14 |
|
|
15 |
=head1 VERSION |
=head1 VERSION |
16 |
|
|
17 |
Version 0.04 |
Version 0.05 |
18 |
|
|
19 |
=cut |
=cut |
20 |
|
|
21 |
our $VERSION = '0.04'; |
our $VERSION = '0.05'; |
22 |
|
|
23 |
=head1 SYNOPSIS |
=head1 SYNOPSIS |
24 |
|
|
65 |
code_page => 'ISO-8859-2', |
code_page => 'ISO-8859-2', |
66 |
low_mem => 1, |
low_mem => 1, |
67 |
recode => 'char pairs', |
recode => 'char pairs', |
68 |
|
no_progress_bar => 1, |
69 |
); |
); |
70 |
|
|
71 |
C<module> is low-level file format module. See L<WebPAC::Input::Isis> and |
C<module> is low-level file format module. See L<WebPAC::Input::Isis> and |
77 |
|
|
78 |
Default is not to use C<low_mem> options (see L<MEMORY USAGE> below). |
Default is not to use C<low_mem> options (see L<MEMORY USAGE> below). |
79 |
|
|
80 |
|
C<recode> is optional string constisting of character or words pairs that |
81 |
|
should be replaced in input stream. |
82 |
|
|
83 |
|
C<no_progress_bar> disables progress bar output on C<STDOUT> |
84 |
|
|
85 |
This function will also call low-level C<init> if it exists with same |
This function will also call low-level C<init> if it exists with same |
86 |
parametars. |
parametars. |
87 |
|
|
160 |
limit => 500, |
limit => 500, |
161 |
offset => 6000, |
offset => 6000, |
162 |
lookup => $lookup_obj, |
lookup => $lookup_obj, |
163 |
|
stats => 1, |
164 |
); |
); |
165 |
|
|
166 |
By default, C<code_page> is assumed to be C<852>. |
By default, C<code_page> is assumed to be C<852>. |
169 |
|
|
170 |
C<limit> is optional parametar to read just C<limit> records from database |
C<limit> is optional parametar to read just C<limit> records from database |
171 |
|
|
172 |
|
C<stats> create optional report about usage of fields and subfields |
173 |
|
|
174 |
Returns size of database, regardless of C<offset> and C<limit> |
Returns size of database, regardless of C<offset> and C<limit> |
175 |
parametars, see also C<size>. |
parametars, see also C<size>. |
176 |
|
|
227 |
filter => $filter_ref, |
filter => $filter_ref, |
228 |
); |
); |
229 |
|
|
230 |
unless ($db) { |
unless (defined($db)) { |
231 |
$log->logwarn("can't open database $arg->{path}, skipping..."); |
$log->logwarn("can't open database $arg->{path}, skipping..."); |
232 |
return; |
return; |
233 |
} |
} |
256 |
# store size for later |
# store size for later |
257 |
$self->{size} = ($to_rec - $from_rec) ? ($to_rec - $from_rec + 1) : 0; |
$self->{size} = ($to_rec - $from_rec) ? ($to_rec - $from_rec + 1) : 0; |
258 |
|
|
259 |
$log->info("processing $self->{size}/$size records [$from_rec-$to_rec] convert $code_page -> $self->{code_page}"); |
$log->info("processing $self->{size}/$size records [$from_rec-$to_rec] convert $code_page -> $self->{code_page}", $self->{stats} ? ' [stats]' : ''); |
260 |
|
|
261 |
# read database |
# read database |
262 |
for (my $pos = $from_rec; $pos <= $to_rec; $pos++) { |
for (my $pos = $from_rec; $pos <= $to_rec; $pos++) { |
282 |
# create lookup |
# create lookup |
283 |
$self->{'lookup'}->add( $rec ) if ($rec && $self->{'lookup'}); |
$self->{'lookup'}->add( $rec ) if ($rec && $self->{'lookup'}); |
284 |
|
|
285 |
$self->progress_bar($pos,$to_rec); |
# update counters for statistics |
286 |
|
if ($self->{stats}) { |
287 |
|
map { |
288 |
|
my $fld = $_; |
289 |
|
$self->{_stats}->{fld}->{ $fld }++; |
290 |
|
if (ref($rec->{ $fld }) eq 'ARRAY') { |
291 |
|
map { |
292 |
|
if (ref($_) eq 'HASH') { |
293 |
|
map { |
294 |
|
$self->{_stats}->{sf}->{ $fld }->{ $_ }++; |
295 |
|
} keys %{ $_ }; |
296 |
|
} else { |
297 |
|
$self->{_stats}->{repeatable}->{ $fld }++; |
298 |
|
} |
299 |
|
} @{ $rec->{$fld} }; |
300 |
|
} |
301 |
|
} keys %{ $rec }; |
302 |
|
} |
303 |
|
|
304 |
|
$self->progress_bar($pos,$to_rec) unless ($self->{no_progress_bar}); |
305 |
|
|
306 |
} |
} |
307 |
|
|
347 |
return; |
return; |
348 |
} |
} |
349 |
|
|
350 |
$self->progress_bar($mfn,$self->{max_pos}); |
$self->progress_bar($mfn,$self->{max_pos}) unless ($self->{no_progress_bar}); |
351 |
|
|
352 |
my $rec; |
my $rec; |
353 |
|
|
422 |
return $self->{pos} = (($pos - 1) || -1); |
return $self->{pos} = (($pos - 1) || -1); |
423 |
} |
} |
424 |
|
|
425 |
|
=head2 stats |
426 |
|
|
427 |
|
Dump statistics about field and subfield usage |
428 |
|
|
429 |
|
print Dumper( $input->stats ); |
430 |
|
|
431 |
|
=cut |
432 |
|
|
433 |
|
sub stats { |
434 |
|
my $self = shift; |
435 |
|
return $self->{_stats}; |
436 |
|
} |
437 |
|
|
438 |
=head1 MEMORY USAGE |
=head1 MEMORY USAGE |
439 |
|
|