1 |
package WebPAC::Normalize; |
package WebPAC::Normalize; |
2 |
|
use Exporter 'import'; |
3 |
|
our @EXPORT = qw/ |
4 |
|
_set_ds _set_lookup |
5 |
|
_set_load_row |
6 |
|
_get_ds _clean_ds |
7 |
|
_debug |
8 |
|
_pack_subfields_hash |
9 |
|
|
10 |
|
to |
11 |
|
search_display search display sorted |
12 |
|
|
13 |
|
rec1 rec2 rec |
14 |
|
frec frec_eq frec_ne |
15 |
|
regex prefix suffix surround |
16 |
|
first lookup join_with |
17 |
|
save_into_lookup |
18 |
|
|
19 |
|
split_rec_on |
20 |
|
|
21 |
|
get set |
22 |
|
count |
23 |
|
|
24 |
|
row |
25 |
|
rec_array |
26 |
|
|
27 |
|
/; |
28 |
|
|
29 |
use warnings; |
use warnings; |
30 |
use strict; |
use strict; |
|
use base 'WebPAC::Common'; |
|
|
use Data::Dumper; |
|
31 |
|
|
32 |
=head1 NAME |
#use base qw/WebPAC::Common/; |
33 |
|
use Data::Dump qw/dump/; |
34 |
|
use Carp qw/confess/; |
35 |
|
|
36 |
|
# debugging warn(s) |
37 |
|
my $debug = 0; |
38 |
|
_debug( $debug ); |
39 |
|
|
40 |
|
# FIXME |
41 |
|
use WebPAC::Normalize::ISBN; |
42 |
|
push @EXPORT, ( 'isbn_10', 'isbn_13' ); |
43 |
|
|
44 |
|
use WebPAC::Normalize::MARC; |
45 |
|
push @EXPORT, ( qw/ |
46 |
|
marc marc_indicators marc_repeatable_subfield |
47 |
|
marc_compose marc_leader marc_fixed |
48 |
|
marc_duplicate marc_remove marc_count |
49 |
|
marc_original_order |
50 |
|
marc_template |
51 |
|
/); |
52 |
|
|
53 |
WebPAC::Normalize - data mungling for normalisation |
use Storable qw/dclone/; |
54 |
|
|
55 |
=head1 VERSION |
=head1 NAME |
56 |
|
|
57 |
Version 0.06 |
WebPAC::Normalize - describe normalisaton rules using sets |
58 |
|
|
59 |
=cut |
=cut |
60 |
|
|
61 |
our $VERSION = '0.06'; |
our $VERSION = '0.36'; |
62 |
|
|
63 |
=head1 SYNOPSIS |
=head1 SYNOPSIS |
64 |
|
|
65 |
This package contains code that mungle data to produce normalized format. |
This module uses C<conf/normalize/*.pl> files to perform normalisation |
66 |
|
from input records using perl functions which are specialized for set |
67 |
|
processing. |
68 |
|
|
69 |
|
Sets are implemented as arrays, and normalisation file is valid perl, which |
70 |
|
means that you check it's validity before running WebPAC using |
71 |
|
C<perl -c normalize.pl>. |
72 |
|
|
73 |
|
Normalisation can generate multiple output normalized data. For now, supported output |
74 |
|
types (on the left side of definition) are: C<search_display>, C<display>, C<search> and |
75 |
|
C<marc>. |
76 |
|
|
77 |
It contains several assumptions: |
=head1 FUNCTIONS |
78 |
|
|
79 |
=over |
Functions which start with C<_> are private and used by WebPAC internally. |
80 |
|
All other functions are available for use within normalisation rules. |
81 |
|
|
82 |
=item * |
=head2 data_structure |
83 |
|
|
84 |
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. |
|
85 |
|
|
86 |
=item * |
my $ds = WebPAC::Normalize::data_structure( |
87 |
|
lookup => $lookup_hash, |
88 |
|
row => $row, |
89 |
|
rules => $normalize_pl_config, |
90 |
|
marc_encoding => 'utf-8', |
91 |
|
config => $config, |
92 |
|
load_row_coderef => sub { |
93 |
|
my ($database,$input,$mfn) = @_; |
94 |
|
$store->load_row( database => $database, input => $input, id => $mfn ); |
95 |
|
}, |
96 |
|
); |
97 |
|
|
98 |
source data records (C<$rec>) have unique identifiers in field C<000> |
Options C<row>, C<rules> and C<log> are mandatory while all |
99 |
|
other are optional. |
100 |
|
|
101 |
=item * |
C<load_row_coderef> is closure only used when executing lookups, so they will |
102 |
|
die if it's not defined. |
103 |
|
|
104 |
optional C<eval{length('v123^a') == 3}> tag at B<beginning of format> will be |
This function will B<die> if normalizastion can't be evaled. |
|
perl code that is evaluated before producing output (value of field will be |
|
|
interpolated before that) |
|
105 |
|
|
106 |
=item * |
Since this function isn't exported you have to call it with |
107 |
|
C<WebPAC::Normalize::data_structure>. |
108 |
|
|
109 |
optional C<filter{filter_name}> at B<begining of format> will apply perl |
=cut |
|
code defined as code ref on format after field substitution to producing |
|
|
output |
|
110 |
|
|
111 |
There is one built-in filter called C<regex> which can be use like this: |
my $load_row_coderef; |
112 |
|
|
113 |
filter{regex(s/foo/bar/)} |
sub data_structure { |
114 |
|
my $arg = {@_}; |
115 |
|
|
116 |
=item * |
die "need row argument" unless ($arg->{row}); |
117 |
|
die "need normalisation argument" unless ($arg->{rules}); |
118 |
|
|
119 |
optional C<lookup{...}> will be then performed. See C<WebPAC::Lookups>. |
_set_lookup( $arg->{lookup} ) if defined($arg->{lookup}); |
120 |
|
_set_ds( $arg->{row} ); |
121 |
|
_set_config( $arg->{config} ) if defined($arg->{config}); |
122 |
|
_clean_ds( %{ $arg } ); |
123 |
|
$load_row_coderef = $arg->{load_row_coderef}; |
124 |
|
|
125 |
=item * |
no strict 'subs'; |
126 |
|
no warnings 'all'; |
127 |
|
eval "$arg->{rules};"; |
128 |
|
die "error evaling $arg->{rules}: $@\n" if ($@); |
129 |
|
|
130 |
at end, optional C<format>s rules are resolved. Format rules are similar to |
return _get_ds(); |
131 |
C<sprintf> and can also contain C<lookup{...}> which is performed after |
} |
|
values are inserted in format. |
|
132 |
|
|
133 |
=back |
=head2 _set_ds |
134 |
|
|
135 |
This also describes order in which transformations are applied (eval, |
Set current record hash |
|
filter, lookup, format) which is important to undestand when deciding how to |
|
|
solve your data mungling and normalisation process. |
|
136 |
|
|
137 |
|
_set_ds( $rec ); |
138 |
|
|
139 |
|
=cut |
140 |
|
|
141 |
|
my $rec; |
142 |
|
|
143 |
=head1 FUNCTIONS |
sub _set_ds { |
144 |
|
$rec = shift or die "no record hash"; |
145 |
|
$WebPAC::Normalize::MARC::rec = $rec; |
146 |
|
} |
147 |
|
|
148 |
=head2 new |
=head2 _get_rec |
149 |
|
|
150 |
Create new normalisation object |
my $rec = _get_rec(); |
151 |
|
|
152 |
my $n = new WebPAC::Normalize::Something( |
=cut |
|
filter => { |
|
|
'filter_name_1' => sub { |
|
|
# filter code |
|
|
return length($_); |
|
|
}, ... |
|
|
}, |
|
|
db => $db_obj, |
|
|
lookup_regex => $lookup->regex, |
|
|
lookup => $lookup_obj, |
|
|
prefix => 'foobar', |
|
|
); |
|
153 |
|
|
154 |
Parametar C<filter> defines user supplied snippets of perl code which can |
sub _get_rec { $rec }; |
|
be use with C<filter{...}> notation. |
|
155 |
|
|
156 |
C<prefix> is used to form filename for database record (to support multiple |
=head2 _set_rec |
|
source files which are joined in one database). |
|
157 |
|
|
158 |
Recommended parametar C<lookup_regex> is used to enable parsing of lookups |
_set_rec( $rec ); |
|
in structures. If you pass this parametar, you must also pass C<lookup> |
|
|
which is C<WebPAC::Lookup> object. |
|
159 |
|
|
160 |
=cut |
=cut |
161 |
|
|
162 |
sub new { |
sub _set_rec { $rec = $_[0] } |
|
my $class = shift; |
|
|
my $self = {@_}; |
|
|
bless($self, $class); |
|
163 |
|
|
164 |
my $r = $self->{'lookup_regex'} ? 1 : 0; |
=head2 _set_config |
|
my $l = $self->{'lookup'} ? 1 : 0; |
|
165 |
|
|
166 |
my $log = $self->_get_logger(); |
Set current config hash |
167 |
|
|
168 |
# those two must be in pair |
_set_config( $config ); |
|
if ( ($r & $l) != ($r || $l) ) { |
|
|
my $log = $self->_get_logger(); |
|
|
$log->logdie("lookup_regex and lookup must be in pair"); |
|
|
} |
|
169 |
|
|
170 |
$log->logdie("lookup must be WebPAC::Lookup object") if ($self->{'lookup'} && ! $self->{'lookup'}->isa('WebPAC::Lookup')); |
Magic keys are: |
171 |
|
|
172 |
$log->warn("no prefix defined. please check that!") unless ($self->{'prefix'}); |
=over 4 |
173 |
|
|
174 |
$log->debug("using lookup regex: ", $self->{lookup_regex}) if ($r && $l); |
=item _ |
175 |
|
|
176 |
if ($self->{filter} && ! $self->{filter}->{regex}) { |
Code of current database |
177 |
$log->debug("adding built-in filter regex"); |
|
178 |
$self->{filter}->{regex} = sub { |
=item _mfn |
179 |
my ($val, $regex) = @_; |
|
180 |
eval "\$val =~ $regex"; |
Current MFN |
181 |
return $val; |
|
182 |
}; |
=back |
183 |
} |
|
184 |
|
=cut |
185 |
|
|
186 |
$self ? return $self : return undef; |
my $config; |
187 |
|
|
188 |
|
sub _set_config { |
189 |
|
$config = shift; |
190 |
} |
} |
191 |
|
|
192 |
|
=head2 _get_ds |
193 |
|
|
194 |
=head2 data_structure |
Return hash formatted as data structure |
195 |
|
|
196 |
Create in-memory data structure which represents normalized layout from |
my $ds = _get_ds(); |
197 |
C<conf/normalize/*.xml>. |
|
198 |
|
=cut |
199 |
|
|
200 |
This structures are used to produce output. |
my $out; |
201 |
|
|
202 |
|
sub _get_ds { |
203 |
|
#warn "## out = ",dump($out); |
204 |
|
return $out; |
205 |
|
} |
206 |
|
|
207 |
my $ds = $webpac->data_structure($rec); |
=head2 _clean_ds |
208 |
|
|
209 |
|
Clean data structure hash for next record |
210 |
|
|
211 |
|
_clean_ds(); |
212 |
|
|
213 |
=cut |
=cut |
214 |
|
|
215 |
sub data_structure { |
sub _clean_ds { |
216 |
my $self = shift; |
my $a = {@_}; |
217 |
|
$out = undef; |
218 |
|
WebPAC::Normalize::MARC::_clean(); |
219 |
|
} |
220 |
|
|
221 |
my $log = $self->_get_logger(); |
=head2 _set_lookup |
222 |
|
|
223 |
my $rec = shift; |
Set current lookup hash |
|
$log->logconfess("need HASH as first argument!") if ($rec !~ /HASH/o); |
|
224 |
|
|
225 |
$log->debug("data_structure rec = ", sub { Dumper($rec) }); |
_set_lookup( $lookup ); |
226 |
|
|
227 |
$log->logdie("need unique ID (mfn) in field 000 of record ", sub { Dumper($rec) } ) unless (defined($rec->{'000'})); |
=cut |
228 |
|
|
229 |
my $id = $rec->{'000'}->[0] || $log->logdie("field 000 isn't array!"); |
my $lookup; |
230 |
|
|
231 |
my $cache_file; |
sub _set_lookup { |
232 |
|
$lookup = shift; |
233 |
|
} |
234 |
|
|
235 |
if ($self->{'db'}) { |
=head2 _get_lookup |
|
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"); |
|
|
} |
|
236 |
|
|
237 |
undef $self->{'currnet_filename'}; |
Get current lookup hash |
|
undef $self->{'headline'}; |
|
238 |
|
|
239 |
my @sorted_tags; |
my $lookup = _get_lookup(); |
|
if ($self->{tags_by_order}) { |
|
|
@sorted_tags = @{$self->{tags_by_order}}; |
|
|
} else { |
|
|
@sorted_tags = sort { $self->_sort_by_order } keys %{$self->{'import_xml'}->{'indexer'}}; |
|
|
$self->{tags_by_order} = \@sorted_tags; |
|
|
} |
|
240 |
|
|
241 |
my $ds; |
=cut |
242 |
|
|
243 |
$log->debug("tags: ",sub { join(", ",@sorted_tags) }); |
sub _get_lookup { |
244 |
|
return $lookup; |
245 |
|
} |
246 |
|
|
247 |
foreach my $field (@sorted_tags) { |
=head2 _set_load_row |
248 |
|
|
249 |
my $row; |
Setup code reference which will return L<data_structure> from |
250 |
|
L<WebPAC::Store> |
251 |
|
|
252 |
#print "field $field [",$self->{'tag'},"] = ",Dumper($self->{'import_xml'}->{'indexer'}->{$field}->{$self->{'tag'}}); |
_set_load_row(sub { |
253 |
|
my ($database,$input,$mfn) = @_; |
254 |
|
$store->load_row( database => $database, input => $input, id => $mfn ); |
255 |
|
}); |
256 |
|
|
257 |
foreach my $tag (@{$self->{'import_xml'}->{'indexer'}->{$field}->{$self->{'tag'}}}) { |
=cut |
|
my $format; |
|
258 |
|
|
259 |
$log->logdie("expected tag HASH and got $tag") unless (ref($tag) eq 'HASH'); |
sub _set_load_row { |
260 |
$format = $tag->{'value'} || $tag->{'content'}; |
my $coderef = shift; |
261 |
|
confess "argument isn't CODE" unless ref($coderef) eq 'CODE'; |
262 |
|
|
263 |
$log->debug("format: $format"); |
$load_row_coderef = $coderef; |
264 |
|
} |
265 |
|
|
266 |
my @v; |
=head2 _debug |
|
if ($self->{'lookup_regex'} && $format =~ $self->{'lookup_regex'}) { |
|
|
@v = $self->fill_in_to_arr($rec,$format); |
|
|
} else { |
|
|
@v = $self->parse_to_arr($rec,$format); |
|
|
} |
|
|
next if (! @v); |
|
267 |
|
|
268 |
if ($tag->{'sort'}) { |
Change level of debug warnings |
|
@v = $self->sort_arr(@v); |
|
|
} |
|
269 |
|
|
270 |
# use format? |
_debug( 2 ); |
|
if ($tag->{'format_name'}) { |
|
|
@v = map { $self->apply_format($tag->{'format_name'},$tag->{'format_delimiter'},$_) } @v; |
|
|
} |
|
271 |
|
|
272 |
# delimiter will join repeatable fields |
=cut |
|
if ($tag->{'delimiter'}) { |
|
|
@v = ( join($tag->{'delimiter'}, @v) ); |
|
|
} |
|
273 |
|
|
274 |
# default types |
sub _debug { |
275 |
my @types = qw(display search); |
my $l = shift; |
276 |
# override by type attribute |
return $debug unless defined($l); |
277 |
@types = ( $tag->{'type'} ) if ($tag->{'type'}); |
warn "debug level $l",$/ if ($l > 0); |
278 |
|
$debug = $l; |
279 |
foreach my $type (@types) { |
$WebPAC::Normalize::MARC::debug = $debug; |
280 |
# 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}}, $last; |
|
281 |
|
|
282 |
} else { |
=head1 Functions to create C<data_structure> |
|
push @{$row->{$type}}, @v; |
|
|
} |
|
|
} |
|
283 |
|
|
284 |
|
Those functions generally have to first in your normalization file. |
285 |
|
|
286 |
} |
=head2 to |
287 |
|
|
288 |
if ($row) { |
Generic way to set values for some name |
|
$row->{'tag'} = $field; |
|
289 |
|
|
290 |
# TODO: name_sigular, name_plural |
to('field-name', 'name-value' => rec('200','a') ); |
|
my $name = $self->{'import_xml'}->{'indexer'}->{$field}->{'name'}; |
|
|
my $row_name = $name ? $self->_x($name) : $field; |
|
|
|
|
|
# post-sort all values in field |
|
|
if ($self->{'import_xml'}->{'indexer'}->{$field}->{'sort'}) { |
|
|
$log->warn("sort at field tag not implemented"); |
|
|
} |
|
291 |
|
|
292 |
$ds->{$row_name} = $row; |
There are many helpers defined below which might be easier to use. |
293 |
|
|
294 |
$log->debug("row $field: ",sub { Dumper($row) }); |
=cut |
|
} |
|
295 |
|
|
296 |
} |
sub to { |
297 |
|
my $type = shift or confess "need type -- BUG?"; |
298 |
|
my $name = shift or confess "needs name as first argument"; |
299 |
|
my @o = grep { defined($_) && $_ ne '' } @_; |
300 |
|
return unless (@o); |
301 |
|
$out->{$name}->{$type} = \@o; |
302 |
|
} |
303 |
|
|
304 |
$self->{'db'}->save_ds( |
=head2 search_display |
|
id => $id, |
|
|
ds => $ds, |
|
|
prefix => $self->{prefix}, |
|
|
) if ($self->{'db'}); |
|
305 |
|
|
306 |
$log->debug("ds: ", sub { Dumper($ds) }); |
Define output for L<search> and L<display> at the same time |
307 |
|
|
308 |
$log->logconfess("data structure returned is not array any more!") if wantarray; |
search_display('Title', rec('200','a') ); |
309 |
|
|
310 |
return $ds; |
=cut |
311 |
|
|
312 |
|
sub search_display { |
313 |
|
my $name = shift or die "search_display needs name as first argument"; |
314 |
|
my @o = grep { defined($_) && $_ ne '' } @_; |
315 |
|
return unless (@o); |
316 |
|
$out->{$name}->{search} = \@o; |
317 |
|
$out->{$name}->{display} = \@o; |
318 |
} |
} |
319 |
|
|
320 |
=head2 parse |
=head2 tag |
321 |
|
|
322 |
Perform smart parsing of string, skipping delimiters for fields which aren't |
Old name for L<search_display>, it will probably be removed at one point. |
|
defined. It can also eval code in format starting with C<eval{...}> and |
|
|
return output or nothing depending on eval code. |
|
323 |
|
|
324 |
my $text = $webpac->parse($rec,'eval{"v901^a" eq "Deskriptor"}descriptor: v250^a', $i); |
=cut |
325 |
|
|
326 |
Filters are implemented here. While simple form of filters looks like this: |
sub tag { |
327 |
|
search_display( @_ ); |
328 |
|
} |
329 |
|
|
330 |
filter{name_of_filter} |
=head2 display |
331 |
|
|
332 |
but, filters can also have variable number of parametars like this: |
Define output just for I<display> |
333 |
|
|
334 |
filter{name_of_filter(param,param,param)} |
@v = display('Title', rec('200','a') ); |
335 |
|
|
336 |
=cut |
=cut |
337 |
|
|
338 |
sub parse { |
sub display { to( 'display', @_ ) } |
|
my $self = shift; |
|
339 |
|
|
340 |
my ($rec, $format_utf8, $i) = @_; |
=head2 search |
341 |
|
|
342 |
return if (! $format_utf8); |
Prepare values just for I<search> |
343 |
|
|
344 |
my $log = $self->_get_logger(); |
@v = search('Title', rec('200','a') ); |
345 |
|
|
346 |
$log->logconfess("need HASH as first argument!") if ($rec !~ /HASH/o); |
=cut |
347 |
|
|
348 |
$i = 0 if (! $i); |
sub search { to( 'search', @_ ) } |
349 |
|
|
350 |
my $format = $self->_x($format_utf8) || $log->logconfess("can't convert '$format_utf8' from UTF-8 to ",$self->{'code_page'}); |
=head2 sorted |
351 |
|
|
352 |
my @out; |
Insert into lists which will be automatically sorted |
353 |
|
|
354 |
$log->debug("format: $format"); |
sorted('Title', rec('200','a') ); |
355 |
|
|
356 |
my $eval_code; |
=cut |
|
# remove eval{...} from beginning |
|
|
$eval_code = $1 if ($format =~ s/^eval{([^}]+)}//s); |
|
357 |
|
|
358 |
my $filter_name; |
sub sorted { to( 'sorted', @_ ) } |
|
# remove filter{...} from beginning |
|
|
$filter_name = $1 if ($format =~ s/^filter{([^}]+)}//s); |
|
359 |
|
|
360 |
my $prefix; |
=head2 row |
|
my $all_found=0; |
|
361 |
|
|
362 |
while ($format =~ s/^(.*?)(v|s)(\d+)(?:\^(\w))?//s) { |
Insert new row of data into output module |
363 |
|
|
364 |
my $del = $1 || ''; |
row( column => 'foo', column2 => 'bar' ); |
|
$prefix ||= $del if ($all_found == 0); |
|
365 |
|
|
366 |
# repeatable index |
=cut |
|
my $r = $i; |
|
|
$r = 0 if (lc("$2") eq 's'); |
|
367 |
|
|
368 |
my $found = 0; |
use Data::Dump qw/dump/; |
|
my $tmp = $self->get_data(\$rec,$3,$4,$r,\$found); |
|
369 |
|
|
370 |
if ($found) { |
sub row { |
371 |
push @out, $del; |
die "array doesn't have odd number of elements but $#_: ",dump( @_ ) if $#_ % 2 == 1; |
372 |
push @out, $tmp; |
my $table = shift @_; |
373 |
$all_found += $found; |
push @{ $out->{'_rows'}->{$table} }, {@_}; |
374 |
} |
} |
|
} |
|
375 |
|
|
|
return if (! $all_found); |
|
376 |
|
|
377 |
my $out = join('',@out); |
=head1 Functions to extract data from input |
378 |
|
|
379 |
if ($out) { |
This function should be used inside functions to create C<data_structure> described |
380 |
# add rest of format (suffix) |
above. |
|
$out .= $format; |
|
381 |
|
|
382 |
# add prefix if not there |
=head2 _pack_subfields_hash |
|
$out = $prefix . $out if ($out !~ m/^\Q$prefix\E/); |
|
383 |
|
|
384 |
$log->debug("result: $out"); |
@subfields = _pack_subfields_hash( $h ); |
385 |
} |
$subfields = _pack_subfields_hash( $h, 1 ); |
386 |
|
|
387 |
if ($eval_code) { |
Return each subfield value in array or pack them all together and return scalar |
388 |
my $eval = $self->fill_in($rec,$eval_code,$i) || return; |
with subfields (denoted by C<^>) and values. |
389 |
$log->debug("about to eval{$eval} format: $out"); |
|
390 |
return if (! $self->_eval($eval)); |
=cut |
391 |
} |
|
392 |
|
sub _pack_subfields_hash { |
393 |
if ($filter_name) { |
|
394 |
my @filter_args; |
warn "## _pack_subfields_hash( ",dump(@_), " )\n" if ($debug > 1); |
395 |
if ($filter_name =~ s/(\w+)\((.*)\)/$1/) { |
|
396 |
@filter_args = split(/,/, $2); |
my ($hash,$include_subfields) = @_; |
397 |
|
|
398 |
|
# sanity and ease of use |
399 |
|
return $hash if (ref($hash) ne 'HASH'); |
400 |
|
|
401 |
|
my $h = dclone( $hash ); |
402 |
|
|
403 |
|
if ( defined($h->{subfields}) ) { |
404 |
|
my $sfs = delete $h->{subfields} || die "no subfields?"; |
405 |
|
my @out; |
406 |
|
while (@$sfs) { |
407 |
|
my $sf = shift @$sfs; |
408 |
|
push @out, '^' . $sf if ($include_subfields); |
409 |
|
my $o = shift @$sfs; |
410 |
|
if ($o == 0 && ref( $h->{$sf} ) ne 'ARRAY' ) { |
411 |
|
# single element subfields are not arrays |
412 |
|
#warn "====> $sf $o / $#$sfs ", dump( $sfs, $h->{$sf} ), "\n"; |
413 |
|
|
414 |
|
push @out, $h->{$sf}; |
415 |
|
} else { |
416 |
|
#warn "====> $sf $o / $#$sfs ", dump( $sfs, $h->{$sf} ), "\n"; |
417 |
|
push @out, $h->{$sf}->[$o]; |
418 |
|
} |
419 |
} |
} |
420 |
if ($self->{'filter'}->{$filter_name}) { |
if ($include_subfields) { |
421 |
$log->debug("about to filter{$filter_name} format: $out with arguments: ", join(",", @filter_args)); |
return join('', @out); |
422 |
unshift @filter_args, $out; |
} else { |
423 |
$out = $self->{'filter'}->{$filter_name}->(@filter_args); |
return @out; |
424 |
return unless(defined($out)); |
} |
425 |
$log->debug("filter result: $out"); |
} else { |
426 |
|
if ($include_subfields) { |
427 |
|
my $out = ''; |
428 |
|
foreach my $sf (sort keys %$h) { |
429 |
|
if (ref($h->{$sf}) eq 'ARRAY') { |
430 |
|
$out .= '^' . $sf . join('^' . $sf, @{ $h->{$sf} }); |
431 |
|
} else { |
432 |
|
$out .= '^' . $sf . $h->{$sf}; |
433 |
|
} |
434 |
|
} |
435 |
|
return $out; |
436 |
} else { |
} else { |
437 |
$log->warn("trying to use undefined filter $filter_name"); |
# FIXME this should probably be in alphabetical order instead of hash order |
438 |
|
values %{$h}; |
439 |
} |
} |
440 |
} |
} |
441 |
|
} |
442 |
|
|
443 |
return $out; |
=head2 rec1 |
444 |
|
|
445 |
|
Return all values in some field |
446 |
|
|
447 |
|
@v = rec1('200') |
448 |
|
|
449 |
|
TODO: order of values is probably same as in source data, need to investigate that |
450 |
|
|
451 |
|
=cut |
452 |
|
|
453 |
|
sub rec1 { |
454 |
|
my $f = shift; |
455 |
|
warn "rec1($f) = ", dump( $rec->{$f} ), $/ if ($debug > 1); |
456 |
|
return unless (defined($rec) && defined($rec->{$f})); |
457 |
|
warn "rec1($f) = ", dump( $rec->{$f} ), $/ if ($debug > 1); |
458 |
|
if (ref($rec->{$f}) eq 'ARRAY') { |
459 |
|
my @out; |
460 |
|
foreach my $h ( @{ $rec->{$f} } ) { |
461 |
|
if (ref($h) eq 'HASH') { |
462 |
|
push @out, ( _pack_subfields_hash( $h ) ); |
463 |
|
} else { |
464 |
|
push @out, $h; |
465 |
|
} |
466 |
|
} |
467 |
|
return @out; |
468 |
|
} elsif( defined($rec->{$f}) ) { |
469 |
|
return $rec->{$f}; |
470 |
|
} |
471 |
} |
} |
472 |
|
|
473 |
=head2 parse_to_arr |
=head2 rec2 |
474 |
|
|
475 |
Similar to C<parse>, but returns array of all repeatable fields |
Return all values in specific field and subfield |
476 |
|
|
477 |
my @arr = $webpac->parse_to_arr($rec,'v250^a'); |
@v = rec2('200','a') |
478 |
|
|
479 |
=cut |
=cut |
480 |
|
|
481 |
sub parse_to_arr { |
sub rec2 { |
482 |
my $self = shift; |
my $f = shift; |
483 |
|
return unless (defined($rec && $rec->{$f})); |
484 |
|
my $sf = shift; |
485 |
|
warn "rec2($f,$sf) = ", dump( $rec->{$f} ), $/ if ($debug > 1); |
486 |
|
return map { |
487 |
|
if (ref($_->{$sf}) eq 'ARRAY') { |
488 |
|
@{ $_->{$sf} }; |
489 |
|
} else { |
490 |
|
$_->{$sf}; |
491 |
|
} |
492 |
|
} grep { ref($_) eq 'HASH' && defined $_->{$sf} } @{ $rec->{$f} }; |
493 |
|
} |
494 |
|
|
495 |
my ($rec, $format_utf8) = @_; |
=head2 rec |
496 |
|
|
497 |
my $log = $self->_get_logger(); |
syntaxtic sugar for |
498 |
|
|
499 |
$log->logconfess("need HASH as first argument!") if ($rec !~ /HASH/o); |
@v = rec('200') |
500 |
return if (! $format_utf8); |
@v = rec('200','a') |
501 |
|
|
502 |
my $i = 0; |
If rec() returns just single value, it will |
503 |
my @arr; |
return scalar, not array. |
504 |
|
|
505 |
|
=cut |
506 |
|
|
507 |
while (my $v = $self->parse($rec,$format_utf8,$i++)) { |
sub rec { |
508 |
push @arr, $v; |
my @out; |
509 |
|
if ($#_ == 0) { |
510 |
|
@out = rec1(@_); |
511 |
|
} elsif ($#_ == 1) { |
512 |
|
@out = rec2(@_); |
513 |
|
} |
514 |
|
if ($#out == 0 && ! wantarray) { |
515 |
|
return $out[0]; |
516 |
|
} elsif (@out) { |
517 |
|
return @out; |
518 |
|
} else { |
519 |
|
return ''; |
520 |
} |
} |
521 |
|
} |
522 |
|
|
523 |
$log->debug("format '$format_utf8' returned ",--$i," elements: ", sub { join(" | ",@arr) }) if (@arr); |
=head2 frec |
524 |
|
|
525 |
return @arr; |
Returns first value from field |
526 |
} |
|
527 |
|
$v = frec('200'); |
528 |
|
$v = frec('200','a'); |
529 |
|
|
530 |
|
=cut |
531 |
|
|
532 |
|
sub frec { |
533 |
|
my @out = rec(@_); |
534 |
|
warn "rec(",dump(@_),") has more than one return value, ignoring\n" if $#out > 0; |
535 |
|
return shift @out; |
536 |
|
} |
537 |
|
|
538 |
=head2 fill_in |
=head2 frec_eq |
539 |
|
|
540 |
Workhourse of all: takes record from in-memory structure of database and |
=head2 frec_ne |
|
strings with placeholders and returns string or array of with substituted |
|
|
values from record. |
|
541 |
|
|
542 |
my $text = $webpac->fill_in($rec,'v250^a'); |
Check if first values from two fields are same or different |
543 |
|
|
544 |
Optional argument is ordinal number for repeatable fields. By default, |
if ( frec_eq( 900 => 'a', 910 => 'c' ) ) { |
545 |
it's assume to be first repeatable field (fields are perl array, so first |
# values are same |
546 |
element is 0). |
} else { |
547 |
Following example will read second value from repeatable field. |
# values are different |
548 |
|
} |
549 |
|
|
550 |
my $text = $webpac->fill_in($rec,'Title: v250^a',1); |
Strictly speaking C<frec_eq> and C<frec_ne> wouldn't be needed if you |
551 |
|
could write something like: |
552 |
|
|
553 |
This function B<does not> perform parsing of format to inteligenty skip |
if ( frec( '900','a' ) eq frec( '910','c' ) ) { |
554 |
delimiters before fields which aren't used. |
# yada tada |
555 |
|
} |
556 |
|
|
557 |
This method will automatically decode UTF-8 string to local code page |
but you can't since our parser L<WebPAC::Parser> will remove all whitespaces |
558 |
if needed. |
in order to parse text and create invalid function C<eqfrec>. |
559 |
|
|
560 |
=cut |
=cut |
561 |
|
|
562 |
sub fill_in { |
sub frec_eq { |
563 |
my $self = shift; |
my ( $f1,$sf1, $f2, $sf2 ) = @_; |
564 |
|
return (rec( $f1, $sf1 ))[0] eq (rec( $f2, $sf2 ))[0]; |
565 |
|
} |
566 |
|
|
567 |
|
sub frec_ne { |
568 |
|
return ! frec_eq( @_ ); |
569 |
|
} |
570 |
|
|
571 |
my $log = $self->_get_logger(); |
=head2 regex |
572 |
|
|
573 |
my $rec = shift || $log->logconfess("need data record"); |
Apply regex to some or all values |
|
my $format = shift || $log->logconfess("need format to parse"); |
|
|
# iteration (for repeatable fields) |
|
|
my $i = shift || 0; |
|
574 |
|
|
575 |
$log->logdie("infitite loop in format $format") if ($i > ($self->{'max_mfn'} || 9999)); |
@v = regex( 's/foo/bar/g', @v ); |
576 |
|
|
577 |
# FIXME remove for speedup? |
=cut |
|
$log->logconfess("need HASH as first argument!") if ($rec !~ /HASH/o); |
|
578 |
|
|
579 |
if (utf8::is_utf8($format)) { |
sub regex { |
580 |
$format = $self->_x($format); |
my $r = shift; |
581 |
|
my @out; |
582 |
|
#warn "r: $r\n", dump(\@_); |
583 |
|
foreach my $t (@_) { |
584 |
|
next unless ($t); |
585 |
|
eval "\$t =~ $r"; |
586 |
|
push @out, $t if ($t && $t ne ''); |
587 |
} |
} |
588 |
|
return @out; |
589 |
|
} |
590 |
|
|
591 |
my $found = 0; |
=head2 prefix |
592 |
|
|
593 |
my $eval_code; |
Prefix all values with a string |
|
# remove eval{...} from beginning |
|
|
$eval_code = $1 if ($format =~ s/^eval{([^}]+)}//s); |
|
594 |
|
|
595 |
my $filter_name; |
@v = prefix( 'my_', @v ); |
|
# remove filter{...} from beginning |
|
|
$filter_name = $1 if ($format =~ s/^filter{([^}]+)}//s); |
|
596 |
|
|
597 |
# do actual replacement of placeholders |
=cut |
|
# 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,$2,0,\$found)/ges; |
|
598 |
|
|
599 |
if ($found) { |
sub prefix { |
600 |
$log->debug("format: $format"); |
my $p = shift; |
601 |
if ($eval_code) { |
return @_ unless defined( $p ); |
602 |
my $eval = $self->fill_in($rec,$eval_code,$i); |
return map { $p . $_ } grep { defined($_) } @_; |
|
return if (! $self->_eval($eval)); |
|
|
} |
|
|
if ($filter_name && $self->{'filter'}->{$filter_name}) { |
|
|
$log->debug("filter '$filter_name' for $format"); |
|
|
$format = $self->{'filter'}->{$filter_name}->($format); |
|
|
return unless(defined($format)); |
|
|
$log->debug("filter result: $format"); |
|
|
} |
|
|
# do we have lookups? |
|
|
if ($self->{'lookup'}) { |
|
|
if ($self->{'lookup'}->can('lookup')) { |
|
|
my @lookup = $self->{lookup}->lookup($format); |
|
|
$log->debug("lookup $format", join(", ", @lookup)); |
|
|
return @lookup; |
|
|
} else { |
|
|
$log->warn("Have lookup object but can't invoke lookup method"); |
|
|
} |
|
|
} else { |
|
|
return $format; |
|
|
} |
|
|
} else { |
|
|
return; |
|
|
} |
|
603 |
} |
} |
604 |
|
|
605 |
|
=head2 suffix |
606 |
|
|
607 |
=head2 fill_in_to_arr |
suffix all values with a string |
608 |
|
|
609 |
Similar to C<fill_in>, but returns array of all repeatable fields. Usable |
@v = suffix( '_my', @v ); |
|
for fields which have lookups, so they shouldn't be parsed but rather |
|
|
C<fill_id>ed. |
|
|
|
|
|
my @arr = $webpac->fill_in_to_arr($rec,'[v900];;[v250^a]'); |
|
610 |
|
|
611 |
=cut |
=cut |
612 |
|
|
613 |
sub fill_in_to_arr { |
sub suffix { |
614 |
my $self = shift; |
my $s = shift; |
615 |
|
return @_ unless defined( $s ); |
616 |
|
return map { $_ . $s } grep { defined($_) } @_; |
617 |
|
} |
618 |
|
|
619 |
my ($rec, $format_utf8) = @_; |
=head2 surround |
620 |
|
|
621 |
my $log = $self->_get_logger(); |
surround all values with a two strings |
622 |
|
|
623 |
$log->logconfess("need HASH as first argument!") if ($rec !~ /HASH/o); |
@v = surround( 'prefix_', '_suffix', @v ); |
|
return if (! $format_utf8); |
|
624 |
|
|
625 |
my $i = 0; |
=cut |
|
my @arr; |
|
626 |
|
|
627 |
while (my @v = $self->fill_in($rec,$format_utf8,$i++)) { |
sub surround { |
628 |
push @arr, @v; |
my $p = shift; |
629 |
} |
my $s = shift; |
630 |
|
$p = '' unless defined( $p ); |
631 |
|
$s = '' unless defined( $s ); |
632 |
|
return map { $p . $_ . $s } grep { defined($_) } @_; |
633 |
|
} |
634 |
|
|
635 |
$log->debug("format '$format_utf8' returned ",--$i," elements: ", sub { join(" | ",@arr) }) if (@arr); |
=head2 first |
636 |
|
|
637 |
return @arr; |
Return first element |
638 |
|
|
639 |
|
$v = first( @v ); |
640 |
|
|
641 |
|
=cut |
642 |
|
|
643 |
|
sub first { |
644 |
|
my $r = shift; |
645 |
|
return $r; |
646 |
} |
} |
647 |
|
|
648 |
|
=head2 lookup |
649 |
|
|
650 |
|
Consult lookup hashes for some value |
651 |
|
|
652 |
=head2 get_data |
@v = lookup( |
653 |
|
sub { |
654 |
|
'ffkk/peri/mfn'.rec('000') |
655 |
|
}, |
656 |
|
'ffkk','peri','200-a-200-e', |
657 |
|
sub { |
658 |
|
first(rec(200,'a')).' '.first(rec('200','e')) |
659 |
|
} |
660 |
|
); |
661 |
|
|
662 |
Returns value from record. |
Code like above will be B<automatically generated> using L<WebPAC::Parse> from |
663 |
|
normal lookup definition in C<conf/lookup/something.pl> which looks like: |
664 |
|
|
665 |
my $text = $self->get_data(\$rec,$f,$sf,$i,\$found); |
lookup( |
666 |
|
# which results to return from record recorded in lookup |
667 |
|
sub { 'ffkk/peri/mfn' . rec('000') }, |
668 |
|
# from which database and input |
669 |
|
'ffkk','peri', |
670 |
|
# such that following values match |
671 |
|
sub { first(rec(200,'a')) . ' ' . first(rec('200','e')) }, |
672 |
|
# if this part is missing, we will try to match same fields |
673 |
|
# from lookup record and current one, or you can override |
674 |
|
# which records to use from current record using |
675 |
|
sub { rec('900','x') . ' ' . rec('900','y') }, |
676 |
|
) |
677 |
|
|
678 |
Arguments are: |
You can think about this lookup as SQL (if that helps): |
|
record reference C<$rec>, |
|
|
field C<$f>, |
|
|
optional subfiled C<$sf>, |
|
|
index for repeatable values C<$i>. |
|
679 |
|
|
680 |
Optinal variable C<$found> will be incremeted if there |
select |
681 |
is field. |
sub { what } |
682 |
|
from |
683 |
|
database, input |
684 |
|
where |
685 |
|
sub { filter from lookuped record } |
686 |
|
having |
687 |
|
sub { optional filter on current record } |
688 |
|
|
689 |
Returns value or empty string. |
Easy as pie, right? |
690 |
|
|
691 |
=cut |
=cut |
692 |
|
|
693 |
sub get_data { |
sub lookup { |
694 |
my $self = shift; |
my ($what, $database, $input, $key, $having) = @_; |
695 |
|
|
696 |
my ($rec,$f,$sf,$i,$found) = @_; |
confess "lookup needs 5 arguments: what, database, input, key, having\n" unless ($#_ == 4); |
697 |
|
|
698 |
if ($$rec->{$f}) { |
warn "## lookup ($database, $input, $key)", $/ if ($debug > 1); |
699 |
return '' if (! $$rec->{$f}->[$i]); |
return unless (defined($lookup->{$database}->{$input}->{$key})); |
700 |
no strict 'refs'; |
|
701 |
if ($sf && $$rec->{$f}->[$i]->{$sf}) { |
confess "lookup really need load_row_coderef added to data_structure\n" unless ($load_row_coderef); |
702 |
$$found++ if (defined($$found)); |
|
703 |
return $$rec->{$f}->[$i]->{$sf}; |
my $mfns; |
704 |
} elsif (! $sf && $$rec->{$f}->[$i]) { |
my @having = $having->(); |
705 |
$$found++ if (defined($$found)); |
|
706 |
# it still might have subfield, just |
warn "## having = ", dump( @having ) if ($debug > 2); |
707 |
# not specified, so we'll dump all |
|
708 |
if ($$rec->{$f}->[$i] =~ /HASH/o) { |
foreach my $h ( @having ) { |
709 |
my $out; |
if (defined($lookup->{$database}->{$input}->{$key}->{$h})) { |
710 |
foreach my $k (keys %{$$rec->{$f}->[$i]}) { |
warn "lookup for $database/$input/$key/$h return ",dump($lookup->{$database}->{$input}->{$key}->{$h}),"\n" if ($debug); |
711 |
$out .= $$rec->{$f}->[$i]->{$k}." "; |
$mfns->{$_}++ foreach keys %{ $lookup->{$database}->{$input}->{$key}->{$h} }; |
|
} |
|
|
return $out; |
|
|
} else { |
|
|
return $$rec->{$f}->[$i]; |
|
|
} |
|
|
} else { |
|
|
return ''; |
|
712 |
} |
} |
713 |
|
} |
714 |
|
|
715 |
|
return unless ($mfns); |
716 |
|
|
717 |
|
my @mfns = sort keys %$mfns; |
718 |
|
|
719 |
|
warn "# lookup loading $database/$input/$key mfn ", join(",",@mfns)," having ",dump(@having),"\n" if ($debug); |
720 |
|
|
721 |
|
my $old_rec = $rec; |
722 |
|
my @out; |
723 |
|
|
724 |
|
foreach my $mfn (@mfns) { |
725 |
|
$rec = $load_row_coderef->( $database, $input, $mfn ); |
726 |
|
|
727 |
|
warn "got $database/$input/$mfn = ", dump($rec), $/ if ($debug); |
728 |
|
|
729 |
|
my @vals = $what->(); |
730 |
|
|
731 |
|
push @out, ( @vals ); |
732 |
|
|
733 |
|
warn "lookup for mfn $mfn returned ", dump(@vals), $/ if ($debug); |
734 |
|
} |
735 |
|
|
736 |
|
# if (ref($lookup->{$k}) eq 'ARRAY') { |
737 |
|
# return @{ $lookup->{$k} }; |
738 |
|
# } else { |
739 |
|
# return $lookup->{$k}; |
740 |
|
# } |
741 |
|
|
742 |
|
$rec = $old_rec; |
743 |
|
|
744 |
|
warn "## lookup returns = ", dump(@out), $/ if ($debug); |
745 |
|
|
746 |
|
if ($#out == 0) { |
747 |
|
return $out[0]; |
748 |
} else { |
} else { |
749 |
return ''; |
return @out; |
750 |
} |
} |
751 |
} |
} |
752 |
|
|
753 |
|
=head2 save_into_lookup |
754 |
|
|
755 |
|
Save value into lookup. It associates current database, input |
756 |
|
and specific keys with one or more values which will be |
757 |
|
associated over MFN. |
758 |
|
|
759 |
=head2 apply_format |
MFN will be extracted from first occurence current of field 000 |
760 |
|
in current record, or if it doesn't exist from L<_set_config> C<_mfn>. |
761 |
|
|
762 |
Apply format specified in tag with C<format_name="name"> and |
my $nr = save_into_lookup($database,$input,$key,sub { |
763 |
C<format_delimiter=";;">. |
# code which produce one or more values |
764 |
|
}); |
765 |
|
|
766 |
my $text = $webpac->apply_format($format_name,$format_delimiter,$data); |
It returns number of items saved. |
767 |
|
|
768 |
Formats can contain C<lookup{...}> if you need them. |
This function shouldn't be called directly, it's called from code created by |
769 |
|
L<WebPAC::Parser>. |
770 |
|
|
771 |
=cut |
=cut |
772 |
|
|
773 |
sub apply_format { |
sub save_into_lookup { |
774 |
my $self = shift; |
my ($database,$input,$key,$coderef) = @_; |
775 |
|
die "save_into_lookup needs database" unless defined($database); |
776 |
|
die "save_into_lookup needs input" unless defined($input); |
777 |
|
die "save_into_lookup needs key" unless defined($key); |
778 |
|
die "save_into_lookup needs CODE" unless ( defined($coderef) && ref($coderef) eq 'CODE' ); |
779 |
|
|
780 |
my ($name,$delimiter,$data) = @_; |
warn "## save_into_lookup rec = ", dump($rec), " config = ", dump($config), $/ if ($debug > 2); |
781 |
|
|
782 |
my $log = $self->_get_logger(); |
my $mfn = |
783 |
|
defined($rec->{'000'}->[0]) ? $rec->{'000'}->[0] : |
784 |
|
defined($config->{_mfn}) ? $config->{_mfn} : |
785 |
|
die "mfn not defined or zero"; |
786 |
|
|
787 |
if (! $self->{'import_xml'}->{'format'}->{$name}) { |
my $nr = 0; |
788 |
$log->warn("<format name=\"$name\"> is not defined in ",$self->{'import_xml_file'}); |
|
789 |
return $data; |
foreach my $v ( $coderef->() ) { |
790 |
|
$lookup->{$database}->{$input}->{$key}->{$v}->{$mfn}++; |
791 |
|
warn "# saved lookup $database/$input/$key [$v] $mfn\n" if ($debug > 1); |
792 |
|
$nr++; |
793 |
} |
} |
794 |
|
|
795 |
$log->warn("no delimiter for format $name") if (! $delimiter); |
return $nr; |
796 |
|
} |
797 |
|
|
798 |
my $format = $self->_x($self->{'import_xml'}->{'format'}->{$name}->{'content'}) || $log->logdie("can't find format '$name'"); |
=head2 config |
799 |
|
|
800 |
my @data = split(/\Q$delimiter\E/, $data); |
Consult config values stored in C<config.yml> |
801 |
|
|
802 |
my $out = sprintf($format, @data); |
# return database code (key under databases in yaml) |
803 |
$log->debug("using format $name [$format] on $data to produce: $out"); |
$database_code = config(); # use _ from hash |
804 |
|
$database_name = config('name'); |
805 |
|
$database_input_name = config('input name'); |
806 |
|
|
807 |
|
Up to three levels are supported. |
808 |
|
|
809 |
|
=cut |
810 |
|
|
811 |
if ($self->{'lookup_regex'} && $out =~ $self->{'lookup_regex'}) { |
sub config { |
812 |
return $self->{'lookup'}->lookup($out); |
return unless ($config); |
813 |
|
|
814 |
|
my $p = shift; |
815 |
|
|
816 |
|
$p ||= ''; |
817 |
|
|
818 |
|
my $v; |
819 |
|
|
820 |
|
warn "### getting config($p)\n" if ($debug > 1); |
821 |
|
|
822 |
|
my @p = split(/\s+/,$p); |
823 |
|
if ($#p < 0) { |
824 |
|
$v = $config->{ '_' }; # special, database code |
825 |
} else { |
} else { |
826 |
return $out; |
|
827 |
|
my $c = dclone( $config ); |
828 |
|
|
829 |
|
foreach my $k (@p) { |
830 |
|
warn "### k: $k c = ",dump($c),$/ if ($debug > 1); |
831 |
|
if (ref($c) eq 'ARRAY') { |
832 |
|
$c = shift @$c; |
833 |
|
warn "config($p) taking first occurence of '$k', probably not what you wanted!\n"; |
834 |
|
last; |
835 |
|
} |
836 |
|
|
837 |
|
if (! defined($c->{$k}) ) { |
838 |
|
$c = undef; |
839 |
|
last; |
840 |
|
} else { |
841 |
|
$c = $c->{$k}; |
842 |
|
} |
843 |
|
} |
844 |
|
$v = $c if ($c); |
845 |
|
|
846 |
} |
} |
847 |
|
|
848 |
|
warn "## config( '$p' ) = ",dump( $v ),$/ if ($v && $debug); |
849 |
|
warn "config( '$p' ) is empty\n" if (! $v); |
850 |
|
|
851 |
|
return $v; |
852 |
} |
} |
853 |
|
|
854 |
=head2 sort_arr |
=head2 id |
855 |
|
|
856 |
|
Returns unique id of this record |
857 |
|
|
858 |
Sort array ignoring case and html in data |
$id = id(); |
859 |
|
|
860 |
my @sorted = $webpac->sort_arr(@unsorted); |
Returns C<42/2> for 2nd occurence of MFN 42. |
861 |
|
|
862 |
=cut |
=cut |
863 |
|
|
864 |
sub sort_arr { |
sub id { |
865 |
my $self = shift; |
my $mfn = $config->{_mfn} || die "no _mfn in config data"; |
866 |
|
return $mfn . ( WebPAC::Normalize::MARC::_created_marc_records() || '' ); |
867 |
|
} |
868 |
|
|
869 |
|
=head2 join_with |
870 |
|
|
871 |
my $log = $self->_get_logger(); |
Joins walues with some delimiter |
872 |
|
|
873 |
# FIXME add Schwartzian Transformation? |
$v = join_with(", ", @v); |
874 |
|
|
875 |
my @sorted = sort { |
=cut |
|
$a =~ s#<[^>]+/*>##; |
|
|
$b =~ s#<[^>]+/*>##; |
|
|
lc($b) cmp lc($a) |
|
|
} @_; |
|
|
$log->debug("sorted values: ",sub { join(", ",@sorted) }); |
|
876 |
|
|
877 |
return @sorted; |
sub join_with { |
878 |
|
my $d = shift; |
879 |
|
warn "### join_with('$d',",dump(@_),")\n" if ($debug > 2); |
880 |
|
my $v = join($d, grep { defined($_) && $_ ne '' } @_); |
881 |
|
return '' unless defined($v); |
882 |
|
return $v; |
883 |
} |
} |
884 |
|
|
885 |
|
=head2 split_rec_on |
886 |
|
|
887 |
|
Split record subfield on some regex and take one of parts out |
888 |
|
|
889 |
=head1 INTERNAL METHODS |
$a_before_semi_column = |
890 |
|
split_rec_on('200','a', /\s*;\s*/, $part); |
891 |
|
|
892 |
=head2 _sort_by_order |
C<$part> is optional number of element. First element is |
893 |
|
B<1>, not 0! |
894 |
|
|
895 |
Sort xml tags data structure accoding to C<order=""> attribute. |
If there is no C<$part> parameter or C<$part> is 0, this function will |
896 |
|
return all values produced by splitting. |
897 |
|
|
898 |
=cut |
=cut |
899 |
|
|
900 |
sub _sort_by_order { |
sub split_rec_on { |
901 |
my $self = shift; |
die "split_rec_on need (fld,sf,regex[,part]" if ($#_ < 2); |
902 |
|
|
903 |
my $va = $self->{'import_xml'}->{'indexer'}->{$a}->{'order'} || |
my ($fld, $sf, $regex, $part) = @_; |
904 |
$self->{'import_xml'}->{'indexer'}->{$a}; |
warn "### regex ", ref($regex), $regex, $/ if ($debug > 2); |
|
my $vb = $self->{'import_xml'}->{'indexer'}->{$b}->{'order'} || |
|
|
$self->{'import_xml'}->{'indexer'}->{$b}; |
|
905 |
|
|
906 |
return $va <=> $vb; |
my @r = rec( $fld, $sf ); |
907 |
|
my $v = shift @r; |
908 |
|
warn "### first rec($fld,$sf) = ",dump($v),$/ if ($debug > 2); |
909 |
|
|
910 |
|
return '' if ( ! defined($v) || $v =~ /^\s*$/); |
911 |
|
|
912 |
|
my @s = split( $regex, $v ); |
913 |
|
warn "## split_rec_on($fld,$sf,$regex,$part) = ",dump(@s),$/ if ($debug > 1); |
914 |
|
if ($part && $part > 0) { |
915 |
|
return $s[ $part - 1 ]; |
916 |
|
} else { |
917 |
|
return @s; |
918 |
|
} |
919 |
} |
} |
920 |
|
|
921 |
=head2 _x |
my $hash; |
922 |
|
|
923 |
Convert strings from C<conf/normalize/*.xml> encoding into application |
=head2 set |
924 |
specific encoding (optinally specified using C<code_page> to C<new> |
|
925 |
constructor). |
set( key => 'value' ); |
926 |
|
|
927 |
|
=cut |
928 |
|
|
929 |
my $text = $n->_x('normalize text string'); |
sub set { |
930 |
|
my ($k,$v) = @_; |
931 |
|
warn "## set ( $k => ", dump($v), " )", $/ if ( $debug ); |
932 |
|
$hash->{$k} = $v; |
933 |
|
}; |
934 |
|
|
935 |
This is a stub so that other modules doesn't have to implement it. |
=head2 get |
936 |
|
|
937 |
|
get( 'key' ); |
938 |
|
|
939 |
=cut |
=cut |
940 |
|
|
941 |
sub _x { |
sub get { |
942 |
my $self = shift; |
my $k = shift || return; |
943 |
return shift; |
my $v = $hash->{$k}; |
944 |
|
warn "## get $k = ", dump( $v ), $/ if ( $debug ); |
945 |
|
return $v; |
946 |
} |
} |
947 |
|
|
948 |
|
=head2 count |
949 |
|
|
950 |
|
if ( count( @result ) == 1 ) { |
951 |
|
# do something if only 1 result is there |
952 |
|
} |
953 |
|
|
954 |
=head1 AUTHOR |
=cut |
955 |
|
|
956 |
Dobrica Pavlinusic, C<< <dpavlin@rot13.org> >> |
sub count { |
957 |
|
warn "## count ",dump(@_),$/ if ( $debug ); |
958 |
|
return @_ . ''; |
959 |
|
} |
960 |
|
|
961 |
=head1 COPYRIGHT & LICENSE |
=head2 rec_array |
962 |
|
|
963 |
Copyright 2005 Dobrica Pavlinusic, All Rights Reserved. |
Always return field as array |
964 |
|
|
965 |
This program is free software; you can redistribute it and/or modify it |
foreach my $d ( rec_array('field') ) { |
966 |
under the same terms as Perl itself. |
warn $d; |
967 |
|
} |
968 |
|
|
969 |
=cut |
=cut |
970 |
|
|
971 |
1; # End of WebPAC::Normalize |
sub rec_array { |
972 |
|
my $d = $rec->{ $_[0] }; |
973 |
|
return @$d if ref($d) eq 'ARRAY'; |
974 |
|
return ($d); |
975 |
|
} |
976 |
|
|
977 |
|
# END |
978 |
|
1; |