2 |
|
|
3 |
use warnings; |
use warnings; |
4 |
use strict; |
use strict; |
5 |
|
use base 'WebPAC::Common'; |
6 |
use Data::Dumper; |
use Data::Dumper; |
7 |
|
|
8 |
=head1 NAME |
=head1 NAME |
9 |
|
|
10 |
WebPAC::Normalize - normalisation of source file |
WebPAC::Normalize - data mungling for normalisation |
11 |
|
|
12 |
=head1 VERSION |
=head1 VERSION |
13 |
|
|
14 |
Version 0.01 |
Version 0.02 |
15 |
|
|
16 |
=cut |
=cut |
17 |
|
|
18 |
our $VERSION = '0.01'; |
our $VERSION = '0.02'; |
19 |
|
|
20 |
=head1 SYNOPSIS |
=head1 SYNOPSIS |
21 |
|
|
22 |
This package contains code that could be helpful in implementing different |
This package contains code that mungle data to produce normalized format. |
23 |
normalisation front-ends. |
|
24 |
|
It contains several assumptions: |
25 |
|
|
26 |
|
=over |
27 |
|
|
28 |
|
=item * |
29 |
|
|
30 |
|
format of fields is defined using C<v123^a> notation for repeatable fields |
31 |
|
or C<s123^a> for single (or first) value, where C<123> is field number and |
32 |
|
C<a> is subfield. |
33 |
|
|
34 |
|
=item * |
35 |
|
|
36 |
|
source data records (C<$rec>) have unique identifiers in field C<000> |
37 |
|
|
38 |
|
=item * |
39 |
|
|
40 |
|
optional C<eval{length('v123^a') == 3}> tag at B<beginning of format> will be |
41 |
|
perl code that is evaluated before producing output (value of field will be |
42 |
|
interpolated before that) |
43 |
|
|
44 |
|
=item * |
45 |
|
|
46 |
|
optional C<filter{filter_name}> at B<begining of format> will apply perl |
47 |
|
code defined as code ref on format after field substitution to producing |
48 |
|
output |
49 |
|
|
50 |
|
=item * |
51 |
|
|
52 |
|
optional C<lookup{...}> will be then performed. See C<WebPAC::Lookups>. |
53 |
|
|
54 |
|
=item * |
55 |
|
|
56 |
|
at end, optional C<format>s rules are resolved. Format rules are similar to |
57 |
|
C<sprintf> and can also contain C<lookup{...}> which is performed after |
58 |
|
values are inserted in format. |
59 |
|
|
60 |
|
=back |
61 |
|
|
62 |
|
This also describes order in which transformations are applied (eval, |
63 |
|
filter, lookup, format) which is important to undestand when deciding how to |
64 |
|
solve your data mungling and normalisation process. |
65 |
|
|
66 |
|
|
67 |
|
|
68 |
|
|
69 |
=head1 FUNCTIONS |
=head1 FUNCTIONS |
70 |
|
|
73 |
Create new normalisation object |
Create new normalisation object |
74 |
|
|
75 |
my $n = new WebPAC::Normalize::Something( |
my $n = new WebPAC::Normalize::Something( |
76 |
cache_data_structure => './cache/ds/', |
filter => { |
77 |
|
'filter_name_1' => sub { |
78 |
|
# filter code |
79 |
|
return length($_); |
80 |
|
}, ... |
81 |
|
}, |
82 |
|
db => $db_obj, |
83 |
lookup_regex => $lookup->regex, |
lookup_regex => $lookup->regex, |
84 |
|
lookup => $lookup_obj, |
85 |
); |
); |
86 |
|
|
87 |
Optional parameter C<cache_data_structure> defines path to directory |
Parametar C<filter> defines user supplied snippets of perl code which can |
88 |
in which cache file for C<data_structure> call will be created. |
be use with C<filter{...}> notation. |
89 |
|
|
90 |
Recommended parametar C<lookup_regex> is used to enable parsing of lookups |
Recommended parametar C<lookup_regex> is used to enable parsing of lookups |
91 |
in structures. |
in structures. If you pass this parametar, you must also pass C<lookup> |
92 |
|
which is C<WebPAC::Lookup> object. |
93 |
|
|
94 |
=cut |
=cut |
95 |
|
|
98 |
my $self = {@_}; |
my $self = {@_}; |
99 |
bless($self, $class); |
bless($self, $class); |
100 |
|
|
101 |
$self->setup_cache_dir( $self->{'cache_data_structure'} ); |
my $r = $self->{'lookup_regex'} ? 1 : 0; |
102 |
|
my $l = $self->{'lookup'} ? 1 : 0; |
|
$self ? return $self : return undef; |
|
|
} |
|
|
|
|
|
=head2 setup_cache_dir |
|
|
|
|
|
Check if specified cache directory exist, and if not, disable caching. |
|
|
|
|
|
$setup_cache_dir('./cache/ds/'); |
|
|
|
|
|
If you pass false or zero value to this function, it will disable |
|
|
cacheing. |
|
|
|
|
|
=cut |
|
|
|
|
|
sub setup_cache_dir { |
|
|
my $self = shift; |
|
|
|
|
|
my $dir = shift; |
|
103 |
|
|
104 |
my $log = $self->_get_logger(); |
my $log = $self->_get_logger(); |
105 |
|
|
106 |
if ($dir) { |
# those two must be in pair |
107 |
my $msg; |
if ( ($r & $l) != ($r || $l) ) { |
108 |
if (! -e $dir) { |
my $log = $self->_get_logger(); |
109 |
$msg = "doesn't exist"; |
$log->logdie("lookup_regex and lookup must be in pair"); |
|
} elsif (! -d $dir) { |
|
|
$msg = "is not directory"; |
|
|
} elsif (! -w $dir) { |
|
|
$msg = "not writable"; |
|
|
} |
|
|
|
|
|
if ($msg) { |
|
|
undef $self->{'cache_data_structure'}; |
|
|
$log->warn("cache_data_structure $dir $msg, disabling..."); |
|
|
} else { |
|
|
$log->debug("using cache dir $dir"); |
|
|
} |
|
|
} else { |
|
|
$log->debug("disabling cache"); |
|
|
undef $self->{'cache_data_structure'}; |
|
110 |
} |
} |
111 |
|
|
112 |
|
$log->logdie("lookup must be WebPAC::Lookup object") if ($self->{'lookup'} && ! $self->{'lookup'}->isa('WebPAC::Lookup')); |
113 |
|
|
114 |
|
$self ? return $self : return undef; |
115 |
} |
} |
116 |
|
|
117 |
|
|
122 |
|
|
123 |
This structures are used to produce output. |
This structures are used to produce output. |
124 |
|
|
125 |
my @ds = $webpac->data_structure($rec); |
my $ds = $webpac->data_structure($rec); |
126 |
|
|
127 |
B<Note: historical oddity follows> |
B<Note: historical oddity follows> |
128 |
|
|
142 |
|
|
143 |
my $cache_file; |
my $cache_file; |
144 |
|
|
145 |
if (my $cache_path = $self->{'cache_data_structure'}) { |
if ($self->{'db'}) { |
146 |
my $id = $rec->{'000'}; |
my $ds = $self->{'db'}->load_ds($rec); |
147 |
$id = $rec->{'000'}->[0] if ($id =~ m/^ARRAY/o); |
$log->debug("load_ds( rec = ", sub { Dumper($rec) }, ") = ", sub { Dumper($ds) }); |
148 |
unless (defined($id)) { |
return $ds if ($ds); |
149 |
$log->warn("Can't use cache_data_structure on records without unique identifier in field 000"); |
$log->debug("cache miss, creating"); |
|
undef $self->{'cache_data_structure'}; |
|
|
} else { |
|
|
$cache_file = "$cache_path/$id"; |
|
|
if (-r $cache_file) { |
|
|
my $ds_ref = retrieve($cache_file); |
|
|
if ($ds_ref) { |
|
|
$log->debug("cache hit: $cache_file"); |
|
|
my $ok = 1; |
|
|
foreach my $f (qw(current_filename headline)) { |
|
|
if ($ds_ref->{$f}) { |
|
|
$self->{$f} = $ds_ref->{$f}; |
|
|
} else { |
|
|
$ok = 0; |
|
|
} |
|
|
}; |
|
|
if ($ok && $ds_ref->{'ds'}) { |
|
|
return @{ $ds_ref->{'ds'} }; |
|
|
} else { |
|
|
$log->warn("cache_data_structure $cache_path corrupt. Use rm $cache_path/* to re-create it on next run!"); |
|
|
undef $self->{'cache_data_structure'}; |
|
|
} |
|
|
} |
|
|
} |
|
|
} |
|
150 |
} |
} |
151 |
|
|
152 |
undef $self->{'currnet_filename'}; |
undef $self->{'currnet_filename'}; |
160 |
$self->{tags_by_order} = \@sorted_tags; |
$self->{tags_by_order} = \@sorted_tags; |
161 |
} |
} |
162 |
|
|
163 |
my @ds; |
my $ds; |
164 |
|
|
165 |
$log->debug("tags: ",sub { join(", ",@sorted_tags) }); |
$log->debug("tags: ",sub { join(", ",@sorted_tags) }); |
166 |
|
|
171 |
#print "field $field [",$self->{'tag'},"] = ",Dumper($self->{'import_xml'}->{'indexer'}->{$field}->{$self->{'tag'}}); |
#print "field $field [",$self->{'tag'},"] = ",Dumper($self->{'import_xml'}->{'indexer'}->{$field}->{$self->{'tag'}}); |
172 |
|
|
173 |
foreach my $tag (@{$self->{'import_xml'}->{'indexer'}->{$field}->{$self->{'tag'}}}) { |
foreach my $tag (@{$self->{'import_xml'}->{'indexer'}->{$field}->{$self->{'tag'}}}) { |
174 |
my $format = $tag->{'value'} || $tag->{'content'}; |
my $format; |
175 |
|
|
176 |
|
$log->logdie("expected tag HASH and got $tag") unless (ref($tag) eq 'HASH'); |
177 |
|
$format = $tag->{'value'} || $tag->{'content'}; |
178 |
|
|
179 |
$log->debug("format: $format"); |
$log->debug("format: $format"); |
180 |
|
|
243 |
|
|
244 |
# TODO: name_sigular, name_plural |
# TODO: name_sigular, name_plural |
245 |
my $name = $self->{'import_xml'}->{'indexer'}->{$field}->{'name'}; |
my $name = $self->{'import_xml'}->{'indexer'}->{$field}->{'name'}; |
246 |
$row->{'name'} = $name ? $self->_x($name) : $field; |
my $row_name = $name ? $self->_x($name) : $field; |
247 |
|
|
248 |
# post-sort all values in field |
# post-sort all values in field |
249 |
if ($self->{'import_xml'}->{'indexer'}->{$field}->{'sort'}) { |
if ($self->{'import_xml'}->{'indexer'}->{$field}->{'sort'}) { |
250 |
$log->warn("sort at field tag not implemented"); |
$log->warn("sort at field tag not implemented"); |
251 |
} |
} |
252 |
|
|
253 |
push @ds, $row; |
$ds->{$row_name} = $row; |
254 |
|
|
255 |
$log->debug("row $field: ",sub { Dumper($row) }); |
$log->debug("row $field: ",sub { Dumper($row) }); |
256 |
} |
} |
257 |
|
|
258 |
} |
} |
259 |
|
|
260 |
if ($cache_file) { |
$log->logdie("there is no current_filename defined! Do you have filename tag in conf/normalize/?.xml") unless ($self->{'current_filename'}); |
|
store { |
|
|
ds => \@ds, |
|
|
current_filename => $self->{'current_filename'}, |
|
|
headline => $self->{'headline'}, |
|
|
}, $cache_file; |
|
|
$log->debug("created storable cache file $cache_file"); |
|
|
} |
|
|
|
|
|
return @ds; |
|
261 |
|
|
262 |
} |
$self->{'db'}->save_ds( |
263 |
|
ds => $ds, |
264 |
=head2 apply_format |
current_filename => $self->{'current_filename'}, |
265 |
|
headline => $self->{'headline'}, |
266 |
Apply format specified in tag with C<format_name="name"> and |
) if ($self->{'db'}); |
|
C<format_delimiter=";;">. |
|
|
|
|
|
my $text = $webpac->apply_format($format_name,$format_delimiter,$data); |
|
|
|
|
|
Formats can contain C<lookup{...}> if you need them. |
|
|
|
|
|
=cut |
|
|
|
|
|
sub apply_format { |
|
|
my $self = shift; |
|
|
|
|
|
my ($name,$delimiter,$data) = @_; |
|
|
|
|
|
my $log = $self->_get_logger(); |
|
|
|
|
|
if (! $self->{'import_xml'}->{'format'}->{$name}) { |
|
|
$log->warn("<format name=\"$name\"> is not defined in ",$self->{'import_xml_file'}); |
|
|
return $data; |
|
|
} |
|
|
|
|
|
$log->warn("no delimiter for format $name") if (! $delimiter); |
|
267 |
|
|
268 |
my $format = $self->_x($self->{'import_xml'}->{'format'}->{$name}->{'content'}) || $log->logdie("can't find format '$name'"); |
$log->debug("ds: ", sub { Dumper($ds) }); |
|
|
|
|
my @data = split(/\Q$delimiter\E/, $data); |
|
269 |
|
|
270 |
my $out = sprintf($format, @data); |
$log->logconfess("data structure returned is not array any more!") if wantarray; |
|
$log->debug("using format $name [$format] on $data to produce: $out"); |
|
271 |
|
|
272 |
if ($self->{'lookup_regex'} && $out =~ $self->{'lookup_regex'}) { |
return $ds; |
|
return $self->lookup($out); |
|
|
} else { |
|
|
return $out; |
|
|
} |
|
273 |
|
|
274 |
} |
} |
275 |
|
|
392 |
return @arr; |
return @arr; |
393 |
} |
} |
394 |
|
|
395 |
|
|
396 |
|
=head2 fill_in |
397 |
|
|
398 |
|
Workhourse of all: takes record from in-memory structure of database and |
399 |
|
strings with placeholders and returns string or array of with substituted |
400 |
|
values from record. |
401 |
|
|
402 |
|
my $text = $webpac->fill_in($rec,'v250^a'); |
403 |
|
|
404 |
|
Optional argument is ordinal number for repeatable fields. By default, |
405 |
|
it's assume to be first repeatable field (fields are perl array, so first |
406 |
|
element is 0). |
407 |
|
Following example will read second value from repeatable field. |
408 |
|
|
409 |
|
my $text = $webpac->fill_in($rec,'Title: v250^a',1); |
410 |
|
|
411 |
|
This function B<does not> perform parsing of format to inteligenty skip |
412 |
|
delimiters before fields which aren't used. |
413 |
|
|
414 |
|
This method will automatically decode UTF-8 string to local code page |
415 |
|
if needed. |
416 |
|
|
417 |
|
=cut |
418 |
|
|
419 |
|
sub fill_in { |
420 |
|
my $self = shift; |
421 |
|
|
422 |
|
my $log = $self->_get_logger(); |
423 |
|
|
424 |
|
my $rec = shift || $log->logconfess("need data record"); |
425 |
|
my $format = shift || $log->logconfess("need format to parse"); |
426 |
|
# iteration (for repeatable fields) |
427 |
|
my $i = shift || 0; |
428 |
|
|
429 |
|
$log->logdie("infitite loop in format $format") if ($i > ($self->{'max_mfn'} || 9999)); |
430 |
|
|
431 |
|
# FIXME remove for speedup? |
432 |
|
$log->logconfess("need HASH as first argument!") if ($rec !~ /HASH/o); |
433 |
|
|
434 |
|
if (utf8::is_utf8($format)) { |
435 |
|
$format = $self->_x($format); |
436 |
|
} |
437 |
|
|
438 |
|
my $found = 0; |
439 |
|
|
440 |
|
my $eval_code; |
441 |
|
# remove eval{...} from beginning |
442 |
|
$eval_code = $1 if ($format =~ s/^eval{([^}]+)}//s); |
443 |
|
|
444 |
|
my $filter_name; |
445 |
|
# remove filter{...} from beginning |
446 |
|
$filter_name = $1 if ($format =~ s/^filter{([^}]+)}//s); |
447 |
|
|
448 |
|
# do actual replacement of placeholders |
449 |
|
# repeatable fields |
450 |
|
$format =~ s/v(\d+)(?:\^(\w))?/$self->get_data(\$rec,$1,$2,$i,\$found)/ges; |
451 |
|
# non-repeatable fields |
452 |
|
$format =~ s/s(\d+)(?:\^(\w))?/$self->get_data(\$rec,$1,$2,0,\$found)/ges; |
453 |
|
|
454 |
|
if ($found) { |
455 |
|
$log->debug("format: $format"); |
456 |
|
if ($eval_code) { |
457 |
|
my $eval = $self->fill_in($rec,$eval_code,$i); |
458 |
|
return if (! $self->_eval($eval)); |
459 |
|
} |
460 |
|
if ($filter_name && $self->{'filter'}->{$filter_name}) { |
461 |
|
$log->debug("filter '$filter_name' for $format"); |
462 |
|
$format = $self->{'filter'}->{$filter_name}->($format); |
463 |
|
return unless(defined($format)); |
464 |
|
$log->debug("filter result: $format"); |
465 |
|
} |
466 |
|
# do we have lookups? |
467 |
|
if ($self->{'lookup'}) { |
468 |
|
if ($self->{'lookup'}->can('lookup')) { |
469 |
|
return $self->{'lookup'}->lookup($format); |
470 |
|
} else { |
471 |
|
$log->warn("Have lookup object but can't invoke lookup method"); |
472 |
|
} |
473 |
|
} else { |
474 |
|
return $format; |
475 |
|
} |
476 |
|
} else { |
477 |
|
return; |
478 |
|
} |
479 |
|
} |
480 |
|
|
481 |
|
|
482 |
=head2 fill_in_to_arr |
=head2 fill_in_to_arr |
483 |
|
|
484 |
Similar to C<fill_in>, but returns array of all repeatable fields. Usable |
Similar to C<fill_in>, but returns array of all repeatable fields. Usable |
511 |
return @arr; |
return @arr; |
512 |
} |
} |
513 |
|
|
514 |
|
|
515 |
|
=head2 get_data |
516 |
|
|
517 |
|
Returns value from record. |
518 |
|
|
519 |
|
my $text = $self->get_data(\$rec,$f,$sf,$i,\$found); |
520 |
|
|
521 |
|
Arguments are: |
522 |
|
record reference C<$rec>, |
523 |
|
field C<$f>, |
524 |
|
optional subfiled C<$sf>, |
525 |
|
index for repeatable values C<$i>. |
526 |
|
|
527 |
|
Optinal variable C<$found> will be incremeted if there |
528 |
|
is field. |
529 |
|
|
530 |
|
Returns value or empty string. |
531 |
|
|
532 |
|
=cut |
533 |
|
|
534 |
|
sub get_data { |
535 |
|
my $self = shift; |
536 |
|
|
537 |
|
my ($rec,$f,$sf,$i,$found) = @_; |
538 |
|
|
539 |
|
if ($$rec->{$f}) { |
540 |
|
return '' if (! $$rec->{$f}->[$i]); |
541 |
|
no strict 'refs'; |
542 |
|
if ($sf && $$rec->{$f}->[$i]->{$sf}) { |
543 |
|
$$found++ if (defined($$found)); |
544 |
|
return $$rec->{$f}->[$i]->{$sf}; |
545 |
|
} elsif (! $sf && $$rec->{$f}->[$i]) { |
546 |
|
$$found++ if (defined($$found)); |
547 |
|
# it still might have subfield, just |
548 |
|
# not specified, so we'll dump all |
549 |
|
if ($$rec->{$f}->[$i] =~ /HASH/o) { |
550 |
|
my $out; |
551 |
|
foreach my $k (keys %{$$rec->{$f}->[$i]}) { |
552 |
|
$out .= $$rec->{$f}->[$i]->{$k}." "; |
553 |
|
} |
554 |
|
return $out; |
555 |
|
} else { |
556 |
|
return $$rec->{$f}->[$i]; |
557 |
|
} |
558 |
|
} else { |
559 |
|
return ''; |
560 |
|
} |
561 |
|
} else { |
562 |
|
return ''; |
563 |
|
} |
564 |
|
} |
565 |
|
|
566 |
|
|
567 |
|
=head2 apply_format |
568 |
|
|
569 |
|
Apply format specified in tag with C<format_name="name"> and |
570 |
|
C<format_delimiter=";;">. |
571 |
|
|
572 |
|
my $text = $webpac->apply_format($format_name,$format_delimiter,$data); |
573 |
|
|
574 |
|
Formats can contain C<lookup{...}> if you need them. |
575 |
|
|
576 |
|
=cut |
577 |
|
|
578 |
|
sub apply_format { |
579 |
|
my $self = shift; |
580 |
|
|
581 |
|
my ($name,$delimiter,$data) = @_; |
582 |
|
|
583 |
|
my $log = $self->_get_logger(); |
584 |
|
|
585 |
|
if (! $self->{'import_xml'}->{'format'}->{$name}) { |
586 |
|
$log->warn("<format name=\"$name\"> is not defined in ",$self->{'import_xml_file'}); |
587 |
|
return $data; |
588 |
|
} |
589 |
|
|
590 |
|
$log->warn("no delimiter for format $name") if (! $delimiter); |
591 |
|
|
592 |
|
my $format = $self->_x($self->{'import_xml'}->{'format'}->{$name}->{'content'}) || $log->logdie("can't find format '$name'"); |
593 |
|
|
594 |
|
my @data = split(/\Q$delimiter\E/, $data); |
595 |
|
|
596 |
|
my $out = sprintf($format, @data); |
597 |
|
$log->debug("using format $name [$format] on $data to produce: $out"); |
598 |
|
|
599 |
|
if ($self->{'lookup_regex'} && $out =~ $self->{'lookup_regex'}) { |
600 |
|
return $self->{'lookup'}->lookup($out); |
601 |
|
} else { |
602 |
|
return $out; |
603 |
|
} |
604 |
|
|
605 |
|
} |
606 |
|
|
607 |
=head2 sort_arr |
=head2 sort_arr |
608 |
|
|
609 |
Sort array ignoring case and html in data |
Sort array ignoring case and html in data |
630 |
} |
} |
631 |
|
|
632 |
|
|
633 |
|
=head1 INTERNAL METHODS |
634 |
|
|
635 |
=head2 _sort_by_order |
=head2 _sort_by_order |
636 |
|
|
637 |
Sort xml tags data structure accoding to C<order=""> attribute. |
Sort xml tags data structure accoding to C<order=""> attribute. |
651 |
|
|
652 |
=head2 _x |
=head2 _x |
653 |
|
|
654 |
Convert strings from C<conf/normalize> encoding into application specific |
Convert strings from C<conf/normalize/*.xml> encoding into application |
655 |
(optinally specified using C<code_page> to C<new> constructor. |
specific encoding (optinally specified using C<code_page> to C<new> |
656 |
|
constructor). |
657 |
|
|
658 |
my $text = $n->_x('normalize text string'); |
my $text = $n->_x('normalize text string'); |
659 |
|
|