1 |
package WebPAC::Normalize; |
package WebPAC::Normalize; |
2 |
use Exporter 'import'; |
use Exporter 'import'; |
3 |
@EXPORT = qw/ |
our @EXPORT = qw/ |
4 |
_set_rec _set_lookup |
_set_ds _set_lookup |
5 |
|
_set_load_row |
6 |
_get_ds _clean_ds |
_get_ds _clean_ds |
7 |
_debug |
_debug |
8 |
|
_pack_subfields_hash |
9 |
|
|
10 |
tag search display |
to |
11 |
marc marc_indicators marc_repeatable_subfield |
search_display search display sorted |
|
marc_compose marc_leader |
|
|
marc_duplicate marc_remove |
|
12 |
|
|
13 |
rec1 rec2 rec |
rec1 rec2 rec |
14 |
|
frec frec_eq frec_ne |
15 |
regex prefix suffix surround |
regex prefix suffix surround |
16 |
first lookup join_with |
first lookup join_with |
17 |
|
save_into_lookup |
18 |
|
|
19 |
split_rec_on |
split_rec_on |
20 |
|
|
21 |
|
get set |
22 |
|
count |
23 |
|
|
24 |
/; |
/; |
25 |
|
|
26 |
use warnings; |
use warnings; |
28 |
|
|
29 |
#use base qw/WebPAC::Common/; |
#use base qw/WebPAC::Common/; |
30 |
use Data::Dump qw/dump/; |
use Data::Dump qw/dump/; |
31 |
use Encode qw/from_to/; |
use Carp qw/confess/; |
|
use Storable qw/dclone/; |
|
32 |
|
|
33 |
# debugging warn(s) |
# debugging warn(s) |
34 |
my $debug = 0; |
my $debug = 0; |
35 |
|
_debug( $debug ); |
36 |
|
|
37 |
|
# FIXME |
38 |
|
use WebPAC::Normalize::ISBN; |
39 |
|
push @EXPORT, ( 'isbn_10', 'isbn_13' ); |
40 |
|
|
41 |
=head1 NAME |
use WebPAC::Normalize::MARC; |
42 |
|
push @EXPORT, ( qw/ |
43 |
|
marc marc_indicators marc_repeatable_subfield |
44 |
|
marc_compose marc_leader marc_fixed |
45 |
|
marc_duplicate marc_remove marc_count |
46 |
|
marc_original_order |
47 |
|
marc_template |
48 |
|
/); |
49 |
|
|
50 |
WebPAC::Normalize - describe normalisaton rules using sets |
use Storable qw/dclone/; |
51 |
|
|
52 |
=head1 VERSION |
=head1 NAME |
53 |
|
|
54 |
Version 0.15 |
WebPAC::Normalize - describe normalisaton rules using sets |
55 |
|
|
56 |
=cut |
=cut |
57 |
|
|
58 |
our $VERSION = '0.15'; |
our $VERSION = '0.36'; |
59 |
|
|
60 |
=head1 SYNOPSIS |
=head1 SYNOPSIS |
61 |
|
|
68 |
C<perl -c normalize.pl>. |
C<perl -c normalize.pl>. |
69 |
|
|
70 |
Normalisation can generate multiple output normalized data. For now, supported output |
Normalisation can generate multiple output normalized data. For now, supported output |
71 |
types (on the left side of definition) are: C<tag>, C<display>, C<search> and |
types (on the left side of definition) are: C<search_display>, C<display>, C<search> and |
72 |
C<marc>. |
C<marc>. |
73 |
|
|
74 |
=head1 FUNCTIONS |
=head1 FUNCTIONS |
81 |
Return data structure |
Return data structure |
82 |
|
|
83 |
my $ds = WebPAC::Normalize::data_structure( |
my $ds = WebPAC::Normalize::data_structure( |
84 |
lookup => $lookup->lookup_hash, |
lookup => $lookup_hash, |
85 |
row => $row, |
row => $row, |
86 |
rules => $normalize_pl_config, |
rules => $normalize_pl_config, |
87 |
marc_encoding => 'utf-8', |
marc_encoding => 'utf-8', |
88 |
config => $config, |
config => $config, |
89 |
|
load_row_coderef => sub { |
90 |
|
my ($database,$input,$mfn) = @_; |
91 |
|
$store->load_row( database => $database, input => $input, id => $mfn ); |
92 |
|
}, |
93 |
); |
); |
94 |
|
|
95 |
Options C<lookup>, C<row>, C<rules> and C<log> are mandatory while all |
Options C<row>, C<rules> and C<log> are mandatory while all |
96 |
other are optional. |
other are optional. |
97 |
|
|
98 |
|
C<load_row_coderef> is closure only used when executing lookups, so they will |
99 |
|
die if it's not defined. |
100 |
|
|
101 |
This function will B<die> if normalizastion can't be evaled. |
This function will B<die> if normalizastion can't be evaled. |
102 |
|
|
103 |
Since this function isn't exported you have to call it with |
Since this function isn't exported you have to call it with |
105 |
|
|
106 |
=cut |
=cut |
107 |
|
|
108 |
|
my $load_row_coderef; |
109 |
|
|
110 |
sub data_structure { |
sub data_structure { |
111 |
my $arg = {@_}; |
my $arg = {@_}; |
112 |
|
|
113 |
die "need row argument" unless ($arg->{row}); |
die "need row argument" unless ($arg->{row}); |
114 |
die "need normalisation argument" unless ($arg->{rules}); |
die "need normalisation argument" unless ($arg->{rules}); |
115 |
|
|
116 |
no strict 'subs'; |
_set_lookup( $arg->{lookup} ) if defined($arg->{lookup}); |
117 |
_set_lookup( $arg->{lookup} ); |
_set_ds( $arg->{row} ); |
118 |
_set_rec( $arg->{row} ); |
_set_config( $arg->{config} ) if defined($arg->{config}); |
|
_set_config( $arg->{config} ); |
|
119 |
_clean_ds( %{ $arg } ); |
_clean_ds( %{ $arg } ); |
120 |
eval "$arg->{rules}"; |
$load_row_coderef = $arg->{load_row_coderef}; |
121 |
|
|
122 |
|
no strict 'subs'; |
123 |
|
no warnings 'redefine'; |
124 |
|
eval "$arg->{rules};"; |
125 |
die "error evaling $arg->{rules}: $@\n" if ($@); |
die "error evaling $arg->{rules}: $@\n" if ($@); |
126 |
|
|
127 |
return _get_ds(); |
return _get_ds(); |
128 |
} |
} |
129 |
|
|
130 |
=head2 _set_rec |
=head2 _set_ds |
131 |
|
|
132 |
Set current record hash |
Set current record hash |
133 |
|
|
134 |
_set_rec( $rec ); |
_set_ds( $rec ); |
135 |
|
|
136 |
=cut |
=cut |
137 |
|
|
138 |
my $rec; |
my $rec; |
139 |
|
|
140 |
sub _set_rec { |
sub _set_ds { |
141 |
$rec = shift or die "no record hash"; |
$rec = shift or die "no record hash"; |
142 |
|
$WebPAC::Normalize::MARC::rec = $rec; |
143 |
} |
} |
144 |
|
|
145 |
|
=head2 |
146 |
|
|
147 |
|
my $rec = _get_rec(); |
148 |
|
|
149 |
|
=cut |
150 |
|
|
151 |
|
sub _get_rec { $rec }; |
152 |
|
|
153 |
=head2 _set_config |
=head2 _set_config |
154 |
|
|
155 |
Set current config hash |
Set current config hash |
186 |
|
|
187 |
=cut |
=cut |
188 |
|
|
189 |
my ($out, $marc_record, $marc_encoding, $marc_repeatable_subfield, $marc_indicators); |
my $out; |
|
my ($marc_record_offset, $marc_fetch_offset) = (0, 0); |
|
190 |
|
|
191 |
sub _get_ds { |
sub _get_ds { |
192 |
|
#warn "## out = ",dump($out); |
193 |
return $out; |
return $out; |
194 |
} |
} |
195 |
|
|
203 |
|
|
204 |
sub _clean_ds { |
sub _clean_ds { |
205 |
my $a = {@_}; |
my $a = {@_}; |
206 |
($out,$marc_record, $marc_encoding, $marc_repeatable_subfield, $marc_indicators) = (); |
$out = undef; |
207 |
($marc_record_offset, $marc_fetch_offset) = (0,0); |
WebPAC::Normalize::MARC::_clean(); |
|
$marc_encoding = $a->{marc_encoding}; |
|
208 |
} |
} |
209 |
|
|
210 |
=head2 _set_lookup |
=head2 _set_lookup |
221 |
$lookup = shift; |
$lookup = shift; |
222 |
} |
} |
223 |
|
|
224 |
=head2 _get_marc_fields |
=head2 _get_lookup |
|
|
|
|
Get all fields defined by calls to C<marc> |
|
|
|
|
|
$marc->add_fields( WebPAC::Normalize:_get_marc_fields() ); |
|
|
|
|
|
We are using I<magic> which detect repeatable fields only from |
|
|
sequence of field/subfield data generated by normalization. |
|
|
|
|
|
Repeatable field is created when there is second occurence of same subfield or |
|
|
if any of indicators are different. |
|
|
|
|
|
This is sane for most cases. Something like: |
|
|
|
|
|
900a-1 900b-1 900c-1 |
|
|
900a-2 900b-2 |
|
|
900a-3 |
|
|
|
|
|
will be created from any combination of: |
|
|
|
|
|
900a-1 900a-2 900a-3 900b-1 900b-2 900c-1 |
|
|
|
|
|
and following rules: |
|
|
|
|
|
marc('900','a', rec('200','a') ); |
|
|
marc('900','b', rec('200','b') ); |
|
|
marc('900','c', rec('200','c') ); |
|
|
|
|
|
which might not be what you have in mind. If you need repeatable subfield, |
|
|
define it using C<marc_repeatable_subfield> like this: |
|
225 |
|
|
226 |
marc_repeatable_subfield('900','a'); |
Get current lookup hash |
|
marc('900','a', rec('200','a') ); |
|
|
marc('900','b', rec('200','b') ); |
|
|
marc('900','c', rec('200','c') ); |
|
227 |
|
|
228 |
will create: |
my $lookup = _get_lookup(); |
|
|
|
|
900a-1 900a-2 900a-3 900b-1 900c-1 |
|
|
900b-2 |
|
|
|
|
|
There is also support for returning next or specific using: |
|
|
|
|
|
while (my $mf = WebPAC::Normalize:_get_marc_fields( fetch_next => 1 ) ) { |
|
|
# do something with $mf |
|
|
} |
|
|
|
|
|
will always return fields from next MARC record or |
|
|
|
|
|
my $mf = WebPAC::Normalize::_get_marc_fields( offset => 42 ); |
|
|
|
|
|
will return 42th copy record (if it exists). |
|
229 |
|
|
230 |
=cut |
=cut |
231 |
|
|
232 |
sub _get_marc_fields { |
sub _get_lookup { |
233 |
|
return $lookup; |
234 |
my $arg = {@_}; |
} |
|
warn "### _get_marc_fields arg: ", dump($arg), $/ if ($debug > 2); |
|
|
my $offset = $marc_fetch_offset; |
|
|
if ($arg->{offset}) { |
|
|
$offset = $arg->{offset}; |
|
|
} elsif($arg->{fetch_next}) { |
|
|
$marc_fetch_offset++; |
|
|
} |
|
|
|
|
|
return if (! $marc_record || ref($marc_record) ne 'ARRAY'); |
|
|
|
|
|
warn "### full marc_record = ", dump( @{ $marc_record }), $/ if ($debug > 2); |
|
|
|
|
|
my $marc_rec = $marc_record->[ $offset ]; |
|
|
|
|
|
warn "## _get_marc_fields (at offset: $offset) -- marc_record = ", dump( @$marc_rec ), $/ if ($debug > 1); |
|
|
|
|
|
return if (! $marc_rec || ref($marc_rec) ne 'ARRAY' || $#{ $marc_rec } < 0); |
|
|
|
|
|
# first, sort all existing fields |
|
|
# XXX might not be needed, but modern perl might randomize elements in hash |
|
|
my @sorted_marc_record = sort { |
|
|
$a->[0] . ( $a->[3] || '' ) cmp $b->[0] . ( $b->[3] || '') |
|
|
} @{ $marc_rec }; |
|
|
|
|
|
@sorted_marc_record = @{ $marc_rec }; ### FIXME disable sorting |
|
|
|
|
|
# output marc fields |
|
|
my @m; |
|
|
|
|
|
# count unique field-subfields (used for offset when walking to next subfield) |
|
|
my $u; |
|
|
map { $u->{ $_->[0] . ( $_->[3] || '') }++ } @sorted_marc_record; |
|
|
|
|
|
if ($debug) { |
|
|
warn "## marc_repeatable_subfield = ", dump( $marc_repeatable_subfield ), $/ if ( $marc_repeatable_subfield ); |
|
|
warn "## marc_record[$offset] = ", dump( $marc_rec ), $/; |
|
|
warn "## sorted_marc_record = ", dump( \@sorted_marc_record ), $/; |
|
|
warn "## subfield count = ", dump( $u ), $/; |
|
|
} |
|
|
|
|
|
my $len = $#sorted_marc_record; |
|
|
my $visited; |
|
|
my $i = 0; |
|
|
my $field; |
|
|
|
|
|
foreach ( 0 .. $len ) { |
|
|
|
|
|
# find next element which isn't visited |
|
|
while ($visited->{$i}) { |
|
|
$i = ($i + 1) % ($len + 1); |
|
|
} |
|
|
|
|
|
# mark it visited |
|
|
$visited->{$i}++; |
|
|
|
|
|
my $row = dclone( $sorted_marc_record[$i] ); |
|
|
|
|
|
# field and subfield which is key for |
|
|
# marc_repeatable_subfield and u |
|
|
my $fsf = $row->[0] . ( $row->[3] || '' ); |
|
|
|
|
|
if ($debug > 1) { |
|
|
|
|
|
print "### field so far [", $#$field, "] : ", dump( $field ), " ", $field ? 'T' : 'F', $/; |
|
|
print "### this [$i]: ", dump( $row ),$/; |
|
|
print "### sf: ", $row->[3], " vs ", $field->[3], |
|
|
$marc_repeatable_subfield->{ $row->[0] . $row->[3] } ? ' (repeatable)' : '', $/, |
|
|
if ($#$field >= 0); |
|
235 |
|
|
236 |
} |
=head2 _set_load_row |
237 |
|
|
238 |
# if field exists |
Setup code reference which will return L<data_structure> from |
239 |
if ( $#$field >= 0 ) { |
L<WebPAC::Store> |
|
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 = $row; |
|
240 |
|
|
241 |
} else { |
_set_load_row(sub { |
242 |
# append new subfields to existing field |
my ($database,$input,$mfn) = @_; |
243 |
push @$field, ( $row->[3], $row->[4] ); |
$store->load_row( database => $database, input => $input, id => $mfn ); |
244 |
} |
}); |
|
} else { |
|
|
# insert first field |
|
|
$field = $row; |
|
|
} |
|
245 |
|
|
246 |
if (! $marc_repeatable_subfield->{ $fsf }) { |
=cut |
|
# make step to next subfield |
|
|
$i = ($i + $u->{ $fsf } ) % ($len + 1); |
|
|
} |
|
|
} |
|
247 |
|
|
248 |
if ($#$field >= 0) { |
sub _set_load_row { |
249 |
push @m, $field; |
my $coderef = shift; |
250 |
warn "## saved/3 ", dump( $field ),$/ if ($debug); |
confess "argument isn't CODE" unless ref($coderef) eq 'CODE'; |
|
} |
|
251 |
|
|
252 |
return \@m; |
$load_row_coderef = $coderef; |
253 |
} |
} |
254 |
|
|
255 |
=head2 _debug |
=head2 _debug |
265 |
return $debug unless defined($l); |
return $debug unless defined($l); |
266 |
warn "debug level $l",$/ if ($l > 0); |
warn "debug level $l",$/ if ($l > 0); |
267 |
$debug = $l; |
$debug = $l; |
268 |
|
$WebPAC::Normalize::MARC::debug = $debug; |
269 |
} |
} |
270 |
|
|
271 |
=head1 Functions to create C<data_structure> |
=head1 Functions to create C<data_structure> |
272 |
|
|
273 |
Those functions generally have to first in your normalization file. |
Those functions generally have to first in your normalization file. |
274 |
|
|
275 |
=head2 tag |
=head2 to |
276 |
|
|
277 |
Define new tag for I<search> and I<display>. |
Generic way to set values for some name |
278 |
|
|
279 |
tag('Title', rec('200','a') ); |
to('field-name', 'name-value' => rec('200','a') ); |
280 |
|
|
281 |
|
There are many helpers defined below which might be easier to use. |
282 |
|
|
283 |
=cut |
=cut |
284 |
|
|
285 |
sub tag { |
sub to { |
286 |
my $name = shift or die "tag needs name as first argument"; |
my $type = shift or confess "need type -- BUG?"; |
287 |
|
my $name = shift or confess "needs name as first argument"; |
288 |
my @o = grep { defined($_) && $_ ne '' } @_; |
my @o = grep { defined($_) && $_ ne '' } @_; |
289 |
return unless (@o); |
return unless (@o); |
290 |
$out->{$name}->{tag} = $name; |
$out->{$name}->{$type} = \@o; |
|
$out->{$name}->{search} = \@o; |
|
|
$out->{$name}->{display} = \@o; |
|
291 |
} |
} |
292 |
|
|
293 |
=head2 display |
=head2 search_display |
294 |
|
|
295 |
Define tag just for I<display> |
Define output for L<search> and L<display> at the same time |
296 |
|
|
297 |
@v = display('Title', rec('200','a') ); |
search_display('Title', rec('200','a') ); |
298 |
|
|
299 |
=cut |
=cut |
300 |
|
|
301 |
sub display { |
sub search_display { |
302 |
my $name = shift or die "display needs name as first argument"; |
my $name = shift or die "search_display needs name as first argument"; |
303 |
my @o = grep { defined($_) && $_ ne '' } @_; |
my @o = grep { defined($_) && $_ ne '' } @_; |
304 |
return unless (@o); |
return unless (@o); |
|
$out->{$name}->{tag} = $name; |
|
|
$out->{$name}->{display} = \@o; |
|
|
} |
|
|
|
|
|
=head2 search |
|
|
|
|
|
Prepare values just for I<search> |
|
|
|
|
|
@v = search('Title', rec('200','a') ); |
|
|
|
|
|
=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; |
|
305 |
$out->{$name}->{search} = \@o; |
$out->{$name}->{search} = \@o; |
306 |
|
$out->{$name}->{display} = \@o; |
307 |
} |
} |
308 |
|
|
309 |
=head2 marc_leader |
=head2 tag |
|
|
|
|
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->{' leader'}; |
|
|
} |
|
|
} |
|
|
|
|
|
=head2 marc |
|
|
|
|
|
Save value for MARC field |
|
310 |
|
|
311 |
marc('900','a', rec('200','a') ); |
Old name for L<search_display>, it will probably be removed at one point. |
|
marc('001', rec('000') ); |
|
312 |
|
|
313 |
=cut |
=cut |
314 |
|
|
315 |
sub marc { |
sub tag { |
316 |
my $f = shift or die "marc needs field"; |
search_display( @_ ); |
|
die "marc field must be numer" unless ($f =~ /^\d+$/); |
|
|
|
|
|
my $sf; |
|
|
if ($f >= 10) { |
|
|
$sf = shift or die "marc needs subfield"; |
|
|
} |
|
|
|
|
|
foreach (@_) { |
|
|
my $v = $_; # make var read-write for Encode |
|
|
next unless (defined($v) && $v !~ /^\s*$/); |
|
|
from_to($v, 'iso-8859-2', $marc_encoding) if ($marc_encoding); |
|
|
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, $v ]; |
|
|
} |
|
|
} |
|
317 |
} |
} |
318 |
|
|
319 |
=head2 marc_repeatable_subfield |
=head2 display |
320 |
|
|
321 |
Save values for MARC repetable subfield |
Define output just for I<display> |
322 |
|
|
323 |
marc_repeatable_subfield('910', 'z', rec('909') ); |
@v = display('Title', rec('200','a') ); |
324 |
|
|
325 |
=cut |
=cut |
326 |
|
|
327 |
sub marc_repeatable_subfield { |
sub display { to( 'display', @_ ) } |
|
my ($f,$sf) = @_; |
|
|
die "marc_repeatable_subfield need field and subfield!\n" unless ($f && $sf); |
|
|
$marc_repeatable_subfield->{ $f . $sf }++; |
|
|
marc(@_); |
|
|
} |
|
|
|
|
|
=head2 marc_indicators |
|
328 |
|
|
329 |
Set both indicators for MARC field |
=head2 search |
330 |
|
|
331 |
marc_indicators('900', ' ', 1); |
Prepare values just for I<search> |
332 |
|
|
333 |
Any indicator value other than C<0-9> will be treated as undefined. |
@v = search('Title', rec('200','a') ); |
334 |
|
|
335 |
=cut |
=cut |
336 |
|
|
337 |
sub marc_indicators { |
sub search { to( 'search', @_ ) } |
|
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)); |
|
338 |
|
|
339 |
$i1 = ' ' if ($i1 !~ /^\d$/); |
=head2 sorted |
|
$i2 = ' ' if ($i2 !~ /^\d$/); |
|
|
@{ $marc_indicators->{$f} } = ($i1,$i2); |
|
|
} |
|
340 |
|
|
341 |
=head2 marc_compose |
Insert into lists which will be automatically sorted |
342 |
|
|
343 |
Save values for each MARC subfield explicitly |
sorted('Title', rec('200','a') ); |
|
|
|
|
marc_compose('900', |
|
|
'a', rec('200','a') |
|
|
'b', rec('201','a') |
|
|
'a', rec('200','b') |
|
|
'c', rec('200','c') |
|
|
); |
|
|
|
|
|
If you specify C<+> for subfield, value will be appended |
|
|
to previous defined subfield. |
|
344 |
|
|
345 |
=cut |
=cut |
346 |
|
|
347 |
sub marc_compose { |
sub sorted { to( 'sorted', @_ ) } |
|
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, $i2 ]; |
|
348 |
|
|
|
warn "### marc_compose input subfields = ", dump(@_),$/ if ($debug > 2); |
|
349 |
|
|
350 |
while (@_) { |
=head1 Functions to extract data from input |
|
my $sf = shift or die "marc_compose $f needs subfield"; |
|
|
my $v = shift; |
|
|
|
|
|
next unless (defined($v) && $v !~ /^\s*$/); |
|
|
from_to($v, 'iso-8859-2', $marc_encoding) if ($marc_encoding); |
|
|
warn "## ++ marc_compose($f,$sf,$v) ", dump( $m ),$/ if ($debug > 1); |
|
|
if ($sf ne '+') { |
|
|
push @$m, ( $sf, $v ); |
|
|
} else { |
|
|
$m->[ $#$m ] .= $v; |
|
|
} |
|
|
} |
|
|
|
|
|
warn "## marc_compose current marc = ", dump( $m ),$/ if ($debug > 1); |
|
|
|
|
|
push @{ $marc_record->[ $marc_record_offset ] }, $m if ($#{$m} > 2); |
|
|
} |
|
351 |
|
|
352 |
=head2 marc_duplicate |
This function should be used inside functions to create C<data_structure> described |
353 |
|
above. |
354 |
|
|
355 |
Generate copy of current MARC record and continue working on copy |
=head2 _pack_subfields_hash |
356 |
|
|
357 |
marc_duplicate(); |
@subfields = _pack_subfields_hash( $h ); |
358 |
|
$subfields = _pack_subfields_hash( $h, 1 ); |
359 |
|
|
360 |
Copies can be accessed using C<< _get_marc_fields( fetch_next => 1 ) >> or |
Return each subfield value in array or pack them all together and return scalar |
361 |
C<< _get_marc_fields( offset => 42 ) >>. |
with subfields (denoted by C<^>) and values. |
362 |
|
|
363 |
=cut |
=cut |
364 |
|
|
365 |
sub marc_duplicate { |
sub _pack_subfields_hash { |
|
my $m = $marc_record->[ -1 ]; |
|
|
die "can't duplicate record which isn't defined" unless ($m); |
|
|
push @{ $marc_record }, dclone( $m ); |
|
|
warn "## marc_duplicate = ", dump(@$marc_record), $/ if ($debug > 1); |
|
|
$marc_record_offset = $#{ $marc_record }; |
|
|
warn "## marc_record_offset = $marc_record_offset", $/ if ($debug > 1); |
|
|
} |
|
|
|
|
|
=head2 marc_remove |
|
|
|
|
|
Remove some field or subfield from MARC record. |
|
|
|
|
|
marc_remove('200'); |
|
|
marc_remove('200','a'); |
|
|
|
|
|
This will erase field C<200> or C<200^a> from current MARC record. |
|
366 |
|
|
367 |
This is useful after calling C<marc_duplicate> or on it's own (but, you |
warn "## _pack_subfields_hash( ",dump(@_), " )\n" if ($debug > 1); |
|
should probably just remove that subfield definition if you are not |
|
|
using C<marc_duplicate>). |
|
368 |
|
|
369 |
FIXME: support fields < 10. |
my ($hash,$include_subfields) = @_; |
370 |
|
|
371 |
=cut |
# sanity and ease of use |
372 |
|
return $hash if (ref($hash) ne 'HASH'); |
|
sub marc_remove { |
|
|
my ($f, $sf) = @_; |
|
373 |
|
|
374 |
die "marc_remove needs record number" unless defined($f); |
my $h = dclone( $hash ); |
375 |
|
|
376 |
my $marc = $marc_record->[ $marc_record_offset ]; |
if ( defined($h->{subfields}) ) { |
377 |
|
my $sfs = delete $h->{subfields} || die "no subfields?"; |
378 |
|
my @out; |
379 |
|
while (@$sfs) { |
380 |
|
my $sf = shift @$sfs; |
381 |
|
push @out, '^' . $sf if ($include_subfields); |
382 |
|
my $o = shift @$sfs; |
383 |
|
if ($o == 0 && ref( $h->{$sf} ) ne 'ARRAY' ) { |
384 |
|
# single element subfields are not arrays |
385 |
|
#warn "====> $sf $o / $#$sfs ", dump( $sfs, $h->{$sf} ), "\n"; |
386 |
|
|
387 |
warn "### marc_remove before = ", dump( $marc ), $/ if ($debug > 2); |
push @out, $h->{$sf}; |
|
|
|
|
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--; |
|
388 |
} else { |
} else { |
389 |
foreach my $j ( 0 .. (( $#{ $marc->[$i] } - 3 ) / 2) ) { |
#warn "====> $sf $o / $#$sfs ", dump( $sfs, $h->{$sf} ), "\n"; |
390 |
my $o = ($j * 2) + 3; |
push @out, $h->{$sf}->[$o]; |
391 |
if ($marc->[$i]->[$o] eq $sf) { |
} |
392 |
# remove subfield |
} |
393 |
splice @{$marc->[$i]}, $o, 2; |
if ($include_subfields) { |
394 |
warn "#### slice \@{\$marc->[$i]}, $o, 2 = ", dump( @{ $marc }), $/ if ($debug > 3); |
return join('', @out); |
395 |
# is record now empty? |
} else { |
396 |
if ($#{ $marc->[$i] } == 2) { |
return @out; |
397 |
splice @$marc, $i, 1; |
} |
398 |
warn "#### slice \@\$marc, $i, 1 = ", dump( @{ $marc }), $/ if ($debug > 3); |
} else { |
399 |
$i--; |
if ($include_subfields) { |
400 |
}; |
my $out = ''; |
401 |
} |
foreach my $sf (sort keys %$h) { |
402 |
|
if (ref($h->{$sf}) eq 'ARRAY') { |
403 |
|
$out .= '^' . $sf . join('^' . $sf, @{ $h->{$sf} }); |
404 |
|
} else { |
405 |
|
$out .= '^' . $sf . $h->{$sf}; |
406 |
} |
} |
407 |
} |
} |
408 |
|
return $out; |
409 |
|
} else { |
410 |
|
# FIXME this should probably be in alphabetical order instead of hash order |
411 |
|
values %{$h}; |
412 |
} |
} |
|
$i++; |
|
413 |
} |
} |
|
|
|
|
warn "### marc_remove($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); |
|
414 |
} |
} |
415 |
|
|
|
=head1 Functions to extract data from input |
|
|
|
|
|
This function should be used inside functions to create C<data_structure> described |
|
|
above. |
|
|
|
|
416 |
=head2 rec1 |
=head2 rec1 |
417 |
|
|
418 |
Return all values in some field |
Return all values in some field |
429 |
return unless (defined($rec) && defined($rec->{$f})); |
return unless (defined($rec) && defined($rec->{$f})); |
430 |
warn "rec1($f) = ", dump( $rec->{$f} ), $/ if ($debug > 1); |
warn "rec1($f) = ", dump( $rec->{$f} ), $/ if ($debug > 1); |
431 |
if (ref($rec->{$f}) eq 'ARRAY') { |
if (ref($rec->{$f}) eq 'ARRAY') { |
432 |
return map { |
my @out; |
433 |
if (ref($_) eq 'HASH') { |
foreach my $h ( @{ $rec->{$f} } ) { |
434 |
values %{$_}; |
if (ref($h) eq 'HASH') { |
435 |
|
push @out, ( _pack_subfields_hash( $h ) ); |
436 |
} else { |
} else { |
437 |
$_; |
push @out, $h; |
438 |
} |
} |
439 |
} @{ $rec->{$f} }; |
} |
440 |
|
return @out; |
441 |
} elsif( defined($rec->{$f}) ) { |
} elsif( defined($rec->{$f}) ) { |
442 |
return $rec->{$f}; |
return $rec->{$f}; |
443 |
} |
} |
462 |
} else { |
} else { |
463 |
$_->{$sf}; |
$_->{$sf}; |
464 |
} |
} |
465 |
} grep { ref($_) eq 'HASH' && $_->{$sf} } @{ $rec->{$f} }; |
} grep { ref($_) eq 'HASH' && defined $_->{$sf} } @{ $rec->{$f} }; |
466 |
} |
} |
467 |
|
|
468 |
=head2 rec |
=head2 rec |
472 |
@v = rec('200') |
@v = rec('200') |
473 |
@v = rec('200','a') |
@v = rec('200','a') |
474 |
|
|
475 |
|
If rec() returns just single value, it will |
476 |
|
return scalar, not array. |
477 |
|
|
478 |
=cut |
=cut |
479 |
|
|
480 |
sub rec { |
sub rec { |
484 |
} elsif ($#_ == 1) { |
} elsif ($#_ == 1) { |
485 |
@out = rec2(@_); |
@out = rec2(@_); |
486 |
} |
} |
487 |
if (@out) { |
if ($#out == 0 && ! wantarray) { |
488 |
|
return $out[0]; |
489 |
|
} elsif (@out) { |
490 |
return @out; |
return @out; |
491 |
} else { |
} else { |
492 |
return ''; |
return ''; |
493 |
} |
} |
494 |
} |
} |
495 |
|
|
496 |
|
=head2 frec |
497 |
|
|
498 |
|
Returns first value from field |
499 |
|
|
500 |
|
$v = frec('200'); |
501 |
|
$v = frec('200','a'); |
502 |
|
|
503 |
|
=cut |
504 |
|
|
505 |
|
sub frec { |
506 |
|
my @out = rec(@_); |
507 |
|
warn "rec(",dump(@_),") has more than one return value, ignoring\n" if $#out > 0; |
508 |
|
return shift @out; |
509 |
|
} |
510 |
|
|
511 |
|
=head2 frec_eq |
512 |
|
|
513 |
|
=head2 frec_ne |
514 |
|
|
515 |
|
Check if first values from two fields are same or different |
516 |
|
|
517 |
|
if ( frec_eq( 900 => 'a', 910 => 'c' ) ) { |
518 |
|
# values are same |
519 |
|
} else { |
520 |
|
# values are different |
521 |
|
} |
522 |
|
|
523 |
|
Strictly speaking C<frec_eq> and C<frec_ne> wouldn't be needed if you |
524 |
|
could write something like: |
525 |
|
|
526 |
|
if ( frec( '900','a' ) eq frec( '910','c' ) ) { |
527 |
|
# yada tada |
528 |
|
} |
529 |
|
|
530 |
|
but you can't since our parser L<WebPAC::Parser> will remove all whitespaces |
531 |
|
in order to parse text and create invalid function C<eqfrec>. |
532 |
|
|
533 |
|
=cut |
534 |
|
|
535 |
|
sub frec_eq { |
536 |
|
my ( $f1,$sf1, $f2, $sf2 ) = @_; |
537 |
|
return (rec( $f1, $sf1 ))[0] eq (rec( $f2, $sf2 ))[0]; |
538 |
|
} |
539 |
|
|
540 |
|
sub frec_ne { |
541 |
|
return ! frec_eq( @_ ); |
542 |
|
} |
543 |
|
|
544 |
=head2 regex |
=head2 regex |
545 |
|
|
546 |
Apply regex to some or all values |
Apply regex to some or all values |
570 |
=cut |
=cut |
571 |
|
|
572 |
sub prefix { |
sub prefix { |
573 |
my $p = shift or return; |
my $p = shift; |
574 |
|
return @_ unless defined( $p ); |
575 |
return map { $p . $_ } grep { defined($_) } @_; |
return map { $p . $_ } grep { defined($_) } @_; |
576 |
} |
} |
577 |
|
|
584 |
=cut |
=cut |
585 |
|
|
586 |
sub suffix { |
sub suffix { |
587 |
my $s = shift or die "suffix needs string as first argument"; |
my $s = shift; |
588 |
|
return @_ unless defined( $s ); |
589 |
return map { $_ . $s } grep { defined($_) } @_; |
return map { $_ . $s } grep { defined($_) } @_; |
590 |
} |
} |
591 |
|
|
598 |
=cut |
=cut |
599 |
|
|
600 |
sub surround { |
sub surround { |
601 |
my $p = shift or die "surround need prefix as first argument"; |
my $p = shift; |
602 |
my $s = shift or die "surround needs suffix as second argument"; |
my $s = shift; |
603 |
|
$p = '' unless defined( $p ); |
604 |
|
$s = '' unless defined( $s ); |
605 |
return map { $p . $_ . $s } grep { defined($_) } @_; |
return map { $p . $_ . $s } grep { defined($_) } @_; |
606 |
} |
} |
607 |
|
|
622 |
|
|
623 |
Consult lookup hashes for some value |
Consult lookup hashes for some value |
624 |
|
|
625 |
@v = lookup( $v ); |
@v = lookup( |
626 |
@v = lookup( @v ); |
sub { |
627 |
|
'ffkk/peri/mfn'.rec('000') |
628 |
|
}, |
629 |
|
'ffkk','peri','200-a-200-e', |
630 |
|
sub { |
631 |
|
first(rec(200,'a')).' '.first(rec('200','e')) |
632 |
|
} |
633 |
|
); |
634 |
|
|
635 |
|
Code like above will be B<automatically generated> using L<WebPAC::Parse> from |
636 |
|
normal lookup definition in C<conf/lookup/something.pl> which looks like: |
637 |
|
|
638 |
|
lookup( |
639 |
|
# which results to return from record recorded in lookup |
640 |
|
sub { 'ffkk/peri/mfn' . rec('000') }, |
641 |
|
# from which database and input |
642 |
|
'ffkk','peri', |
643 |
|
# such that following values match |
644 |
|
sub { first(rec(200,'a')) . ' ' . first(rec('200','e')) }, |
645 |
|
# if this part is missing, we will try to match same fields |
646 |
|
# from lookup record and current one, or you can override |
647 |
|
# which records to use from current record using |
648 |
|
sub { rec('900','x') . ' ' . rec('900','y') }, |
649 |
|
) |
650 |
|
|
651 |
|
You can think about this lookup as SQL (if that helps): |
652 |
|
|
653 |
|
select |
654 |
|
sub { what } |
655 |
|
from |
656 |
|
database, input |
657 |
|
where |
658 |
|
sub { filter from lookuped record } |
659 |
|
having |
660 |
|
sub { optional filter on current record } |
661 |
|
|
662 |
|
Easy as pie, right? |
663 |
|
|
664 |
=cut |
=cut |
665 |
|
|
666 |
sub lookup { |
sub lookup { |
667 |
my $k = shift or return; |
my ($what, $database, $input, $key, $having) = @_; |
668 |
return unless (defined($lookup->{$k})); |
|
669 |
if (ref($lookup->{$k}) eq 'ARRAY') { |
confess "lookup needs 5 arguments: what, database, input, key, having\n" unless ($#_ == 4); |
670 |
return @{ $lookup->{$k} }; |
|
671 |
|
warn "## lookup ($database, $input, $key)", $/ if ($debug > 1); |
672 |
|
return unless (defined($lookup->{$database}->{$input}->{$key})); |
673 |
|
|
674 |
|
confess "lookup really need load_row_coderef added to data_structure\n" unless ($load_row_coderef); |
675 |
|
|
676 |
|
my $mfns; |
677 |
|
my @having = $having->(); |
678 |
|
|
679 |
|
warn "## having = ", dump( @having ) if ($debug > 2); |
680 |
|
|
681 |
|
foreach my $h ( @having ) { |
682 |
|
if (defined($lookup->{$database}->{$input}->{$key}->{$h})) { |
683 |
|
warn "lookup for $database/$input/$key/$h return ",dump($lookup->{$database}->{$input}->{$key}->{$h}),"\n" if ($debug); |
684 |
|
$mfns->{$_}++ foreach keys %{ $lookup->{$database}->{$input}->{$key}->{$h} }; |
685 |
|
} |
686 |
|
} |
687 |
|
|
688 |
|
return unless ($mfns); |
689 |
|
|
690 |
|
my @mfns = sort keys %$mfns; |
691 |
|
|
692 |
|
warn "# lookup loading $database/$input/$key mfn ", join(",",@mfns)," having ",dump(@having),"\n" if ($debug); |
693 |
|
|
694 |
|
my $old_rec = $rec; |
695 |
|
my @out; |
696 |
|
|
697 |
|
foreach my $mfn (@mfns) { |
698 |
|
$rec = $load_row_coderef->( $database, $input, $mfn ); |
699 |
|
|
700 |
|
warn "got $database/$input/$mfn = ", dump($rec), $/ if ($debug); |
701 |
|
|
702 |
|
my @vals = $what->(); |
703 |
|
|
704 |
|
push @out, ( @vals ); |
705 |
|
|
706 |
|
warn "lookup for mfn $mfn returned ", dump(@vals), $/ if ($debug); |
707 |
|
} |
708 |
|
|
709 |
|
# if (ref($lookup->{$k}) eq 'ARRAY') { |
710 |
|
# return @{ $lookup->{$k} }; |
711 |
|
# } else { |
712 |
|
# return $lookup->{$k}; |
713 |
|
# } |
714 |
|
|
715 |
|
$rec = $old_rec; |
716 |
|
|
717 |
|
warn "## lookup returns = ", dump(@out), $/ if ($debug); |
718 |
|
|
719 |
|
if ($#out == 0) { |
720 |
|
return $out[0]; |
721 |
} else { |
} else { |
722 |
return $lookup->{$k}; |
return @out; |
723 |
|
} |
724 |
|
} |
725 |
|
|
726 |
|
=head2 save_into_lookup |
727 |
|
|
728 |
|
Save value into lookup. It associates current database, input |
729 |
|
and specific keys with one or more values which will be |
730 |
|
associated over MFN. |
731 |
|
|
732 |
|
MFN will be extracted from first occurence current of field 000 |
733 |
|
in current record, or if it doesn't exist from L<_set_config> C<_mfn>. |
734 |
|
|
735 |
|
my $nr = save_into_lookup($database,$input,$key,sub { |
736 |
|
# code which produce one or more values |
737 |
|
}); |
738 |
|
|
739 |
|
It returns number of items saved. |
740 |
|
|
741 |
|
This function shouldn't be called directly, it's called from code created by |
742 |
|
L<WebPAC::Parser>. |
743 |
|
|
744 |
|
=cut |
745 |
|
|
746 |
|
sub save_into_lookup { |
747 |
|
my ($database,$input,$key,$coderef) = @_; |
748 |
|
die "save_into_lookup needs database" unless defined($database); |
749 |
|
die "save_into_lookup needs input" unless defined($input); |
750 |
|
die "save_into_lookup needs key" unless defined($key); |
751 |
|
die "save_into_lookup needs CODE" unless ( defined($coderef) && ref($coderef) eq 'CODE' ); |
752 |
|
|
753 |
|
warn "## save_into_lookup rec = ", dump($rec), " config = ", dump($config), $/ if ($debug > 2); |
754 |
|
|
755 |
|
my $mfn = |
756 |
|
defined($rec->{'000'}->[0]) ? $rec->{'000'}->[0] : |
757 |
|
defined($config->{_mfn}) ? $config->{_mfn} : |
758 |
|
die "mfn not defined or zero"; |
759 |
|
|
760 |
|
my $nr = 0; |
761 |
|
|
762 |
|
foreach my $v ( $coderef->() ) { |
763 |
|
$lookup->{$database}->{$input}->{$key}->{$v}->{$mfn}++; |
764 |
|
warn "# saved lookup $database/$input/$key [$v] $mfn\n" if ($debug > 1); |
765 |
|
$nr++; |
766 |
} |
} |
767 |
|
|
768 |
|
return $nr; |
769 |
} |
} |
770 |
|
|
771 |
=head2 config |
=head2 config |
776 |
$database_code = config(); # use _ from hash |
$database_code = config(); # use _ from hash |
777 |
$database_name = config('name'); |
$database_name = config('name'); |
778 |
$database_input_name = config('input name'); |
$database_input_name = config('input name'); |
|
$tag = config('input normalize tag'); |
|
779 |
|
|
780 |
Up to three levels are supported. |
Up to three levels are supported. |
781 |
|
|
836 |
|
|
837 |
sub id { |
sub id { |
838 |
my $mfn = $config->{_mfn} || die "no _mfn in config data"; |
my $mfn = $config->{_mfn} || die "no _mfn in config data"; |
839 |
return $mfn . $#{$marc_record} ? $#{$marc_record} + 1 : ''; |
return $mfn . ( WebPAC::Normalize::MARC::_created_marc_records() || '' ); |
840 |
} |
} |
841 |
|
|
842 |
=head2 join_with |
=head2 join_with |
880 |
my $v = shift @r; |
my $v = shift @r; |
881 |
warn "### first rec($fld,$sf) = ",dump($v),$/ if ($debug > 2); |
warn "### first rec($fld,$sf) = ",dump($v),$/ if ($debug > 2); |
882 |
|
|
883 |
return '' if( ! defined($v) || $v =~ /^\s*$/); |
return '' if ( ! defined($v) || $v =~ /^\s*$/); |
884 |
|
|
885 |
my @s = split( $regex, $v ); |
my @s = split( $regex, $v ); |
886 |
warn "## split_rec_on($fld,$sf,$regex,$part) = ",dump(@s),$/ if ($debug > 1); |
warn "## split_rec_on($fld,$sf,$regex,$part) = ",dump(@s),$/ if ($debug > 1); |
891 |
} |
} |
892 |
} |
} |
893 |
|
|
894 |
|
my $hash; |
895 |
|
|
896 |
|
=head2 set |
897 |
|
|
898 |
|
set( key => 'value' ); |
899 |
|
|
900 |
|
=cut |
901 |
|
|
902 |
|
sub set { |
903 |
|
my ($k,$v) = @_; |
904 |
|
warn "## set ( $k => ", dump($v), " )", $/ if ( $debug ); |
905 |
|
$hash->{$k} = $v; |
906 |
|
}; |
907 |
|
|
908 |
|
=head2 get |
909 |
|
|
910 |
|
get( 'key' ); |
911 |
|
|
912 |
|
=cut |
913 |
|
|
914 |
|
sub get { |
915 |
|
my $k = shift || return; |
916 |
|
my $v = $hash->{$k}; |
917 |
|
warn "## get $k = ", dump( $v ), $/ if ( $debug ); |
918 |
|
return $v; |
919 |
|
} |
920 |
|
|
921 |
|
=head2 count |
922 |
|
|
923 |
|
if ( count( @result ) == 1 ) { |
924 |
|
# do something if only 1 result is there |
925 |
|
} |
926 |
|
|
927 |
|
=cut |
928 |
|
|
929 |
|
sub count { |
930 |
|
warn "## count ",dump(@_),$/ if ( $debug ); |
931 |
|
return @_ . ''; |
932 |
|
} |
933 |
|
|
934 |
# END |
# END |
935 |
1; |
1; |