webpac2/trunk/run.pl

#!/usr/bin/perl -w

use strict;

use Cwd qw/abs_path/;
use File::Temp qw/tempdir/;
use lib './lib';

use WebPAC::Common 0.02;
use WebPAC::Lookup 0.03;
use WebPAC::Input 0.07;
use WebPAC::Store 0.03;
use WebPAC::Normalize 0.11;
use WebPAC::Output::TT;
use WebPAC::Validate;
use WebPAC::Output::MARC;
use YAML qw/LoadFile/;
use Getopt::Long;
use File::Path;
use Time::HiRes qw/time/;
use File::Slurp;
use Data::Dump qw/dump/;
use Storable qw/dclone/;

use Proc::Queue size => 1;
use POSIX ":sys_wait_h"; # imports WNOHANG

=head1 NAME

run.pl - start WebPAC indexing

B<this command will probably go away. Don't get used to it!>

Options:

=over 4

=item --offset 42

start loading (all) databases at offset 42

=item --limit 100

limit loading to 100 records

=item --clean

remove database and Hyper Estraier index before indexing

=item --only=database_name/input_filter

reindex just single database (legacy name is --one)

C</input_filter> is optional part which can be C<name>
or C<type> from input

=item --config conf/config.yml

path to YAML configuration file

=item --stats

disable indexing and dump statistics about field and subfield
usage for each input

=item --validate path/to/validation_file

turn on extra validation of imput records, see L<WebPAC::Validation>

=item --marc-normalize conf/normalize/mapping.pl

This option specifies normalisation file for MARC creation

=item --marc-output out/marc/test.marc

Optional path to output file

=item --marc-lint

By default turned on if C<--marc-normalize> is used. You can disable lint
messages with C<--no-marc-lint>.

=item --marc-dump

Force dump or input and marc record for debugging.

=item --parallel 4

Run databases in parallel (aproximatly same as number of processors in
machine if you want to use full load)

=back

=cut

my $offset;
my $limit;

my $clean = 0;
my $config = 'conf/config.yml';
my $debug = 0;
my $only_filter;
my $stats = 0;
my $validate_path;
my ($marc_normalize, $marc_output);
my $marc_lint = 1;
my $marc_dump = 0;

my $parallel = 0;

GetOptions(
        "limit=i" => \$limit,
        "offset=i" => \$offset,
        "clean" => \$clean,
        "one=s" => \$only_filter,
        "only=s" => \$only_filter,
        "config" => \$config,
        "debug+" => \$debug,
        "stats" => \$stats,
        "validate=s" => \$validate_path,
        "marc-normalize=s" => \$marc_normalize,
        "marc-output=s" => \$marc_output,
        "marc-lint!" => \$marc_lint,
        "marc-dump!" => \$marc_dump,
        "parallel=i" => \$parallel,
);

$config = LoadFile($config);

print "config = ",dump($config) if ($debug);

die "no databases in config file!\n" unless ($config->{databases});

my $log = _new WebPAC::Common()->_get_logger();
$log->info( "-" x 79 );

my $validate;
$validate = new WebPAC::Validate(
        path => $validate_path,
) if ($validate_path);

my $use_indexer = $config->{use_indexer} || 'hyperestraier';
if ($stats) {
        $log->debug("option --stats disables update of indexing engine...");
        $use_indexer = undef;
} else {
        $log->info("using $use_indexer indexing engine...");
}

# disable indexing when creating marc
$use_indexer = undef if ($marc_normalize);

my $total_rows = 0;
my $start_t = time();

my @links;
my $indexer;

if ($parallel) {
        $log->info("Using $parallel processes for speedup");
        Proc::Queue::size($parallel);
}

while (my ($database, $db_config) = each %{ $config->{databases} }) {

        my ($only_database,$only_input) = split(m#/#, $only_filter) if ($only_filter);
        next if ($only_database && $database !~ m/$only_database/i);

        if ($parallel) {
                my $f=fork;
                if(defined ($f) and $f==0) {
                        $log->info("Created processes $$ for speedup");
                } else {
                        next;
                }
        }

        if ($use_indexer) {
                my $indexer_config = $config->{$use_indexer} || $log->logdie("can't find '$use_indexer' part in confguration");
                $indexer_config->{database} = $database;
                $indexer_config->{clean} = $clean;
                $indexer_config->{label} = $db_config->{name};

                if ($use_indexer eq 'hyperestraier') {

                        # open Hyper Estraier database
                        use WebPAC::Output::Estraier '0.10';
                        $indexer = new WebPAC::Output::Estraier( %{ $indexer_config } );
                
                } elsif ($use_indexer eq 'kinosearch') {

                        # open KinoSearch
                        use WebPAC::Output::KinoSearch;
                        $indexer_config->{clean} = 1 unless (-e $indexer_config->{index_path});
                        $indexer = new WebPAC::Output::KinoSearch( %{ $indexer_config } );

                } else {
                        $log->logdie("unknown use_indexer: $use_indexer");
                }

                $log->logide("can't continue without valid indexer") unless ($indexer);
        }


        #
        # now WebPAC::Store
        #
        my $abs_path = abs_path($0);
        $abs_path =~ s#/[^/]*$#/#;

        my $db_path = $config->{webpac}->{db_path} . '/' . $database;

        if ($clean) {
                $log->info("creating new database '$database' in $db_path");
                rmtree( $db_path ) || $log->warn("can't remove $db_path: $!");
        } else {
                $log->info("working on database '$database' in $db_path");
        }

        my $db = new WebPAC::Store(
                path => $db_path,
                database => $database,
                debug => $debug,
        );


        #
        # now, iterate through input formats
        #

        my @inputs;
        if (ref($db_config->{input}) eq 'ARRAY') {
                @inputs = @{ $db_config->{input} };
        } elsif ($db_config->{input}) {
                push @inputs, $db_config->{input};
        } else {
                $log->info("database $database doesn't have inputs defined");
        }

        my @supported_inputs = keys %{ $config->{webpac}->{inputs} };

        foreach my $input (@inputs) {

                next if ($only_input && ($input->{name} !~ m#$only_input#i && $input->{type} !~ m#$only_input#i));

                my $type = lc($input->{type});

                die "I know only how to handle input types ", join(",", @supported_inputs), " not '$type'!\n" unless (grep(/$type/, @supported_inputs));

                my $lookup;
                if ($input->{lookup}) {
                        $lookup = new WebPAC::Lookup(
                                lookup_file => $input->{lookup},
                        );
                        delete( $input->{lookup} );
                }

                my $input_module = $config->{webpac}->{inputs}->{$type};

                $log->info("working on input '$input->{name}' in $input->{path} [type: $input->{type}] using $input_module",
                        $input->{lookup} ? "lookup '$input->{lookup}'" : ""
                );

                my $input_db = new WebPAC::Input(
                        module => $input_module,
                        encoding => $config->{webpac}->{webpac_encoding},
                        limit => $limit || $input->{limit},
                        offset => $offset,
                        lookup_coderef => sub {
                                my $rec = shift || return;
                                $lookup->add( $rec );
                        },
                        recode => $input->{recode},
                        stats => $stats,
                        modify_records => $input->{modify_records},
                );
                $log->logdie("can't create input using $input_module") unless ($input);

                my $maxmfn = $input_db->open(
                        path => $input->{path},
                        code_page => $input->{encoding},        # database encoding
                        %{ $input },
                );

                my @norm_array = ref($input->{normalize}) eq 'ARRAY' ?
                        @{ $input->{normalize} } : ( $input->{normalize} );

                if ($marc_normalize) {
                        @norm_array = ( {
                                path => $marc_normalize,
                                output => $marc_output || 'out/marc/' . $database . '-' . $input->{name} . '.marc',
                        } );
                }

                foreach my $normalize (@norm_array) {

                        my $normalize_path = $normalize->{path} || $log->logdie("can't find normalize path in config");

                        $log->logdie("Found '$normalize_path' as normalization file which isn't supported any more!") unless ( $normalize_path =~ m!\.pl$!i );

                        my $rules = read_file( $normalize_path ) or die "can't open $normalize_path: $!";

                        $log->info("Using $normalize_path for normalization...");

                        my $marc = new WebPAC::Output::MARC(
                                path => $normalize->{output},
                                lint => $marc_lint,
                                dump => $marc_dump,
                        ) if ($normalize->{output});

                        # reset position in database
                        $input_db->seek(1);

                        foreach my $pos ( 0 ... $input_db->size ) {

                                my $row = $input_db->fetch || next;

                                my $mfn = $row->{'000'}->[0];

                                if (! $mfn || $mfn !~ m#^\d+$#) {
                                        $log->warn("record $pos doesn't have valid MFN but '$mfn', using $pos");
                                        $mfn = $pos;
                                        push @{ $row->{'000'} }, $pos;
                                }


                                if ($validate) {
                                        my @errors = $validate->validate_errors( $row );
                                        $log->error( "MFN $mfn validation errors:\n", join("\n", @errors) ) if (@errors);
                                }

                                my $ds_config = dclone($db_config);

                                # default values -> database key
                                $ds_config->{_} = $database;

                                # current mfn
                                $ds_config->{_mfn} = $mfn;

                                # attach current input
                                $ds_config->{input} = $input;

                                my $ds = WebPAC::Normalize::data_structure(
                                        row => $row,
                                        rules => $rules,
                                        lookup => $lookup ? $lookup->lookup_hash : undef,
                                        config => $ds_config,
                                        marc_encoding => 'utf-8',
                                );

                                $db->save_ds(
                                        id => $mfn,
                                        ds => $ds,
                                        prefix => $input->{name},
                                ) if ($ds && !$stats);

                                $indexer->add(
                                        id => $input->{name} . "/" . $mfn,
                                        ds => $ds,
                                        type => $config->{$use_indexer}->{type},
                                ) if ($indexer && $ds);

                                if ($marc) {
                                        my $i = 0;

                                        while (my $fields = WebPAC::Normalize::_get_marc_fields( fetch_next => 1 ) ) {
                                                $marc->add(
                                                        id => $mfn . ( $i ? "/$i" : '' ),
                                                        fields => $fields,
                                                        leader => WebPAC::Normalize::marc_leader(),
                                                        row => $row,
                                                );
                                                $i++;
                                        }

                                        $log->info("Created $i instances of MFN $mfn\n") if ($i > 1);
                                }

                                $total_rows++;
                        }

                        $log->info("statistics of fields usage:\n", $input_db->stats) if ($stats);

                        # close MARC file
                        $marc->finish if ($marc);

                }

        }

        eval { $indexer->finish } if ($indexer && $indexer->can('finish'));

        my $dt = time() - $start_t;
        $log->info("$total_rows records ", $indexer ? "indexed " : "",
                sprintf("in %.2f sec [%.2f rec/sec]",
                        $dt, ($total_rows / $dt)
                )
        );

        #
        # add Hyper Estraier links to other databases
        #
        if (ref($db_config->{links}) eq 'ARRAY' && $use_indexer) {
                foreach my $link (@{ $db_config->{links} }) {
                        if ($use_indexer eq 'hyperestraier') {
                                $log->info("saving link $database -> $link->{to} [$link->{credit}]");
                                push @links, {
                                        from => $database,
                                        to => $link->{to},
                                        credit => $link->{credit},
                                };
                        } else {
                                $log->warn("NOT IMPLEMENTED WITH $use_indexer: adding link $database -> $link->{to} [$link->{credit}]");
                        }
                }
        }

        # end forked process
        if ($parallel) {
                $log->info("parallel process $$ finished");
                exit(0);
        }

}

if ($parallel) {
        # wait all children to finish
        sleep(1) while wait != -1;
        $log->info("all parallel processes finished");
}

foreach my $link (@links) {
        $log->info("adding link $link->{from} -> $link->{to} [$link->{credit}]");
        $indexer->add_link( %{ $link } );
}

1	dpavlin	74	#!/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	dpavlin	255	use WebPAC::Common 0.02;
10	dpavlin	588	use WebPAC::Lookup 0.03;
11	dpavlin	585	use WebPAC::Input 0.07;
12	dpavlin	209	use WebPAC::Store 0.03;
13	dpavlin	581	use WebPAC::Normalize 0.11;
14	dpavlin	74	use WebPAC::Output::TT;
15	dpavlin	516	use WebPAC::Validate;
16	dpavlin	578	use WebPAC::Output::MARC;
17	dpavlin	141	use YAML qw/LoadFile/;
18	dpavlin	301	use Getopt::Long;
19			use File::Path;
20	dpavlin	389	use Time::HiRes qw/time/;
21	dpavlin	492	use File::Slurp;
22	dpavlin	556	use Data::Dump qw/dump/;
23	dpavlin	595	use Storable qw/dclone/;
24	dpavlin	74
25	dpavlin	606	use Proc::Queue size => 1;
26			use POSIX ":sys_wait_h"; # imports WNOHANG
27
28	dpavlin	301	=head1 NAME
29	dpavlin	76
30	dpavlin	301	run.pl - start WebPAC indexing
31	dpavlin	141
32	dpavlin	301	B<this command will probably go away. Don't get used to it!>
33	dpavlin	141
34	dpavlin	301	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	dpavlin	510	=item --only=database_name/input_filter
51	dpavlin	335
52	dpavlin	423	reindex just single database (legacy name is --one)
53	dpavlin	335
54	dpavlin	510	C</input_filter> is optional part which can be C<name>
55			or C<type> from input
56
57	dpavlin	301	=item --config conf/config.yml
58
59			path to YAML configuration file
60
61	dpavlin	507	=item --stats
62
63	dpavlin	509	disable indexing and dump statistics about field and subfield
64			usage for each input
65	dpavlin	507
66	dpavlin	516	=item --validate path/to/validation_file
67
68			turn on extra validation of imput records, see L<WebPAC::Validation>
69
70	dpavlin	552	=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	dpavlin	556	=item --marc-lint
79
80			By default turned on if C<--marc-normalize> is used. You can disable lint
81	dpavlin	558	messages with C<--no-marc-lint>.
82	dpavlin	556
83	dpavlin	559	=item --marc-dump
84
85			Force dump or input and marc record for debugging.
86
87	dpavlin	606	=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	dpavlin	301	=back
93
94			=cut
95
96			my $offset;
97			my $limit;
98
99			my $clean = 0;
100			my $config = 'conf/config.yml';
101			my $debug = 0;
102	dpavlin	510	my $only_filter;
103	dpavlin	507	my $stats = 0;
104	dpavlin	516	my $validate_path;
105	dpavlin	552	my ($marc_normalize, $marc_output);
106	dpavlin	556	my $marc_lint = 1;
107	dpavlin	559	my $marc_dump = 0;
108	dpavlin	301
109	dpavlin	606	my $parallel = 0;
110
111	dpavlin	301	GetOptions(
112			"limit=i" => \$limit,
113			"offset=i" => \$offset,
114			"clean" => \$clean,
115	dpavlin	510	"one=s" => \$only_filter,
116			"only=s" => \$only_filter,
117	dpavlin	301	"config" => \$config,
118	dpavlin	560	"debug+" => \$debug,
119	dpavlin	507	"stats" => \$stats,
120	dpavlin	516	"validate=s" => \$validate_path,
121	dpavlin	552	"marc-normalize=s" => \$marc_normalize,
122			"marc-output=s" => \$marc_output,
123	dpavlin	556	"marc-lint!" => \$marc_lint,
124	dpavlin	559	"marc-dump!" => \$marc_dump,
125	dpavlin	606	"parallel=i" => \$parallel,
126	dpavlin	301	);
127
128			$config = LoadFile($config);
129
130	dpavlin	556	print "config = ",dump($config) if ($debug);
131	dpavlin	301
132	dpavlin	210	die "no databases in config file!\n" unless ($config->{databases});
133	dpavlin	431
134			my $log = _new WebPAC::Common()->_get_logger();
135	dpavlin	509	$log->info( "-" x 79 );
136	dpavlin	431
137	dpavlin	516	my $validate;
138			$validate = new WebPAC::Validate(
139			path => $validate_path,
140			) if ($validate_path);
141
142	dpavlin	430	my $use_indexer = $config->{use_indexer} \|\| 'hyperestraier';
143	dpavlin	509	if ($stats) {
144			$log->debug("option --stats disables update of indexing engine...");
145			$use_indexer = undef;
146			} else {
147			$log->info("using $use_indexer indexing engine...");
148			}
149	dpavlin	141
150	dpavlin	552	# disable indexing when creating marc
151			$use_indexer = undef if ($marc_normalize);
152
153	dpavlin	213	my $total_rows = 0;
154	dpavlin	389	my $start_t = time();
155	dpavlin	213
156	dpavlin	529	my @links;
157			my $indexer;
158
159	dpavlin	606	if ($parallel) {
160			$log->info("Using $parallel processes for speedup");
161			Proc::Queue::size($parallel);
162			}
163
164	dpavlin	210	while (my ($database, $db_config) = each %{ $config->{databases} }) {
165	dpavlin	141
166	dpavlin	512	my ($only_database,$only_input) = split(m#/#, $only_filter) if ($only_filter);
167	dpavlin	510	next if ($only_database && $database !~ m/$only_database/i);
168	dpavlin	335
169	dpavlin	606	if ($parallel) {
170			my $f=fork;
171			if(defined ($f) and $f==0) {
172			$log->info("Created processes $$ for speedup");
173			} else {
174			next;
175			}
176			}
177
178	dpavlin	509	if ($use_indexer) {
179			my $indexer_config = $config->{$use_indexer} \|\| $log->logdie("can't find '$use_indexer' part in confguration");
180			$indexer_config->{database} = $database;
181			$indexer_config->{clean} = $clean;
182			$indexer_config->{label} = $db_config->{name};
183	dpavlin	431
184	dpavlin	509	if ($use_indexer eq 'hyperestraier') {
185	dpavlin	255
186	dpavlin	509	# open Hyper Estraier database
187			use WebPAC::Output::Estraier '0.10';
188			$indexer = new WebPAC::Output::Estraier( %{ $indexer_config } );
189
190			} elsif ($use_indexer eq 'kinosearch') {
191	dpavlin	430
192	dpavlin	509	# open KinoSearch
193			use WebPAC::Output::KinoSearch;
194			$indexer_config->{clean} = 1 unless (-e $indexer_config->{index_path});
195			$indexer = new WebPAC::Output::KinoSearch( %{ $indexer_config } );
196	dpavlin	431
197	dpavlin	509	} else {
198			$log->logdie("unknown use_indexer: $use_indexer");
199			}
200
201			$log->logide("can't continue without valid indexer") unless ($indexer);
202	dpavlin	430	}
203
204
205	dpavlin	255	#
206			# now WebPAC::Store
207			#
208	dpavlin	210	my $abs_path = abs_path($0);
209			$abs_path =~ s#/[^/]*$#/#;
210	dpavlin	141
211	dpavlin	210	my $db_path = $config->{webpac}->{db_path} . '/' . $database;
212	dpavlin	74
213	dpavlin	301	if ($clean) {
214	dpavlin	511	$log->info("creating new database '$database' in $db_path");
215	dpavlin	301	rmtree( $db_path ) \|\| $log->warn("can't remove $db_path: $!");
216			} else {
217	dpavlin	511	$log->info("working on database '$database' in $db_path");
218	dpavlin	301	}
219	dpavlin	255
220	dpavlin	210	my $db = new WebPAC::Store(
221			path => $db_path,
222	dpavlin	217	database => $database,
223	dpavlin	301	debug => $debug,
224	dpavlin	210	);
225	dpavlin	74
226	dpavlin	233
227	dpavlin	213	#
228			# now, iterate through input formats
229			#
230	dpavlin	74
231	dpavlin	213	my @inputs;
232			if (ref($db_config->{input}) eq 'ARRAY') {
233			@inputs = @{ $db_config->{input} };
234	dpavlin	255	} elsif ($db_config->{input}) {
235			push @inputs, $db_config->{input};
236	dpavlin	213	} else {
237	dpavlin	255	$log->info("database $database doesn't have inputs defined");
238	dpavlin	213	}
239	dpavlin	74
240	dpavlin	286	my @supported_inputs = keys %{ $config->{webpac}->{inputs} };
241
242	dpavlin	213	foreach my $input (@inputs) {
243	dpavlin	233
244	dpavlin	512	next if ($only_input && ($input->{name} !~ m#$only_input#i && $input->{type} !~ m#$only_input#i));
245	dpavlin	510
246	dpavlin	233	my $type = lc($input->{type});
247
248	dpavlin	286	die "I know only how to handle input types ", join(",", @supported_inputs), " not '$type'!\n" unless (grep(/$type/, @supported_inputs));
249	dpavlin	233
250	dpavlin	588	my $lookup;
251			if ($input->{lookup}) {
252			$lookup = new WebPAC::Lookup(
253			lookup_file => $input->{lookup},
254			);
255			delete( $input->{lookup} );
256			}
257	dpavlin	233
258	dpavlin	286	my $input_module = $config->{webpac}->{inputs}->{$type};
259	dpavlin	74
260	dpavlin	523	$log->info("working on input '$input->{name}' in $input->{path} [type: $input->{type}] using $input_module",
261			$input->{lookup} ? "lookup '$input->{lookup}'" : ""
262			);
263	dpavlin	286
264	dpavlin	287	my $input_db = new WebPAC::Input(
265			module => $input_module,
266	dpavlin	585	encoding => $config->{webpac}->{webpac_encoding},
267	dpavlin	301	limit => $limit \|\| $input->{limit},
268			offset => $offset,
269	dpavlin	585	lookup_coderef => sub {
270			my $rec = shift \|\| return;
271			$lookup->add( $rec );
272			},
273	dpavlin	416	recode => $input->{recode},
274	dpavlin	507	stats => $stats,
275	dpavlin	600	modify_records => $input->{modify_records},
276	dpavlin	287	);
277	dpavlin	286	$log->logdie("can't create input using $input_module") unless ($input);
278	dpavlin	113
279	dpavlin	287	my $maxmfn = $input_db->open(
280	dpavlin	285	path => $input->{path},
281	dpavlin	213	code_page => $input->{encoding}, # database encoding
282	dpavlin	523	%{ $input },
283	dpavlin	213	);
284	dpavlin	113
285	dpavlin	539	my @norm_array = ref($input->{normalize}) eq 'ARRAY' ?
286			@{ $input->{normalize} } : ( $input->{normalize} );
287	dpavlin	210
288	dpavlin	552	if ($marc_normalize) {
289			@norm_array = ( {
290			path => $marc_normalize,
291			output => $marc_output \|\| 'out/marc/' . $database . '-' . $input->{name} . '.marc',
292			} );
293			}
294
295	dpavlin	539	foreach my $normalize (@norm_array) {
296	dpavlin	492
297	dpavlin	539	my $normalize_path = $normalize->{path} \|\| $log->logdie("can't find normalize path in config");
298	dpavlin	269
299	dpavlin	539	$log->logdie("Found '$normalize_path' as normalization file which isn't supported any more!") unless ( $normalize_path =~ m!\.pl$!i );
300	dpavlin	210
301	dpavlin	539	my $rules = read_file( $normalize_path ) or die "can't open $normalize_path: $!";
302	dpavlin	213
303	dpavlin	539	$log->info("Using $normalize_path for normalization...");
304	dpavlin	213
305	dpavlin	578	my $marc = new WebPAC::Output::MARC(
306			path => $normalize->{output},
307			lint => $marc_lint,
308			dump => $marc_dump,
309			) if ($normalize->{output});
310	dpavlin	540
311	dpavlin	539	# reset position in database
312			$input_db->seek(1);
313	dpavlin	290
314	dpavlin	539	foreach my $pos ( 0 ... $input_db->size ) {
315	dpavlin	516
316	dpavlin	539	my $row = $input_db->fetch \|\| next;
317	dpavlin	516
318	dpavlin	539	my $mfn = $row->{'000'}->[0];
319	dpavlin	213
320	dpavlin	539	if (! $mfn \|\| $mfn !~ m#^\d+$#) {
321			$log->warn("record $pos doesn't have valid MFN but '$mfn', using $pos");
322			$mfn = $pos;
323			push @{ $row->{'000'} }, $pos;
324			}
325	dpavlin	508
326	dpavlin	213
327	dpavlin	539	if ($validate) {
328			my @errors = $validate->validate_errors( $row );
329			$log->error( "MFN $mfn validation errors:\n", join("\n", @errors) ) if (@errors);
330			}
331	dpavlin	213
332	dpavlin	595	my $ds_config = dclone($db_config);
333
334			# default values -> database key
335			$ds_config->{_} = $database;
336
337			# current mfn
338			$ds_config->{_mfn} = $mfn;
339
340			# attach current input
341			$ds_config->{input} = $input;
342
343	dpavlin	539	my $ds = WebPAC::Normalize::data_structure(
344			row => $row,
345			rules => $rules,
346			lookup => $lookup ? $lookup->lookup_hash : undef,
347	dpavlin	595	config => $ds_config,
348	dpavlin	541	marc_encoding => 'utf-8',
349	dpavlin	539	);
350	dpavlin	507
351	dpavlin	539	$db->save_ds(
352			id => $mfn,
353			ds => $ds,
354			prefix => $input->{name},
355			) if ($ds && !$stats);
356	dpavlin	210
357	dpavlin	539	$indexer->add(
358			id => $input->{name} . "/" . $mfn,
359			ds => $ds,
360			type => $config->{$use_indexer}->{type},
361	dpavlin	540	) if ($indexer && $ds);
362	dpavlin	539
363	dpavlin	580	if ($marc) {
364			my $i = 0;
365	dpavlin	559
366	dpavlin	580	while (my $fields = WebPAC::Normalize::_get_marc_fields( fetch_next => 1 ) ) {
367			$marc->add(
368			id => $mfn . ( $i ? "/$i" : '' ),
369			fields => $fields,
370			leader => WebPAC::Normalize::marc_leader(),
371			row => $row,
372			);
373			$i++;
374			}
375
376			$log->info("Created $i instances of MFN $mfn\n") if ($i > 1);
377			}
378
379	dpavlin	539	$total_rows++;
380			}
381
382			$log->info("statistics of fields usage:\n", $input_db->stats) if ($stats);
383
384	dpavlin	540	# close MARC file
385	dpavlin	578	$marc->finish if ($marc);
386	dpavlin	539
387	dpavlin	540	}
388
389	dpavlin	539	}
390
391	dpavlin	509	eval { $indexer->finish } if ($indexer && $indexer->can('finish'));
392	dpavlin	434
393	dpavlin	389	my $dt = time() - $start_t;
394	dpavlin	511	$log->info("$total_rows records ", $indexer ? "indexed " : "",
395			sprintf("in %.2f sec [%.2f rec/sec]",
396	dpavlin	389	$dt, ($total_rows / $dt)
397			)
398			);
399	dpavlin	255
400	dpavlin	431	#
401			# add Hyper Estraier links to other databases
402			#
403	dpavlin	552	if (ref($db_config->{links}) eq 'ARRAY' && $use_indexer) {
404	dpavlin	431	foreach my $link (@{ $db_config->{links} }) {
405			if ($use_indexer eq 'hyperestraier') {
406	dpavlin	529	$log->info("saving link $database -> $link->{to} [$link->{credit}]");
407			push @links, {
408	dpavlin	430	from => $database,
409			to => $link->{to},
410			credit => $link->{credit},
411	dpavlin	529	};
412	dpavlin	431	} else {
413			$log->warn("NOT IMPLEMENTED WITH $use_indexer: adding link $database -> $link->{to} [$link->{credit}]");
414	dpavlin	430	}
415	dpavlin	255	}
416			}
417
418	dpavlin	606	# end forked process
419			if ($parallel) {
420			$log->info("parallel process $$ finished");
421			exit(0);
422			}
423
424	dpavlin	210	}
425
426	dpavlin	606	if ($parallel) {
427			# wait all children to finish
428			sleep(1) while wait != -1;
429			$log->info("all parallel processes finished");
430			}
431
432	dpavlin	529	foreach my $link (@links) {
433			$log->info("adding link $link->{from} -> $link->{to} [$link->{credit}]");
434			$indexer->add_link( %{ $link } );
435			}
436	dpavlin	606