| 1 |
package WebPAC::Normalize; |
package WebPAC::Normalize; |
| 2 |
|
use Exporter 'import'; |
| 3 |
|
@EXPORT = qw/ |
| 4 |
|
_set_rec _set_lookup |
| 5 |
|
_get_ds _clean_ds |
| 6 |
|
|
| 7 |
|
tag search display |
| 8 |
|
marc21 |
| 9 |
|
|
| 10 |
|
rec1 rec2 rec |
| 11 |
|
regex prefix suffix surround |
| 12 |
|
first lookup join_with |
| 13 |
|
/; |
| 14 |
|
|
| 15 |
use warnings; |
use warnings; |
| 16 |
use strict; |
use strict; |
| 17 |
use blib; |
|
| 18 |
use WebPAC::Common; |
#use base qw/WebPAC::Common/; |
|
use base 'WebPAC::Common'; |
|
| 19 |
use Data::Dumper; |
use Data::Dumper; |
| 20 |
|
use Encode qw/from_to/; |
| 21 |
|
|
| 22 |
=head1 NAME |
=head1 NAME |
| 23 |
|
|
| 24 |
WebPAC::Normalize - data mungling for normalisation |
WebPAC::Normalize - describe normalisaton rules using sets |
| 25 |
|
|
| 26 |
=head1 VERSION |
=head1 VERSION |
| 27 |
|
|
| 28 |
Version 0.08 |
Version 0.06 |
| 29 |
|
|
| 30 |
=cut |
=cut |
| 31 |
|
|
| 32 |
our $VERSION = '0.08'; |
our $VERSION = '0.06'; |
| 33 |
|
|
| 34 |
=head1 SYNOPSIS |
=head1 SYNOPSIS |
| 35 |
|
|
| 36 |
This package contains code that mungle data to produce normalized format. |
This module uses C<conf/normalize/*.pl> files to perform normalisation |
| 37 |
|
from input records using perl functions which are specialized for set |
| 38 |
It contains several assumptions: |
processing. |
| 39 |
|
|
| 40 |
=over |
Sets are implemented as arrays, and normalisation file is valid perl, which |
| 41 |
|
means that you check it's validity before running WebPAC using |
| 42 |
=item * |
C<perl -c normalize.pl>. |
| 43 |
|
|
| 44 |
format of fields is defined using C<v123^a> notation for repeatable fields |
Normalisation can generate multiple output normalized data. For now, supported output |
| 45 |
or C<s123^a> for single (or first) value, where C<123> is field number and |
types (on the left side of definition) are: C<tag>, C<display>, C<search> and |
| 46 |
C<a> is subfield. |
C<marc21>. |
|
|
|
|
=item * |
|
|
|
|
|
source data records (C<$rec>) have unique identifiers in field C<000> |
|
|
|
|
|
=item * |
|
|
|
|
|
optional C<eval{length('v123^a') == 3}> tag at B<beginning of format> will be |
|
|
perl code that is evaluated before producing output (value of field will be |
|
|
interpolated before that) |
|
|
|
|
|
=item * |
|
|
|
|
|
optional C<filter{filter_name}> at B<begining of format> will apply perl |
|
|
code defined as code ref on format after field substitution to producing |
|
|
output |
|
|
|
|
|
There is one built-in filter called C<regex> which can be use like this: |
|
|
|
|
|
filter{regex(s/foo/bar/)} |
|
|
|
|
|
=item * |
|
|
|
|
|
optional C<lookup{...}> will be then performed. See C<WebPAC::Lookups>. |
|
|
|
|
|
=item * |
|
|
|
|
|
at end, optional C<format>s rules are resolved. Format rules are similar to |
|
|
C<sprintf> and can also contain C<lookup{...}> which is performed after |
|
|
values are inserted in format. |
|
|
|
|
|
=back |
|
|
|
|
|
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. |
|
|
|
|
|
|
|
|
|
|
| 47 |
|
|
| 48 |
=head1 FUNCTIONS |
=head1 FUNCTIONS |
| 49 |
|
|
| 50 |
=head2 new |
Functions which start with C<_> are private and used by WebPAC internally. |
| 51 |
|
All other functions are available for use within normalisation rules. |
| 52 |
|
|
| 53 |
|
=head2 data_structure |
| 54 |
|
|
| 55 |
Create new normalisation object |
Return data structure |
| 56 |
|
|
| 57 |
my $n = new WebPAC::Normalize::Something( |
my $ds = WebPAC::Normalize::data_structure( |
| 58 |
filter => { |
lookup => $lookup->lookup_hash, |
| 59 |
'filter_name_1' => sub { |
row => $row, |
| 60 |
# filter code |
rules => $normalize_pl_config, |
| 61 |
return length($_); |
marc_encoding => 'utf-8', |
|
}, ... |
|
|
}, |
|
|
db => $db_obj, |
|
|
lookup_regex => $lookup->regex, |
|
|
lookup => $lookup_obj, |
|
|
prefix => 'foobar', |
|
| 62 |
); |
); |
| 63 |
|
|
| 64 |
Parametar C<filter> defines user supplied snippets of perl code which can |
Options C<lookup>, C<row>, C<rules> and C<log> are mandatory while all |
| 65 |
be use with C<filter{...}> notation. |
other are optional. |
| 66 |
|
|
| 67 |
C<prefix> is used to form filename for database record (to support multiple |
This function will B<die> if normalizastion can't be evaled. |
|
source files which are joined in one database). |
|
| 68 |
|
|
| 69 |
Recommended parametar C<lookup_regex> is used to enable parsing of lookups |
Since this function isn't exported you have to call it with |
| 70 |
in structures. If you pass this parametar, you must also pass C<lookup> |
C<WebPAC::Normalize::data_structure>. |
|
which is C<WebPAC::Lookup> object. |
|
| 71 |
|
|
| 72 |
=cut |
=cut |
| 73 |
|
|
| 74 |
sub new { |
sub data_structure { |
| 75 |
my $class = shift; |
my $arg = {@_}; |
|
my $self = {@_}; |
|
|
bless($self, $class); |
|
|
|
|
|
my $r = $self->{'lookup_regex'} ? 1 : 0; |
|
|
my $l = $self->{'lookup'} ? 1 : 0; |
|
|
|
|
|
my $log = $self->_get_logger(); |
|
|
|
|
|
# those two must be in pair |
|
|
if ( ($r & $l) != ($r || $l) ) { |
|
|
my $log = $self->_get_logger(); |
|
|
$log->logdie("lookup_regex and lookup must be in pair"); |
|
|
} |
|
|
|
|
|
$log->logdie("lookup must be WebPAC::Lookup object") if ($self->{'lookup'} && ! $self->{'lookup'}->isa('WebPAC::Lookup')); |
|
|
|
|
|
$log->warn("no prefix defined. please check that!") unless ($self->{'prefix'}); |
|
| 76 |
|
|
| 77 |
$log->debug("using lookup regex: ", $self->{lookup_regex}) if ($r && $l); |
die "need row argument" unless ($arg->{row}); |
| 78 |
|
die "need normalisation argument" unless ($arg->{rules}); |
| 79 |
|
|
| 80 |
if (! $self->{filter} || ! $self->{filter}->{regex}) { |
no strict 'subs'; |
| 81 |
$log->debug("adding built-in filter regex"); |
_set_lookup( $arg->{lookup} ); |
| 82 |
$self->{filter}->{regex} = sub { |
_set_rec( $arg->{row} ); |
| 83 |
my ($val, $regex) = @_; |
_clean_ds( %{ $arg } ); |
| 84 |
eval "\$val =~ $regex"; |
eval "$arg->{rules}"; |
| 85 |
return $val; |
die "error evaling $arg->{rules}: $@\n" if ($@); |
|
}; |
|
|
} |
|
| 86 |
|
|
| 87 |
$self ? return $self : return undef; |
return _get_ds(); |
| 88 |
} |
} |
| 89 |
|
|
| 90 |
|
=head2 _set_rec |
| 91 |
|
|
| 92 |
=head2 data_structure |
Set current record hash |
|
|
|
|
Create in-memory data structure which represents normalized layout from |
|
|
C<conf/normalize/*.xml>. |
|
|
|
|
|
This structures are used to produce output. |
|
| 93 |
|
|
| 94 |
my $ds = $webpac->data_structure($rec); |
_set_rec( $rec ); |
| 95 |
|
|
| 96 |
=cut |
=cut |
| 97 |
|
|
| 98 |
sub data_structure { |
my $rec; |
|
my $self = shift; |
|
|
|
|
|
my $log = $self->_get_logger(); |
|
| 99 |
|
|
| 100 |
my $rec = shift; |
sub _set_rec { |
| 101 |
$log->logconfess("need HASH as first argument!") if ($rec !~ /HASH/o); |
$rec = shift or die "no record hash"; |
| 102 |
|
} |
| 103 |
|
|
| 104 |
$log->debug("data_structure rec = ", sub { Dumper($rec) }); |
=head2 _get_ds |
| 105 |
|
|
| 106 |
$log->logdie("need unique ID (mfn) in field 000 of record " . Dumper($rec) ) unless (defined($rec->{'000'})); |
Return hash formatted as data structure |
| 107 |
|
|
| 108 |
my $id = $rec->{'000'}->[0] || $log->logdie("field 000 isn't array!"); |
my $ds = _get_ds(); |
| 109 |
|
|
| 110 |
my $cache_file; |
=cut |
| 111 |
|
|
| 112 |
if ($self->{'db'}) { |
my $out; |
| 113 |
my $ds = $self->{'db'}->load_ds( id => $id, prefix => $self->{prefix} ); |
my $marc21; |
| 114 |
$log->debug("load_ds( rec = ", sub { Dumper($rec) }, ") = ", sub { Dumper($ds) }); |
my $marc_encoding; |
|
return $ds if ($ds); |
|
|
$log->debug("cache miss, creating"); |
|
|
} |
|
| 115 |
|
|
| 116 |
my @sorted_tags; |
sub _get_ds { |
| 117 |
if ($self->{tags_by_order}) { |
return $out; |
| 118 |
@sorted_tags = @{$self->{tags_by_order}}; |
} |
|
} else { |
|
|
@sorted_tags = sort { $self->_sort_by_order } keys %{$self->{'import_xml'}->{'indexer'}}; |
|
|
$self->{tags_by_order} = \@sorted_tags; |
|
|
} |
|
|
|
|
|
my $ds; |
|
| 119 |
|
|
| 120 |
$log->debug("tags: ",sub { join(", ",@sorted_tags) }); |
=head2 _clean_ds |
| 121 |
|
|
| 122 |
foreach my $field (@sorted_tags) { |
Clean data structure hash for next record |
| 123 |
|
|
| 124 |
my $row; |
_clean_ds(); |
| 125 |
|
|
| 126 |
#print "field $field [",$self->{'tag'},"] = ",Dumper($self->{'import_xml'}->{'indexer'}->{$field}->{$self->{'tag'}}); |
=cut |
| 127 |
|
|
| 128 |
foreach my $tag (@{$self->{'import_xml'}->{'indexer'}->{$field}->{$self->{'tag'}}}) { |
sub _clean_ds { |
| 129 |
my $format; |
my $a = {@_}; |
| 130 |
|
$out = undef; |
| 131 |
|
$marc21 = undef; |
| 132 |
|
$marc_encoding = $a->{marc_encoding}; |
| 133 |
|
} |
| 134 |
|
|
| 135 |
$log->logdie("expected tag HASH and got $tag") unless (ref($tag) eq 'HASH'); |
=head2 _set_lookup |
|
$format = $tag->{'value'} || $tag->{'content'}; |
|
| 136 |
|
|
| 137 |
my @v; |
Set current lookup hash |
|
if ($self->{'lookup_regex'} && $format =~ $self->{'lookup_regex'}) { |
|
|
@v = $self->fill_in_to_arr($rec,$format); |
|
|
} else { |
|
|
@v = $self->parse_to_arr($rec,$format); |
|
|
} |
|
|
if (! @v) { |
|
|
$log->debug("$field <",$self->{tag},"> format: $format no values"); |
|
|
# next; |
|
|
} else { |
|
|
$log->debug("$field <",$self->{tag},"> format: $format values: ", join(",", @v)); |
|
|
} |
|
| 138 |
|
|
| 139 |
if ($tag->{'sort'}) { |
_set_lookup( $lookup ); |
|
@v = $self->sort_arr(@v); |
|
|
} |
|
| 140 |
|
|
| 141 |
# use format? |
=cut |
|
if ($tag->{'format_name'}) { |
|
|
@v = map { $self->apply_format($tag->{'format_name'},$tag->{'format_delimiter'},$_) } @v; |
|
|
} |
|
| 142 |
|
|
| 143 |
# delimiter will join repeatable fields |
my $lookup; |
|
if ($tag->{'delimiter'}) { |
|
|
@v = ( join($tag->{'delimiter'}, @v) ); |
|
|
} |
|
| 144 |
|
|
| 145 |
# default types |
sub _set_lookup { |
| 146 |
my @types = qw(display search); |
$lookup = shift; |
| 147 |
# override by type attribute |
} |
|
@types = ( $tag->{'type'} ) if ($tag->{'type'}); |
|
|
|
|
|
foreach my $type (@types) { |
|
|
# append to previous line? |
|
|
$log->debug("tag $field / $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->{$type}}, @v; |
|
|
} |
|
|
} |
|
| 148 |
|
|
| 149 |
|
=head2 _get_marc21_fields |
| 150 |
|
|
| 151 |
} |
Get all fields defined by calls to C<marc21> |
| 152 |
|
|
| 153 |
if ($row) { |
$marc->add_fields( WebPAC::Normalize:_get_marc21_fields() ); |
|
$row->{'tag'} = $field; |
|
| 154 |
|
|
| 155 |
# TODO: name_sigular, name_plural |
B<TODO>: implement exceptions to magic which unrolls repeated subfields |
| 156 |
my $name = $self->{'import_xml'}->{'indexer'}->{$field}->{'name'}; |
as new field with that subfield. |
|
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"); |
|
|
} |
|
| 157 |
|
|
| 158 |
$ds->{$row_name} = $row; |
=cut |
| 159 |
|
|
| 160 |
$log->debug("row $field: ",sub { Dumper($row) }); |
sub _get_marc21_fields { |
| 161 |
|
my @m; |
| 162 |
|
my $last; |
| 163 |
|
foreach my $row (@{ $marc21 }) { |
| 164 |
|
if ($last && ( |
| 165 |
|
$last->[0] eq $row->[0] || # check if field is same |
| 166 |
|
$last->[1] eq $row->[1] || # check for i1 |
| 167 |
|
$last->[2] eq $row->[2] # and for i2 |
| 168 |
|
) ) { |
| 169 |
|
$last->[3]->{ $row->[3] } = $row->[4]; |
| 170 |
|
warn "## ++ added $row->[0] ^$row->[3]\n"; |
| 171 |
|
next; |
| 172 |
|
} elsif ($last) { |
| 173 |
|
push @m, $last; |
| 174 |
} |
} |
| 175 |
|
|
| 176 |
|
push @m, $row; |
| 177 |
} |
} |
| 178 |
|
|
| 179 |
$self->{'db'}->save_ds( |
push @m, $last if ($last); |
|
id => $id, |
|
|
ds => $ds, |
|
|
prefix => $self->{prefix}, |
|
|
) if ($self->{'db'}); |
|
|
|
|
|
$log->debug("ds: ", sub { Dumper($ds) }); |
|
|
|
|
|
$log->logconfess("data structure returned is not array any more!") if wantarray; |
|
|
|
|
|
return $ds; |
|
| 180 |
|
|
| 181 |
|
return @m; |
| 182 |
} |
} |
| 183 |
|
|
| 184 |
=head2 parse |
=head1 Functions to create C<data_structure> |
|
|
|
|
Perform smart parsing of string, skipping delimiters for fields which aren't |
|
|
defined. It can also eval code in format starting with C<eval{...}> and |
|
|
return output or nothing depending on eval code. |
|
| 185 |
|
|
| 186 |
my $text = $webpac->parse($rec,'eval{"v901^a" eq "Deskriptor"}descriptor: v250^a', $i); |
Those functions generally have to first in your normalization file. |
| 187 |
|
|
| 188 |
Filters are implemented here. While simple form of filters looks like this: |
=head2 tag |
| 189 |
|
|
| 190 |
filter{name_of_filter} |
Define new tag for I<search> and I<display>. |
| 191 |
|
|
| 192 |
but, filters can also have variable number of parametars like this: |
tag('Title', rec('200','a') ); |
| 193 |
|
|
|
filter{name_of_filter(param,param,param)} |
|
| 194 |
|
|
| 195 |
=cut |
=cut |
| 196 |
|
|
| 197 |
my $warn_once; |
sub tag { |
| 198 |
|
my $name = shift or die "tag needs name as first argument"; |
| 199 |
sub parse { |
my @o = grep { defined($_) && $_ ne '' } @_; |
| 200 |
my $self = shift; |
return unless (@o); |
| 201 |
|
$out->{$name}->{tag} = $name; |
| 202 |
my ($rec, $format_utf8, $i, $rec_size) = @_; |
$out->{$name}->{search} = \@o; |
| 203 |
|
$out->{$name}->{display} = \@o; |
| 204 |
return if (! $format_utf8); |
} |
|
|
|
|
my $log = $self->_get_logger(); |
|
|
|
|
|
$log->logconfess("need HASH as first argument!") if ($rec !~ /HASH/o); |
|
|
|
|
|
$i = 0 if (! $i); |
|
|
|
|
|
my $format = $self->_x($format_utf8) || $log->logconfess("can't convert '$format_utf8' from UTF-8 to ",$self->{'code_page'}); |
|
|
|
|
|
my @out; |
|
|
|
|
|
$log->debug("format: $format [$i]"); |
|
|
|
|
|
my $eval_code; |
|
|
# remove eval{...} from beginning |
|
|
$eval_code = $1 if ($format =~ s/^eval{([^}]+)}//s); |
|
|
|
|
|
my $filter_name; |
|
|
# remove filter{...} from beginning |
|
|
$filter_name = $1 if ($format =~ s/^filter{([^}]+)}//s); |
|
|
|
|
|
# did we found any (att all) field from format in row? |
|
|
my $found_any; |
|
|
# prefix before first field which we preserve it $found_any |
|
|
my $prefix; |
|
|
|
|
|
my $f_step = 1; |
|
|
|
|
|
while ($format =~ s/^(.*?)(v|s)(\d+)(?:\^(\w))?//s) { |
|
|
|
|
|
my $del = $1 || ''; |
|
|
$prefix = $del if ($f_step == 1); |
|
|
|
|
|
my $fld_type = lc($2); |
|
| 205 |
|
|
| 206 |
# repeatable index |
=head2 display |
|
my $r = $i; |
|
|
if ($fld_type eq 's') { |
|
|
if ($found_any->{'v'}) { |
|
|
$r = 0; |
|
|
} else { |
|
|
return; |
|
|
} |
|
|
} |
|
| 207 |
|
|
| 208 |
my $found = 0; |
Define tag just for I<display> |
|
my $tmp = $self->get_data(\$rec,$3,$4,$r,\$found,$rec_size); |
|
| 209 |
|
|
| 210 |
if ($found) { |
@v = display('Title', rec('200','a') ); |
|
$found_any->{$fld_type} += $found; |
|
| 211 |
|
|
| 212 |
# we will skip delimiter before first occurence of field! |
=cut |
|
push @out, $del unless($found_any->{$fld_type} == 1); |
|
|
push @out, $tmp; |
|
|
} |
|
|
$f_step++; |
|
|
} |
|
| 213 |
|
|
| 214 |
# test if any fields found? |
sub display { |
| 215 |
return if (! $found_any->{'v'} && ! $found_any->{'s'}); |
my $name = shift or die "display needs name as first argument"; |
| 216 |
|
my @o = grep { defined($_) && $_ ne '' } @_; |
| 217 |
|
return unless (@o); |
| 218 |
|
$out->{$name}->{tag} = $name; |
| 219 |
|
$out->{$name}->{display} = \@o; |
| 220 |
|
} |
| 221 |
|
|
| 222 |
my $out = join('',@out); |
=head2 search |
| 223 |
|
|
| 224 |
if ($out) { |
Prepare values just for I<search> |
|
# add rest of format (suffix) |
|
|
$out .= $format; |
|
| 225 |
|
|
| 226 |
# add prefix if not there |
@v = search('Title', rec('200','a') ); |
|
$out = $prefix . $out if ($out !~ m/^\Q$prefix\E/); |
|
| 227 |
|
|
| 228 |
$log->debug("result: $out"); |
=cut |
|
} |
|
|
|
|
|
if ($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)); |
|
|
} |
|
|
|
|
|
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}++; |
|
|
} |
|
|
} |
|
| 229 |
|
|
| 230 |
return $out; |
sub search { |
| 231 |
|
my $name = shift or die "search needs name as first argument"; |
| 232 |
|
my @o = grep { defined($_) && $_ ne '' } @_; |
| 233 |
|
return unless (@o); |
| 234 |
|
$out->{$name}->{tag} = $name; |
| 235 |
|
$out->{$name}->{search} = \@o; |
| 236 |
} |
} |
| 237 |
|
|
| 238 |
=head2 parse_to_arr |
=head2 marc21 |
| 239 |
|
|
| 240 |
Similar to C<parse>, but returns array of all repeatable fields |
Save value for MARC field |
| 241 |
|
|
| 242 |
my @arr = $webpac->parse_to_arr($rec,'v250^a'); |
marc21('900','a', rec('200','a') ); |
| 243 |
|
|
| 244 |
=cut |
=cut |
| 245 |
|
|
| 246 |
sub parse_to_arr { |
sub marc21 { |
| 247 |
my $self = shift; |
my $f = shift or die "marc21 needs field"; |
| 248 |
|
die "marc21 field must be numer" unless ($f =~ /^\d+$/); |
|
my ($rec, $format_utf8) = @_; |
|
| 249 |
|
|
| 250 |
my $log = $self->_get_logger(); |
my $sf = shift or die "marc21 needs subfield"; |
| 251 |
|
|
| 252 |
$log->logconfess("need HASH as first argument!") if ($rec !~ /HASH/o); |
foreach (@_) { |
| 253 |
return if (! $format_utf8); |
my $v = $_; # make var read-write for Encode |
| 254 |
|
next unless (defined($v) && $v !~ /^\s+$/); |
| 255 |
my $i = 0; |
from_to($v, 'iso-8859-2', $marc_encoding) if ($marc_encoding); |
| 256 |
my @arr; |
push @{ $marc21 }, [ $f, ' ', ' ', $sf => $v ]; |
|
|
|
|
my $rec_size = { '_' => '_' }; |
|
|
|
|
|
while (my $v = $self->parse($rec,$format_utf8,$i++,\$rec_size)) { |
|
|
push @arr, $v; |
|
|
warn "parse rec_size = ", Dumper($rec_size); |
|
| 257 |
} |
} |
|
|
|
|
$log->debug("format '$format_utf8' returned ",--$i," elements: ", sub { join(" | ",@arr) }) if (@arr); |
|
|
|
|
|
return @arr; |
|
| 258 |
} |
} |
| 259 |
|
|
| 260 |
|
=head1 Functions to extract data from input |
| 261 |
|
|
| 262 |
=head2 fill_in |
This function should be used inside functions to create C<data_structure> described |
| 263 |
|
above. |
|
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. |
|
| 264 |
|
|
| 265 |
my $text = $webpac->fill_in($rec,'v250^a'); |
=head2 rec1 |
| 266 |
|
|
| 267 |
Optional argument is ordinal number for repeatable fields. By default, |
Return all values in some field |
|
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. |
|
| 268 |
|
|
| 269 |
my $text = $webpac->fill_in($rec,'Title: v250^a',1); |
@v = rec1('200') |
| 270 |
|
|
| 271 |
This function B<does not> perform parsing of format to inteligenty skip |
TODO: order of values is probably same as in source data, need to investigate that |
|
delimiters before fields which aren't used. |
|
|
|
|
|
This method will automatically decode UTF-8 string to local code page |
|
|
if needed. |
|
|
|
|
|
There is optional parametar C<$record_size> which can be used to get sizes of |
|
|
all C<field^subfield> combinations in this format. |
|
|
|
|
|
my $text = $webpac->fill_in($rec,'got: v900^a v900^x',0,\$rec_size); |
|
| 272 |
|
|
| 273 |
=cut |
=cut |
| 274 |
|
|
| 275 |
sub fill_in { |
sub rec1 { |
| 276 |
my $self = shift; |
my $f = shift; |
| 277 |
|
return unless (defined($rec) && defined($rec->{$f})); |
| 278 |
my $log = $self->_get_logger(); |
if (ref($rec->{$f}) eq 'ARRAY') { |
| 279 |
|
return map { |
| 280 |
my ($rec,$format,$i,$rec_size) = @_; |
if (ref($_) eq 'HASH') { |
| 281 |
|
values %{$_}; |
|
$log->logconfess("need data record") unless ($rec); |
|
|
$log->logconfess("need format to parse") unless($format); |
|
|
|
|
|
# iteration (for repeatable fields) |
|
|
$i ||= 0; |
|
|
|
|
|
$log->logdie("infitite loop in format $format") if ($i > ($self->{'max_mfn'} || 9999)); |
|
|
|
|
|
# FIXME remove for speedup? |
|
|
$log->logconfess("need HASH as first argument!") if ($rec !~ /HASH/o); |
|
|
|
|
|
if (utf8::is_utf8($format)) { |
|
|
$format = $self->_x($format); |
|
|
} |
|
|
|
|
|
my $found = 0; |
|
|
my $just_single = 1; |
|
|
|
|
|
my $eval_code; |
|
|
# remove eval{...} from beginning |
|
|
$eval_code = $1 if ($format =~ s/^eval{([^}]+)}//s); |
|
|
|
|
|
my $filter_name; |
|
|
# remove filter{...} from beginning |
|
|
$filter_name = $1 if ($format =~ s/^filter{([^}]+)}//s); |
|
|
|
|
|
# do actual replacement of placeholders |
|
|
# repeatable fields |
|
|
if ($format =~ s/v(\d+)(?:\^(\w))?/$self->get_data(\$rec,$1,$2,$i,\$found,$rec_size)/ges) { |
|
|
$just_single = 0; |
|
|
} |
|
|
|
|
|
# non-repeatable fields |
|
|
if ($format =~ s/s(\d+)(?:\^(\w))?/$self->get_data(\$rec,$1,$2,0,\$found,$rec_size)/ges) { |
|
|
return if ($i > 0 && $just_single); |
|
|
} |
|
|
|
|
|
if ($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; |
|
| 282 |
} else { |
} else { |
| 283 |
$log->warn("Have lookup object but can't invoke lookup method"); |
$_; |
| 284 |
} |
} |
| 285 |
} else { |
} @{ $rec->{$f} }; |
| 286 |
return $format; |
} elsif( defined($rec->{$f}) ) { |
| 287 |
} |
return $rec->{$f}; |
|
} else { |
|
|
return; |
|
| 288 |
} |
} |
| 289 |
} |
} |
| 290 |
|
|
| 291 |
|
=head2 rec2 |
| 292 |
|
|
| 293 |
=head2 fill_in_to_arr |
Return all values in specific field and subfield |
|
|
|
|
Similar to C<fill_in>, but returns array of all repeatable fields. Usable |
|
|
for fields which have lookups, so they shouldn't be parsed but rather |
|
|
C<fill_id>ed. |
|
| 294 |
|
|
| 295 |
my @arr = $webpac->fill_in_to_arr($rec,'[v900];;[v250^a]'); |
@v = rec2('200','a') |
| 296 |
|
|
| 297 |
=cut |
=cut |
| 298 |
|
|
| 299 |
sub fill_in_to_arr { |
sub rec2 { |
| 300 |
my $self = shift; |
my $f = shift; |
| 301 |
|
return unless (defined($rec && $rec->{$f})); |
| 302 |
my ($rec, $format_utf8) = @_; |
my $sf = shift; |
| 303 |
|
return map { $_->{$sf} } grep { ref($_) eq 'HASH' && $_->{$sf} } @{ $rec->{$f} }; |
| 304 |
|
} |
| 305 |
|
|
| 306 |
my $log = $self->_get_logger(); |
=head2 rec |
| 307 |
|
|
| 308 |
$log->logconfess("need HASH as first argument!") if ($rec !~ /HASH/o); |
syntaxtic sugar for |
|
return if (! $format_utf8); |
|
| 309 |
|
|
| 310 |
my $i = 0; |
@v = rec('200') |
| 311 |
my @arr; |
@v = rec('200','a') |
| 312 |
|
|
| 313 |
my $rec_size; |
=cut |
| 314 |
|
|
| 315 |
while (my $v = $self->fill_in($rec,$format_utf8,$i,\$rec_size)) { |
sub rec { |
| 316 |
push @arr, $v; |
if ($#_ == 0) { |
| 317 |
warn "rec_size = ", Dumper($rec_size); |
return rec1(@_); |
| 318 |
|
} elsif ($#_ == 1) { |
| 319 |
|
return rec2(@_); |
| 320 |
} |
} |
|
|
|
|
$log->debug("format '$format_utf8' returned ",--$i," elements: ", sub { join(" | ",@arr) }) if (@arr); |
|
|
|
|
|
return @arr; |
|
| 321 |
} |
} |
| 322 |
|
|
| 323 |
|
=head2 regex |
| 324 |
|
|
| 325 |
=head2 get_data |
Apply regex to some or all values |
|
|
|
|
Returns value from record. |
|
|
|
|
|
my $text = $self->get_data(\$rec,$f,$sf,$i,\$found,\$rec_size); |
|
|
|
|
|
Required arguments are: |
|
|
|
|
|
=over 8 |
|
|
|
|
|
=item C<$rec> |
|
|
|
|
|
record reference |
|
|
|
|
|
=item C<$f> |
|
| 326 |
|
|
| 327 |
field |
@v = regex( 's/foo/bar/g', @v ); |
|
|
|
|
=item C<$sf> |
|
|
|
|
|
optional subfield |
|
|
|
|
|
=item C<$i> |
|
|
|
|
|
index offset for repeatable values ( 0 ... $rec_size->{'400^a'} ) |
|
|
|
|
|
=item C<$found> |
|
|
|
|
|
optional variable that will be incremeted if preset |
|
|
|
|
|
=item C<$rec_size> |
|
|
|
|
|
hash to hold maximum occurances of C<field^subfield> combinations |
|
|
(which can be accessed using keys in same format) |
|
|
|
|
|
=back |
|
|
|
|
|
Returns value or empty string, updates C<$found> and C<rec_size> |
|
|
if present. |
|
| 328 |
|
|
| 329 |
=cut |
=cut |
| 330 |
|
|
| 331 |
sub get_data { |
sub regex { |
| 332 |
my $self = shift; |
my $r = shift; |
| 333 |
|
my @out; |
| 334 |
my ($rec,$f,$sf,$i,$found,$cache) = @_; |
#warn "r: $r\n",Dumper(\@_); |
| 335 |
|
foreach my $t (@_) { |
| 336 |
return '' unless ($$rec->{$f} && ref($$rec->{$f}) eq 'ARRAY'); |
next unless ($t); |
| 337 |
|
eval "\$t =~ $r"; |
| 338 |
if (defined($$cache)) { |
push @out, $t if ($t && $t ne ''); |
|
$$cache->{ $f . ( $sf ? '^' . $sf : '' ) } ||= scalar @{ $$rec->{$f} }; |
|
|
} |
|
|
|
|
|
return '' unless ($$rec->{$f}->[$i]); |
|
|
|
|
|
{ |
|
|
no strict 'refs'; |
|
|
if (defined($sf)) { |
|
|
$$found++ if (defined($$found) && $$rec->{$f}->[$i]->{$sf}); |
|
|
return $$rec->{$f}->[$i]->{$sf}; |
|
|
} else { |
|
|
$$found++ if (defined($$found)); |
|
|
# it still might have subfields, just |
|
|
# not specified, so we'll dump some debug info |
|
|
if ($$rec->{$f}->[$i] =~ /HASH/o) { |
|
|
my $out; |
|
|
foreach my $k (keys %{$$rec->{$f}->[$i]}) { |
|
|
$out .= '$' . $k .':' . $$rec->{$f}->[$i]->{$k}." "; |
|
|
} |
|
|
return $out; |
|
|
} else { |
|
|
return $$rec->{$f}->[$i]; |
|
|
} |
|
|
} |
|
| 339 |
} |
} |
| 340 |
|
return @out; |
| 341 |
} |
} |
| 342 |
|
|
| 343 |
|
=head2 prefix |
| 344 |
|
|
| 345 |
=head2 apply_format |
Prefix all values with a string |
|
|
|
|
Apply format specified in tag with C<format_name="name"> and |
|
|
C<format_delimiter=";;">. |
|
| 346 |
|
|
| 347 |
my $text = $webpac->apply_format($format_name,$format_delimiter,$data); |
@v = prefix( 'my_', @v ); |
|
|
|
|
Formats can contain C<lookup{...}> if you need them. |
|
| 348 |
|
|
| 349 |
=cut |
=cut |
| 350 |
|
|
| 351 |
sub apply_format { |
sub prefix { |
| 352 |
my $self = shift; |
my $p = shift or die "prefix needs string as first argument"; |
| 353 |
|
return map { $p . $_ } grep { defined($_) } @_; |
| 354 |
my ($name,$delimiter,$data) = @_; |
} |
|
|
|
|
my $log = $self->_get_logger(); |
|
|
|
|
|
if (! $self->{'import_xml'}->{'format'}->{$name}) { |
|
|
$log->warn("<format name=\"$name\"> is not defined in ",$self->{'import_xml_file'}); |
|
|
return $data; |
|
|
} |
|
|
|
|
|
$log->warn("no delimiter for format $name") if (! $delimiter); |
|
| 355 |
|
|
| 356 |
my $format = $self->_x($self->{'import_xml'}->{'format'}->{$name}->{'content'}) || $log->logdie("can't find format '$name'"); |
=head2 suffix |
| 357 |
|
|
| 358 |
my @data = split(/\Q$delimiter\E/, $data); |
suffix all values with a string |
| 359 |
|
|
| 360 |
my $out = sprintf($format, @data); |
@v = suffix( '_my', @v ); |
|
$log->debug("using format $name [$format] on $data to produce: $out"); |
|
| 361 |
|
|
| 362 |
if ($self->{'lookup_regex'} && $out =~ $self->{'lookup_regex'}) { |
=cut |
|
return $self->{'lookup'}->lookup($out); |
|
|
} else { |
|
|
return $out; |
|
|
} |
|
| 363 |
|
|
| 364 |
|
sub suffix { |
| 365 |
|
my $s = shift or die "suffix needs string as first argument"; |
| 366 |
|
return map { $_ . $s } grep { defined($_) } @_; |
| 367 |
} |
} |
| 368 |
|
|
| 369 |
=head2 sort_arr |
=head2 surround |
| 370 |
|
|
| 371 |
Sort array ignoring case and html in data |
surround all values with a two strings |
| 372 |
|
|
| 373 |
my @sorted = $webpac->sort_arr(@unsorted); |
@v = surround( 'prefix_', '_suffix', @v ); |
| 374 |
|
|
| 375 |
=cut |
=cut |
| 376 |
|
|
| 377 |
sub sort_arr { |
sub surround { |
| 378 |
my $self = shift; |
my $p = shift or die "surround need prefix as first argument"; |
| 379 |
|
my $s = shift or die "surround needs suffix as second argument"; |
| 380 |
my $log = $self->_get_logger(); |
return map { $p . $_ . $s } grep { defined($_) } @_; |
|
|
|
|
# FIXME add Schwartzian Transformation? |
|
|
|
|
|
my @sorted = sort { |
|
|
$a =~ s#<[^>]+/*>##; |
|
|
$b =~ s#<[^>]+/*>##; |
|
|
lc($b) cmp lc($a) |
|
|
} @_; |
|
|
$log->debug("sorted values: ",sub { join(", ",@sorted) }); |
|
|
|
|
|
return @sorted; |
|
| 381 |
} |
} |
| 382 |
|
|
| 383 |
|
=head2 first |
| 384 |
|
|
| 385 |
=head1 INTERNAL METHODS |
Return first element |
| 386 |
|
|
| 387 |
=head2 _sort_by_order |
$v = first( @v ); |
|
|
|
|
Sort xml tags data structure accoding to C<order=""> attribute. |
|
| 388 |
|
|
| 389 |
=cut |
=cut |
| 390 |
|
|
| 391 |
sub _sort_by_order { |
sub first { |
| 392 |
my $self = shift; |
my $r = shift; |
| 393 |
|
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}; |
|
|
|
|
|
return $va <=> $vb; |
|
| 394 |
} |
} |
| 395 |
|
|
| 396 |
=head2 _x |
=head2 lookup |
| 397 |
|
|
| 398 |
Convert strings from C<conf/normalize/*.xml> encoding into application |
Consult lookup hashes for some value |
|
specific encoding (optinally specified using C<code_page> to C<new> |
|
|
constructor). |
|
| 399 |
|
|
| 400 |
my $text = $n->_x('normalize text string'); |
@v = lookup( $v ); |
| 401 |
|
@v = lookup( @v ); |
|
This is a stub so that other modules doesn't have to implement it. |
|
| 402 |
|
|
| 403 |
=cut |
=cut |
| 404 |
|
|
| 405 |
sub _x { |
sub lookup { |
| 406 |
my $self = shift; |
my $k = shift or return; |
| 407 |
return shift; |
return unless (defined($lookup->{$k})); |
| 408 |
|
if (ref($lookup->{$k}) eq 'ARRAY') { |
| 409 |
|
return @{ $lookup->{$k} }; |
| 410 |
|
} else { |
| 411 |
|
return $lookup->{$k}; |
| 412 |
|
} |
| 413 |
} |
} |
| 414 |
|
|
| 415 |
|
=head2 join_with |
| 416 |
|
|
| 417 |
=head1 AUTHOR |
Joins walues with some delimiter |
| 418 |
|
|
| 419 |
Dobrica Pavlinusic, C<< <dpavlin@rot13.org> >> |
$v = join_with(", ", @v); |
|
|
|
|
=head1 COPYRIGHT & LICENSE |
|
|
|
|
|
Copyright 2005 Dobrica Pavlinusic, All Rights Reserved. |
|
|
|
|
|
This program is free software; you can redistribute it and/or modify it |
|
|
under the same terms as Perl itself. |
|
| 420 |
|
|
| 421 |
=cut |
=cut |
| 422 |
|
|
| 423 |
1; # End of WebPAC::Normalize |
sub join_with { |
| 424 |
|
my $d = shift; |
| 425 |
|
return join($d, grep { defined($_) && $_ ne '' } @_); |
| 426 |
|
} |
| 427 |
|
|
| 428 |
|
# END |
| 429 |
|
1; |
Parent Directory
| 
Revision Log
| 
Patch