| 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 0.06; |
| 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 |
} elsif ($use_indexer eq 'kinosearch') { |
| 236 |
|
| 237 |
# open KinoSearch |
| 238 |
use WebPAC::Output::KinoSearch; |
| 239 |
$indexer_config->{clean} = 1 unless (-e $indexer_config->{index_path}); |
| 240 |
$indexer = new WebPAC::Output::KinoSearch( %{ $indexer_config } ); |
| 241 |
|
| 242 |
} else { |
| 243 |
$log->logdie("unknown use_indexer: $use_indexer"); |
| 244 |
} |
| 245 |
|
| 246 |
$log->logide("can't continue without valid indexer") unless ($indexer); |
| 247 |
} |
| 248 |
|
| 249 |
|
| 250 |
# |
| 251 |
# store Hyper Estraier links to other databases |
| 252 |
# |
| 253 |
if (ref($db_config->{links}) eq 'ARRAY' && $use_indexer) { |
| 254 |
foreach my $link (@{ $db_config->{links} }) { |
| 255 |
if ($use_indexer eq 'hyperestraier') { |
| 256 |
if ($merge) { |
| 257 |
print $estcmd_fh 'sudo -u www-data estcmd merge ' . $database . ' ' . $link->{to},$/; |
| 258 |
} else { |
| 259 |
$log->info("saving link $database -> $link->{to} [$link->{credit}]"); |
| 260 |
push @links, sub { |
| 261 |
$log->info("adding link $database -> $link->{to} [$link->{credit}]"); |
| 262 |
$indexer->add_link( |
| 263 |
from => $database, |
| 264 |
to => $link->{to}, |
| 265 |
credit => $link->{credit}, |
| 266 |
); |
| 267 |
}; |
| 268 |
} |
| 269 |
} else { |
| 270 |
$log->warn("NOT IMPLEMENTED WITH $use_indexer: adding link $database -> $link->{to} [$link->{credit}]"); |
| 271 |
} |
| 272 |
} |
| 273 |
} |
| 274 |
next if ($only_links); |
| 275 |
|
| 276 |
|
| 277 |
# |
| 278 |
# now WebPAC::Store |
| 279 |
# |
| 280 |
my $abs_path = abs_path($0); |
| 281 |
$abs_path =~ s#/[^/]*$#/#; |
| 282 |
|
| 283 |
my $db_path = $config->{webpac}->{db_path} . '/' . $database; |
| 284 |
|
| 285 |
if ($clean) { |
| 286 |
$log->info("creating new database '$database' in $db_path"); |
| 287 |
rmtree( $db_path ) || $log->warn("can't remove $db_path: $!"); |
| 288 |
} else { |
| 289 |
$log->info("working on database '$database' in $db_path"); |
| 290 |
} |
| 291 |
|
| 292 |
my $db = new WebPAC::Store( |
| 293 |
path => $db_path, |
| 294 |
database => $database, |
| 295 |
debug => $debug, |
| 296 |
); |
| 297 |
|
| 298 |
|
| 299 |
# |
| 300 |
# now, iterate through input formats |
| 301 |
# |
| 302 |
|
| 303 |
my @inputs; |
| 304 |
if (ref($db_config->{input}) eq 'ARRAY') { |
| 305 |
@inputs = @{ $db_config->{input} }; |
| 306 |
} elsif ($db_config->{input}) { |
| 307 |
push @inputs, $db_config->{input}; |
| 308 |
} else { |
| 309 |
$log->info("database $database doesn't have inputs defined"); |
| 310 |
} |
| 311 |
|
| 312 |
my @supported_inputs = keys %{ $config->{webpac}->{inputs} }; |
| 313 |
|
| 314 |
foreach my $input (@inputs) { |
| 315 |
|
| 316 |
next if ($only_input && ($input->{name} !~ m#$only_input#i && $input->{type} !~ m#$only_input#i)); |
| 317 |
|
| 318 |
my $type = lc($input->{type}); |
| 319 |
|
| 320 |
die "I know only how to handle input types ", join(",", @supported_inputs), " not '$type'!\n" unless (grep(/$type/, @supported_inputs)); |
| 321 |
|
| 322 |
my $lookup; |
| 323 |
if ($input->{lookup}) { |
| 324 |
$lookup = new WebPAC::Lookup( |
| 325 |
lookup_file => $input->{lookup}, |
| 326 |
); |
| 327 |
delete( $input->{lookup} ); |
| 328 |
} |
| 329 |
|
| 330 |
my $input_module = $config->{webpac}->{inputs}->{$type}; |
| 331 |
|
| 332 |
$log->info("working on input '$input->{name}' in $input->{path} [type: $input->{type}] using $input_module", |
| 333 |
$input->{lookup} ? "lookup '$input->{lookup}'" : "" |
| 334 |
); |
| 335 |
|
| 336 |
if ($stats) { |
| 337 |
# disable modification of records if --stats is in use |
| 338 |
delete($input->{modify_records}); |
| 339 |
delete($input->{modify_file}); |
| 340 |
} |
| 341 |
|
| 342 |
my $input_db = new WebPAC::Input( |
| 343 |
module => $input_module, |
| 344 |
encoding => $config->{webpac}->{webpac_encoding}, |
| 345 |
limit => $limit || $input->{limit}, |
| 346 |
offset => $offset, |
| 347 |
lookup_coderef => sub { |
| 348 |
my $rec = shift || return; |
| 349 |
$lookup->add( $rec ); |
| 350 |
}, |
| 351 |
recode => $input->{recode}, |
| 352 |
stats => $stats, |
| 353 |
modify_records => $input->{modify_records}, |
| 354 |
modify_file => $input->{modify_file}, |
| 355 |
); |
| 356 |
$log->logdie("can't create input using $input_module") unless ($input); |
| 357 |
|
| 358 |
my $maxmfn = $input_db->open( |
| 359 |
path => $input->{path}, |
| 360 |
code_page => $input->{encoding}, # database encoding |
| 361 |
%{ $input }, |
| 362 |
); |
| 363 |
|
| 364 |
my $report_fh; |
| 365 |
if ($stats || $validate) { |
| 366 |
my $path = "out/report/" . $database . '-' . $input->{name} . '.txt'; |
| 367 |
open($report_fh, '>', $path) || $log->logdie("can't open $path: $!"); |
| 368 |
|
| 369 |
print $report_fh "Report for database '$database' input '$input->{name}' records ", |
| 370 |
$offset || 1, "-", $limit || $input->{limit} || $maxmfn, "\n\n"; |
| 371 |
$log->info("Generating report file $path"); |
| 372 |
} |
| 373 |
|
| 374 |
my @norm_array = ref($input->{normalize}) eq 'ARRAY' ? |
| 375 |
@{ $input->{normalize} } : ( $input->{normalize} ); |
| 376 |
|
| 377 |
if ($marc_normalize) { |
| 378 |
@norm_array = ( { |
| 379 |
path => $marc_normalize, |
| 380 |
output => $marc_output || 'out/marc/' . $database . '-' . $input->{name} . '.marc', |
| 381 |
} ); |
| 382 |
} |
| 383 |
|
| 384 |
foreach my $normalize (@norm_array) { |
| 385 |
|
| 386 |
my $normalize_path = $normalize->{path} || $log->logdie("can't find normalize path in config"); |
| 387 |
|
| 388 |
$log->logdie("Found '$normalize_path' as normalization file which isn't supported any more!") unless ( $normalize_path =~ m!\.pl$!i ); |
| 389 |
|
| 390 |
my $rules = read_file( $normalize_path ) or die "can't open $normalize_path: $!"; |
| 391 |
|
| 392 |
$log->info("Using $normalize_path for normalization..."); |
| 393 |
|
| 394 |
my $marc = new WebPAC::Output::MARC( |
| 395 |
path => $normalize->{output}, |
| 396 |
lint => $marc_lint, |
| 397 |
dump => $marc_dump, |
| 398 |
) if ($normalize->{output}); |
| 399 |
|
| 400 |
# reset position in database |
| 401 |
$input_db->seek(1); |
| 402 |
|
| 403 |
# generate name of config key for indexer (strip everything after -) |
| 404 |
my $indexer_config = $use_indexer; |
| 405 |
$indexer_config =~ s/^(\w+)-?.*$/$1/g if ($indexer_config); |
| 406 |
|
| 407 |
foreach my $pos ( 0 ... $input_db->size ) { |
| 408 |
|
| 409 |
my $row = $input_db->fetch || next; |
| 410 |
|
| 411 |
my $mfn = $row->{'000'}->[0]; |
| 412 |
|
| 413 |
if (! $mfn || $mfn !~ m#^\d+$#) { |
| 414 |
$log->warn("record $pos doesn't have valid MFN but '$mfn', using $pos"); |
| 415 |
$mfn = $pos; |
| 416 |
push @{ $row->{'000'} }, $pos; |
| 417 |
} |
| 418 |
|
| 419 |
|
| 420 |
if ($validate) { |
| 421 |
if ( my $errors = $validate->validate_errors( $row, $input_db->dump ) ) { |
| 422 |
$log->error( "MFN $mfn validation error:\n", |
| 423 |
$validate->report_error( $errors ) |
| 424 |
); |
| 425 |
} |
| 426 |
} |
| 427 |
|
| 428 |
my $ds_config = dclone($db_config); |
| 429 |
|
| 430 |
# default values -> database key |
| 431 |
$ds_config->{_} = $database; |
| 432 |
|
| 433 |
# current mfn |
| 434 |
$ds_config->{_mfn} = $mfn; |
| 435 |
|
| 436 |
# attach current input |
| 437 |
$ds_config->{input} = $input; |
| 438 |
|
| 439 |
my $ds = WebPAC::Normalize::data_structure( |
| 440 |
row => $row, |
| 441 |
rules => $rules, |
| 442 |
lookup => $lookup ? $lookup->lookup_hash : undef, |
| 443 |
config => $ds_config, |
| 444 |
marc_encoding => 'utf-8', |
| 445 |
); |
| 446 |
|
| 447 |
$db->save_ds( |
| 448 |
id => $mfn, |
| 449 |
ds => $ds, |
| 450 |
prefix => $input->{name}, |
| 451 |
) if ($ds && !$stats); |
| 452 |
|
| 453 |
$indexer->add( |
| 454 |
id => $input->{name} . "/" . $mfn, |
| 455 |
ds => $ds, |
| 456 |
type => $config->{$indexer_config}->{type}, |
| 457 |
) if ($indexer && $ds); |
| 458 |
|
| 459 |
if ($marc) { |
| 460 |
my $i = 0; |
| 461 |
|
| 462 |
while (my $fields = WebPAC::Normalize::_get_marc_fields( fetch_next => 1 ) ) { |
| 463 |
$marc->add( |
| 464 |
id => $mfn . ( $i ? "/$i" : '' ), |
| 465 |
fields => $fields, |
| 466 |
leader => WebPAC::Normalize::marc_leader(), |
| 467 |
row => $row, |
| 468 |
); |
| 469 |
$i++; |
| 470 |
} |
| 471 |
|
| 472 |
$log->info("Created $i instances of MFN $mfn\n") if ($i > 1); |
| 473 |
} |
| 474 |
|
| 475 |
$total_rows++; |
| 476 |
} |
| 477 |
|
| 478 |
if ($validate) { |
| 479 |
my $errors = $validate->report; |
| 480 |
if ($errors) { |
| 481 |
$log->info("validation errors:\n$errors\n" ); |
| 482 |
print $report_fh "$errors\n" if ($report_fh); |
| 483 |
} |
| 484 |
} |
| 485 |
|
| 486 |
if ($stats) { |
| 487 |
my $s = $input_db->stats; |
| 488 |
$log->info("statistics of fields usage:\n$s"); |
| 489 |
print $report_fh "Statistics of fields usage:\n$s" if ($report_fh); |
| 490 |
} |
| 491 |
|
| 492 |
# close MARC file |
| 493 |
$marc->finish if ($marc); |
| 494 |
|
| 495 |
# close report |
| 496 |
close($report_fh) if ($report_fh) |
| 497 |
} |
| 498 |
|
| 499 |
} |
| 500 |
|
| 501 |
eval { $indexer->finish } if ($indexer && $indexer->can('finish')); |
| 502 |
|
| 503 |
my $dt = time() - $start_t; |
| 504 |
$log->info("$total_rows records ", $indexer ? "indexed " : "", |
| 505 |
sprintf("in %.2f sec [%.2f rec/sec]", |
| 506 |
$dt, ($total_rows / $dt) |
| 507 |
) |
| 508 |
); |
| 509 |
|
| 510 |
|
| 511 |
# end forked process |
| 512 |
if ($parallel) { |
| 513 |
$log->info("parallel process $$ finished"); |
| 514 |
exit(0); |
| 515 |
} |
| 516 |
|
| 517 |
} |
| 518 |
|
| 519 |
if ($parallel) { |
| 520 |
# wait all children to finish |
| 521 |
sleep(1) while wait != -1; |
| 522 |
$log->info("all parallel processes finished"); |
| 523 |
} |
| 524 |
|
| 525 |
# |
| 526 |
# handle links or merge after indexing |
| 527 |
# |
| 528 |
|
| 529 |
if ($merge) { |
| 530 |
print $estcmd_fh 'sudo /etc/init.d/hyperestraier start',$/; |
| 531 |
close($estcmd_fh); |
| 532 |
chmod 0700, $estcmd_path || $log->warn("can't chmod 0700 $estcmd_path: $!"); |
| 533 |
system $estcmd_path; |
| 534 |
} else { |
| 535 |
foreach my $link (@links) { |
| 536 |
$log->logdie("coderef in link ", Dumper($link), " is ", ref($link), " and not CODE") unless (ref($link) eq 'CODE'); |
| 537 |
$link->(); |
| 538 |
} |
| 539 |
} |
Parent Directory
| 
Revision Log