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 |
|
_debug |
7 |
|
|
8 |
|
tag search display |
9 |
|
marc marc_indicators marc_repeatable_subfield |
10 |
|
marc_compose marc_leader |
11 |
|
marc_duplicate marc_remove |
12 |
|
marc_original_order |
13 |
|
|
14 |
|
rec1 rec2 rec |
15 |
|
regex prefix suffix surround |
16 |
|
first lookup join_with |
17 |
|
|
18 |
|
split_rec_on |
19 |
|
/; |
20 |
|
|
21 |
use warnings; |
use warnings; |
22 |
use strict; |
use strict; |
23 |
use base 'WebPAC::Common'; |
|
24 |
use Data::Dumper; |
#use base qw/WebPAC::Common/; |
25 |
|
use Data::Dump qw/dump/; |
26 |
|
use Encode qw/from_to/; |
27 |
|
use Storable qw/dclone/; |
28 |
|
|
29 |
|
# debugging warn(s) |
30 |
|
my $debug = 0; |
31 |
|
|
32 |
|
|
33 |
=head1 NAME |
=head1 NAME |
34 |
|
|
35 |
WebPAC::Normalize - data mungling for normalisation |
WebPAC::Normalize - describe normalisaton rules using sets |
36 |
|
|
37 |
=head1 VERSION |
=head1 VERSION |
38 |
|
|
39 |
Version 0.06 |
Version 0.16 |
40 |
|
|
41 |
=cut |
=cut |
42 |
|
|
43 |
our $VERSION = '0.06'; |
our $VERSION = '0.16'; |
44 |
|
|
45 |
=head1 SYNOPSIS |
=head1 SYNOPSIS |
46 |
|
|
47 |
This package contains code that mungle data to produce normalized format. |
This module uses C<conf/normalize/*.pl> files to perform normalisation |
48 |
|
from input records using perl functions which are specialized for set |
49 |
|
processing. |
50 |
|
|
51 |
|
Sets are implemented as arrays, and normalisation file is valid perl, which |
52 |
|
means that you check it's validity before running WebPAC using |
53 |
|
C<perl -c normalize.pl>. |
54 |
|
|
55 |
|
Normalisation can generate multiple output normalized data. For now, supported output |
56 |
|
types (on the left side of definition) are: C<tag>, C<display>, C<search> and |
57 |
|
C<marc>. |
58 |
|
|
59 |
It contains several assumptions: |
=head1 FUNCTIONS |
60 |
|
|
61 |
=over |
Functions which start with C<_> are private and used by WebPAC internally. |
62 |
|
All other functions are available for use within normalisation rules. |
63 |
|
|
64 |
=item * |
=head2 data_structure |
65 |
|
|
66 |
format of fields is defined using C<v123^a> notation for repeatable fields |
Return data structure |
|
or C<s123^a> for single (or first) value, where C<123> is field number and |
|
|
C<a> is subfield. |
|
67 |
|
|
68 |
=item * |
my $ds = WebPAC::Normalize::data_structure( |
69 |
|
lookup => $lookup->lookup_hash, |
70 |
|
row => $row, |
71 |
|
rules => $normalize_pl_config, |
72 |
|
marc_encoding => 'utf-8', |
73 |
|
config => $config, |
74 |
|
); |
75 |
|
|
76 |
source data records (C<$rec>) have unique identifiers in field C<000> |
Options C<lookup>, C<row>, C<rules> and C<log> are mandatory while all |
77 |
|
other are optional. |
78 |
|
|
79 |
=item * |
This function will B<die> if normalizastion can't be evaled. |
80 |
|
|
81 |
optional C<eval{length('v123^a') == 3}> tag at B<beginning of format> will be |
Since this function isn't exported you have to call it with |
82 |
perl code that is evaluated before producing output (value of field will be |
C<WebPAC::Normalize::data_structure>. |
|
interpolated before that) |
|
83 |
|
|
84 |
=item * |
=cut |
85 |
|
|
86 |
optional C<filter{filter_name}> at B<begining of format> will apply perl |
sub data_structure { |
87 |
code defined as code ref on format after field substitution to producing |
my $arg = {@_}; |
|
output |
|
88 |
|
|
89 |
There is one built-in filter called C<regex> which can be use like this: |
die "need row argument" unless ($arg->{row}); |
90 |
|
die "need normalisation argument" unless ($arg->{rules}); |
91 |
|
|
92 |
filter{regex(s/foo/bar/)} |
no strict 'subs'; |
93 |
|
_set_lookup( $arg->{lookup} ); |
94 |
|
_set_rec( $arg->{row} ); |
95 |
|
_set_config( $arg->{config} ); |
96 |
|
_clean_ds( %{ $arg } ); |
97 |
|
eval "$arg->{rules}"; |
98 |
|
die "error evaling $arg->{rules}: $@\n" if ($@); |
99 |
|
|
100 |
=item * |
return _get_ds(); |
101 |
|
} |
102 |
|
|
103 |
optional C<lookup{...}> will be then performed. See C<WebPAC::Lookups>. |
=head2 _set_rec |
104 |
|
|
105 |
=item * |
Set current record hash |
106 |
|
|
107 |
at end, optional C<format>s rules are resolved. Format rules are similar to |
_set_rec( $rec ); |
|
C<sprintf> and can also contain C<lookup{...}> which is performed after |
|
|
values are inserted in format. |
|
108 |
|
|
109 |
=back |
=cut |
110 |
|
|
111 |
This also describes order in which transformations are applied (eval, |
my $rec; |
|
filter, lookup, format) which is important to undestand when deciding how to |
|
|
solve your data mungling and normalisation process. |
|
112 |
|
|
113 |
|
sub _set_rec { |
114 |
|
$rec = shift or die "no record hash"; |
115 |
|
} |
116 |
|
|
117 |
|
=head2 _set_config |
118 |
|
|
119 |
|
Set current config hash |
120 |
|
|
121 |
=head1 FUNCTIONS |
_set_config( $config ); |
122 |
|
|
123 |
=head2 new |
Magic keys are: |
124 |
|
|
125 |
Create new normalisation object |
=over 4 |
126 |
|
|
127 |
my $n = new WebPAC::Normalize::Something( |
=item _ |
128 |
filter => { |
|
129 |
'filter_name_1' => sub { |
Code of current database |
|
# filter code |
|
|
return length($_); |
|
|
}, ... |
|
|
}, |
|
|
db => $db_obj, |
|
|
lookup_regex => $lookup->regex, |
|
|
lookup => $lookup_obj, |
|
|
prefix => 'foobar', |
|
|
); |
|
130 |
|
|
131 |
Parametar C<filter> defines user supplied snippets of perl code which can |
=item _mfn |
|
be use with C<filter{...}> notation. |
|
132 |
|
|
133 |
C<prefix> is used to form filename for database record (to support multiple |
Current MFN |
|
source files which are joined in one database). |
|
134 |
|
|
135 |
Recommended parametar C<lookup_regex> is used to enable parsing of lookups |
=back |
|
in structures. If you pass this parametar, you must also pass C<lookup> |
|
|
which is C<WebPAC::Lookup> object. |
|
136 |
|
|
137 |
=cut |
=cut |
138 |
|
|
139 |
sub new { |
my $config; |
|
my $class = shift; |
|
|
my $self = {@_}; |
|
|
bless($self, $class); |
|
140 |
|
|
141 |
my $r = $self->{'lookup_regex'} ? 1 : 0; |
sub _set_config { |
142 |
my $l = $self->{'lookup'} ? 1 : 0; |
$config = shift; |
143 |
|
} |
144 |
|
|
145 |
my $log = $self->_get_logger(); |
=head2 _get_ds |
146 |
|
|
147 |
# those two must be in pair |
Return hash formatted as data structure |
|
if ( ($r & $l) != ($r || $l) ) { |
|
|
my $log = $self->_get_logger(); |
|
|
$log->logdie("lookup_regex and lookup must be in pair"); |
|
|
} |
|
148 |
|
|
149 |
$log->logdie("lookup must be WebPAC::Lookup object") if ($self->{'lookup'} && ! $self->{'lookup'}->isa('WebPAC::Lookup')); |
my $ds = _get_ds(); |
150 |
|
|
151 |
$log->warn("no prefix defined. please check that!") unless ($self->{'prefix'}); |
=cut |
152 |
|
|
153 |
$log->debug("using lookup regex: ", $self->{lookup_regex}) if ($r && $l); |
my ($out, $marc_record, $marc_encoding, $marc_repeatable_subfield, $marc_indicators); |
154 |
|
my ($marc_record_offset, $marc_fetch_offset) = (0, 0); |
155 |
|
|
156 |
if ($self->{filter} && ! $self->{filter}->{regex}) { |
sub _get_ds { |
157 |
$log->debug("adding built-in filter regex"); |
return $out; |
|
$self->{filter}->{regex} = sub { |
|
|
my ($val, $regex) = @_; |
|
|
eval "\$val =~ $regex"; |
|
|
return $val; |
|
|
}; |
|
|
} |
|
|
|
|
|
$self ? return $self : return undef; |
|
158 |
} |
} |
159 |
|
|
160 |
|
=head2 _clean_ds |
161 |
|
|
162 |
=head2 data_structure |
Clean data structure hash for next record |
163 |
|
|
164 |
Create in-memory data structure which represents normalized layout from |
_clean_ds(); |
|
C<conf/normalize/*.xml>. |
|
165 |
|
|
166 |
This structures are used to produce output. |
=cut |
167 |
|
|
168 |
|
sub _clean_ds { |
169 |
|
my $a = {@_}; |
170 |
|
($out,$marc_record, $marc_encoding, $marc_repeatable_subfield, $marc_indicators) = (); |
171 |
|
($marc_record_offset, $marc_fetch_offset) = (0,0); |
172 |
|
$marc_encoding = $a->{marc_encoding}; |
173 |
|
} |
174 |
|
|
175 |
my $ds = $webpac->data_structure($rec); |
=head2 _set_lookup |
176 |
|
|
177 |
|
Set current lookup hash |
178 |
|
|
179 |
|
_set_lookup( $lookup ); |
180 |
|
|
181 |
=cut |
=cut |
182 |
|
|
183 |
sub data_structure { |
my $lookup; |
|
my $self = shift; |
|
184 |
|
|
185 |
my $log = $self->_get_logger(); |
sub _set_lookup { |
186 |
|
$lookup = shift; |
187 |
|
} |
188 |
|
|
189 |
my $rec = shift; |
=head2 _get_marc_fields |
|
$log->logconfess("need HASH as first argument!") if ($rec !~ /HASH/o); |
|
190 |
|
|
191 |
$log->debug("data_structure rec = ", sub { Dumper($rec) }); |
Get all fields defined by calls to C<marc> |
192 |
|
|
193 |
$log->logdie("need unique ID (mfn) in field 000 of record ", sub { Dumper($rec) } ) unless (defined($rec->{'000'})); |
$marc->add_fields( WebPAC::Normalize:_get_marc_fields() ); |
194 |
|
|
195 |
my $id = $rec->{'000'}->[0] || $log->logdie("field 000 isn't array!"); |
We are using I<magic> which detect repeatable fields only from |
196 |
|
sequence of field/subfield data generated by normalization. |
197 |
|
|
198 |
my $cache_file; |
Repeatable field is created when there is second occurence of same subfield or |
199 |
|
if any of indicators are different. |
200 |
|
|
201 |
if ($self->{'db'}) { |
This is sane for most cases. Something like: |
|
my $ds = $self->{'db'}->load_ds( id => $id, prefix => $self->{prefix} ); |
|
|
$log->debug("load_ds( rec = ", sub { Dumper($rec) }, ") = ", sub { Dumper($ds) }); |
|
|
return $ds if ($ds); |
|
|
$log->debug("cache miss, creating"); |
|
|
} |
|
202 |
|
|
203 |
my @sorted_tags; |
900a-1 900b-1 900c-1 |
204 |
if ($self->{tags_by_order}) { |
900a-2 900b-2 |
205 |
@sorted_tags = @{$self->{tags_by_order}}; |
900a-3 |
|
} else { |
|
|
@sorted_tags = sort { $self->_sort_by_order } keys %{$self->{'import_xml'}->{'indexer'}}; |
|
|
$self->{tags_by_order} = \@sorted_tags; |
|
|
} |
|
206 |
|
|
207 |
my $ds; |
will be created from any combination of: |
208 |
|
|
209 |
$log->debug("tags: ",sub { join(", ",@sorted_tags) }); |
900a-1 900a-2 900a-3 900b-1 900b-2 900c-1 |
210 |
|
|
211 |
foreach my $field (@sorted_tags) { |
and following rules: |
212 |
|
|
213 |
my $row; |
marc('900','a', rec('200','a') ); |
214 |
|
marc('900','b', rec('200','b') ); |
215 |
|
marc('900','c', rec('200','c') ); |
216 |
|
|
217 |
#print "field $field [",$self->{'tag'},"] = ",Dumper($self->{'import_xml'}->{'indexer'}->{$field}->{$self->{'tag'}}); |
which might not be what you have in mind. If you need repeatable subfield, |
218 |
|
define it using C<marc_repeatable_subfield> like this: |
219 |
|
|
220 |
foreach my $tag (@{$self->{'import_xml'}->{'indexer'}->{$field}->{$self->{'tag'}}}) { |
marc_repeatable_subfield('900','a'); |
221 |
my $format; |
marc('900','a', rec('200','a') ); |
222 |
|
marc('900','b', rec('200','b') ); |
223 |
|
marc('900','c', rec('200','c') ); |
224 |
|
|
225 |
$log->logdie("expected tag HASH and got $tag") unless (ref($tag) eq 'HASH'); |
will create: |
|
$format = $tag->{'value'} || $tag->{'content'}; |
|
226 |
|
|
227 |
$log->debug("format: $format"); |
900a-1 900a-2 900a-3 900b-1 900c-1 |
228 |
|
900b-2 |
229 |
|
|
230 |
my @v; |
There is also support for returning next or specific using: |
|
if ($self->{'lookup_regex'} && $format =~ $self->{'lookup_regex'}) { |
|
|
@v = $self->fill_in_to_arr($rec,$format); |
|
|
} else { |
|
|
@v = $self->parse_to_arr($rec,$format); |
|
|
} |
|
|
next if (! @v); |
|
231 |
|
|
232 |
if ($tag->{'sort'}) { |
while (my $mf = WebPAC::Normalize:_get_marc_fields( fetch_next => 1 ) ) { |
233 |
@v = $self->sort_arr(@v); |
# do something with $mf |
234 |
} |
} |
235 |
|
|
236 |
# use format? |
will always return fields from next MARC record or |
|
if ($tag->{'format_name'}) { |
|
|
@v = map { $self->apply_format($tag->{'format_name'},$tag->{'format_delimiter'},$_) } @v; |
|
|
} |
|
237 |
|
|
238 |
# delimiter will join repeatable fields |
my $mf = WebPAC::Normalize::_get_marc_fields( offset => 42 ); |
|
if ($tag->{'delimiter'}) { |
|
|
@v = ( join($tag->{'delimiter'}, @v) ); |
|
|
} |
|
239 |
|
|
240 |
# default types |
will return 42th copy record (if it exists). |
|
my @types = qw(display search); |
|
|
# override by type attribute |
|
|
@types = ( $tag->{'type'} ) if ($tag->{'type'}); |
|
|
|
|
|
foreach my $type (@types) { |
|
|
# append to previous line? |
|
|
$log->debug("type: $type ",sub { join(" ",@v) }, " ", $row->{'append'} || 'no append'); |
|
|
if ($tag->{'append'}) { |
|
|
|
|
|
# I will delimit appended part with |
|
|
# delimiter (or ,) |
|
|
my $d = $tag->{'delimiter'}; |
|
|
# default delimiter |
|
|
$d ||= " "; |
|
|
|
|
|
my $last = pop @{$row->{$type}}; |
|
|
$d = "" if (! $last); |
|
|
$last .= $d . join($d, @v); |
|
|
push @{$row->{$type}}, $last; |
|
241 |
|
|
242 |
} else { |
=cut |
243 |
push @{$row->{$type}}, @v; |
|
244 |
} |
sub _get_marc_fields { |
245 |
} |
|
246 |
|
my $arg = {@_}; |
247 |
|
warn "### _get_marc_fields arg: ", dump($arg), $/ if ($debug > 2); |
248 |
|
my $offset = $marc_fetch_offset; |
249 |
|
if ($arg->{offset}) { |
250 |
|
$offset = $arg->{offset}; |
251 |
|
} elsif($arg->{fetch_next}) { |
252 |
|
$marc_fetch_offset++; |
253 |
|
} |
254 |
|
|
255 |
|
return if (! $marc_record || ref($marc_record) ne 'ARRAY'); |
256 |
|
|
257 |
|
warn "### full marc_record = ", dump( @{ $marc_record }), $/ if ($debug > 2); |
258 |
|
|
259 |
|
my $marc_rec = $marc_record->[ $offset ]; |
260 |
|
|
261 |
|
warn "## _get_marc_fields (at offset: $offset) -- marc_record = ", dump( @$marc_rec ), $/ if ($debug > 1); |
262 |
|
|
263 |
|
return if (! $marc_rec || ref($marc_rec) ne 'ARRAY' || $#{ $marc_rec } < 0); |
264 |
|
|
265 |
|
# first, sort all existing fields |
266 |
|
# XXX might not be needed, but modern perl might randomize elements in hash |
267 |
|
my @sorted_marc_record = sort { |
268 |
|
$a->[0] . ( $a->[3] || '' ) cmp $b->[0] . ( $b->[3] || '') |
269 |
|
} @{ $marc_rec }; |
270 |
|
|
271 |
|
@sorted_marc_record = @{ $marc_rec }; ### FIXME disable sorting |
272 |
|
|
273 |
|
# output marc fields |
274 |
|
my @m; |
275 |
|
|
276 |
|
# count unique field-subfields (used for offset when walking to next subfield) |
277 |
|
my $u; |
278 |
|
map { $u->{ $_->[0] . ( $_->[3] || '') }++ } @sorted_marc_record; |
279 |
|
|
280 |
|
if ($debug) { |
281 |
|
warn "## marc_repeatable_subfield = ", dump( $marc_repeatable_subfield ), $/ if ( $marc_repeatable_subfield ); |
282 |
|
warn "## marc_record[$offset] = ", dump( $marc_rec ), $/; |
283 |
|
warn "## sorted_marc_record = ", dump( \@sorted_marc_record ), $/; |
284 |
|
warn "## subfield count = ", dump( $u ), $/; |
285 |
|
} |
286 |
|
|
287 |
|
my $len = $#sorted_marc_record; |
288 |
|
my $visited; |
289 |
|
my $i = 0; |
290 |
|
my $field; |
291 |
|
|
292 |
|
foreach ( 0 .. $len ) { |
293 |
|
|
294 |
|
# find next element which isn't visited |
295 |
|
while ($visited->{$i}) { |
296 |
|
$i = ($i + 1) % ($len + 1); |
297 |
} |
} |
298 |
|
|
299 |
if ($row) { |
# mark it visited |
300 |
$row->{'tag'} = $field; |
$visited->{$i}++; |
301 |
|
|
302 |
# TODO: name_sigular, name_plural |
my $row = dclone( $sorted_marc_record[$i] ); |
303 |
my $name = $self->{'import_xml'}->{'indexer'}->{$field}->{'name'}; |
|
304 |
my $row_name = $name ? $self->_x($name) : $field; |
# field and subfield which is key for |
305 |
|
# marc_repeatable_subfield and u |
306 |
# post-sort all values in field |
my $fsf = $row->[0] . ( $row->[3] || '' ); |
307 |
if ($self->{'import_xml'}->{'indexer'}->{$field}->{'sort'}) { |
|
308 |
$log->warn("sort at field tag not implemented"); |
if ($debug > 1) { |
309 |
} |
|
310 |
|
print "### field so far [", $#$field, "] : ", dump( $field ), " ", $field ? 'T' : 'F', $/; |
311 |
|
print "### this [$i]: ", dump( $row ),$/; |
312 |
|
print "### sf: ", $row->[3], " vs ", $field->[3], |
313 |
|
$marc_repeatable_subfield->{ $row->[0] . $row->[3] } ? ' (repeatable)' : '', $/, |
314 |
|
if ($#$field >= 0); |
315 |
|
|
316 |
$ds->{$row_name} = $row; |
} |
317 |
|
|
318 |
|
# if field exists |
319 |
|
if ( $#$field >= 0 ) { |
320 |
|
if ( |
321 |
|
$row->[0] ne $field->[0] || # field |
322 |
|
$row->[1] ne $field->[1] || # i1 |
323 |
|
$row->[2] ne $field->[2] # i2 |
324 |
|
) { |
325 |
|
push @m, $field; |
326 |
|
warn "## saved/1 ", dump( $field ),$/ if ($debug); |
327 |
|
$field = $row; |
328 |
|
|
329 |
|
} elsif ( |
330 |
|
( $row->[3] lt $field->[-2] ) # subfield which is not next (e.g. a after c) |
331 |
|
|| |
332 |
|
( $row->[3] eq $field->[-2] && # same subfield, but not repeatable |
333 |
|
! $marc_repeatable_subfield->{ $fsf } |
334 |
|
) |
335 |
|
) { |
336 |
|
push @m, $field; |
337 |
|
warn "## saved/2 ", dump( $field ),$/ if ($debug); |
338 |
|
$field = $row; |
339 |
|
|
340 |
|
} else { |
341 |
|
# append new subfields to existing field |
342 |
|
push @$field, ( $row->[3], $row->[4] ); |
343 |
|
} |
344 |
|
} else { |
345 |
|
# insert first field |
346 |
|
$field = $row; |
347 |
|
} |
348 |
|
|
349 |
$log->debug("row $field: ",sub { Dumper($row) }); |
if (! $marc_repeatable_subfield->{ $fsf }) { |
350 |
|
# make step to next subfield |
351 |
|
$i = ($i + $u->{ $fsf } ) % ($len + 1); |
352 |
} |
} |
353 |
|
} |
354 |
|
|
355 |
|
if ($#$field >= 0) { |
356 |
|
push @m, $field; |
357 |
|
warn "## saved/3 ", dump( $field ),$/ if ($debug); |
358 |
} |
} |
359 |
|
|
360 |
$self->{'db'}->save_ds( |
return \@m; |
361 |
id => $id, |
} |
|
ds => $ds, |
|
|
prefix => $self->{prefix}, |
|
|
) if ($self->{'db'}); |
|
362 |
|
|
363 |
$log->debug("ds: ", sub { Dumper($ds) }); |
=head2 _debug |
364 |
|
|
365 |
$log->logconfess("data structure returned is not array any more!") if wantarray; |
Change level of debug warnings |
366 |
|
|
367 |
return $ds; |
_debug( 2 ); |
368 |
|
|
369 |
} |
=cut |
370 |
|
|
371 |
=head2 parse |
sub _debug { |
372 |
|
my $l = shift; |
373 |
|
return $debug unless defined($l); |
374 |
|
warn "debug level $l",$/ if ($l > 0); |
375 |
|
$debug = $l; |
376 |
|
} |
377 |
|
|
378 |
Perform smart parsing of string, skipping delimiters for fields which aren't |
=head1 Functions to create C<data_structure> |
|
defined. It can also eval code in format starting with C<eval{...}> and |
|
|
return output or nothing depending on eval code. |
|
379 |
|
|
380 |
my $text = $webpac->parse($rec,'eval{"v901^a" eq "Deskriptor"}descriptor: v250^a', $i); |
Those functions generally have to first in your normalization file. |
381 |
|
|
382 |
Filters are implemented here. While simple form of filters looks like this: |
=head2 tag |
383 |
|
|
384 |
filter{name_of_filter} |
Define new tag for I<search> and I<display>. |
385 |
|
|
386 |
but, filters can also have variable number of parametars like this: |
tag('Title', rec('200','a') ); |
387 |
|
|
|
filter{name_of_filter(param,param,param)} |
|
388 |
|
|
389 |
=cut |
=cut |
390 |
|
|
391 |
my $warn_once; |
sub tag { |
392 |
|
my $name = shift or die "tag needs name as first argument"; |
393 |
sub parse { |
my @o = grep { defined($_) && $_ ne '' } @_; |
394 |
my $self = shift; |
return unless (@o); |
395 |
|
$out->{$name}->{tag} = $name; |
396 |
|
$out->{$name}->{search} = \@o; |
397 |
|
$out->{$name}->{display} = \@o; |
398 |
|
} |
399 |
|
|
400 |
my ($rec, $format_utf8, $i) = @_; |
=head2 display |
401 |
|
|
402 |
return if (! $format_utf8); |
Define tag just for I<display> |
403 |
|
|
404 |
my $log = $self->_get_logger(); |
@v = display('Title', rec('200','a') ); |
405 |
|
|
406 |
$log->logconfess("need HASH as first argument!") if ($rec !~ /HASH/o); |
=cut |
407 |
|
|
408 |
$i = 0 if (! $i); |
sub display { |
409 |
|
my $name = shift or die "display needs name as first argument"; |
410 |
|
my @o = grep { defined($_) && $_ ne '' } @_; |
411 |
|
return unless (@o); |
412 |
|
$out->{$name}->{tag} = $name; |
413 |
|
$out->{$name}->{display} = \@o; |
414 |
|
} |
415 |
|
|
416 |
my $format = $self->_x($format_utf8) || $log->logconfess("can't convert '$format_utf8' from UTF-8 to ",$self->{'code_page'}); |
=head2 search |
417 |
|
|
418 |
my @out; |
Prepare values just for I<search> |
419 |
|
|
420 |
$log->debug("format: $format"); |
@v = search('Title', rec('200','a') ); |
421 |
|
|
422 |
my $eval_code; |
=cut |
|
# remove eval{...} from beginning |
|
|
$eval_code = $1 if ($format =~ s/^eval{([^}]+)}//s); |
|
423 |
|
|
424 |
my $filter_name; |
sub search { |
425 |
# remove filter{...} from beginning |
my $name = shift or die "search needs name as first argument"; |
426 |
$filter_name = $1 if ($format =~ s/^filter{([^}]+)}//s); |
my @o = grep { defined($_) && $_ ne '' } @_; |
427 |
|
return unless (@o); |
428 |
|
$out->{$name}->{tag} = $name; |
429 |
|
$out->{$name}->{search} = \@o; |
430 |
|
} |
431 |
|
|
432 |
my $prefix; |
=head2 marc_leader |
|
my $all_found=0; |
|
433 |
|
|
434 |
while ($format =~ s/^(.*?)(v|s)(\d+)(?:\^(\w))?//s) { |
Setup fields within MARC leader or get leader |
435 |
|
|
436 |
my $del = $1 || ''; |
marc_leader('05','c'); |
437 |
$prefix ||= $del if ($all_found == 0); |
my $leader = marc_leader(); |
438 |
|
|
439 |
# repeatable index |
=cut |
|
my $r = $i; |
|
|
$r = 0 if (lc("$2") eq 's'); |
|
440 |
|
|
441 |
my $found = 0; |
sub marc_leader { |
442 |
my $tmp = $self->get_data(\$rec,$3,$4,$r,\$found); |
my ($offset,$value) = @_; |
443 |
|
|
444 |
if ($found) { |
if ($offset) { |
445 |
push @out, $del; |
$out->{' leader'}->{ $offset } = $value; |
446 |
push @out, $tmp; |
} else { |
447 |
$all_found += $found; |
return $out->{' leader'}; |
|
} |
|
448 |
} |
} |
449 |
|
} |
450 |
|
|
451 |
return if (! $all_found); |
=head2 marc |
452 |
|
|
453 |
my $out = join('',@out); |
Save value for MARC field |
454 |
|
|
455 |
if ($out) { |
marc('900','a', rec('200','a') ); |
456 |
# add rest of format (suffix) |
marc('001', rec('000') ); |
|
$out .= $format; |
|
457 |
|
|
458 |
# add prefix if not there |
=cut |
|
$out = $prefix . $out if ($out !~ m/^\Q$prefix\E/); |
|
459 |
|
|
460 |
$log->debug("result: $out"); |
sub marc { |
461 |
} |
my $f = shift or die "marc needs field"; |
462 |
|
die "marc field must be numer" unless ($f =~ /^\d+$/); |
463 |
|
|
464 |
if ($eval_code) { |
my $sf; |
465 |
my $eval = $self->fill_in($rec,$eval_code,$i) || return; |
if ($f >= 10) { |
466 |
$log->debug("about to eval{$eval} format: $out"); |
$sf = shift or die "marc needs subfield"; |
|
return if (! $self->_eval($eval)); |
|
467 |
} |
} |
468 |
|
|
469 |
if ($filter_name) { |
foreach (@_) { |
470 |
my @filter_args; |
my $v = $_; # make var read-write for Encode |
471 |
if ($filter_name =~ s/(\w+)\((.*)\)/$1/) { |
next unless (defined($v) && $v !~ /^\s*$/); |
472 |
@filter_args = split(/,/, $2); |
from_to($v, 'iso-8859-2', $marc_encoding) if ($marc_encoding); |
473 |
} |
my ($i1,$i2) = defined($marc_indicators->{$f}) ? @{ $marc_indicators->{$f} } : (' ',' '); |
474 |
if ($self->{'filter'}->{$filter_name}) { |
if (defined $sf) { |
475 |
$log->debug("about to filter{$filter_name} format: $out with arguments: ", join(",", @filter_args)); |
push @{ $marc_record->[ $marc_record_offset ] }, [ $f, $i1, $i2, $sf => $v ]; |
476 |
unshift @filter_args, $out; |
} else { |
477 |
$out = $self->{'filter'}->{$filter_name}->(@filter_args); |
push @{ $marc_record->[ $marc_record_offset ] }, [ $f, $v ]; |
|
return unless(defined($out)); |
|
|
$log->debug("filter result: $out"); |
|
|
} elsif (! $warn_once->{$filter_name}) { |
|
|
$log->warn("trying to use undefined filter $filter_name"); |
|
|
$warn_once->{$filter_name}++; |
|
478 |
} |
} |
479 |
} |
} |
|
|
|
|
return $out; |
|
480 |
} |
} |
481 |
|
|
482 |
=head2 parse_to_arr |
=head2 marc_repeatable_subfield |
483 |
|
|
484 |
Similar to C<parse>, but returns array of all repeatable fields |
Save values for MARC repetable subfield |
485 |
|
|
486 |
my @arr = $webpac->parse_to_arr($rec,'v250^a'); |
marc_repeatable_subfield('910', 'z', rec('909') ); |
487 |
|
|
488 |
=cut |
=cut |
489 |
|
|
490 |
sub parse_to_arr { |
sub marc_repeatable_subfield { |
491 |
my $self = shift; |
my ($f,$sf) = @_; |
492 |
|
die "marc_repeatable_subfield need field and subfield!\n" unless ($f && $sf); |
493 |
|
$marc_repeatable_subfield->{ $f . $sf }++; |
494 |
|
marc(@_); |
495 |
|
} |
496 |
|
|
497 |
my ($rec, $format_utf8) = @_; |
=head2 marc_indicators |
498 |
|
|
499 |
my $log = $self->_get_logger(); |
Set both indicators for MARC field |
500 |
|
|
501 |
$log->logconfess("need HASH as first argument!") if ($rec !~ /HASH/o); |
marc_indicators('900', ' ', 1); |
|
return if (! $format_utf8); |
|
502 |
|
|
503 |
my $i = 0; |
Any indicator value other than C<0-9> will be treated as undefined. |
|
my @arr; |
|
504 |
|
|
505 |
while (my $v = $self->parse($rec,$format_utf8,$i++)) { |
=cut |
|
push @arr, $v; |
|
|
} |
|
506 |
|
|
507 |
$log->debug("format '$format_utf8' returned ",--$i," elements: ", sub { join(" | ",@arr) }) if (@arr); |
sub marc_indicators { |
508 |
|
my $f = shift || die "marc_indicators need field!\n"; |
509 |
|
my ($i1,$i2) = @_; |
510 |
|
die "marc_indicators($f, ...) need i1!\n" unless(defined($i1)); |
511 |
|
die "marc_indicators($f, $i1, ...) need i2!\n" unless(defined($i2)); |
512 |
|
|
513 |
return @arr; |
$i1 = ' ' if ($i1 !~ /^\d$/); |
514 |
|
$i2 = ' ' if ($i2 !~ /^\d$/); |
515 |
|
@{ $marc_indicators->{$f} } = ($i1,$i2); |
516 |
} |
} |
517 |
|
|
518 |
|
=head2 marc_compose |
519 |
|
|
520 |
|
Save values for each MARC subfield explicitly |
521 |
|
|
522 |
|
marc_compose('900', |
523 |
|
'a', rec('200','a') |
524 |
|
'b', rec('201','a') |
525 |
|
'a', rec('200','b') |
526 |
|
'c', rec('200','c') |
527 |
|
); |
528 |
|
|
529 |
|
If you specify C<+> for subfield, value will be appended |
530 |
|
to previous defined subfield. |
531 |
|
|
532 |
|
=cut |
533 |
|
|
534 |
|
sub marc_compose { |
535 |
|
my $f = shift or die "marc_compose needs field"; |
536 |
|
die "marc_compose field must be numer" unless ($f =~ /^\d+$/); |
537 |
|
|
538 |
|
my ($i1,$i2) = defined($marc_indicators->{$f}) ? @{ $marc_indicators->{$f} } : (' ',' '); |
539 |
|
my $m = [ $f, $i1, $i2 ]; |
540 |
|
|
541 |
|
warn "### marc_compose input subfields = ", dump(@_),$/ if ($debug > 2); |
542 |
|
|
543 |
|
if ($#_ % 2 != 1) { |
544 |
|
die "ERROR: marc_compose",dump($f,@_)," not valid (must be even).\nDo you need to add first() or join() around some argument?\n"; |
545 |
|
} |
546 |
|
|
547 |
|
while (@_) { |
548 |
|
my $sf = shift; |
549 |
|
my $v = shift; |
550 |
|
|
551 |
=head2 fill_in |
next unless (defined($v) && $v !~ /^\s*$/); |
552 |
|
from_to($v, 'iso-8859-2', $marc_encoding) if ($marc_encoding); |
553 |
|
warn "## ++ marc_compose($f,$sf,$v) ", dump( $m ),$/ if ($debug > 1); |
554 |
|
if ($sf ne '+') { |
555 |
|
push @$m, ( $sf, $v ); |
556 |
|
} else { |
557 |
|
$m->[ $#$m ] .= $v; |
558 |
|
} |
559 |
|
} |
560 |
|
|
561 |
Workhourse of all: takes record from in-memory structure of database and |
warn "## marc_compose current marc = ", dump( $m ),$/ if ($debug > 1); |
|
strings with placeholders and returns string or array of with substituted |
|
|
values from record. |
|
562 |
|
|
563 |
my $text = $webpac->fill_in($rec,'v250^a'); |
push @{ $marc_record->[ $marc_record_offset ] }, $m if ($#{$m} > 2); |
564 |
|
} |
565 |
|
|
566 |
Optional argument is ordinal number for repeatable fields. By default, |
=head2 marc_duplicate |
|
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. |
|
567 |
|
|
568 |
my $text = $webpac->fill_in($rec,'Title: v250^a',1); |
Generate copy of current MARC record and continue working on copy |
569 |
|
|
570 |
This function B<does not> perform parsing of format to inteligenty skip |
marc_duplicate(); |
|
delimiters before fields which aren't used. |
|
571 |
|
|
572 |
This method will automatically decode UTF-8 string to local code page |
Copies can be accessed using C<< _get_marc_fields( fetch_next => 1 ) >> or |
573 |
if needed. |
C<< _get_marc_fields( offset => 42 ) >>. |
574 |
|
|
575 |
=cut |
=cut |
576 |
|
|
577 |
sub fill_in { |
sub marc_duplicate { |
578 |
my $self = shift; |
my $m = $marc_record->[ -1 ]; |
579 |
|
die "can't duplicate record which isn't defined" unless ($m); |
580 |
|
push @{ $marc_record }, dclone( $m ); |
581 |
|
warn "## marc_duplicate = ", dump(@$marc_record), $/ if ($debug > 1); |
582 |
|
$marc_record_offset = $#{ $marc_record }; |
583 |
|
warn "## marc_record_offset = $marc_record_offset", $/ if ($debug > 1); |
584 |
|
} |
585 |
|
|
586 |
my $log = $self->_get_logger(); |
=head2 marc_remove |
587 |
|
|
588 |
my $rec = shift || $log->logconfess("need data record"); |
Remove some field or subfield from MARC record. |
|
my $format = shift || $log->logconfess("need format to parse"); |
|
|
# iteration (for repeatable fields) |
|
|
my $i = shift || 0; |
|
589 |
|
|
590 |
$log->logdie("infitite loop in format $format") if ($i > ($self->{'max_mfn'} || 9999)); |
marc_remove('200'); |
591 |
|
marc_remove('200','a'); |
592 |
|
|
593 |
# FIXME remove for speedup? |
This will erase field C<200> or C<200^a> from current MARC record. |
|
$log->logconfess("need HASH as first argument!") if ($rec !~ /HASH/o); |
|
594 |
|
|
595 |
if (utf8::is_utf8($format)) { |
This is useful after calling C<marc_duplicate> or on it's own (but, you |
596 |
$format = $self->_x($format); |
should probably just remove that subfield definition if you are not |
597 |
} |
using C<marc_duplicate>). |
598 |
|
|
599 |
|
FIXME: support fields < 10. |
600 |
|
|
601 |
my $found = 0; |
=cut |
602 |
|
|
603 |
my $eval_code; |
sub marc_remove { |
604 |
# remove eval{...} from beginning |
my ($f, $sf) = @_; |
|
$eval_code = $1 if ($format =~ s/^eval{([^}]+)}//s); |
|
605 |
|
|
606 |
my $filter_name; |
die "marc_remove needs record number" unless defined($f); |
|
# remove filter{...} from beginning |
|
|
$filter_name = $1 if ($format =~ s/^filter{([^}]+)}//s); |
|
607 |
|
|
608 |
# do actual replacement of placeholders |
my $marc = $marc_record->[ $marc_record_offset ]; |
|
# repeatable fields |
|
|
$format =~ s/v(\d+)(?:\^(\w))?/$self->get_data(\$rec,$1,$2,$i,\$found)/ges; |
|
|
# non-repeatable fields |
|
|
$format =~ s/s(\d+)(?:\^(\w))?/$self->get_data(\$rec,$1,$2,0,\$found)/ges; |
|
609 |
|
|
610 |
if ($found) { |
warn "### marc_remove before = ", dump( $marc ), $/ if ($debug > 2); |
611 |
$log->debug("format: $format"); |
|
612 |
if ($eval_code) { |
my $i = 0; |
613 |
my $eval = $self->fill_in($rec,$eval_code,$i); |
foreach ( 0 .. $#{ $marc } ) { |
614 |
return if (! $self->_eval($eval)); |
last unless (defined $marc->[$i]); |
615 |
} |
warn "#### working on ",dump( @{ $marc->[$i] }), $/ if ($debug > 3); |
616 |
if ($filter_name && $self->{'filter'}->{$filter_name}) { |
if ($marc->[$i]->[0] eq $f) { |
617 |
$log->debug("filter '$filter_name' for $format"); |
if (! defined $sf) { |
618 |
$format = $self->{'filter'}->{$filter_name}->($format); |
# remove whole field |
619 |
return unless(defined($format)); |
splice @$marc, $i, 1; |
620 |
$log->debug("filter result: $format"); |
warn "#### slice \@\$marc, $i, 1 = ",dump( @{ $marc }), $/ if ($debug > 3); |
621 |
} |
$i--; |
|
# do we have lookups? |
|
|
if ($self->{'lookup'}) { |
|
|
if ($self->{'lookup'}->can('lookup')) { |
|
|
my @lookup = $self->{lookup}->lookup($format); |
|
|
$log->debug("lookup $format", join(", ", @lookup)); |
|
|
return @lookup; |
|
622 |
} else { |
} else { |
623 |
$log->warn("Have lookup object but can't invoke lookup method"); |
foreach my $j ( 0 .. (( $#{ $marc->[$i] } - 3 ) / 2) ) { |
624 |
|
my $o = ($j * 2) + 3; |
625 |
|
if ($marc->[$i]->[$o] eq $sf) { |
626 |
|
# remove subfield |
627 |
|
splice @{$marc->[$i]}, $o, 2; |
628 |
|
warn "#### slice \@{\$marc->[$i]}, $o, 2 = ", dump( @{ $marc }), $/ if ($debug > 3); |
629 |
|
# is record now empty? |
630 |
|
if ($#{ $marc->[$i] } == 2) { |
631 |
|
splice @$marc, $i, 1; |
632 |
|
warn "#### slice \@\$marc, $i, 1 = ", dump( @{ $marc }), $/ if ($debug > 3); |
633 |
|
$i--; |
634 |
|
}; |
635 |
|
} |
636 |
|
} |
637 |
} |
} |
|
} else { |
|
|
return $format; |
|
638 |
} |
} |
639 |
} else { |
$i++; |
|
return; |
|
640 |
} |
} |
641 |
|
|
642 |
|
warn "### marc_remove($f", $sf ? ",$sf" : "", ") after = ", dump( $marc ), $/ if ($debug > 2); |
643 |
|
|
644 |
|
$marc_record->[ $marc_record_offset ] = $marc; |
645 |
|
|
646 |
|
warn "## full marc_record = ", dump( @{ $marc_record }), $/ if ($debug > 1); |
647 |
} |
} |
648 |
|
|
649 |
|
=head2 marc_original_order |
650 |
|
|
651 |
=head2 fill_in_to_arr |
Copy all subfields preserving original order to marc field. |
652 |
|
|
653 |
Similar to C<fill_in>, but returns array of all repeatable fields. Usable |
marc_original_order( marc_field_number, original_input_field_number ); |
654 |
for fields which have lookups, so they shouldn't be parsed but rather |
|
655 |
C<fill_id>ed. |
Please note that field numbers are consistent with other commands (marc |
656 |
|
field number first), but somewhat counter-intuitive (destination and then |
657 |
|
source). |
658 |
|
|
659 |
|
You might want to use this command if you are just renaming subfields or |
660 |
|
using pre-processing modify_record in C<config.yml> and don't need any |
661 |
|
post-processing or want to preserve order of original subfields. |
662 |
|
|
|
my @arr = $webpac->fill_in_to_arr($rec,'[v900];;[v250^a]'); |
|
663 |
|
|
664 |
=cut |
=cut |
665 |
|
|
666 |
sub fill_in_to_arr { |
sub marc_original_order { |
|
my $self = shift; |
|
667 |
|
|
668 |
my ($rec, $format_utf8) = @_; |
my ($to, $from) = @_; |
669 |
|
die "marc_original_order needs from and to fields\n" unless ($from && $to); |
670 |
|
|
671 |
my $log = $self->_get_logger(); |
return unless defined($rec->{$from}); |
672 |
|
|
673 |
$log->logconfess("need HASH as first argument!") if ($rec !~ /HASH/o); |
my $r = $rec->{$from}; |
674 |
return if (! $format_utf8); |
die "record field $from isn't array\n" unless (ref($r) eq 'ARRAY'); |
675 |
|
|
676 |
my $i = 0; |
my ($i1,$i2) = defined($marc_indicators->{$to}) ? @{ $marc_indicators->{$to} } : (' ',' '); |
677 |
my @arr; |
warn "## marc_original_order($to,$from) source = ", dump( $r ),$/ if ($debug > 1); |
678 |
|
|
679 |
while (my @v = $self->fill_in($rec,$format_utf8,$i++)) { |
foreach my $d (@$r) { |
|
push @arr, @v; |
|
|
} |
|
680 |
|
|
681 |
$log->debug("format '$format_utf8' returned ",--$i," elements: ", sub { join(" | ",@arr) }) if (@arr); |
if (! defined($d->{subfields}) && ref($d->{subfields}) ne 'ARRAY') { |
682 |
|
warn "# marc_original_order($to,$from): field $from doesn't have subfields specification\n"; |
683 |
|
next; |
684 |
|
} |
685 |
|
|
686 |
|
my @sfs = @{ $d->{subfields} }; |
687 |
|
|
688 |
return @arr; |
die "field $from doesn't have even number of subfields specifications\n" unless($#sfs % 2 == 1); |
|
} |
|
689 |
|
|
690 |
|
warn "#--> d: ",dump($d), "\n#--> sfs: ",dump(@sfs),$/ if ($debug > 2); |
691 |
|
|
692 |
=head2 get_data |
my $m = [ $to, $i1, $i2 ]; |
693 |
|
|
694 |
Returns value from record. |
while (my $sf = shift @sfs) { |
695 |
|
|
696 |
my $text = $self->get_data(\$rec,$f,$sf,$i,\$found); |
warn "#--> sf: ",dump($sf), $/ if ($debug > 2); |
697 |
|
my $offset = shift @sfs; |
698 |
|
die "corrupted sufields specification for field $from\n" unless defined($offset); |
699 |
|
|
700 |
Arguments are: |
my $v; |
701 |
record reference C<$rec>, |
if (ref($d->{$sf}) eq 'ARRAY') { |
702 |
field C<$f>, |
$v = $d->{$sf}->[$offset] if (defined($d->{$sf}->[$offset])); |
703 |
optional subfiled C<$sf>, |
} elsif ($offset == 0) { |
704 |
index for repeatable values C<$i>. |
$v = $d->{$sf}; |
705 |
|
} else { |
706 |
|
die "field $from subfield '$sf' need occurence $offset which doesn't exist", dump($d->{$sf}); |
707 |
|
} |
708 |
|
push @$m, ( $sf, $v ) if (defined($v)); |
709 |
|
} |
710 |
|
|
711 |
Optinal variable C<$found> will be incremeted if there |
if ($#{$m} > 2) { |
712 |
is field. |
push @{ $marc_record->[ $marc_record_offset ] }, $m; |
713 |
|
} |
714 |
|
} |
715 |
|
|
716 |
Returns value or empty string. |
warn "## marc_record = ", dump( $marc_record ),$/ if ($debug > 1); |
717 |
|
} |
718 |
|
|
|
=cut |
|
719 |
|
|
720 |
sub get_data { |
=head1 Functions to extract data from input |
|
my $self = shift; |
|
721 |
|
|
722 |
my ($rec,$f,$sf,$i,$found) = @_; |
This function should be used inside functions to create C<data_structure> described |
723 |
|
above. |
724 |
|
|
725 |
if ($$rec->{$f}) { |
=head2 rec1 |
726 |
return '' if (! $$rec->{$f}->[$i]); |
|
727 |
no strict 'refs'; |
Return all values in some field |
728 |
if ($sf && $$rec->{$f}->[$i]->{$sf}) { |
|
729 |
$$found++ if (defined($$found)); |
@v = rec1('200') |
730 |
return $$rec->{$f}->[$i]->{$sf}; |
|
731 |
} elsif (! $sf && $$rec->{$f}->[$i]) { |
TODO: order of values is probably same as in source data, need to investigate that |
732 |
$$found++ if (defined($$found)); |
|
733 |
# it still might have subfield, just |
=cut |
734 |
# not specified, so we'll dump all |
|
735 |
if ($$rec->{$f}->[$i] =~ /HASH/o) { |
sub rec1 { |
736 |
my $out; |
my $f = shift; |
737 |
foreach my $k (keys %{$$rec->{$f}->[$i]}) { |
warn "rec1($f) = ", dump( $rec->{$f} ), $/ if ($debug > 1); |
738 |
$out .= $$rec->{$f}->[$i]->{$k}." "; |
return unless (defined($rec) && defined($rec->{$f})); |
739 |
} |
warn "rec1($f) = ", dump( $rec->{$f} ), $/ if ($debug > 1); |
740 |
return $out; |
if (ref($rec->{$f}) eq 'ARRAY') { |
741 |
|
return map { |
742 |
|
if (ref($_) eq 'HASH') { |
743 |
|
values %{$_}; |
744 |
} else { |
} else { |
745 |
return $$rec->{$f}->[$i]; |
$_; |
746 |
} |
} |
747 |
|
} @{ $rec->{$f} }; |
748 |
|
} elsif( defined($rec->{$f}) ) { |
749 |
|
return $rec->{$f}; |
750 |
|
} |
751 |
|
} |
752 |
|
|
753 |
|
=head2 rec2 |
754 |
|
|
755 |
|
Return all values in specific field and subfield |
756 |
|
|
757 |
|
@v = rec2('200','a') |
758 |
|
|
759 |
|
=cut |
760 |
|
|
761 |
|
sub rec2 { |
762 |
|
my $f = shift; |
763 |
|
return unless (defined($rec && $rec->{$f})); |
764 |
|
my $sf = shift; |
765 |
|
warn "rec2($f,$sf) = ", dump( $rec->{$f} ), $/ if ($debug > 1); |
766 |
|
return map { |
767 |
|
if (ref($_->{$sf}) eq 'ARRAY') { |
768 |
|
@{ $_->{$sf} }; |
769 |
} else { |
} else { |
770 |
return ''; |
$_->{$sf}; |
771 |
} |
} |
772 |
|
} grep { ref($_) eq 'HASH' && $_->{$sf} } @{ $rec->{$f} }; |
773 |
|
} |
774 |
|
|
775 |
|
=head2 rec |
776 |
|
|
777 |
|
syntaxtic sugar for |
778 |
|
|
779 |
|
@v = rec('200') |
780 |
|
@v = rec('200','a') |
781 |
|
|
782 |
|
=cut |
783 |
|
|
784 |
|
sub rec { |
785 |
|
my @out; |
786 |
|
if ($#_ == 0) { |
787 |
|
@out = rec1(@_); |
788 |
|
} elsif ($#_ == 1) { |
789 |
|
@out = rec2(@_); |
790 |
|
} |
791 |
|
if (@out) { |
792 |
|
return @out; |
793 |
} else { |
} else { |
794 |
return ''; |
return ''; |
795 |
} |
} |
796 |
} |
} |
797 |
|
|
798 |
|
=head2 regex |
799 |
|
|
800 |
=head2 apply_format |
Apply regex to some or all values |
801 |
|
|
802 |
Apply format specified in tag with C<format_name="name"> and |
@v = regex( 's/foo/bar/g', @v ); |
|
C<format_delimiter=";;">. |
|
803 |
|
|
804 |
my $text = $webpac->apply_format($format_name,$format_delimiter,$data); |
=cut |
805 |
|
|
806 |
Formats can contain C<lookup{...}> if you need them. |
sub regex { |
807 |
|
my $r = shift; |
808 |
|
my @out; |
809 |
|
#warn "r: $r\n", dump(\@_); |
810 |
|
foreach my $t (@_) { |
811 |
|
next unless ($t); |
812 |
|
eval "\$t =~ $r"; |
813 |
|
push @out, $t if ($t && $t ne ''); |
814 |
|
} |
815 |
|
return @out; |
816 |
|
} |
817 |
|
|
818 |
|
=head2 prefix |
819 |
|
|
820 |
|
Prefix all values with a string |
821 |
|
|
822 |
|
@v = prefix( 'my_', @v ); |
823 |
|
|
824 |
=cut |
=cut |
825 |
|
|
826 |
sub apply_format { |
sub prefix { |
827 |
my $self = shift; |
my $p = shift or return; |
828 |
|
return map { $p . $_ } grep { defined($_) } @_; |
829 |
|
} |
830 |
|
|
831 |
my ($name,$delimiter,$data) = @_; |
=head2 suffix |
832 |
|
|
833 |
my $log = $self->_get_logger(); |
suffix all values with a string |
834 |
|
|
835 |
if (! $self->{'import_xml'}->{'format'}->{$name}) { |
@v = suffix( '_my', @v ); |
836 |
$log->warn("<format name=\"$name\"> is not defined in ",$self->{'import_xml_file'}); |
|
837 |
return $data; |
=cut |
|
} |
|
838 |
|
|
839 |
$log->warn("no delimiter for format $name") if (! $delimiter); |
sub suffix { |
840 |
|
my $s = shift or die "suffix needs string as first argument"; |
841 |
|
return map { $_ . $s } grep { defined($_) } @_; |
842 |
|
} |
843 |
|
|
844 |
my $format = $self->_x($self->{'import_xml'}->{'format'}->{$name}->{'content'}) || $log->logdie("can't find format '$name'"); |
=head2 surround |
845 |
|
|
846 |
my @data = split(/\Q$delimiter\E/, $data); |
surround all values with a two strings |
847 |
|
|
848 |
my $out = sprintf($format, @data); |
@v = surround( 'prefix_', '_suffix', @v ); |
|
$log->debug("using format $name [$format] on $data to produce: $out"); |
|
849 |
|
|
850 |
if ($self->{'lookup_regex'} && $out =~ $self->{'lookup_regex'}) { |
=cut |
|
return $self->{'lookup'}->lookup($out); |
|
|
} else { |
|
|
return $out; |
|
|
} |
|
851 |
|
|
852 |
|
sub surround { |
853 |
|
my $p = shift or die "surround need prefix as first argument"; |
854 |
|
my $s = shift or die "surround needs suffix as second argument"; |
855 |
|
return map { $p . $_ . $s } grep { defined($_) } @_; |
856 |
} |
} |
857 |
|
|
858 |
=head2 sort_arr |
=head2 first |
859 |
|
|
860 |
Sort array ignoring case and html in data |
Return first element |
861 |
|
|
862 |
my @sorted = $webpac->sort_arr(@unsorted); |
$v = first( @v ); |
863 |
|
|
864 |
=cut |
=cut |
865 |
|
|
866 |
sub sort_arr { |
sub first { |
867 |
my $self = shift; |
my $r = shift; |
868 |
|
return $r; |
869 |
|
} |
870 |
|
|
871 |
my $log = $self->_get_logger(); |
=head2 lookup |
872 |
|
|
873 |
# FIXME add Schwartzian Transformation? |
Consult lookup hashes for some value |
874 |
|
|
875 |
my @sorted = sort { |
@v = lookup( $v ); |
876 |
$a =~ s#<[^>]+/*>##; |
@v = lookup( @v ); |
|
$b =~ s#<[^>]+/*>##; |
|
|
lc($b) cmp lc($a) |
|
|
} @_; |
|
|
$log->debug("sorted values: ",sub { join(", ",@sorted) }); |
|
877 |
|
|
878 |
return @sorted; |
=cut |
879 |
|
|
880 |
|
sub lookup { |
881 |
|
my $k = shift or return; |
882 |
|
return unless (defined($lookup->{$k})); |
883 |
|
if (ref($lookup->{$k}) eq 'ARRAY') { |
884 |
|
return @{ $lookup->{$k} }; |
885 |
|
} else { |
886 |
|
return $lookup->{$k}; |
887 |
|
} |
888 |
} |
} |
889 |
|
|
890 |
|
=head2 config |
891 |
|
|
892 |
=head1 INTERNAL METHODS |
Consult config values stored in C<config.yml> |
893 |
|
|
894 |
=head2 _sort_by_order |
# return database code (key under databases in yaml) |
895 |
|
$database_code = config(); # use _ from hash |
896 |
|
$database_name = config('name'); |
897 |
|
$database_input_name = config('input name'); |
898 |
|
$tag = config('input normalize tag'); |
899 |
|
|
900 |
Sort xml tags data structure accoding to C<order=""> attribute. |
Up to three levels are supported. |
901 |
|
|
902 |
=cut |
=cut |
903 |
|
|
904 |
sub _sort_by_order { |
sub config { |
905 |
my $self = shift; |
return unless ($config); |
906 |
|
|
907 |
|
my $p = shift; |
908 |
|
|
909 |
|
$p ||= ''; |
910 |
|
|
911 |
|
my $v; |
912 |
|
|
913 |
my $va = $self->{'import_xml'}->{'indexer'}->{$a}->{'order'} || |
warn "### getting config($p)\n" if ($debug > 1); |
|
$self->{'import_xml'}->{'indexer'}->{$a}; |
|
|
my $vb = $self->{'import_xml'}->{'indexer'}->{$b}->{'order'} || |
|
|
$self->{'import_xml'}->{'indexer'}->{$b}; |
|
914 |
|
|
915 |
return $va <=> $vb; |
my @p = split(/\s+/,$p); |
916 |
|
if ($#p < 0) { |
917 |
|
$v = $config->{ '_' }; # special, database code |
918 |
|
} else { |
919 |
|
|
920 |
|
my $c = dclone( $config ); |
921 |
|
|
922 |
|
foreach my $k (@p) { |
923 |
|
warn "### k: $k c = ",dump($c),$/ if ($debug > 1); |
924 |
|
if (ref($c) eq 'ARRAY') { |
925 |
|
$c = shift @$c; |
926 |
|
warn "config($p) taking first occurence of '$k', probably not what you wanted!\n"; |
927 |
|
last; |
928 |
|
} |
929 |
|
|
930 |
|
if (! defined($c->{$k}) ) { |
931 |
|
$c = undef; |
932 |
|
last; |
933 |
|
} else { |
934 |
|
$c = $c->{$k}; |
935 |
|
} |
936 |
|
} |
937 |
|
$v = $c if ($c); |
938 |
|
|
939 |
|
} |
940 |
|
|
941 |
|
warn "## config( '$p' ) = ",dump( $v ),$/ if ($v && $debug); |
942 |
|
warn "config( '$p' ) is empty\n" if (! $v); |
943 |
|
|
944 |
|
return $v; |
945 |
} |
} |
946 |
|
|
947 |
=head2 _x |
=head2 id |
948 |
|
|
949 |
Convert strings from C<conf/normalize/*.xml> encoding into application |
Returns unique id of this record |
|
specific encoding (optinally specified using C<code_page> to C<new> |
|
|
constructor). |
|
950 |
|
|
951 |
my $text = $n->_x('normalize text string'); |
$id = id(); |
952 |
|
|
953 |
This is a stub so that other modules doesn't have to implement it. |
Returns C<42/2> for 2nd occurence of MFN 42. |
954 |
|
|
955 |
=cut |
=cut |
956 |
|
|
957 |
sub _x { |
sub id { |
958 |
my $self = shift; |
my $mfn = $config->{_mfn} || die "no _mfn in config data"; |
959 |
return shift; |
return $mfn . $#{$marc_record} ? $#{$marc_record} + 1 : ''; |
960 |
} |
} |
961 |
|
|
962 |
|
=head2 join_with |
963 |
|
|
964 |
|
Joins walues with some delimiter |
965 |
|
|
966 |
=head1 AUTHOR |
$v = join_with(", ", @v); |
967 |
|
|
968 |
Dobrica Pavlinusic, C<< <dpavlin@rot13.org> >> |
=cut |
969 |
|
|
970 |
|
sub join_with { |
971 |
|
my $d = shift; |
972 |
|
warn "### join_with('$d',",dump(@_),")\n" if ($debug > 2); |
973 |
|
my $v = join($d, grep { defined($_) && $_ ne '' } @_); |
974 |
|
return '' unless defined($v); |
975 |
|
return $v; |
976 |
|
} |
977 |
|
|
978 |
|
=head2 split_rec_on |
979 |
|
|
980 |
=head1 COPYRIGHT & LICENSE |
Split record subfield on some regex and take one of parts out |
981 |
|
|
982 |
Copyright 2005 Dobrica Pavlinusic, All Rights Reserved. |
$a_before_semi_column = |
983 |
|
split_rec_on('200','a', /\s*;\s*/, $part); |
984 |
|
|
985 |
This program is free software; you can redistribute it and/or modify it |
C<$part> is optional number of element. First element is |
986 |
under the same terms as Perl itself. |
B<1>, not 0! |
987 |
|
|
988 |
|
If there is no C<$part> parameter or C<$part> is 0, this function will |
989 |
|
return all values produced by splitting. |
990 |
|
|
991 |
=cut |
=cut |
992 |
|
|
993 |
1; # End of WebPAC::Normalize |
sub split_rec_on { |
994 |
|
die "split_rec_on need (fld,sf,regex[,part]" if ($#_ < 2); |
995 |
|
|
996 |
|
my ($fld, $sf, $regex, $part) = @_; |
997 |
|
warn "### regex ", ref($regex), $regex, $/ if ($debug > 2); |
998 |
|
|
999 |
|
my @r = rec( $fld, $sf ); |
1000 |
|
my $v = shift @r; |
1001 |
|
warn "### first rec($fld,$sf) = ",dump($v),$/ if ($debug > 2); |
1002 |
|
|
1003 |
|
return '' if ( ! defined($v) || $v =~ /^\s*$/); |
1004 |
|
|
1005 |
|
my @s = split( $regex, $v ); |
1006 |
|
warn "## split_rec_on($fld,$sf,$regex,$part) = ",dump(@s),$/ if ($debug > 1); |
1007 |
|
if ($part && $part > 0) { |
1008 |
|
return $s[ $part - 1 ]; |
1009 |
|
} else { |
1010 |
|
return @s; |
1011 |
|
} |
1012 |
|
} |
1013 |
|
|
1014 |
|
# END |
1015 |
|
1; |