1 |
#!/usr/bin/perl -w |
2 |
|
3 |
use strict; |
4 |
|
5 |
use Cwd qw/abs_path/; |
6 |
use File::Temp qw/tempdir/; |
7 |
use lib './lib'; |
8 |
|
9 |
use WebPAC::Common 0.02; |
10 |
use WebPAC::Lookup 0.03; |
11 |
use WebPAC::Input 0.11; |
12 |
use WebPAC::Store 0.03; |
13 |
use WebPAC::Normalize 0.11; |
14 |
use WebPAC::Output::TT; |
15 |
use WebPAC::Validate; |
16 |
use WebPAC::Output::MARC; |
17 |
use YAML qw/LoadFile/; |
18 |
use Getopt::Long; |
19 |
use File::Path; |
20 |
use Time::HiRes qw/time/; |
21 |
use File::Slurp; |
22 |
use Data::Dump qw/dump/; |
23 |
use Storable qw/dclone/; |
24 |
|
25 |
use Proc::Queue size => 1; |
26 |
use POSIX ":sys_wait_h"; # imports WNOHANG |
27 |
|
28 |
=head1 NAME |
29 |
|
30 |
run.pl - start WebPAC indexing |
31 |
|
32 |
B<this command will probably go away. Don't get used to it!> |
33 |
|
34 |
Options: |
35 |
|
36 |
=over 4 |
37 |
|
38 |
=item --offset 42 |
39 |
|
40 |
start loading (all) databases at offset 42 |
41 |
|
42 |
=item --limit 100 |
43 |
|
44 |
limit loading to 100 records |
45 |
|
46 |
=item --clean |
47 |
|
48 |
remove database and Hyper Estraier index before indexing |
49 |
|
50 |
=item --only=database_name/input_filter |
51 |
|
52 |
reindex just single database (legacy name is --one) |
53 |
|
54 |
C</input_filter> is optional part which can be C<name> |
55 |
or C<type> from input |
56 |
|
57 |
=item --config conf/config.yml |
58 |
|
59 |
path to YAML configuration file |
60 |
|
61 |
=item --stats |
62 |
|
63 |
disable indexing, modify_* in configuration and dump statistics about field |
64 |
and subfield usage for each input |
65 |
|
66 |
=item --validate path/to/validation_file |
67 |
|
68 |
turn on extra validation of imput records, see L<WebPAC::Validation> |
69 |
|
70 |
=item --marc-normalize conf/normalize/mapping.pl |
71 |
|
72 |
This option specifies normalisation file for MARC creation |
73 |
|
74 |
=item --marc-output out/marc/test.marc |
75 |
|
76 |
Optional path to output file |
77 |
|
78 |
=item --marc-lint |
79 |
|
80 |
By default turned on if C<--marc-normalize> is used. You can disable lint |
81 |
messages with C<--no-marc-lint>. |
82 |
|
83 |
=item --marc-dump |
84 |
|
85 |
Force dump or input and marc record for debugging. |
86 |
|
87 |
=item --parallel 4 |
88 |
|
89 |
Run databases in parallel (aproximatly same as number of processors in |
90 |
machine if you want to use full load) |
91 |
|
92 |
=item --only-links |
93 |
|
94 |
Create just links |
95 |
|
96 |
=item --merge |
97 |
|
98 |
Create merged index of databases which have links |
99 |
|
100 |
=back |
101 |
|
102 |
=cut |
103 |
|
104 |
my $offset; |
105 |
my $limit; |
106 |
|
107 |
my $clean = 0; |
108 |
my $config = 'conf/config.yml'; |
109 |
my $debug = 0; |
110 |
my $only_filter; |
111 |
my $stats = 0; |
112 |
my $validate_path; |
113 |
my ($marc_normalize, $marc_output); |
114 |
my $marc_lint = 1; |
115 |
my $marc_dump = 0; |
116 |
my $parallel = 0; |
117 |
my $only_links = 0; |
118 |
my $merge = 0; |
119 |
|
120 |
my $log = _new WebPAC::Common()->_get_logger(); |
121 |
|
122 |
my $hostname = `hostname`; |
123 |
chomp($hostname); |
124 |
$hostname =~ s/\..+$//; |
125 |
if (-e "conf/$hostname.yml") { |
126 |
$config = "conf/$hostname.yml"; |
127 |
$log->info("using host configuration file: $config"); |
128 |
} |
129 |
|
130 |
GetOptions( |
131 |
"limit=i" => \$limit, |
132 |
"offset=i" => \$offset, |
133 |
"clean" => \$clean, |
134 |
"one=s" => \$only_filter, |
135 |
"only=s" => \$only_filter, |
136 |
"config" => \$config, |
137 |
"debug+" => \$debug, |
138 |
"stats" => \$stats, |
139 |
"validate=s" => \$validate_path, |
140 |
"marc-normalize=s" => \$marc_normalize, |
141 |
"marc-output=s" => \$marc_output, |
142 |
"marc-lint!" => \$marc_lint, |
143 |
"marc-dump!" => \$marc_dump, |
144 |
"parallel=i" => \$parallel, |
145 |
"only-links!" => \$only_links, |
146 |
"merge" => \$merge, |
147 |
); |
148 |
|
149 |
$config = LoadFile($config); |
150 |
|
151 |
#print "config = ",dump($config) if ($debug); |
152 |
|
153 |
die "no databases in config file!\n" unless ($config->{databases}); |
154 |
|
155 |
$log->info( "-" x 79 ); |
156 |
|
157 |
|
158 |
my $estcmd_fh; |
159 |
my $estcmd_path = './estcmd-merge.sh'; |
160 |
if ($merge) { |
161 |
open($estcmd_fh, '>', $estcmd_path) || $log->logdie("can't open $estcmd_path: $!"); |
162 |
print $estcmd_fh 'cd /data/estraier/_node/ || exit 1',$/; |
163 |
print $estcmd_fh 'sudo /etc/init.d/hyperestraier stop',$/; |
164 |
$log->info("created merge batch file $estcmd_path"); |
165 |
} |
166 |
|
167 |
|
168 |
my $validate; |
169 |
$validate = new WebPAC::Validate( |
170 |
path => $validate_path, |
171 |
) if ($validate_path); |
172 |
|
173 |
|
174 |
my $use_indexer = $config->{use_indexer} || 'hyperestraier'; |
175 |
if ($stats) { |
176 |
$log->debug("option --stats disables update of indexing engine..."); |
177 |
$use_indexer = undef; |
178 |
} else { |
179 |
$log->info("using $use_indexer indexing engine..."); |
180 |
} |
181 |
|
182 |
# disable indexing when creating marc |
183 |
$use_indexer = undef if ($marc_normalize); |
184 |
|
185 |
my $total_rows = 0; |
186 |
my $start_t = time(); |
187 |
|
188 |
my @links; |
189 |
|
190 |
if ($parallel) { |
191 |
$log->info("Using $parallel processes for speedup"); |
192 |
Proc::Queue::size($parallel); |
193 |
} |
194 |
|
195 |
while (my ($database, $db_config) = each %{ $config->{databases} }) { |
196 |
|
197 |
my ($only_database,$only_input) = split(m#/#, $only_filter) if ($only_filter); |
198 |
next if ($only_database && $database !~ m/$only_database/i); |
199 |
|
200 |
if ($parallel) { |
201 |
my $f=fork; |
202 |
if(defined ($f) and $f==0) { |
203 |
$log->info("Created processes $$ for speedup"); |
204 |
} else { |
205 |
next; |
206 |
} |
207 |
} |
208 |
|
209 |
my $indexer; |
210 |
if ($use_indexer) { |
211 |
|
212 |
my $cfg_name = $use_indexer; |
213 |
$cfg_name =~ s/\-.*$//; |
214 |
|
215 |
my $indexer_config = $config->{$cfg_name} || $log->logdie("can't find '$cfg_name' part in confguration"); |
216 |
$indexer_config->{database} = $database; |
217 |
$indexer_config->{clean} = $clean; |
218 |
$indexer_config->{label} = $db_config->{name}; |
219 |
|
220 |
# force clean if database has links |
221 |
$indexer_config->{clean} = 1 if ($db_config->{links}); |
222 |
|
223 |
if ($use_indexer eq 'hyperestraier') { |
224 |
|
225 |
# open Hyper Estraier database |
226 |
use WebPAC::Output::Estraier '0.10'; |
227 |
$indexer = new WebPAC::Output::Estraier( %{ $indexer_config } ); |
228 |
|
229 |
} elsif ($use_indexer eq 'hyperestraier-native') { |
230 |
|
231 |
# open Hyper Estraier database |
232 |
use WebPAC::Output::EstraierNative; |
233 |
$indexer = new WebPAC::Output::EstraierNative( %{ $indexer_config } ); |
234 |
|
235 |
$use_indexer = 'hyperestraier'; |
236 |
|
237 |
} elsif ($use_indexer eq 'kinosearch') { |
238 |
|
239 |
# open KinoSearch |
240 |
use WebPAC::Output::KinoSearch; |
241 |
$indexer_config->{clean} = 1 unless (-e $indexer_config->{index_path}); |
242 |
$indexer = new WebPAC::Output::KinoSearch( %{ $indexer_config } ); |
243 |
|
244 |
} else { |
245 |
$log->logdie("unknown use_indexer: $use_indexer"); |
246 |
} |
247 |
|
248 |
$log->logide("can't continue without valid indexer") unless ($indexer); |
249 |
} |
250 |
|
251 |
|
252 |
# |
253 |
# store Hyper Estraier links to other databases |
254 |
# |
255 |
if (ref($db_config->{links}) eq 'ARRAY' && $use_indexer) { |
256 |
foreach my $link (@{ $db_config->{links} }) { |
257 |
if ($use_indexer eq 'hyperestraier') { |
258 |
if ($merge) { |
259 |
print $estcmd_fh 'sudo -u www-data estcmd merge ' . $database . ' ' . $link->{to},$/; |
260 |
} else { |
261 |
$log->info("saving link $database -> $link->{to} [$link->{credit}]"); |
262 |
push @links, sub { |
263 |
$log->info("adding link $database -> $link->{to} [$link->{credit}]"); |
264 |
$indexer->add_link( |
265 |
from => $database, |
266 |
to => $link->{to}, |
267 |
credit => $link->{credit}, |
268 |
); |
269 |
}; |
270 |
} |
271 |
} else { |
272 |
$log->warn("NOT IMPLEMENTED WITH $use_indexer: adding link $database -> $link->{to} [$link->{credit}]"); |
273 |
} |
274 |
} |
275 |
} |
276 |
next if ($only_links); |
277 |
|
278 |
|
279 |
# |
280 |
# now WebPAC::Store |
281 |
# |
282 |
my $abs_path = abs_path($0); |
283 |
$abs_path =~ s#/[^/]*$#/#; |
284 |
|
285 |
my $db_path = $config->{webpac}->{db_path} . '/' . $database; |
286 |
|
287 |
if ($clean) { |
288 |
$log->info("creating new database '$database' in $db_path"); |
289 |
rmtree( $db_path ) || $log->warn("can't remove $db_path: $!"); |
290 |
} else { |
291 |
$log->info("working on database '$database' in $db_path"); |
292 |
} |
293 |
|
294 |
my $db = new WebPAC::Store( |
295 |
path => $db_path, |
296 |
database => $database, |
297 |
debug => $debug, |
298 |
); |
299 |
|
300 |
|
301 |
# |
302 |
# now, iterate through input formats |
303 |
# |
304 |
|
305 |
my @inputs; |
306 |
if (ref($db_config->{input}) eq 'ARRAY') { |
307 |
@inputs = @{ $db_config->{input} }; |
308 |
} elsif ($db_config->{input}) { |
309 |
push @inputs, $db_config->{input}; |
310 |
} else { |
311 |
$log->info("database $database doesn't have inputs defined"); |
312 |
} |
313 |
|
314 |
my @supported_inputs = keys %{ $config->{webpac}->{inputs} }; |
315 |
|
316 |
foreach my $input (@inputs) { |
317 |
|
318 |
next if ($only_input && ($input->{name} !~ m#$only_input#i && $input->{type} !~ m#$only_input#i)); |
319 |
|
320 |
my $type = lc($input->{type}); |
321 |
|
322 |
die "I know only how to handle input types ", join(",", @supported_inputs), " not '$type'!\n" unless (grep(/$type/, @supported_inputs)); |
323 |
|
324 |
my $lookup; |
325 |
if ($input->{lookup}) { |
326 |
$lookup = new WebPAC::Lookup( |
327 |
lookup_file => $input->{lookup}, |
328 |
); |
329 |
delete( $input->{lookup} ); |
330 |
} |
331 |
|
332 |
my $input_module = $config->{webpac}->{inputs}->{$type}; |
333 |
|
334 |
$log->info("working on input '$input->{name}' in $input->{path} [type: $input->{type}] using $input_module", |
335 |
$input->{lookup} ? "lookup '$input->{lookup}'" : "" |
336 |
); |
337 |
|
338 |
if ($stats) { |
339 |
# disable modification of records if --stats is in use |
340 |
delete($input->{modify_records}); |
341 |
delete($input->{modify_file}); |
342 |
} |
343 |
|
344 |
my $input_db = new WebPAC::Input( |
345 |
module => $input_module, |
346 |
encoding => $config->{webpac}->{webpac_encoding}, |
347 |
limit => $limit || $input->{limit}, |
348 |
offset => $offset, |
349 |
lookup_coderef => sub { |
350 |
my $rec = shift || return; |
351 |
$lookup->add( $rec ); |
352 |
}, |
353 |
recode => $input->{recode}, |
354 |
stats => $stats, |
355 |
modify_records => $input->{modify_records}, |
356 |
modify_file => $input->{modify_file}, |
357 |
); |
358 |
$log->logdie("can't create input using $input_module") unless ($input); |
359 |
|
360 |
my $maxmfn = $input_db->open( |
361 |
path => $input->{path}, |
362 |
code_page => $input->{encoding}, # database encoding |
363 |
%{ $input }, |
364 |
); |
365 |
|
366 |
my @norm_array = ref($input->{normalize}) eq 'ARRAY' ? |
367 |
@{ $input->{normalize} } : ( $input->{normalize} ); |
368 |
|
369 |
if ($marc_normalize) { |
370 |
@norm_array = ( { |
371 |
path => $marc_normalize, |
372 |
output => $marc_output || 'out/marc/' . $database . '-' . $input->{name} . '.marc', |
373 |
} ); |
374 |
} |
375 |
|
376 |
foreach my $normalize (@norm_array) { |
377 |
|
378 |
my $normalize_path = $normalize->{path} || $log->logdie("can't find normalize path in config"); |
379 |
|
380 |
$log->logdie("Found '$normalize_path' as normalization file which isn't supported any more!") unless ( $normalize_path =~ m!\.pl$!i ); |
381 |
|
382 |
my $rules = read_file( $normalize_path ) or die "can't open $normalize_path: $!"; |
383 |
|
384 |
$log->info("Using $normalize_path for normalization..."); |
385 |
|
386 |
my $marc = new WebPAC::Output::MARC( |
387 |
path => $normalize->{output}, |
388 |
lint => $marc_lint, |
389 |
dump => $marc_dump, |
390 |
) if ($normalize->{output}); |
391 |
|
392 |
# reset position in database |
393 |
$input_db->seek(1); |
394 |
|
395 |
foreach my $pos ( 0 ... $input_db->size ) { |
396 |
|
397 |
my $row = $input_db->fetch || next; |
398 |
|
399 |
my $mfn = $row->{'000'}->[0]; |
400 |
|
401 |
if (! $mfn || $mfn !~ m#^\d+$#) { |
402 |
$log->warn("record $pos doesn't have valid MFN but '$mfn', using $pos"); |
403 |
$mfn = $pos; |
404 |
push @{ $row->{'000'} }, $pos; |
405 |
} |
406 |
|
407 |
|
408 |
if ($validate) { |
409 |
my @errors = $validate->validate_errors( $row ); |
410 |
$log->error( "MFN $mfn validation errors:\n", join("\n", @errors) ) if (@errors); |
411 |
} |
412 |
|
413 |
my $ds_config = dclone($db_config); |
414 |
|
415 |
# default values -> database key |
416 |
$ds_config->{_} = $database; |
417 |
|
418 |
# current mfn |
419 |
$ds_config->{_mfn} = $mfn; |
420 |
|
421 |
# attach current input |
422 |
$ds_config->{input} = $input; |
423 |
|
424 |
my $ds = WebPAC::Normalize::data_structure( |
425 |
row => $row, |
426 |
rules => $rules, |
427 |
lookup => $lookup ? $lookup->lookup_hash : undef, |
428 |
config => $ds_config, |
429 |
marc_encoding => 'utf-8', |
430 |
); |
431 |
|
432 |
$db->save_ds( |
433 |
id => $mfn, |
434 |
ds => $ds, |
435 |
prefix => $input->{name}, |
436 |
) if ($ds && !$stats); |
437 |
|
438 |
$indexer->add( |
439 |
id => $input->{name} . "/" . $mfn, |
440 |
ds => $ds, |
441 |
type => $config->{$use_indexer}->{type}, |
442 |
) if ($indexer && $ds); |
443 |
|
444 |
if ($marc) { |
445 |
my $i = 0; |
446 |
|
447 |
while (my $fields = WebPAC::Normalize::_get_marc_fields( fetch_next => 1 ) ) { |
448 |
$marc->add( |
449 |
id => $mfn . ( $i ? "/$i" : '' ), |
450 |
fields => $fields, |
451 |
leader => WebPAC::Normalize::marc_leader(), |
452 |
row => $row, |
453 |
); |
454 |
$i++; |
455 |
} |
456 |
|
457 |
$log->info("Created $i instances of MFN $mfn\n") if ($i > 1); |
458 |
} |
459 |
|
460 |
$total_rows++; |
461 |
} |
462 |
|
463 |
$log->info("statistics of fields usage:\n", $input_db->stats) if ($stats); |
464 |
|
465 |
# close MARC file |
466 |
$marc->finish if ($marc); |
467 |
|
468 |
} |
469 |
|
470 |
} |
471 |
|
472 |
eval { $indexer->finish } if ($indexer && $indexer->can('finish')); |
473 |
|
474 |
my $dt = time() - $start_t; |
475 |
$log->info("$total_rows records ", $indexer ? "indexed " : "", |
476 |
sprintf("in %.2f sec [%.2f rec/sec]", |
477 |
$dt, ($total_rows / $dt) |
478 |
) |
479 |
); |
480 |
|
481 |
|
482 |
# end forked process |
483 |
if ($parallel) { |
484 |
$log->info("parallel process $$ finished"); |
485 |
exit(0); |
486 |
} |
487 |
|
488 |
} |
489 |
|
490 |
if ($parallel) { |
491 |
# wait all children to finish |
492 |
sleep(1) while wait != -1; |
493 |
$log->info("all parallel processes finished"); |
494 |
} |
495 |
|
496 |
# |
497 |
# handle links or merge after indexing |
498 |
# |
499 |
|
500 |
if ($merge) { |
501 |
print $estcmd_fh 'sudo /etc/init.d/hyperestraier start',$/; |
502 |
close($estcmd_fh); |
503 |
chmod 0700, $estcmd_path || $log->warn("can't chmod 0700 $estcmd_path: $!"); |
504 |
system $estcmd_path; |
505 |
} else { |
506 |
foreach my $link (@links) { |
507 |
$log->logdie("coderef in link ", Dumper($link), " is ", ref($link), " and not CODE") unless (ref($link) eq 'CODE'); |
508 |
$link->(); |
509 |
} |
510 |
} |