--- trunk/lib/WebPAC/Normalize.pm 2005/07/16 23:56:14 13 +++ trunk/lib/WebPAC/Normalize.pm 2006/07/02 21:17:54 566 @@ -1,534 +1,679 @@ package WebPAC::Normalize; +use Exporter 'import'; +@EXPORT = qw/ + _set_rec _set_lookup + _get_ds _clean_ds + _debug + + tag search display + marc marc_indicators marc_repeatable_subfield + marc_compose marc_leader + + rec1 rec2 rec + regex prefix suffix surround + first lookup join_with + + split_rec_on +/; use warnings; use strict; -use Data::Dumper; + +#use base qw/WebPAC::Common/; +use Data::Dump qw/dump/; +use Encode qw/from_to/; + +# debugging warn(s) +my $debug = 0; + =head1 NAME -WebPAC::Normalize - normalisation of source file +WebPAC::Normalize - describe normalisaton rules using sets =head1 VERSION -Version 0.01 +Version 0.09 =cut -our $VERSION = '0.01'; +our $VERSION = '0.09'; =head1 SYNOPSIS -This package contains code that could be helpful in implementing different -normalisation front-ends. +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. =head1 FUNCTIONS -=head2 new +Functions which start with C<_> are private and used by WebPAC internally. +All other functions are available for use within normalisation rules. -Create new normalisation object +=head2 data_structure - my $n = new WebPAC::Normalize::Something( - cache_data_structure => './cache/ds/', - lookup_regex => $lookup->regex, +Return data structure + + my $ds = WebPAC::Normalize::data_structure( + lookup => $lookup->lookup_hash, + row => $row, + rules => $normalize_pl_config, + marc_encoding => 'utf-8', ); -Optional parameter C defines path to directory -in which cache file for C call will be created. +Options C, C, C and C are mandatory while all +other are optional. + +This function will B if normalizastion can't be evaled. -Recommended parametar C is used to enable parsing of lookups -in structures. +Since this function isn't exported you have to call it with +C. =cut -sub new { - my $class = shift; - my $self = {@_}; - bless($self, $class); +sub data_structure { + my $arg = {@_}; - $self->setup_cache_dir( $self->{'cache_data_structure'} ); + die "need row argument" unless ($arg->{row}); + die "need normalisation argument" unless ($arg->{rules}); - $self ? return $self : return undef; -} + no strict 'subs'; + _set_lookup( $arg->{lookup} ); + _set_rec( $arg->{row} ); + _clean_ds( %{ $arg } ); + eval "$arg->{rules}"; + die "error evaling $arg->{rules}: $@\n" if ($@); -=head2 setup_cache_dir + return _get_ds(); +} -Check if specified cache directory exist, and if not, disable caching. +=head2 _set_rec - $setup_cache_dir('./cache/ds/'); +Set current record hash -If you pass false or zero value to this function, it will disable -cacheing. + _set_rec( $rec ); =cut -sub setup_cache_dir { - my $self = shift; +my $rec; - my $dir = shift; +sub _set_rec { + $rec = shift or die "no record hash"; +} - my $log = $self->_get_logger(); +=head2 _get_ds - if ($dir) { - my $msg; - if (! -e $dir) { - $msg = "doesn't exist"; - } elsif (! -d $dir) { - $msg = "is not directory"; - } elsif (! -w $dir) { - $msg = "not writable"; - } +Return hash formatted as data structure - if ($msg) { - undef $self->{'cache_data_structure'}; - $log->warn("cache_data_structure $dir $msg, disabling..."); - } else { - $log->debug("using cache dir $dir"); - } - } else { - $log->debug("disabling cache"); - undef $self->{'cache_data_structure'}; - } + my $ds = _get_ds(); + +=cut + +my ($out,$marc_record, $marc_encoding, $marc_repeatable_subfield, $marc_indicators); + +sub _get_ds { + return $out; } +=head2 _clean_ds -=head2 data_structure +Clean data structure hash for next record -Create in-memory data structure which represents normalized layout from -C. + _clean_ds(); -This structures are used to produce output. +=cut + +sub _clean_ds { + my $a = {@_}; + ($out,$marc_record, $marc_encoding, $marc_repeatable_subfield, $marc_indicators) = (); + $marc_encoding = $a->{marc_encoding}; +} - my @ds = $webpac->data_structure($rec); +=head2 _set_lookup -B +Set current lookup hash -This method will also set C<< $webpac->{'currnet_filename'} >> if there is -C<< >> tag and C<< $webpac->{'headline'} >> if there is -C<< >> tag. + _set_lookup( $lookup ); =cut -sub data_structure { - my $self = shift; +my $lookup; - my $log = $self->_get_logger(); +sub _set_lookup { + $lookup = shift; +} - my $rec = shift; - $log->logconfess("need HASH as first argument!") if ($rec !~ /HASH/o); +=head2 _get_marc_fields - my $cache_file; +Get all fields defined by calls to C - if (my $cache_path = $self->{'cache_data_structure'}) { - my $id = $rec->{'000'}; - $id = $rec->{'000'}->[0] if ($id =~ m/^ARRAY/o); - unless (defined($id)) { - $log->warn("Can't use cache_data_structure on records without unique identifier in field 000"); - undef $self->{'cache_data_structure'}; - } else { - $cache_file = "$cache_path/$id"; - if (-r $cache_file) { - my $ds_ref = retrieve($cache_file); - if ($ds_ref) { - $log->debug("cache hit: $cache_file"); - my $ok = 1; - foreach my $f (qw(current_filename headline)) { - if ($ds_ref->{$f}) { - $self->{$f} = $ds_ref->{$f}; - } else { - $ok = 0; - } - }; - if ($ok && $ds_ref->{'ds'}) { - return @{ $ds_ref->{'ds'} }; - } else { - $log->warn("cache_data_structure $cache_path corrupt. Use rm $cache_path/* to re-create it on next run!"); - undef $self->{'cache_data_structure'}; - } - } - } - } - } + $marc->add_fields( WebPAC::Normalize:_get_marc_fields() ); - undef $self->{'currnet_filename'}; - undef $self->{'headline'}; +We are using I which detect repeatable fields only from +sequence of field/subfield data generated by normalization. - my @sorted_tags; - if ($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; - } +Repeatable field is created when there is second occurence of same subfield or +if any of indicators are different. - my @ds; +This is sane for most cases. Something like: - $log->debug("tags: ",sub { join(", ",@sorted_tags) }); + 900a-1 900b-1 900c-1 + 900a-2 900b-2 + 900a-3 - foreach my $field (@sorted_tags) { +will be created from any combination of: - my $row; + 900a-1 900a-2 900a-3 900b-1 900b-2 900c-1 -#print "field $field [",$self->{'tag'},"] = ",Dumper($self->{'import_xml'}->{'indexer'}->{$field}->{$self->{'tag'}}); +and following rules: - foreach my $tag (@{$self->{'import_xml'}->{'indexer'}->{$field}->{$self->{'tag'}}}) { - my $format = $tag->{'value'} || $tag->{'content'}; + marc('900','a', rec('200','a') ); + marc('900','b', rec('200','b') ); + marc('900','c', rec('200','c') ); - $log->debug("format: $format"); +which might not be what you have in mind. If you need repeatable subfield, +define it using C like this: - my @v; - if ($self->{'lookup_regex'} && $format =~ $self->{'lookup_regex'}) { - @v = $self->fill_in_to_arr($rec,$format); - } else { - @v = $self->parse_to_arr($rec,$format); - } - next if (! @v); +.... - if ($tag->{'sort'}) { - @v = $self->sort_arr(@v); - } +=cut - # use format? - if ($tag->{'format_name'}) { - @v = map { $self->apply_format($tag->{'format_name'},$tag->{'format_delimiter'},$_) } @v; - } +sub _get_marc_fields { - if ($field eq 'filename') { - $self->{'current_filename'} = join('',@v); - $log->debug("filename: ",$self->{'current_filename'}); - } elsif ($field eq 'headline') { - $self->{'headline'} .= join('',@v); - $log->debug("headline: ",$self->{'headline'}); - next; # don't return headline in data_structure! - } + return if (! $marc_record || ref($marc_record) ne 'ARRAY' || $#{ $marc_record } < 0); - # delimiter will join repeatable fields - if ($tag->{'delimiter'}) { - @v = ( join($tag->{'delimiter'}, @v) ); - } + # 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 }; - # default types - my @types = qw(display swish); - # 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}}, $last; - - } else { - push @{$row->{$type}}, @v; - } - } + @sorted_marc_record = @{ $marc_record }; ### 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 ), $/; + warn "## marc_record ", dump( $marc_record ), $/; + 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'}; - $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 = $sorted_marc_record[$i]; - push @ds, $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); - $log->debug("row $field: ",sub { Dumper($row) }); } + # 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 ($cache_file) { - store { - ds => \@ds, - current_filename => $self->{'current_filename'}, - headline => $self->{'headline'}, - }, $cache_file; - $log->debug("created storable cache file $cache_file"); + if ($#$field >= 0) { + push @m, $field; + warn "## saved/3 ", dump( $field ),$/ if ($debug); } - return @ds; + return @m; +} + +=head2 _debug + +Change level of debug warnings + + _debug( 2 ); + +=cut +sub _debug { + my $l = shift; + return $debug unless defined($l); + warn "debug level $l" if ($l > 0); + $debug = $l; } -=head2 apply_format +=head1 Functions to create C -Apply format specified in tag with C and -C. +Those functions generally have to first in your normalization file. - my $text = $webpac->apply_format($format_name,$format_delimiter,$data); +=head2 tag + +Define new tag for I and I. + + tag('Title', rec('200','a') ); -Formats can contain C if you need them. =cut -sub apply_format { - my $self = shift; +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; +} - my ($name,$delimiter,$data) = @_; +=head2 display - my $log = $self->_get_logger(); +Define tag just for I - if (! $self->{'import_xml'}->{'format'}->{$name}) { - $log->warn(" is not defined in ",$self->{'import_xml_file'}); - return $data; - } + @v = display('Title', rec('200','a') ); - $log->warn("no delimiter for format $name") if (! $delimiter); +=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; +} - my $format = $self->_x($self->{'import_xml'}->{'format'}->{$name}->{'content'}) || $log->logdie("can't find format '$name'"); +=head2 search - my @data = split(/\Q$delimiter\E/, $data); +Prepare values just for I - my $out = sprintf($format, @data); - $log->debug("using format $name [$format] on $data to produce: $out"); + @v = search('Title', rec('200','a') ); - if ($self->{'lookup_regex'} && $out =~ $self->{'lookup_regex'}) { - return $self->lookup($out); +=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; +} + +=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) { + $out->{' leader'}->{ $offset } = $value; } else { - return $out; + return $out->{' leader'}; } - } -=head2 parse +=head2 marc -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. +Save value for MARC field - my $text = $webpac->parse($rec,'eval{"v901^a" eq "Deskriptor"}descriptor: v250^a', $i); + marc('900','a', rec('200','a') ); =cut -sub parse { - my $self = shift; +sub marc { + my $f = shift or die "marc needs field"; + die "marc field must be numer" unless ($f =~ /^\d+$/); + + my $sf = shift or die "marc needs subfield"; - my ($rec, $format_utf8, $i) = @_; + 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 ]; + } +} - return if (! $format_utf8); +=head2 marc_repeatable_subfield - my $log = $self->_get_logger(); +Save values for MARC repetable subfield - $log->logconfess("need HASH as first argument!") if ($rec !~ /HASH/o); + marc_repeatable_subfield('910', 'z', rec('909') ); - $i = 0 if (! $i); +=cut - my $format = $self->_x($format_utf8) || $log->logconfess("can't convert '$format_utf8' from UTF-8 to ",$self->{'code_page'}); +sub marc_repeatable_subfield { + my ($f,$sf) = @_; + die "marc_repeatable_subfield need field and subfield!\n" unless ($f && $sf); + $marc_repeatable_subfield->{ $f . $sf }++; + marc(@_); +} - my @out; +=head2 marc_indicators - $log->debug("format: $format"); +Set both indicators for MARC field - my $eval_code; - # remove eval{...} from beginning - $eval_code = $1 if ($format =~ s/^eval{([^}]+)}//s); + marc_indicators('900', ' ', 1); - my $filter_name; - # remove filter{...} from beginning - $filter_name = $1 if ($format =~ s/^filter{([^}]+)}//s); +Any indicator value other than C<0-9> will be treated as undefined. - my $prefix; - my $all_found=0; +=cut - while ($format =~ s/^(.*?)(v|s)(\d+)(?:\^(\w))?//s) { +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)); - my $del = $1 || ''; - $prefix ||= $del if ($all_found == 0); + $i1 = ' ' if ($i1 !~ /^\d$/); + $i2 = ' ' if ($i2 !~ /^\d$/); + @{ $marc_indicators->{$f} } = ($i1,$i2); +} - # repeatable index - my $r = $i; - $r = 0 if (lc("$2") eq 's'); +=head2 marc_compose - my $found = 0; - my $tmp = $self->get_data(\$rec,$3,$4,$r,\$found); +Save values for each MARC subfield explicitly - if ($found) { - push @out, $del; - push @out, $tmp; - $all_found += $found; - } - } + marc_compose('900', + 'a', rec('200','a') + 'b', rec('201','a') + 'a', rec('200','b') + 'c', rec('200','c') + ); - return if (! $all_found); +=cut - my $out = join('',@out); +sub marc_compose { + my $f = shift or die "marc_compose needs field"; + die "marc_compose field must be numer" unless ($f =~ /^\d+$/); - if ($out) { - # add rest of format (suffix) - $out .= $format; + my ($i1,$i2) = defined($marc_indicators->{$f}) ? @{ $marc_indicators->{$f} } : (' ',' '); + my $m = [ $f, $i1, $i2 ]; - # add prefix if not there - $out = $prefix . $out if ($out !~ m/^\Q$prefix\E/); + while (@_) { + my $sf = shift or die "marc_compose $f needs subfield"; + my $v = shift; - $log->debug("result: $out"); + next unless (defined($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 ($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)); - } - - if ($filter_name && $self->{'filter'}->{$filter_name}) { - $log->debug("about to filter{$filter_name} format: $out"); - $out = $self->{'filter'}->{$filter_name}->($out); - return unless(defined($out)); - $log->debug("filter result: $out"); - } + warn "## marc_compose(d) ", dump( $m ) if ($debug > 1); - return $out; + push @{ $marc_record }, $m if ($#{$m} > 2); } -=head2 parse_to_arr -Similar to C, but returns array of all repeatable fields +=head1 Functions to extract data from input - my @arr = $webpac->parse_to_arr($rec,'v250^a'); +This function should be used inside functions to create C described +above. -=cut - -sub parse_to_arr { - my $self = shift; +=head2 rec1 - my ($rec, $format_utf8) = @_; +Return all values in some field - my $log = $self->_get_logger(); + @v = rec1('200') - $log->logconfess("need HASH as first argument!") if ($rec !~ /HASH/o); - return if (! $format_utf8); +TODO: order of values is probably same as in source data, need to investigate that - my $i = 0; - my @arr; +=cut - while (my $v = $self->parse($rec,$format_utf8,$i++)) { - push @arr, $v; +sub rec1 { + my $f = shift; + return unless (defined($rec) && defined($rec->{$f})); + if (ref($rec->{$f}) eq 'ARRAY') { + return map { + if (ref($_) eq 'HASH') { + values %{$_}; + } else { + $_; + } + } @{ $rec->{$f} }; + } elsif( defined($rec->{$f}) ) { + return $rec->{$f}; } +} + +=head2 rec2 - $log->debug("format '$format_utf8' returned ",--$i," elements: ", sub { join(" | ",@arr) }) if (@arr); +Return all values in specific field and subfield - return @arr; + @v = rec2('200','a') + +=cut + +sub rec2 { + my $f = shift; + return unless (defined($rec && $rec->{$f})); + my $sf = shift; + return map { $_->{$sf} } grep { ref($_) eq 'HASH' && $_->{$sf} } @{ $rec->{$f} }; } -=head2 fill_in_to_arr +=head2 rec -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. +syntaxtic sugar for - my @arr = $webpac->fill_in_to_arr($rec,'[v900];;[v250^a]'); + @v = rec('200') + @v = rec('200','a') =cut -sub fill_in_to_arr { - my $self = shift; +sub rec { + if ($#_ == 0) { + return rec1(@_); + } elsif ($#_ == 1) { + return rec2(@_); + } +} - my ($rec, $format_utf8) = @_; +=head2 regex - my $log = $self->_get_logger(); +Apply regex to some or all values - $log->logconfess("need HASH as first argument!") if ($rec !~ /HASH/o); - return if (! $format_utf8); + @v = regex( 's/foo/bar/g', @v ); - my $i = 0; - my @arr; +=cut - while (my @v = $self->fill_in($rec,$format_utf8,$i++)) { - push @arr, @v; +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; +} - $log->debug("format '$format_utf8' returned ",--$i," elements: ", sub { join(" | ",@arr) }) if (@arr); +=head2 prefix - return @arr; +Prefix all values with a string + + @v = prefix( 'my_', @v ); + +=cut + +sub prefix { + my $p = shift or die "prefix needs string as first argument"; + return map { $p . $_ } grep { defined($_) } @_; } -=head2 sort_arr +=head2 suffix -Sort array ignoring case and html in data +suffix all values with a string - my @sorted = $webpac->sort_arr(@unsorted); + @v = suffix( '_my', @v ); =cut -sub sort_arr { - my $self = shift; +sub suffix { + my $s = shift or die "suffix needs string as first argument"; + return map { $_ . $s } grep { defined($_) } @_; +} + +=head2 surround - my $log = $self->_get_logger(); +surround all values with a two strings - # FIXME add Schwartzian Transformation? + @v = surround( 'prefix_', '_suffix', @v ); - my @sorted = sort { - $a =~ s#<[^>]+/*>##; - $b =~ s#<[^>]+/*>##; - lc($b) cmp lc($a) - } @_; - $log->debug("sorted values: ",sub { join(", ",@sorted) }); +=cut - return @sorted; +sub surround { + my $p = shift or die "surround need prefix as first argument"; + my $s = shift or die "surround needs suffix as second argument"; + return map { $p . $_ . $s } grep { defined($_) } @_; } +=head2 first -=head2 _sort_by_order +Return first element -Sort xml tags data structure accoding to C attribute. + $v = first( @v ); =cut -sub _sort_by_order { - my $self = shift; +sub first { + my $r = shift; + return $r; +} - 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}; +=head2 lookup - return $va <=> $vb; -} +Consult lookup hashes for some value + + @v = lookup( $v ); + @v = lookup( @v ); + +=cut -=head2 _x +sub lookup { + my $k = shift or return; + return unless (defined($lookup->{$k})); + if (ref($lookup->{$k}) eq 'ARRAY') { + return @{ $lookup->{$k} }; + } else { + return $lookup->{$k}; + } +} -Convert strings from C encoding into application specific -(optinally specified using C to C constructor. +=head2 join_with - my $text = $n->_x('normalize text string'); +Joins walues with some delimiter -This is a stub so that other modules doesn't have to implement it. + $v = join_with(", ", @v); =cut -sub _x { - my $self = shift; - return shift; +sub join_with { + my $d = shift; + return join($d, grep { defined($_) && $_ ne '' } @_); } +=head2 split_rec_on -=head1 AUTHOR - -Dobrica Pavlinusic, C<< >> +Split record subfield on some regex and take one of parts out -=head1 COPYRIGHT & LICENSE + $a_before_semi_column = + split_rec_on('200','a', /\s*;\s*/, $part); -Copyright 2005 Dobrica Pavlinusic, All Rights Reserved. +C<$part> is optional number of element. First element is +B<1>, not 0! -This program is free software; you can redistribute it and/or modify it -under the same terms as Perl itself. +If there is no C<$part> parameter or C<$part> is 0, this function will +return all values produced by splitting. =cut -1; # End of WebPAC::DB +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 ]; + } +} + +# END +1;