1 |
package WebPAC::Input; |
2 |
|
3 |
use warnings; |
4 |
use strict; |
5 |
|
6 |
use blib; |
7 |
|
8 |
use WebPAC::Common; |
9 |
use base qw/WebPAC::Common/; |
10 |
use Text::Iconv; |
11 |
|
12 |
=head1 NAME |
13 |
|
14 |
WebPAC::Input - read different file formats into WebPAC |
15 |
|
16 |
=head1 VERSION |
17 |
|
18 |
Version 0.03 |
19 |
|
20 |
=cut |
21 |
|
22 |
our $VERSION = '0.03'; |
23 |
|
24 |
=head1 SYNOPSIS |
25 |
|
26 |
This module implements input as database which have fixed and known |
27 |
I<size> while indexing and single unique numeric identifier for database |
28 |
position ranging from 1 to I<size>. |
29 |
|
30 |
Simply, something that is indexed by unmber from 1 .. I<size>. |
31 |
|
32 |
Examples of such databases are CDS/ISIS files, MARC files, lines in |
33 |
text file, and so on. |
34 |
|
35 |
Specific file formats are implemented using low-level interface modules, |
36 |
located in C<WebPAC::Input::*> namespace which export C<open_db>, |
37 |
C<fetch_rec> and optional C<init> functions. |
38 |
|
39 |
Perhaps a little code snippet. |
40 |
|
41 |
use WebPAC::Input; |
42 |
|
43 |
my $db = WebPAC::Input->new( |
44 |
module => 'WebPAC::Input::ISIS', |
45 |
config => $config, |
46 |
lookup => $lookup_obj, |
47 |
low_mem => 1, |
48 |
); |
49 |
|
50 |
$db->open('/path/to/database'); |
51 |
print "database size: ",$db->size,"\n"; |
52 |
while (my $rec = $db->fetch) { |
53 |
} |
54 |
|
55 |
|
56 |
|
57 |
=head1 FUNCTIONS |
58 |
|
59 |
=head2 new |
60 |
|
61 |
Create new input database object. |
62 |
|
63 |
my $db = new WebPAC::Input( |
64 |
module => 'WebPAC::Input::MARC', |
65 |
code_page => 'ISO-8859-2', |
66 |
low_mem => 1, |
67 |
); |
68 |
|
69 |
C<module> is low-level file format module. See L<WebPAC::Input::Isis> and |
70 |
L<WebPAC::Input::MARC>. |
71 |
|
72 |
Optional parametar C<code_page> specify application code page (which will be |
73 |
used internally). This should probably be your terminal encoding, and by |
74 |
default, it C<ISO-8859-2>. |
75 |
|
76 |
Default is not to use C<low_mem> options (see L<MEMORY USAGE> below). |
77 |
|
78 |
This function will also call low-level C<init> if it exists with same |
79 |
parametars. |
80 |
|
81 |
=cut |
82 |
|
83 |
sub new { |
84 |
my $class = shift; |
85 |
my $self = {@_}; |
86 |
bless($self, $class); |
87 |
|
88 |
my $log = $self->_get_logger; |
89 |
|
90 |
$log->logconfess("specify low-level file format module") unless ($self->{module}); |
91 |
my $module = $self->{module}; |
92 |
$module =~ s#::#/#g; |
93 |
$module .= '.pm'; |
94 |
$log->debug("require low-level module $self->{module} from $module"); |
95 |
|
96 |
require $module; |
97 |
#eval $self->{module} .'->import'; |
98 |
|
99 |
# check if required subclasses are implemented |
100 |
foreach my $subclass (qw/open_db fetch_rec init/) { |
101 |
my $n = $self->{module} . '::' . $subclass; |
102 |
if (! defined &{ $n }) { |
103 |
my $missing = "missing $subclass in $self->{module}"; |
104 |
$self->{$subclass} = sub { $log->logwarn($missing) }; |
105 |
} else { |
106 |
$self->{$subclass} = \&{ $n }; |
107 |
} |
108 |
} |
109 |
|
110 |
if ($self->{init}) { |
111 |
$log->debug("calling init"); |
112 |
$self->{init}->($self, @_); |
113 |
} |
114 |
|
115 |
$self->{'code_page'} ||= 'ISO-8859-2'; |
116 |
|
117 |
# running with low_mem flag? well, use DBM::Deep then. |
118 |
if ($self->{'low_mem'}) { |
119 |
$log->info("running with low_mem which impacts performance (<32 Mb memory usage)"); |
120 |
|
121 |
my $db_file = "data.db"; |
122 |
|
123 |
if (-e $db_file) { |
124 |
unlink $db_file or $log->logdie("can't remove '$db_file' from last run"); |
125 |
$log->debug("removed '$db_file' from last run"); |
126 |
} |
127 |
|
128 |
require DBM::Deep; |
129 |
|
130 |
my $db = new DBM::Deep $db_file; |
131 |
|
132 |
$log->logdie("DBM::Deep error: $!") unless ($db); |
133 |
|
134 |
if ($db->error()) { |
135 |
$log->logdie("can't open '$db_file' under low_mem: ",$db->error()); |
136 |
} else { |
137 |
$log->debug("using file '$db_file' for DBM::Deep"); |
138 |
} |
139 |
|
140 |
$self->{'db'} = $db; |
141 |
} |
142 |
|
143 |
$self ? return $self : return undef; |
144 |
} |
145 |
|
146 |
=head2 open |
147 |
|
148 |
This function will read whole database in memory and produce lookups. |
149 |
|
150 |
$input->open( |
151 |
path => '/path/to/database/file', |
152 |
code_page => '852', |
153 |
limit => 500, |
154 |
offset => 6000, |
155 |
lookup => $lookup_obj, |
156 |
); |
157 |
|
158 |
By default, C<code_page> is assumed to be C<852>. |
159 |
|
160 |
C<offset> is optional parametar to position at some offset before reading from database. |
161 |
|
162 |
C<limit> is optional parametar to read just C<limit> records from database |
163 |
|
164 |
Returns size of database, regardless of C<offset> and C<limit> |
165 |
parametars, see also C<size>. |
166 |
|
167 |
=cut |
168 |
|
169 |
sub open { |
170 |
my $self = shift; |
171 |
my $arg = {@_}; |
172 |
|
173 |
my $log = $self->_get_logger(); |
174 |
|
175 |
$log->logcroak("need path") if (! $arg->{'path'}); |
176 |
my $code_page = $arg->{'code_page'} || '852'; |
177 |
|
178 |
# store data in object |
179 |
$self->{'input_code_page'} = $code_page; |
180 |
foreach my $v (qw/path offset limit/) { |
181 |
$self->{$v} = $arg->{$v} if ($arg->{$v}); |
182 |
} |
183 |
|
184 |
# create Text::Iconv object |
185 |
$self->{iconv} = Text::Iconv->new($code_page,$self->{'code_page'}); |
186 |
|
187 |
my ($db, $size) = $self->{open_db}->( $self, |
188 |
path => $arg->{path}, |
189 |
); |
190 |
|
191 |
unless ($db) { |
192 |
$log->logwarn("can't open database $arg->{path}, skipping..."); |
193 |
return; |
194 |
} |
195 |
|
196 |
unless ($size) { |
197 |
$log->logwarn("no records in database $arg->{path}, skipping..."); |
198 |
return; |
199 |
} |
200 |
|
201 |
my $offset = 1; |
202 |
my $limit = $size; |
203 |
|
204 |
if (my $s = $self->{offset}) { |
205 |
$log->info("skipping to MFN $s"); |
206 |
$offset = $s; |
207 |
} else { |
208 |
$self->{offset} = $offset; |
209 |
} |
210 |
|
211 |
if ($self->{limit}) { |
212 |
$log->debug("limiting to ",$self->{limit}," records"); |
213 |
$limit = $offset + $self->{limit} - 1; |
214 |
$limit = $size if ($limit > $size); |
215 |
} |
216 |
|
217 |
# store size for later |
218 |
$self->{size} = ($limit - $offset) ? ($limit - $offset + 1) : 0; |
219 |
|
220 |
$log->info("processing $self->{size} records in $code_page, convert to $self->{code_page}"); |
221 |
|
222 |
# read database |
223 |
for (my $pos = $offset; $pos <= $limit; $pos++) { |
224 |
|
225 |
$log->debug("position: $pos\n"); |
226 |
|
227 |
my $rec = $self->{fetch_rec}->($self, $db, $pos ); |
228 |
|
229 |
if (! $rec) { |
230 |
$log->warn("record $pos empty? skipping..."); |
231 |
next; |
232 |
} |
233 |
|
234 |
# store |
235 |
if ($self->{low_mem}) { |
236 |
$self->{db}->put($pos, $rec); |
237 |
} else { |
238 |
$self->{data}->{$pos} = $rec; |
239 |
} |
240 |
|
241 |
# create lookup |
242 |
$self->{'lookup'}->add( $rec ) if ($rec && $self->{'lookup'}); |
243 |
|
244 |
$self->progress_bar($pos,$limit); |
245 |
|
246 |
} |
247 |
|
248 |
$self->{pos} = -1; |
249 |
$self->{last_pcnt} = 0; |
250 |
|
251 |
# store max mfn and return it. |
252 |
$self->{max_pos} = $limit; |
253 |
$log->debug("max_pos: $limit"); |
254 |
|
255 |
return $size; |
256 |
} |
257 |
|
258 |
=head2 fetch |
259 |
|
260 |
Fetch next record from database. It will also displays progress bar. |
261 |
|
262 |
my $rec = $isis->fetch; |
263 |
|
264 |
Record from this function should probably go to C<data_structure> for |
265 |
normalisation. |
266 |
|
267 |
=cut |
268 |
|
269 |
sub fetch { |
270 |
my $self = shift; |
271 |
|
272 |
my $log = $self->_get_logger(); |
273 |
|
274 |
$log->logconfess("it seems that you didn't load database!") unless ($self->{pos}); |
275 |
|
276 |
if ($self->{pos} == -1) { |
277 |
$self->{pos} = $self->{offset}; |
278 |
} else { |
279 |
$self->{pos}++; |
280 |
} |
281 |
|
282 |
my $mfn = $self->{pos}; |
283 |
|
284 |
if ($mfn > $self->{max_pos}) { |
285 |
$self->{pos} = $self->{max_pos}; |
286 |
$log->debug("at EOF"); |
287 |
return; |
288 |
} |
289 |
|
290 |
$self->progress_bar($mfn,$self->{max_pos}); |
291 |
|
292 |
my $rec; |
293 |
|
294 |
if ($self->{low_mem}) { |
295 |
$rec = $self->{db}->get($mfn); |
296 |
} else { |
297 |
$rec = $self->{data}->{$mfn}; |
298 |
} |
299 |
|
300 |
$rec ||= 0E0; |
301 |
} |
302 |
|
303 |
=head2 pos |
304 |
|
305 |
Returns current record number (MFN). |
306 |
|
307 |
print $isis->pos; |
308 |
|
309 |
First record in database has position 1. |
310 |
|
311 |
=cut |
312 |
|
313 |
sub pos { |
314 |
my $self = shift; |
315 |
return $self->{pos}; |
316 |
} |
317 |
|
318 |
|
319 |
=head2 size |
320 |
|
321 |
Returns number of records in database |
322 |
|
323 |
print $isis->size; |
324 |
|
325 |
Result from this function can be used to loop through all records |
326 |
|
327 |
foreach my $mfn ( 1 ... $isis->size ) { ... } |
328 |
|
329 |
because it takes into account C<offset> and C<limit>. |
330 |
|
331 |
=cut |
332 |
|
333 |
sub size { |
334 |
my $self = shift; |
335 |
return $self->{size}; |
336 |
} |
337 |
|
338 |
=head2 seek |
339 |
|
340 |
Seek to specified MFN in file. |
341 |
|
342 |
$isis->seek(42); |
343 |
|
344 |
First record in database has position 1. |
345 |
|
346 |
=cut |
347 |
|
348 |
sub seek { |
349 |
my $self = shift; |
350 |
my $pos = shift || return; |
351 |
|
352 |
my $log = $self->_get_logger(); |
353 |
|
354 |
if ($pos < 1) { |
355 |
$log->warn("seek before first record"); |
356 |
$pos = 1; |
357 |
} elsif ($pos > $self->{max_pos}) { |
358 |
$log->warn("seek beyond last record"); |
359 |
$pos = $self->{max_pos}; |
360 |
} |
361 |
|
362 |
return $self->{pos} = (($pos - 1) || -1); |
363 |
} |
364 |
|
365 |
|
366 |
=head1 MEMORY USAGE |
367 |
|
368 |
C<low_mem> options is double-edged sword. If enabled, WebPAC |
369 |
will run on memory constraint machines (which doesn't have enough |
370 |
physical RAM to create memory structure for whole source database). |
371 |
|
372 |
If your machine has 512Mb or more of RAM and database is around 10000 records, |
373 |
memory shouldn't be an issue. If you don't have enough physical RAM, you |
374 |
might consider using virtual memory (if your operating system is handling it |
375 |
well, like on FreeBSD or Linux) instead of dropping to L<DBM::Deep> to handle |
376 |
parsed structure of ISIS database (this is what C<low_mem> option does). |
377 |
|
378 |
Hitting swap at end of reading source database is probably o.k. However, |
379 |
hitting swap before 90% will dramatically decrease performance and you will |
380 |
be better off with C<low_mem> and using rest of availble memory for |
381 |
operating system disk cache (Linux is particuallary good about this). |
382 |
However, every access to database record will require disk access, so |
383 |
generation phase will be slower 10-100 times. |
384 |
|
385 |
Parsed structures are essential - you just have option to trade RAM memory |
386 |
(which is fast) for disk space (which is slow). Be sure to have planty of |
387 |
disk space if you are using C<low_mem> and thus L<DBM::Deep>. |
388 |
|
389 |
However, when WebPAC is running on desktop machines (or laptops :-), it's |
390 |
highly undesireable for system to start swapping. Using C<low_mem> option can |
391 |
reduce WecPAC memory usage to around 64Mb for same database with lookup |
392 |
fields and sorted indexes which stay in RAM. Performance will suffer, but |
393 |
memory usage will really be minimal. It might be also more confortable to |
394 |
run WebPAC reniced on those machines. |
395 |
|
396 |
|
397 |
=head1 AUTHOR |
398 |
|
399 |
Dobrica Pavlinusic, C<< <dpavlin@rot13.org> >> |
400 |
|
401 |
=head1 COPYRIGHT & LICENSE |
402 |
|
403 |
Copyright 2005 Dobrica Pavlinusic, All Rights Reserved. |
404 |
|
405 |
This program is free software; you can redistribute it and/or modify it |
406 |
under the same terms as Perl itself. |
407 |
|
408 |
=cut |
409 |
|
410 |
1; # End of WebPAC::Input |