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::Parser 0.08;
use WebPAC::Input 0.16;
use WebPAC::Store 0.14;
use WebPAC::Normalize 0.22;
use WebPAC::Output::TT;
use WebPAC::Validate 0.11;
use WebPAC::Output::MARC;
use WebPAC::Config;
use Getopt::Long;
use File::Path;
use Time::HiRes qw/time/;
use File::Slurp;
use Data::Dump qw/dump/;
use Storable qw/dclone/;
use Pod::Usage qw/pod2usage/;

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!>

=head1 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, modify_* in configuration 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 --validate-delimiters path/to/validate_delimiters_file

this option is used with C<--validate> to turn on extra validation of
delimiters. If file is non existant, it will be created on first run.

=item --marc-generate

Generate MARC file. This will automatically be on if file contains C<marc*> directives.
You can use this option as C<--no-marc-generate> to disable MARC generation.

=item --marc-lint

By default turned on if normalisation file has C<marc*> directives. 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)

=item --only-links

Create just links

=item --merge

Create merged index of databases which have links

=back

=cut

my $offset;
my $limit;

my $clean = 0;
my $config_path;
my $debug = 0;
my $only_filter;
my $stats = 0;
my $validate_path;
my $validate_delimiters_path;
my $marc_generate = 1;
my $marc_lint = 1;
my $marc_dump = 0;
my $parallel = 0;
my $only_links = 0;
my $merge = 0;
my $help;

my $log = _new WebPAC::Common()->_get_logger();

GetOptions(
        "limit=i" => \$limit,
        "offset=i" => \$offset,
        "clean" => \$clean,
        "one=s" => \$only_filter,
        "only=s" => \$only_filter,
        "config" => \$config_path,
        "debug+" => \$debug,
        "stats" => \$stats,
        "validate=s" => \$validate_path,
        "validate-delimiters=s" => \$validate_delimiters_path,
        "marc-generate!" => \$marc_generate,
        "marc-lint!" => \$marc_lint,
        "marc-dump!" => \$marc_dump,
        "parallel=i" => \$parallel,
        "only-links!" => \$only_links,
        "merge" => \$merge,
        "help" => \$help,
);

pod2usage(-verbose => 2) if ($help);

my $config = new WebPAC::Config( path => $config_path );

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

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

$log->info( "-" x 79 );

my $log_file = 'log';

if (-e $log_file ) {    # && -s $log_file > 5 * 1024 * 1024) {
        $log->info("moved old log with ", -s $log_file, " bytes to '${log_file}.old'");
        rename $log_file, "${log_file}.old" || $log->logwarn("can't rename $log_file to ${log_file}.old: $!");
}

my $estcmd_fh;
my $estcmd_path = './estcmd-merge.sh';
if ($merge) {
        open($estcmd_fh, '>', $estcmd_path) || $log->logdie("can't open $estcmd_path: $!");
        print $estcmd_fh 'cd /data/estraier/_node/ || exit 1',$/;
        print $estcmd_fh 'sudo /etc/init.d/hyperestraier stop',$/;
        $log->info("created merge batch file $estcmd_path");
}

my $validate;
$validate = new WebPAC::Validate(
        path => $validate_path,
        delimiters => $config->webpac('delimiters'),
        delimiters_path => $validate_delimiters_path,
) if ($validate_path || $validate_delimiters_path);

my $use_indexer = $config->use_indexer;
$stats ||= $validate;
if ($stats) {
        $log->debug("disabled indexing for stats collection");
        $use_indexer = undef;
} else {
        $log->info("using $use_indexer indexing engine...");
}

# parse normalize files and create source files for lookup and normalization

my $parser = new WebPAC::Parser( config => $config );

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

my @links;

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

sub create_ds_config {
        my ($db_config, $database, $input, $mfn) = @_;
        my $c = dclone( $db_config );
        $c->{_} = $database || $log->logconfess("need database");
        $c->{_mfn} = $mfn || $log->logconfess("need mfn");
        $c->{input} = $input || $log->logconfess("need input");
        return $c;
}

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;
                }
        }

        my $indexer;
        if ($use_indexer && $parser->have_rules( 'search', $database )) {

                my $cfg_name = $use_indexer;
                $cfg_name =~ s/\-.*$//;

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

                # force clean if database has links
                $indexer_config->{clean} = 1 if ($db_config->{links});

                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 'hyperestraier-native') {

                        # open Hyper Estraier database
                        use WebPAC::Output::EstraierNative;
                        $indexer = new WebPAC::Output::EstraierNative( %{ $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);
        }


        #
        # store 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') {
                                if ($merge) {
                                        print $estcmd_fh 'sudo -u www-data estcmd merge ' . $database . ' ' . $link->{to},$/;
                                } else {
                                        $log->info("saving link $database -> $link->{to} [$link->{credit}]");
                                        push @links, sub {
                                                $log->info("adding link $database -> $link->{to} [$link->{credit}]");
                                                $indexer->add_link(
                                                        from => $database,
                                                        to => $link->{to},
                                                        credit => $link->{credit},
                                                );
                                        };
                                }
                        } else {
                                $log->warn("NOT IMPLEMENTED WITH $use_indexer: adding link $database -> $link->{to} [$link->{credit}]");
                        }
                }
        }
        next if ($only_links);


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

        my $db_path = $config->webpac('db_path');

        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 $store = new WebPAC::Store(
                path => $db_path,
                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");
        }

        foreach my $input (@inputs) {

                my $input_name = $input->{name} || $log->logdie("input without a name isn't valid: ",dump($input));

                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(",", $config->webpac('inputs') ), " not '$type'!\n" unless (grep(/$type/, $config->webpac('inputs')));

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

                my @lookups = $parser->have_lookup_create($database, $input);

                $log->info("working on input '$input_name' in $input->{path} [type: $input->{type}] using $input_module",
                        @lookups ? " creating lookups: ".join(", ", @lookups) : ""
                );

                if ($stats) {
                        # disable modification of records if --stats is in use
                        delete($input->{modify_records});
                        delete($input->{modify_file});
                }

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

                if (defined( $input->{lookup} )) {
                        $log->warn("$database/$input_name has depriciated lookup definition, removing it...");
                        delete( $input->{lookup} );
                }

                my $lookup_coderef;

                if (@lookups) {

                        my $rules = $parser->lookup_create_rules($database, $input) || $log->logdie("no rules found for $database/$input");

                        $lookup_coderef = sub {
                                my $rec = shift || die "need rec!";
                                my $mfn = $rec->{'000'}->[0] || die "need mfn in 000";

                                WebPAC::Normalize::data_structure(
                                        row => $rec,
                                        rules => $rules,
                                        config => create_ds_config( $db_config, $database, $input, $mfn ),
                                );

                                #warn "current lookup: ", dump(WebPAC::Normalize::_get_lookup());
                        };

                        WebPAC::Normalize::_set_lookup( undef );

                        $log->debug("created lookup_coderef using:\n$rules");

                };

                my $lookup_jar;

                my $maxmfn = $input_db->open(
                        path => $input->{path},
                        code_page => $input->{encoding},        # database encoding
                        lookup_coderef => $lookup_coderef,
                        lookup => $lookup_jar,
                        %{ $input },
                        load_row => sub {
                                my $a = shift;
                                return $store->load_row(
                                        database => $database,
                                        input => $input_name,
                                        id => $a->{id},
                                );
                        },
                        save_row => sub {
                                my $a = shift;
                                return $store->save_row(
                                        database => $database,
                                        input => $input_name,
                                        id => $a->{id},
                                        row => $a->{row},
                                );
                        },

                );

                my $lookup_data = WebPAC::Normalize::_get_lookup();

                if (defined( $lookup_data->{$database}->{$input_name} )) {
                        $log->debug("created following lookups: ", sub { dump( $lookup_data ) } );

                        foreach my $key (keys %{ $lookup_data->{$database}->{$input_name} }) {
                                $store->save_lookup(
                                        database => $database,
                                        input => $input_name,
                                        key => $key,
                                        data => $lookup_data->{$database}->{$input_name}->{$key},
                                );
                        }
                }

                my $report_fh;
                if ($stats || $validate) {
                        my $path = "out/report/${database}-${input_name}.txt";
                        open($report_fh, '>', $path) || $log->logdie("can't open $path: $!");

                        print $report_fh "Report for database '$database' input '$input_name' records ",
                                $offset || 1, "-", $limit || $input->{limit} || $maxmfn, "\n\n";
                        $log->info("Generating report file $path");
                }

                my $marc;
                if ($marc_generate && $parser->have_rules( 'marc', $database, $input_name )) {
                        $marc = new WebPAC::Output::MARC(
                                path => "out/marc/${database}-${input_name}.marc",
                                lint => $marc_lint,
                                dump => $marc_dump,
                        );
                }

                my $rules = $parser->normalize_rules($database,$input_name) || $log->logdie("no normalize rules found for $database/$input_name");
                $log->debug("parsed normalize rules:\n$rules");

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

                # generate name of config key for indexer (strip everything after -)
                my $indexer_config = $use_indexer;
                $indexer_config =~ s/^(\w+)-?.*$/$1/g if ($indexer_config);

                my $lookup_hash;
                my $depends = $parser->depends($database,$input_name);
        
                if ($depends) {
                        $log->debug("$database/$input_name depends on: ", dump($depends)) if ($depends);
                        $log->logdie("parser->depends didn't return HASH") unless (ref($depends) eq 'HASH');

                        foreach my $db (keys %$depends) {
                                foreach my $i (keys %{$depends->{$db}}) {
                                        foreach my $k (keys %{$depends->{$db}->{$i}}) {
                                                my $t = time();
                                                $log->debug("loading lookup $db/$i");
                                                $lookup_hash->{$db}->{$i}->{$k} = $store->load_lookup(
                                                        database => $db,
                                                        input => $i,
                                                        key => $k,
                                                );
                                                $log->debug(sprintf("lookup $db/$i took %.2fs", time() - $t));
                                        }
                                }
                        }

                        $log->debug("lookup_hash = ", sub { dump( $lookup_hash ) });
                }


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

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

                        $total_rows++;

                        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) {
                                if ( my $errors = $validate->validate_rec( $row, $input_db->dump_ascii ) ) {
                                        $log->error( "MFN $mfn validation error:\n",
                                                $validate->report_error( $errors )
                                        );
                                }
                                next;   # validation doesn't create any output
                        }

                        my $ds = WebPAC::Normalize::data_structure(
                                row => $row,
                                rules => $rules,
                                lookup => $lookup_hash,
                                config => create_ds_config( $db_config, $database, $input, $mfn ),
                                marc_encoding => 'utf-8',
                                load_row_coderef => sub {
                                        my ($database,$input,$mfn) = @_;
                                        return $store->load_row(
                                                database => $database,
                                                input => $input,
                                                id => $mfn,
                                        );
                                },
                        );

                        $log->debug("ds = ", sub { dump($ds) }) if ($ds);

                        $store->save_ds(
                                database => $database,
                                input => $input_name,
                                id => $mfn,
                                ds => $ds,
                        ) if ($ds && !$stats);

                        $indexer->add(
                                id => "${input_name}/${mfn}",
                                ds => $ds,
                                type => $config->get($indexer_config)->{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::_get_marc_leader(),
                                                row => $row,
                                        );
                                        $i++;
                                }

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

                if ($validate) {
                        my $errors = $validate->report;
                        if ($errors) {
                                $log->info("validation errors:\n$errors\n" );
                                print $report_fh "$errors\n" if ($report_fh);
                        }

                        print $report_fh "\nAll possible subfields/delimiter templates:\n", $validate->delimiters_templates( report => 1 ), "\n\n";
                }

                if ($stats) {
                        my $s = $input_db->stats;
                        $log->info("statistics of fields usage:\n$s");
                        print $report_fh "Statistics of fields usage:\n$s" if ($report_fh);
                }

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

                # close report
                close($report_fh) if ($report_fh)

                # reset validate stats
                $validate->reset if ( $validate );
        }

        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)
                )
        );


        # 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");
}

# save new delimiters if needed
$validate->save_delimiters_templates if ( $validate );

#
# handle links or merge after indexing
#

if ($merge) {
        print $estcmd_fh 'sudo /etc/init.d/hyperestraier start',$/;
        close($estcmd_fh);
        chmod 0700, $estcmd_path || $log->warn("can't chmod 0700 $estcmd_path: $!");
        system $estcmd_path;
} else {
        foreach my $link (@links) {
                $log->logdie("coderef in link ", Dumper($link), " is ", ref($link), " and not CODE") unless (ref($link) eq 'CODE');
                $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	755	use WebPAC::Parser 0.08;
11	dpavlin	774	use WebPAC::Input 0.16;
12	dpavlin	762	use WebPAC::Store 0.14;
13	dpavlin	736	use WebPAC::Normalize 0.22;
14	dpavlin	74	use WebPAC::Output::TT;
15	dpavlin	833	use WebPAC::Validate 0.11;
16	dpavlin	578	use WebPAC::Output::MARC;
17	dpavlin	684	use WebPAC::Config;
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	790	use Pod::Usage qw/pod2usage/;
25	dpavlin	74
26	dpavlin	606	use Proc::Queue size => 1;
27			use POSIX ":sys_wait_h"; # imports WNOHANG
28
29	dpavlin	301	=head1 NAME
30	dpavlin	76
31	dpavlin	301	run.pl - start WebPAC indexing
32	dpavlin	141
33	dpavlin	301	B<this command will probably go away. Don't get used to it!>
34	dpavlin	141
35	dpavlin	790	=head1 OPTIONS
36	dpavlin	301
37			=over 4
38
39			=item --offset 42
40
41			start loading (all) databases at offset 42
42
43			=item --limit 100
44
45			limit loading to 100 records
46
47			=item --clean
48
49			remove database and Hyper Estraier index before indexing
50
51	dpavlin	510	=item --only=database_name/input_filter
52	dpavlin	335
53	dpavlin	423	reindex just single database (legacy name is --one)
54	dpavlin	335
55	dpavlin	510	C</input_filter> is optional part which can be C<name>
56			or C<type> from input
57
58	dpavlin	301	=item --config conf/config.yml
59
60			path to YAML configuration file
61
62	dpavlin	507	=item --stats
63
64	dpavlin	638	disable indexing, modify_* in configuration and dump statistics about field
65			and subfield usage for each input
66	dpavlin	507
67	dpavlin	516	=item --validate path/to/validation_file
68
69			turn on extra validation of imput records, see L<WebPAC::Validation>
70
71	dpavlin	837	=item --validate-delimiters path/to/validate_delimiters_file
72
73			this option is used with C<--validate> to turn on extra validation of
74			delimiters. If file is non existant, it will be created on first run.
75
76	dpavlin	803	=item --marc-generate
77
78			Generate MARC file. This will automatically be on if file contains C<marc*> directives.
79			You can use this option as C<--no-marc-generate> to disable MARC generation.
80
81	dpavlin	556	=item --marc-lint
82
83	dpavlin	738	By default turned on if normalisation file has C<marc*> directives. You can disable lint
84	dpavlin	558	messages with C<--no-marc-lint>.
85	dpavlin	556
86	dpavlin	559	=item --marc-dump
87
88			Force dump or input and marc record for debugging.
89
90	dpavlin	606	=item --parallel 4
91
92			Run databases in parallel (aproximatly same as number of processors in
93			machine if you want to use full load)
94
95	dpavlin	607	=item --only-links
96
97			Create just links
98
99	dpavlin	608	=item --merge
100
101			Create merged index of databases which have links
102
103	dpavlin	301	=back
104
105			=cut
106
107			my $offset;
108			my $limit;
109
110			my $clean = 0;
111	dpavlin	684	my $config_path;
112	dpavlin	301	my $debug = 0;
113	dpavlin	510	my $only_filter;
114	dpavlin	507	my $stats = 0;
115	dpavlin	516	my $validate_path;
116	dpavlin	837	my $validate_delimiters_path;
117	dpavlin	803	my $marc_generate = 1;
118	dpavlin	556	my $marc_lint = 1;
119	dpavlin	559	my $marc_dump = 0;
120	dpavlin	606	my $parallel = 0;
121	dpavlin	607	my $only_links = 0;
122	dpavlin	608	my $merge = 0;
123	dpavlin	790	my $help;
124	dpavlin	606
125	dpavlin	611	my $log = _new WebPAC::Common()->_get_logger();
126
127	dpavlin	301	GetOptions(
128			"limit=i" => \$limit,
129			"offset=i" => \$offset,
130			"clean" => \$clean,
131	dpavlin	510	"one=s" => \$only_filter,
132			"only=s" => \$only_filter,
133	dpavlin	684	"config" => \$config_path,
134	dpavlin	560	"debug+" => \$debug,
135	dpavlin	507	"stats" => \$stats,
136	dpavlin	516	"validate=s" => \$validate_path,
137	dpavlin	837	"validate-delimiters=s" => \$validate_delimiters_path,
138	dpavlin	803	"marc-generate!" => \$marc_generate,
139	dpavlin	556	"marc-lint!" => \$marc_lint,
140	dpavlin	559	"marc-dump!" => \$marc_dump,
141	dpavlin	606	"parallel=i" => \$parallel,
142	dpavlin	607	"only-links!" => \$only_links,
143	dpavlin	608	"merge" => \$merge,
144	dpavlin	790	"help" => \$help,
145	dpavlin	301	);
146
147	dpavlin	790	pod2usage(-verbose => 2) if ($help);
148
149	dpavlin	684	my $config = new WebPAC::Config( path => $config_path );
150	dpavlin	301
151	dpavlin	611	#print "config = ",dump($config) if ($debug);
152	dpavlin	301
153	dpavlin	684	die "no databases in config file!\n" unless ($config->databases);
154	dpavlin	431
155	dpavlin	509	$log->info( "-" x 79 );
156	dpavlin	431
157	dpavlin	775	my $log_file = 'log';
158	dpavlin	608
159	dpavlin	775	if (-e $log_file ) { # && -s $log_file > 5 * 1024 * 1024) {
160			$log->info("moved old log with ", -s $log_file, " bytes to '${log_file}.old'");
161			rename $log_file, "${log_file}.old" \|\| $log->logwarn("can't rename $log_file to ${log_file}.old: $!");
162			}
163
164	dpavlin	608	my $estcmd_fh;
165			my $estcmd_path = './estcmd-merge.sh';
166			if ($merge) {
167			open($estcmd_fh, '>', $estcmd_path) \|\| $log->logdie("can't open $estcmd_path: $!");
168			print $estcmd_fh 'cd /data/estraier/_node/ \|\| exit 1',$/;
169			print $estcmd_fh 'sudo /etc/init.d/hyperestraier stop',$/;
170			$log->info("created merge batch file $estcmd_path");
171			}
172
173	dpavlin	516	my $validate;
174			$validate = new WebPAC::Validate(
175			path => $validate_path,
176	dpavlin	833	delimiters => $config->webpac('delimiters'),
177	dpavlin	837	delimiters_path => $validate_delimiters_path,
178			) if ($validate_path \|\| $validate_delimiters_path);
179	dpavlin	516
180	dpavlin	684	my $use_indexer = $config->use_indexer;
181	dpavlin	769	$stats \|\|= $validate;
182	dpavlin	509	if ($stats) {
183	dpavlin	769	$log->debug("disabled indexing for stats collection");
184	dpavlin	509	$use_indexer = undef;
185			} else {
186			$log->info("using $use_indexer indexing engine...");
187			}
188	dpavlin	141
189	dpavlin	698	# parse normalize files and create source files for lookup and normalization
190
191			my $parser = new WebPAC::Parser( config => $config );
192
193	dpavlin	213	my $total_rows = 0;
194	dpavlin	389	my $start_t = time();
195	dpavlin	213
196	dpavlin	529	my @links;
197
198	dpavlin	606	if ($parallel) {
199			$log->info("Using $parallel processes for speedup");
200			Proc::Queue::size($parallel);
201			}
202
203	dpavlin	707	sub create_ds_config {
204			my ($db_config, $database, $input, $mfn) = @_;
205			my $c = dclone( $db_config );
206			$c->{_} = $database \|\| $log->logconfess("need database");
207			$c->{_mfn} = $mfn \|\| $log->logconfess("need mfn");
208			$c->{input} = $input \|\| $log->logconfess("need input");
209			return $c;
210			}
211
212	dpavlin	684	while (my ($database, $db_config) = each %{ $config->databases }) {
213	dpavlin	141
214	dpavlin	512	my ($only_database,$only_input) = split(m#/#, $only_filter) if ($only_filter);
215	dpavlin	510	next if ($only_database && $database !~ m/$only_database/i);
216	dpavlin	335
217	dpavlin	606	if ($parallel) {
218			my $f=fork;
219			if(defined ($f) and $f==0) {
220			$log->info("Created processes $$ for speedup");
221			} else {
222			next;
223			}
224			}
225
226	dpavlin	607	my $indexer;
227	dpavlin	755	if ($use_indexer && $parser->have_rules( 'search', $database )) {
228	dpavlin	627
229			my $cfg_name = $use_indexer;
230			$cfg_name =~ s/\-.*$//;
231
232	dpavlin	684	my $indexer_config = $config->get( $cfg_name ) \|\| $log->logdie("can't find '$cfg_name' part in confguration");
233	dpavlin	509	$indexer_config->{database} = $database;
234			$indexer_config->{clean} = $clean;
235			$indexer_config->{label} = $db_config->{name};
236	dpavlin	431
237	dpavlin	608	# force clean if database has links
238			$indexer_config->{clean} = 1 if ($db_config->{links});
239
240	dpavlin	509	if ($use_indexer eq 'hyperestraier') {
241	dpavlin	255
242	dpavlin	509	# open Hyper Estraier database
243			use WebPAC::Output::Estraier '0.10';
244			$indexer = new WebPAC::Output::Estraier( %{ $indexer_config } );
245
246	dpavlin	627	} elsif ($use_indexer eq 'hyperestraier-native') {
247
248			# open Hyper Estraier database
249			use WebPAC::Output::EstraierNative;
250			$indexer = new WebPAC::Output::EstraierNative( %{ $indexer_config } );
251
252	dpavlin	509	} elsif ($use_indexer eq 'kinosearch') {
253	dpavlin	430
254	dpavlin	509	# open KinoSearch
255			use WebPAC::Output::KinoSearch;
256			$indexer_config->{clean} = 1 unless (-e $indexer_config->{index_path});
257			$indexer = new WebPAC::Output::KinoSearch( %{ $indexer_config } );
258	dpavlin	431
259	dpavlin	509	} else {
260			$log->logdie("unknown use_indexer: $use_indexer");
261			}
262
263			$log->logide("can't continue without valid indexer") unless ($indexer);
264	dpavlin	430	}
265
266
267	dpavlin	255	#
268	dpavlin	607	# store Hyper Estraier links to other databases
269			#
270			if (ref($db_config->{links}) eq 'ARRAY' && $use_indexer) {
271			foreach my $link (@{ $db_config->{links} }) {
272			if ($use_indexer eq 'hyperestraier') {
273	dpavlin	608	if ($merge) {
274			print $estcmd_fh 'sudo -u www-data estcmd merge ' . $database . ' ' . $link->{to},$/;
275			} else {
276			$log->info("saving link $database -> $link->{to} [$link->{credit}]");
277			push @links, sub {
278			$log->info("adding link $database -> $link->{to} [$link->{credit}]");
279			$indexer->add_link(
280			from => $database,
281			to => $link->{to},
282			credit => $link->{credit},
283			);
284			};
285			}
286	dpavlin	607	} else {
287			$log->warn("NOT IMPLEMENTED WITH $use_indexer: adding link $database -> $link->{to} [$link->{credit}]");
288			}
289			}
290			}
291			next if ($only_links);
292
293
294			#
295	dpavlin	255	# now WebPAC::Store
296			#
297	dpavlin	210	my $abs_path = abs_path($0);
298			$abs_path =~ s#/[^/]*$#/#;
299	dpavlin	141
300	dpavlin	713	my $db_path = $config->webpac('db_path');
301	dpavlin	74
302	dpavlin	301	if ($clean) {
303	dpavlin	511	$log->info("creating new database '$database' in $db_path");
304	dpavlin	301	rmtree( $db_path ) \|\| $log->warn("can't remove $db_path: $!");
305			} else {
306	dpavlin	511	$log->info("working on database '$database' in $db_path");
307	dpavlin	301	}
308	dpavlin	255
309	dpavlin	710	my $store = new WebPAC::Store(
310	dpavlin	210	path => $db_path,
311	dpavlin	301	debug => $debug,
312	dpavlin	210	);
313	dpavlin	74
314	dpavlin	233
315	dpavlin	213	#
316			# now, iterate through input formats
317			#
318	dpavlin	74
319	dpavlin	213	my @inputs;
320			if (ref($db_config->{input}) eq 'ARRAY') {
321			@inputs = @{ $db_config->{input} };
322	dpavlin	255	} elsif ($db_config->{input}) {
323			push @inputs, $db_config->{input};
324	dpavlin	213	} else {
325	dpavlin	255	$log->info("database $database doesn't have inputs defined");
326	dpavlin	213	}
327	dpavlin	74
328	dpavlin	213	foreach my $input (@inputs) {
329	dpavlin	233
330	dpavlin	710	my $input_name = $input->{name} \|\| $log->logdie("input without a name isn't valid: ",dump($input));
331	dpavlin	510
332	dpavlin	710	next if ($only_input && ($input_name !~ m#$only_input#i && $input->{type} !~ m#$only_input#i));
333
334	dpavlin	233	my $type = lc($input->{type});
335
336	dpavlin	684	die "I know only how to handle input types ", join(",", $config->webpac('inputs') ), " not '$type'!\n" unless (grep(/$type/, $config->webpac('inputs')));
337	dpavlin	233
338	dpavlin	684	my $input_module = $config->webpac('inputs')->{$type};
339	dpavlin	74
340	dpavlin	706	my @lookups = $parser->have_lookup_create($database, $input);
341
342	dpavlin	710	$log->info("working on input '$input_name' in $input->{path} [type: $input->{type}] using $input_module",
343	dpavlin	707	@lookups ? " creating lookups: ".join(", ", @lookups) : ""
344	dpavlin	523	);
345	dpavlin	286
346	dpavlin	638	if ($stats) {
347			# disable modification of records if --stats is in use
348			delete($input->{modify_records});
349			delete($input->{modify_file});
350			}
351
352	dpavlin	287	my $input_db = new WebPAC::Input(
353			module => $input_module,
354	dpavlin	684	encoding => $config->webpac('webpac_encoding'),
355	dpavlin	301	limit => $limit \|\| $input->{limit},
356			offset => $offset,
357	dpavlin	416	recode => $input->{recode},
358	dpavlin	507	stats => $stats,
359	dpavlin	600	modify_records => $input->{modify_records},
360	dpavlin	636	modify_file => $input->{modify_file},
361	dpavlin	287	);
362	dpavlin	286	$log->logdie("can't create input using $input_module") unless ($input);
363	dpavlin	113
364	dpavlin	702	if (defined( $input->{lookup} )) {
365	dpavlin	710	$log->warn("$database/$input_name has depriciated lookup definition, removing it...");
366	dpavlin	702	delete( $input->{lookup} );
367			}
368
369	dpavlin	707	my $lookup_coderef;
370
371			if (@lookups) {
372
373			my $rules = $parser->lookup_create_rules($database, $input) \|\| $log->logdie("no rules found for $database/$input");
374
375			$lookup_coderef = sub {
376			my $rec = shift \|\| die "need rec!";
377			my $mfn = $rec->{'000'}->[0] \|\| die "need mfn in 000";
378
379			WebPAC::Normalize::data_structure(
380			row => $rec,
381	dpavlin	732	rules => $rules,
382	dpavlin	707	config => create_ds_config( $db_config, $database, $input, $mfn ),
383			);
384
385	dpavlin	732	#warn "current lookup: ", dump(WebPAC::Normalize::_get_lookup());
386	dpavlin	707	};
387
388			WebPAC::Normalize::_set_lookup( undef );
389
390			$log->debug("created lookup_coderef using:\n$rules");
391
392			};
393
394	dpavlin	731	my $lookup_jar;
395
396	dpavlin	287	my $maxmfn = $input_db->open(
397	dpavlin	285	path => $input->{path},
398	dpavlin	213	code_page => $input->{encoding}, # database encoding
399	dpavlin	707	lookup_coderef => $lookup_coderef,
400	dpavlin	731	lookup => $lookup_jar,
401	dpavlin	523	%{ $input },
402	dpavlin	762	load_row => sub {
403			my $a = shift;
404			return $store->load_row(
405			database => $database,
406			input => $input_name,
407			id => $a->{id},
408			);
409			},
410			save_row => sub {
411			my $a = shift;
412			return $store->save_row(
413			database => $database,
414			input => $input_name,
415			id => $a->{id},
416			row => $a->{row},
417			);
418			},
419
420	dpavlin	213	);
421	dpavlin	113
422	dpavlin	710	my $lookup_data = WebPAC::Normalize::_get_lookup();
423	dpavlin	707
424	dpavlin	712	if (defined( $lookup_data->{$database}->{$input_name} )) {
425	dpavlin	765	$log->debug("created following lookups: ", sub { dump( $lookup_data ) } );
426	dpavlin	710
427	dpavlin	712	foreach my $key (keys %{ $lookup_data->{$database}->{$input_name} }) {
428	dpavlin	713	$store->save_lookup(
429			database => $database,
430			input => $input_name,
431			key => $key,
432			data => $lookup_data->{$database}->{$input_name}->{$key},
433			);
434	dpavlin	712	}
435	dpavlin	710	}
436
437	dpavlin	671	my $report_fh;
438			if ($stats \|\| $validate) {
439	dpavlin	710	my $path = "out/report/${database}-${input_name}.txt";
440	dpavlin	671	open($report_fh, '>', $path) \|\| $log->logdie("can't open $path: $!");
441
442	dpavlin	710	print $report_fh "Report for database '$database' input '$input_name' records ",
443	dpavlin	671	$offset \|\| 1, "-", $limit \|\| $input->{limit} \|\| $maxmfn, "\n\n";
444			$log->info("Generating report file $path");
445			}
446
447	dpavlin	731	my $marc;
448	dpavlin	803	if ($marc_generate && $parser->have_rules( 'marc', $database, $input_name )) {
449	dpavlin	731	$marc = new WebPAC::Output::MARC(
450	dpavlin	738	path => "out/marc/${database}-${input_name}.marc",
451	dpavlin	731	lint => $marc_lint,
452			dump => $marc_dump,
453			);
454	dpavlin	552	}
455
456	dpavlin	731	my $rules = $parser->normalize_rules($database,$input_name) \|\| $log->logdie("no normalize rules found for $database/$input_name");
457			$log->debug("parsed normalize rules:\n$rules");
458	dpavlin	492
459	dpavlin	731	# reset position in database
460			$input_db->seek(1);
461	dpavlin	269
462	dpavlin	731	# generate name of config key for indexer (strip everything after -)
463			my $indexer_config = $use_indexer;
464			$indexer_config =~ s/^(\w+)-?.*$/$1/g if ($indexer_config);
465	dpavlin	210
466	dpavlin	731	my $lookup_hash;
467			my $depends = $parser->depends($database,$input_name);
468
469			if ($depends) {
470			$log->debug("$database/$input_name depends on: ", dump($depends)) if ($depends);
471			$log->logdie("parser->depends didn't return HASH") unless (ref($depends) eq 'HASH');
472	dpavlin	213
473	dpavlin	731	foreach my $db (keys %$depends) {
474			foreach my $i (keys %{$depends->{$db}}) {
475			foreach my $k (keys %{$depends->{$db}->{$i}}) {
476	dpavlin	762	my $t = time();
477	dpavlin	731	$log->debug("loading lookup $db/$i");
478			$lookup_hash->{$db}->{$i}->{$k} = $store->load_lookup(
479			database => $db,
480			input => $i,
481			key => $k,
482			);
483	dpavlin	762	$log->debug(sprintf("lookup $db/$i took %.2fs", time() - $t));
484	dpavlin	731	}
485			}
486			}
487	dpavlin	213
488	dpavlin	762	$log->debug("lookup_hash = ", sub { dump( $lookup_hash ) });
489	dpavlin	731	}
490	dpavlin	540
491	dpavlin	290
492	dpavlin	731	foreach my $pos ( 0 ... $input_db->size ) {
493	dpavlin	674
494	dpavlin	731	my $row = $input_db->fetch \|\| next;
495	dpavlin	516
496	dpavlin	769	$total_rows++;
497
498	dpavlin	731	my $mfn = $row->{'000'}->[0];
499	dpavlin	516
500	dpavlin	731	if (! $mfn \|\| $mfn !~ m#^\d+$#) {
501			$log->warn("record $pos doesn't have valid MFN but '$mfn', using $pos");
502			$mfn = $pos;
503			push @{ $row->{'000'} }, $pos;
504			}
505	dpavlin	213
506	dpavlin	508
507	dpavlin	731	if ($validate) {
508	dpavlin	774	if ( my $errors = $validate->validate_rec( $row, $input_db->dump_ascii ) ) {
509	dpavlin	731	$log->error( "MFN $mfn validation error:\n",
510			$validate->report_error( $errors )
511			);
512	dpavlin	539	}
513	dpavlin	769	next; # validation doesn't create any output
514	dpavlin	731	}
515	dpavlin	213
516	dpavlin	731	my $ds = WebPAC::Normalize::data_structure(
517			row => $row,
518			rules => $rules,
519			lookup => $lookup_hash,
520			config => create_ds_config( $db_config, $database, $input, $mfn ),
521			marc_encoding => 'utf-8',
522	dpavlin	736	load_row_coderef => sub {
523	dpavlin	731	my ($database,$input,$mfn) = @_;
524	dpavlin	736	return $store->load_row(
525	dpavlin	731	database => $database,
526			input => $input,
527			id => $mfn,
528			);
529			},
530			);
531	dpavlin	507
532	dpavlin	765	$log->debug("ds = ", sub { dump($ds) }) if ($ds);
533	dpavlin	736
534	dpavlin	731	$store->save_ds(
535			database => $database,
536			input => $input_name,
537			id => $mfn,
538			ds => $ds,
539			) if ($ds && !$stats);
540	dpavlin	210
541	dpavlin	731	$indexer->add(
542			id => "${input_name}/${mfn}",
543			ds => $ds,
544			type => $config->get($indexer_config)->{type},
545			) if ($indexer && $ds);
546	dpavlin	539
547	dpavlin	731	if ($marc) {
548			my $i = 0;
549	dpavlin	559
550	dpavlin	731	while (my $fields = WebPAC::Normalize::_get_marc_fields( fetch_next => 1 ) ) {
551			$marc->add(
552			id => $mfn . ( $i ? "/$i" : '' ),
553			fields => $fields,
554	dpavlin	814	leader => WebPAC::Normalize::_get_marc_leader(),
555	dpavlin	731	row => $row,
556			);
557			$i++;
558	dpavlin	580	}
559
560	dpavlin	731	$log->info("Created $i instances of MFN $mfn\n") if ($i > 1);
561	dpavlin	539	}
562	dpavlin	731	}
563	dpavlin	657
564	dpavlin	731	if ($validate) {
565			my $errors = $validate->report;
566			if ($errors) {
567			$log->info("validation errors:\n$errors\n" );
568			print $report_fh "$errors\n" if ($report_fh);
569	dpavlin	671	}
570	dpavlin	833
571	dpavlin	834	print $report_fh "\nAll possible subfields/delimiter templates:\n", $validate->delimiters_templates( report => 1 ), "\n\n";
572	dpavlin	731	}
573	dpavlin	539
574	dpavlin	731	if ($stats) {
575			my $s = $input_db->stats;
576			$log->info("statistics of fields usage:\n$s");
577			print $report_fh "Statistics of fields usage:\n$s" if ($report_fh);
578	dpavlin	540	}
579
580	dpavlin	731	# close MARC file
581			$marc->finish if ($marc);
582
583			# close report
584			close($report_fh) if ($report_fh)
585
586	dpavlin	849	# reset validate stats
587			$validate->reset if ( $validate );
588	dpavlin	539	}
589
590	dpavlin	509	eval { $indexer->finish } if ($indexer && $indexer->can('finish'));
591	dpavlin	434
592	dpavlin	389	my $dt = time() - $start_t;
593	dpavlin	511	$log->info("$total_rows records ", $indexer ? "indexed " : "",
594			sprintf("in %.2f sec [%.2f rec/sec]",
595	dpavlin	389	$dt, ($total_rows / $dt)
596			)
597			);
598	dpavlin	255
599
600	dpavlin	606	# end forked process
601			if ($parallel) {
602			$log->info("parallel process $$ finished");
603			exit(0);
604			}
605
606	dpavlin	210	}
607
608	dpavlin	606	if ($parallel) {
609			# wait all children to finish
610			sleep(1) while wait != -1;
611			$log->info("all parallel processes finished");
612			}
613
614	dpavlin	837	# save new delimiters if needed
615	dpavlin	845	$validate->save_delimiters_templates if ( $validate );
616	dpavlin	837
617	dpavlin	608	#
618			# handle links or merge after indexing
619			#
620	dpavlin	607
621	dpavlin	608	if ($merge) {
622			print $estcmd_fh 'sudo /etc/init.d/hyperestraier start',$/;
623			close($estcmd_fh);
624			chmod 0700, $estcmd_path \|\| $log->warn("can't chmod 0700 $estcmd_path: $!");
625			system $estcmd_path;
626			} else {
627			foreach my $link (@links) {
628			$log->logdie("coderef in link ", Dumper($link), " is ", ref($link), " and not CODE") unless (ref($link) eq 'CODE');
629			$link->();
630			}
631	dpavlin	529	}