7 |
|
|
8 |
use WebPAC::Common; |
use WebPAC::Common; |
9 |
use base qw/WebPAC::Common/; |
use base qw/WebPAC::Common/; |
|
use Text::Iconv; |
|
10 |
use Data::Dumper; |
use Data::Dumper; |
11 |
|
use Encode qw/from_to/; |
12 |
|
|
13 |
=head1 NAME |
=head1 NAME |
14 |
|
|
159 |
|
|
160 |
$input->open( |
$input->open( |
161 |
path => '/path/to/database/file', |
path => '/path/to/database/file', |
162 |
code_page => '852', |
code_page => 'cp852', |
163 |
limit => 500, |
limit => 500, |
164 |
offset => 6000, |
offset => 6000, |
165 |
lookup => $lookup_obj, |
lookup => $lookup_obj, |
174 |
}, |
}, |
175 |
); |
); |
176 |
|
|
177 |
By default, C<code_page> is assumed to be C<852>. |
By default, C<code_page> is assumed to be C<cp852>. |
178 |
|
|
179 |
C<offset> is optional parametar to position at some offset before reading from database. |
C<offset> is optional parametar to position at some offset before reading from database. |
180 |
|
|
205 |
if ($arg->{lookup_coderef} && ref($arg->{lookup_coderef}) ne 'CODE'); |
if ($arg->{lookup_coderef} && ref($arg->{lookup_coderef}) ne 'CODE'); |
206 |
|
|
207 |
$log->logcroak("need path") if (! $arg->{'path'}); |
$log->logcroak("need path") if (! $arg->{'path'}); |
208 |
my $code_page = $arg->{'code_page'} || '852'; |
my $code_page = $arg->{'code_page'} || 'cp852'; |
209 |
|
|
210 |
# store data in object |
# store data in object |
211 |
$self->{'input_code_page'} = $code_page; |
$self->{'input_code_page'} = $code_page; |
213 |
$self->{$v} = $arg->{$v} if ($arg->{$v}); |
$self->{$v} = $arg->{$v} if ($arg->{$v}); |
214 |
} |
} |
215 |
|
|
|
# create Text::Iconv object |
|
|
$self->{iconv} = Text::Iconv->new($code_page,$self->{'encoding'}); ## FIXME remove! |
|
|
|
|
216 |
my $filter_ref; |
my $filter_ref; |
217 |
my $recode_regex; |
my $recode_regex; |
218 |
my $recode_map; |
my $recode_map; |
240 |
|
|
241 |
my ($db, $size) = $self->{open_db}->( $self, |
my ($db, $size) = $self->{open_db}->( $self, |
242 |
path => $arg->{path}, |
path => $arg->{path}, |
243 |
filter => sub { |
# filter => sub { |
244 |
my ($l,$f_nr) = @_; |
# my ($l,$f_nr) = @_; |
245 |
return unless defined($l); |
# return unless defined($l); |
246 |
|
# from_to($l, $code_page, $self->{'encoding'}); |
247 |
## FIXME remove iconv! |
# $l =~ s/($recode_regex)/$recode_map->{$1}/g if ($recode_regex && $recode_map); |
248 |
$l = $self->{iconv}->convert($l) if ($self->{iconv}); |
# return $l; |
249 |
|
# }, |
|
$l =~ s/($recode_regex)/$recode_map->{$1}/g if ($recode_regex && $recode_map); |
|
|
|
|
|
## FIXME remove this warning when we are sure that none of API is calling |
|
|
## this wrongly |
|
|
warn "filter called without field number" unless ($f_nr); |
|
|
|
|
|
return $l; |
|
|
}, |
|
250 |
%{ $arg }, |
%{ $arg }, |
251 |
); |
); |
252 |
|
|
288 |
|
|
289 |
my $rec = $self->{fetch_rec}->($self, $db, $pos, sub { |
my $rec = $self->{fetch_rec}->($self, $db, $pos, sub { |
290 |
my ($l,$f_nr) = @_; |
my ($l,$f_nr) = @_; |
291 |
return unless defined($l); |
# return unless defined($l); |
292 |
return $l unless ($rec_regex && $f_nr); |
# return $l unless ($rec_regex && $f_nr); |
293 |
|
|
294 |
|
$log->debug("-=> $f_nr ## $l"); |
295 |
|
|
296 |
|
# codepage conversion and recode_regex |
297 |
|
from_to($l, $code_page, $self->{'encoding'}); |
298 |
|
$l =~ s/($recode_regex)/$recode_map->{$1}/g if ($recode_regex && $recode_map); |
299 |
|
|
300 |
# apply regexps |
# apply regexps |
301 |
if ($rec_regex && defined($rec_regex->{$f_nr})) { |
if ($rec_regex && defined($rec_regex->{$f_nr})) { |
302 |
$log->logconfess("regexps->{$f_nr} must be ARRAY") if (ref($rec_regex->{$f_nr}) ne 'ARRAY'); |
$log->logconfess("regexps->{$f_nr} must be ARRAY") if (ref($rec_regex->{$f_nr}) ne 'ARRAY'); |
303 |
my $c = 0; |
my $c = 0; |
304 |
foreach my $r (@{ $rec_regex->{$f_nr} }) { |
foreach my $r (@{ $rec_regex->{$f_nr} }) { |
305 |
#$log->debug("\$l = $l\neval \$l =~ $r"); |
my $old_l = $l; |
306 |
eval '$l =~ ' . $r; |
eval '$l =~ ' . $r; |
307 |
|
if ($old_l ne $l) { |
308 |
|
$log->debug("REGEX on $f_nr eval \$l =~ $r\n## old l: [$old_l]\n## new l: [$l]"); |
309 |
|
} |
310 |
$log->error("error applying regex: $r") if ($@); |
$log->error("error applying regex: $r") if ($@); |
311 |
} |
} |
312 |
} |
} |
313 |
|
|
314 |
|
$log->debug("<=- $f_nr ## $l"); |
315 |
return $l; |
return $l; |
316 |
}); |
}); |
317 |
|
|
551 |
|
|
552 |
my $regexpes; |
my $regexpes; |
553 |
|
|
554 |
|
my $log = $self->_get_logger(); |
555 |
|
|
556 |
foreach my $f (keys %$modify_record) { |
foreach my $f (keys %$modify_record) { |
557 |
warn "--- f: $f\n"; |
$log->debug("field: $f"); |
558 |
|
|
559 |
foreach my $sf (keys %{ $modify_record->{$f} }) { |
foreach my $sf (keys %{ $modify_record->{$f} }) { |
560 |
warn "---- sf: $sf\n"; |
$log->debug("subfield: $sf"); |
561 |
|
|
562 |
foreach my $from (keys %{ $modify_record->{$f}->{$sf} }) { |
foreach my $from (keys %{ $modify_record->{$f}->{$sf} }) { |
563 |
my $to = $modify_record->{$f}->{$sf}->{$from}; |
my $to = $modify_record->{$f}->{$sf}->{$from}; |
564 |
#die "no field?" unless defined($to); |
#die "no field?" unless defined($to); |
565 |
warn "----- transform: |$from| -> |$to|\n"; |
$log->debug("transform: |$from| -> |$to|"); |
566 |
|
|
567 |
if ($sf =~ /^\^/) { |
if ($sf =~ /^\^/) { |
568 |
my $regex = |
my $regex = |
569 |
's/\Q'. $sf .'\E([^\^]+)\Q'. $from .'\E([^\^]+)/'. $sf .'$1'. $to .'$2/g'; |
's/\Q'. $sf .'\E(.*?)\Q'. $from .'\E(.*?)/'. $sf .'$1'. $to .'$2/g'; |
570 |
push @{ $regexpes->{$f} }, $regex; |
push @{ $regexpes->{$f} }, $regex; |
571 |
warn ">>>>> $regex [sf]\n"; |
$log->debug(">>>>> $regex [sf]"); |
572 |
} else { |
} else { |
573 |
my $regex = |
my $regex = |
574 |
's/\Q'. $from .'\E/'. $to .'/g'; |
's/\Q'. $from .'\E/'. $to .'/g'; |
575 |
push @{ $regexpes->{$f} }, $regex; |
push @{ $regexpes->{$f} }, $regex; |
576 |
warn ">>>>> $regex [global]\n"; |
$log->debug(">>>>> $regex [global]"); |
577 |
} |
} |
578 |
|
|
579 |
} |
} |