument!") if ($rec !~ /HASH/o); - return if (! $format_utf8); + marc_compose('900', + 'a', rec('200','a') + 'b', rec('201','a') + 'a', rec('200','b') + 'c', rec('200','c') + ); - my $i = 0; - my @arr; +If you specify C<+> for subfield, value will be appended +to previous defined subfield. - while (my $v = $self->parse($rec,$format_utf8,$i++)) { - push @arr, $v; +=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; + } } - $log->debug("format '$format_utf8' returned ",--$i," elements: ", sub { join(" | ",@arr) }) if (@arr); + warn "## marc_compose current marc = ", dump( $m ),$/ if ($debug > 1); - return @arr; + push @{ $marc_record->[ $marc_record_offset ] }, $m if ($#{$m} > 2); } +=head2 marc_duplicate -=head2 fill_in +Generate copy of current MARC record and continue working on copy -Workhourse of all: takes record from in-memory structure of database and -strings with placeholders and returns string or array of with substituted -values from record. + marc_duplicate(); - my $text = $webpac->fill_in($rec,'v250^a'); +Copies can be accessed using C<< _get_marc_fields( fetch_next => 1 ) >> or +C<< _get_marc_fields( offset => 42 ) >>. -Optional argument is ordinal number for repeatable fields. By default, -it's assume to be first repeatable field (fields are perl array, so first -element is 0). -Following example will read second value from repeatable field. +=cut - my $text = $webpac->fill_in($rec,'Title: v250^a',1); +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); -This function B perform parsing of format to inteligenty skip -delimiters before fields which aren't used. +} -This method will automatically decode UTF-8 string to local code page -if needed. +=head2 marc_remove -=cut +Remove some field or subfield from MARC record. -sub fill_in { - my $self = shift; + marc_remove('200'); + marc_remove('200','a'); - my $log = $self->_get_logger(); +This will erase field C<200> or C<200^a> from current MARC record. - my $rec = shift || $log->logconfess("need data record"); - my $format = shift || $log->logconfess("need format to parse"); - # iteration (for repeatable fields) - my $i = shift || 0; + marc_remove('*'); - $log->logdie("infitite loop in format $format") if ($i > ($self->{'max_mfn'} || 9999)); +Will remove all fields in current MARC record. - # FIXME remove for speedup? - $log->logconfess("need HASH as first argument!") if ($rec !~ /HASH/o); +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). - if (utf8::is_utf8($format)) { - $format = $self->_x($format); - } +FIXME: support fields < 10. - my $found = 0; +=cut - my $eval_code; - # remove eval{...} from beginning - $eval_code = $1 if ($format =~ s/^eval{([^}]+)}//s); +sub marc_remove { + my ($f, $sf) = @_; - my $filter_name; - # remove filter{...} from beginning - $filter_name = $1 if ($format =~ s/^filter{([^}]+)}//s); + die "marc_remove needs record number" unless defined($f); - # do actual replacement of placeholders - # repeatable fields - $format =~ s/v(\d+)(?:\^(\w))?/$self->get_data(\$rec,$1,$2,$i,\$found)/ges; - # non-repeatable fields - $format =~ s/s(\d+)(?:\^(\w))?/$self->get_data(\$rec,$1,$2,0,\$found)/ges; + my $marc = $marc_record->[ $marc_record_offset ]; - if ($found) { - $log->debug("format: $format"); - if ($eval_code) { - my $eval = $self->fill_in($rec,$eval_code,$i); - return if (! $self->_eval($eval)); - } - if ($filter_name && $self->{'filter'}->{$filter_name}) { - $log->debug("filter '$filter_name' for $format"); - $format = $self->{'filter'}->{$filter_name}->($format); - return unless(defined($format)); - $log->debug("filter result: $format"); - } - # do we have lookups? - if ($self->{'lookup'}) { - if ($self->{'lookup'}->can('lookup')) { - my @lookup = $self->{lookup}->lookup($format); - $log->debug("lookup $format", join(", ", @lookup)); - return @lookup; - } else { - $log->warn("Have lookup object but can't invoke lookup method"); + 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--; + }; + } + } + } } - } else { - return $format; + $i++; } - } else { - return; + + warn "### marc_remove($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 ); -=head2 fill_in_to_arr +Please note that field numbers are consistent with other commands (marc +field number first), but somewhat counter-intuitive (destination and then +source). -Similar to C, but returns array of all repeatable fields. Usable -for fields which have lookups, so they shouldn't be parsed but rather -Ced. +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. - my @arr = $webpac->fill_in_to_arr($rec,'[v900];;[v250^a]'); =cut -sub fill_in_to_arr { - my $self = shift; +sub marc_original_order { - my ($rec, $format_utf8) = @_; + my ($to, $from) = @_; + die "marc_original_order needs from and to fields\n" unless ($from && $to); - my $log = $self->_get_logger(); + return unless defined($rec->{$from}); - $log->logconfess("need HASH as first argument!") if ($rec !~ /HASH/o); - return if (! $format_utf8); + my $r = $rec->{$from}; + die "record field $from isn't array\n" unless (ref($r) eq 'ARRAY'); - my $i = 0; - my @arr; + my ($i1,$i2) = defined($marc_indicators->{$to}) ? @{ $marc_indicators->{$to} } : (' ',' '); + warn "## marc_original_order($to,$from) source = ", dump( $r ),$/ if ($debug > 1); - while (my @v = $self->fill_in($rec,$format_utf8,$i++)) { - push @arr, @v; - } + 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); - $log->debug("format '$format_utf8' returned ",--$i," elements: ", sub { join(" | ",@arr) }) if (@arr); + warn "#--> d: ",dump($d), "\n#--> sfs: ",dump(@sfs),$/ if ($debug > 2); - return @arr; + 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_count -=head2 get_data +Return number of MARC records created using L. + + print "created ", marc_count(), " records"; + +=cut + +sub marc_count { + return $#{ $marc_record }; +} -Returns value from record. - my $text = $self->get_data(\$rec,$f,$sf,$i,\$found); +=head1 Functions to extract data from input -Arguments are: -record reference C<$rec>, -field C<$f>, -optional subfiled C<$sf>, -index for repeatable values C<$i>. +This function should be used inside functions to create C described +above. -Optinal variable C<$found> will be incremeted if there -is field. +=head2 _pack_subfields_hash -Returns value or empty string. + @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 get_data { - my $self = shift; +sub _pack_subfields_hash { + + warn "## _pack_subfields_hash( ",dump(@_), " )\n" if ($debug > 1); + + my ($h,$include_subfields) = @_; - my ($rec,$f,$sf,$i,$found) = @_; + # sanity and ease of use + return $h if (ref($h) ne 'HASH'); - if ($$rec->{$f}) { - return '' if (! $$rec->{$f}->[$i]); - no strict 'refs'; - if ($sf && $$rec->{$f}->[$i]->{$sf}) { - $$found++ if (defined($$found)); - return $$rec->{$f}->[$i]->{$sf}; - } elsif (! $sf && $$rec->{$f}->[$i]) { - $$found++ if (defined($$found)); - # it still might have subfield, just - # not specified, so we'll dump all - if ($$rec->{$f}->[$i] =~ /HASH/o) { - my $out; - foreach my $k (keys %{$$rec->{$f}->[$i]}) { - $out .= $$rec->{$f}->[$i]->{$k}." "; + if ( defined($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; + } + 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 + + @v = rec1('200') + +TODO: order of values is probably same as in source data, need to investigate that + +=cut + +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') { + my @out; + foreach my $h ( @{ $rec->{$f} } ) { + if (ref($h) eq 'HASH') { + push @out, ( _pack_subfields_hash( $h ) ); } else { - return $$rec->{$f}->[$i]; + push @out, $h; } + } + return @out; + } elsif( defined($rec->{$f}) ) { + return $rec->{$f}; + } +} + +=head2 rec2 + +Return all values in specific field and subfield + + @v = rec2('200','a') + +=cut + +sub rec2 { + my $f = shift; + return unless (defined($rec && $rec->{$f})); + my $sf = shift; + warn "rec2($f,$sf) = ", dump( $rec->{$f} ), $/ if ($debug > 1); + return map { + if (ref($_->{$sf}) eq 'ARRAY') { + @{ $_->{$sf} }; } else { - return ''; + $_->{$sf}; } + } grep { ref($_) eq 'HASH' && $_->{$sf} } @{ $rec->{$f} }; +} + +=head2 rec + +syntaxtic sugar for + + @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 ($#_ == 0) { + @out = rec1(@_); + } elsif ($#_ == 1) { + @out = rec2(@_); + } + if ($#out == 0 && ! wantarray) { + return $out[0]; + } elsif (@out) { + return @out; } else { return ''; } } +=head2 regex + +Apply regex to some or all values + + @v = regex( 's/foo/bar/g', @v ); + +=cut + +sub regex { + my $r = shift; + my @out; + #warn "r: $r\n", dump(\@_); + foreach my $t (@_) { + next unless ($t); + eval "\$t =~ $r"; + push @out, $t if ($t && $t ne ''); + } + return @out; +} + +=head2 prefix -=head2 apply_format +Prefix all values with a string -Apply format specified in tag with C and -C. + @v = prefix( 'my_', @v ); - my $text = $webpac->apply_format($format_name,$format_delimiter,$data); +=cut + +sub prefix { + my $p = shift; + return @_ unless defined( $p ); + return map { $p . $_ } grep { defined($_) } @_; +} + +=head2 suffix + +suffix all values with a string -Formats can contain C if you need them. + @v = suffix( '_my', @v ); =cut -sub apply_format { - my $self = shift; +sub suffix { + my $s = shift; + return @_ unless defined( $s ); + return map { $_ . $s } grep { defined($_) } @_; +} + +=head2 surround - my ($name,$delimiter,$data) = @_; +surround all values with a two strings - my $log = $self->_get_logger(); + @v = surround( 'prefix_', '_suffix', @v ); - if (! $self->{'import_xml'}->{'format'}->{$name}) { - $log->warn(" is not defined in ",$self->{'import_xml_file'}); - return $data; +=cut + +sub surround { + my $p = shift; + my $s = shift; + $p = '' unless defined( $p ); + $s = '' unless defined( $s ); + return map { $p . $_ . $s } grep { defined($_) } @_; +} + +=head2 first + +Return first element + + $v = first( @v ); + +=cut + +sub first { + my $r = shift; + return $r; +} + +=head2 lookup + +Consult lookup hashes for some value + + @v = lookup( + sub { + 'ffkk/peri/mfn'.rec('000') + }, + 'ffkk','peri','200-a-200-e', + sub { + first(rec(200,'a')).' '.first(rec('200','e')) } + ); + +Code like above will be B using L from +normal lookup definition in C which looks like: + + lookup( + # which results to return from record recorded in lookup + sub { 'ffkk/peri/mfn' . rec('000') }, + # from which database and input + 'ffkk','peri', + # such that following values match + sub { first(rec(200,'a')) . ' ' . first(rec('200','e')) }, + # if this part is missing, we will try to match same fields + # from lookup record and current one, or you can override + # which records to use from current record using + sub { rec('900','x') . ' ' . rec('900','y') }, + ) + +You can think about this lookup as SQL (if that helps): + + select + sub { what } + from + database, input + where + sub { filter from lookuped record } + having + sub { optional filter on current record } + +Easy as pie, right? + +=cut + +sub lookup { + my ($what, $database, $input, $key, $having) = @_; + + confess "lookup needs 5 arguments: what, database, input, key, having\n" unless ($#_ == 4); - $log->warn("no delimiter for format $name") if (! $delimiter); + warn "## lookup ($database, $input, $key)", $/ if ($debug > 1); + return unless (defined($lookup->{$database}->{$input}->{$key})); - my $format = $self->_x($self->{'import_xml'}->{'format'}->{$name}->{'content'}) || $log->logdie("can't find format '$name'"); + confess "lookup really need load_row_coderef added to data_structure\n" unless ($load_row_coderef); - my @data = split(/\Q$delimiter\E/, $data); + my $mfns; + my @having = $having->(); - my $out = sprintf($format, @data); - $log->debug("using format $name [$format] on $data to produce: $out"); + warn "## having = ", dump( @having ) if ($debug > 2); - if ($self->{'lookup_regex'} && $out =~ $self->{'lookup_regex'}) { - return $self->{'lookup'}->lookup($out); + 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; + 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. + +=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 sort_arr +=head2 config + +Consult config values stored in C -Sort array ignoring case and html in data + # return database code (key under databases in yaml) + $database_code = config(); # use _ from hash + $database_name = config('name'); + $database_input_name = config('input name'); + $tag = config('input normalize tag'); - my @sorted = $webpac->sort_arr(@unsorted); +Up to three levels are supported. =cut -sub sort_arr { - my $self = shift; +sub config { + return unless ($config); + + my $p = shift; + + $p ||= ''; + + my $v; - my $log = $self->_get_logger(); + warn "### getting config($p)\n" if ($debug > 1); - # FIXME add Schwartzian Transformation? + my @p = split(/\s+/,$p); + if ($#p < 0) { + $v = $config->{ '_' }; # special, database code + } else { + + 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); + + } - my @sorted = sort { - $a =~ s#<[^>]+/*>##; - $b =~ s#<[^>]+/*>##; - lc($b) cmp lc($a) - } @_; - $log->debug("sorted values: ",sub { join(", ",@sorted) }); + warn "## config( '$p' ) = ",dump( $v ),$/ if ($v && $debug); + warn "config( '$p' ) is empty\n" if (! $v); - return @sorted; + return $v; } +=head2 id -=head1 INTERNAL METHODS +Returns unique id of this record -=head2 _sort_by_order + $id = id(); -Sort xml tags data structure accoding to C attribute. +Returns C<42/2> for 2nd occurence of MFN 42. =cut -sub _sort_by_order { - my $self = shift; +sub id { + my $mfn = $config->{_mfn} || die "no _mfn in config data"; + return $mfn . $#{$marc_record} ? $#{$marc_record} + 1 : ''; +} + +=head2 join_with + +Joins walues with some delimiter - my $va = $self->{'import_xml'}->{'indexer'}->{$a}->{'order'} || - $self->{'import_xml'}->{'indexer'}->{$a}; - my $vb = $self->{'import_xml'}->{'indexer'}->{$b}->{'order'} || - $self->{'import_xml'}->{'indexer'}->{$b}; + $v = join_with(", ", @v); + +=cut - return $va <=> $vb; +sub join_with { + my $d = shift; + warn "### join_with('$d',",dump(@_),")\n" if ($debug > 2); + my $v = join($d, grep { defined($_) && $_ ne '' } @_); + return '' unless defined($v); + return $v; } -=head2 _x +=head2 split_rec_on -Convert strings from C encoding into application -specific encoding (optinally specified using C to C -constructor). +Split record subfield on some regex and take one of parts out - my $text = $n->_x('normalize text string'); + $a_before_semi_column = + split_rec_on('200','a', /\s*;\s*/, $part); -This is a stub so that other modules doesn't have to implement it. +C<$part> is optional number of element. First element is +B<1>, not 0! + +If there is no C<$part> parameter or C<$part> is 0, this function will +return all values produced by splitting. =cut -sub _x { - my $self = shift; - return shift; +sub split_rec_on { + die "split_rec_on need (fld,sf,regex[,part]" if ($#_ < 2); + + my ($fld, $sf, $regex, $part) = @_; + warn "### regex ", ref($regex), $regex, $/ if ($debug > 2); + + my @r = rec( $fld, $sf ); + my $v = shift @r; + warn "### first rec($fld,$sf) = ",dump($v),$/ if ($debug > 2); + + 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; + } } +my $hash; -=head1 AUTHOR +=head2 set -Dobrica Pavlinusic, C<< >> + set( key => 'value' ); -=head1 COPYRIGHT & LICENSE +=cut + +sub set { + my ($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 $v; +} -Copyright 2005 Dobrica Pavlinusic, All Rights Reserved. +=head2 count -This program is free software; you can redistribute it and/or modify it -under the same terms as Perl itself. + if ( count( @result ) == 1 ) { + # do something if only 1 result is there + } =cut -1; # End of WebPAC::Normalize +sub count { + warn "## count ",dump(@_),$/ if ( $debug ); + return @_ . ''; +} + +# END +1;