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)

=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 = '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;
my $only_links = 0;
my $merge = 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,
        "only-links!" => \$only_links,
        "merge" => \$merge,
);

$config = LoadFile($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 $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 => $validate_path,
) if ($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;

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;
                }
        }

        my $indexer;
        if ($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};

                # 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 '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} . '/' . $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)
                )
        );


        # 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");
}

#
# 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	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	607	=item --only-links
93
94			Create just links
95
96	dpavlin	608	=item --merge
97
98			Create merged index of databases which have links
99
100	dpavlin	301	=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	dpavlin	510	my $only_filter;
111	dpavlin	507	my $stats = 0;
112	dpavlin	516	my $validate_path;
113	dpavlin	552	my ($marc_normalize, $marc_output);
114	dpavlin	556	my $marc_lint = 1;
115	dpavlin	559	my $marc_dump = 0;
116	dpavlin	606	my $parallel = 0;
117	dpavlin	607	my $only_links = 0;
118	dpavlin	608	my $merge = 0;
119	dpavlin	606
120	dpavlin	301	GetOptions(
121			"limit=i" => \$limit,
122			"offset=i" => \$offset,
123			"clean" => \$clean,
124	dpavlin	510	"one=s" => \$only_filter,
125			"only=s" => \$only_filter,
126	dpavlin	301	"config" => \$config,
127	dpavlin	560	"debug+" => \$debug,
128	dpavlin	507	"stats" => \$stats,
129	dpavlin	516	"validate=s" => \$validate_path,
130	dpavlin	552	"marc-normalize=s" => \$marc_normalize,
131			"marc-output=s" => \$marc_output,
132	dpavlin	556	"marc-lint!" => \$marc_lint,
133	dpavlin	559	"marc-dump!" => \$marc_dump,
134	dpavlin	606	"parallel=i" => \$parallel,
135	dpavlin	607	"only-links!" => \$only_links,
136	dpavlin	608	"merge" => \$merge,
137	dpavlin	301	);
138
139			$config = LoadFile($config);
140
141	dpavlin	556	print "config = ",dump($config) if ($debug);
142	dpavlin	301
143	dpavlin	210	die "no databases in config file!\n" unless ($config->{databases});
144	dpavlin	431
145			my $log = _new WebPAC::Common()->_get_logger();
146	dpavlin	509	$log->info( "-" x 79 );
147	dpavlin	431
148	dpavlin	608
149			my $estcmd_fh;
150			my $estcmd_path = './estcmd-merge.sh';
151			if ($merge) {
152			open($estcmd_fh, '>', $estcmd_path) \|\| $log->logdie("can't open $estcmd_path: $!");
153			print $estcmd_fh 'cd /data/estraier/_node/ \|\| exit 1',$/;
154			print $estcmd_fh 'sudo /etc/init.d/hyperestraier stop',$/;
155			$log->info("created merge batch file $estcmd_path");
156			}
157
158
159	dpavlin	516	my $validate;
160			$validate = new WebPAC::Validate(
161			path => $validate_path,
162			) if ($validate_path);
163
164	dpavlin	608
165	dpavlin	430	my $use_indexer = $config->{use_indexer} \|\| 'hyperestraier';
166	dpavlin	509	if ($stats) {
167			$log->debug("option --stats disables update of indexing engine...");
168			$use_indexer = undef;
169			} else {
170			$log->info("using $use_indexer indexing engine...");
171			}
172	dpavlin	141
173	dpavlin	552	# disable indexing when creating marc
174			$use_indexer = undef if ($marc_normalize);
175
176	dpavlin	213	my $total_rows = 0;
177	dpavlin	389	my $start_t = time();
178	dpavlin	213
179	dpavlin	529	my @links;
180
181	dpavlin	606	if ($parallel) {
182			$log->info("Using $parallel processes for speedup");
183			Proc::Queue::size($parallel);
184			}
185
186	dpavlin	210	while (my ($database, $db_config) = each %{ $config->{databases} }) {
187	dpavlin	141
188	dpavlin	512	my ($only_database,$only_input) = split(m#/#, $only_filter) if ($only_filter);
189	dpavlin	510	next if ($only_database && $database !~ m/$only_database/i);
190	dpavlin	335
191	dpavlin	606	if ($parallel) {
192			my $f=fork;
193			if(defined ($f) and $f==0) {
194			$log->info("Created processes $$ for speedup");
195			} else {
196			next;
197			}
198			}
199
200	dpavlin	607	my $indexer;
201	dpavlin	509	if ($use_indexer) {
202			my $indexer_config = $config->{$use_indexer} \|\| $log->logdie("can't find '$use_indexer' part in confguration");
203			$indexer_config->{database} = $database;
204			$indexer_config->{clean} = $clean;
205			$indexer_config->{label} = $db_config->{name};
206	dpavlin	431
207	dpavlin	608	# force clean if database has links
208			$indexer_config->{clean} = 1 if ($db_config->{links});
209
210	dpavlin	509	if ($use_indexer eq 'hyperestraier') {
211	dpavlin	255
212	dpavlin	509	# open Hyper Estraier database
213			use WebPAC::Output::Estraier '0.10';
214			$indexer = new WebPAC::Output::Estraier( %{ $indexer_config } );
215
216			} elsif ($use_indexer eq 'kinosearch') {
217	dpavlin	430
218	dpavlin	509	# open KinoSearch
219			use WebPAC::Output::KinoSearch;
220			$indexer_config->{clean} = 1 unless (-e $indexer_config->{index_path});
221			$indexer = new WebPAC::Output::KinoSearch( %{ $indexer_config } );
222	dpavlin	431
223	dpavlin	509	} else {
224			$log->logdie("unknown use_indexer: $use_indexer");
225			}
226
227			$log->logide("can't continue without valid indexer") unless ($indexer);
228	dpavlin	430	}
229
230
231	dpavlin	255	#
232	dpavlin	607	# store Hyper Estraier links to other databases
233			#
234			if (ref($db_config->{links}) eq 'ARRAY' && $use_indexer) {
235			foreach my $link (@{ $db_config->{links} }) {
236			if ($use_indexer eq 'hyperestraier') {
237	dpavlin	608	if ($merge) {
238			print $estcmd_fh 'sudo -u www-data estcmd merge ' . $database . ' ' . $link->{to},$/;
239			} else {
240			$log->info("saving link $database -> $link->{to} [$link->{credit}]");
241			push @links, sub {
242			$log->info("adding link $database -> $link->{to} [$link->{credit}]");
243			$indexer->add_link(
244			from => $database,
245			to => $link->{to},
246			credit => $link->{credit},
247			);
248			};
249			}
250	dpavlin	607	} else {
251			$log->warn("NOT IMPLEMENTED WITH $use_indexer: adding link $database -> $link->{to} [$link->{credit}]");
252			}
253			}
254			}
255			next if ($only_links);
256
257
258			#
259	dpavlin	255	# now WebPAC::Store
260			#
261	dpavlin	210	my $abs_path = abs_path($0);
262			$abs_path =~ s#/[^/]*$#/#;
263	dpavlin	141
264	dpavlin	210	my $db_path = $config->{webpac}->{db_path} . '/' . $database;
265	dpavlin	74
266	dpavlin	301	if ($clean) {
267	dpavlin	511	$log->info("creating new database '$database' in $db_path");
268	dpavlin	301	rmtree( $db_path ) \|\| $log->warn("can't remove $db_path: $!");
269			} else {
270	dpavlin	511	$log->info("working on database '$database' in $db_path");
271	dpavlin	301	}
272	dpavlin	255
273	dpavlin	210	my $db = new WebPAC::Store(
274			path => $db_path,
275	dpavlin	217	database => $database,
276	dpavlin	301	debug => $debug,
277	dpavlin	210	);
278	dpavlin	74
279	dpavlin	233
280	dpavlin	213	#
281			# now, iterate through input formats
282			#
283	dpavlin	74
284	dpavlin	213	my @inputs;
285			if (ref($db_config->{input}) eq 'ARRAY') {
286			@inputs = @{ $db_config->{input} };
287	dpavlin	255	} elsif ($db_config->{input}) {
288			push @inputs, $db_config->{input};
289	dpavlin	213	} else {
290	dpavlin	255	$log->info("database $database doesn't have inputs defined");
291	dpavlin	213	}
292	dpavlin	74
293	dpavlin	286	my @supported_inputs = keys %{ $config->{webpac}->{inputs} };
294
295	dpavlin	213	foreach my $input (@inputs) {
296	dpavlin	233
297	dpavlin	512	next if ($only_input && ($input->{name} !~ m#$only_input#i && $input->{type} !~ m#$only_input#i));
298	dpavlin	510
299	dpavlin	233	my $type = lc($input->{type});
300
301	dpavlin	286	die "I know only how to handle input types ", join(",", @supported_inputs), " not '$type'!\n" unless (grep(/$type/, @supported_inputs));
302	dpavlin	233
303	dpavlin	588	my $lookup;
304			if ($input->{lookup}) {
305			$lookup = new WebPAC::Lookup(
306			lookup_file => $input->{lookup},
307			);
308			delete( $input->{lookup} );
309			}
310	dpavlin	233
311	dpavlin	286	my $input_module = $config->{webpac}->{inputs}->{$type};
312	dpavlin	74
313	dpavlin	523	$log->info("working on input '$input->{name}' in $input->{path} [type: $input->{type}] using $input_module",
314			$input->{lookup} ? "lookup '$input->{lookup}'" : ""
315			);
316	dpavlin	286
317	dpavlin	287	my $input_db = new WebPAC::Input(
318			module => $input_module,
319	dpavlin	585	encoding => $config->{webpac}->{webpac_encoding},
320	dpavlin	301	limit => $limit \|\| $input->{limit},
321			offset => $offset,
322	dpavlin	585	lookup_coderef => sub {
323			my $rec = shift \|\| return;
324			$lookup->add( $rec );
325			},
326	dpavlin	416	recode => $input->{recode},
327	dpavlin	507	stats => $stats,
328	dpavlin	600	modify_records => $input->{modify_records},
329	dpavlin	287	);
330	dpavlin	286	$log->logdie("can't create input using $input_module") unless ($input);
331	dpavlin	113
332	dpavlin	287	my $maxmfn = $input_db->open(
333	dpavlin	285	path => $input->{path},
334	dpavlin	213	code_page => $input->{encoding}, # database encoding
335	dpavlin	523	%{ $input },
336	dpavlin	213	);
337	dpavlin	113
338	dpavlin	539	my @norm_array = ref($input->{normalize}) eq 'ARRAY' ?
339			@{ $input->{normalize} } : ( $input->{normalize} );
340	dpavlin	210
341	dpavlin	552	if ($marc_normalize) {
342			@norm_array = ( {
343			path => $marc_normalize,
344			output => $marc_output \|\| 'out/marc/' . $database . '-' . $input->{name} . '.marc',
345			} );
346			}
347
348	dpavlin	539	foreach my $normalize (@norm_array) {
349	dpavlin	492
350	dpavlin	539	my $normalize_path = $normalize->{path} \|\| $log->logdie("can't find normalize path in config");
351	dpavlin	269
352	dpavlin	539	$log->logdie("Found '$normalize_path' as normalization file which isn't supported any more!") unless ( $normalize_path =~ m!\.pl$!i );
353	dpavlin	210
354	dpavlin	539	my $rules = read_file( $normalize_path ) or die "can't open $normalize_path: $!";
355	dpavlin	213
356	dpavlin	539	$log->info("Using $normalize_path for normalization...");
357	dpavlin	213
358	dpavlin	578	my $marc = new WebPAC::Output::MARC(
359			path => $normalize->{output},
360			lint => $marc_lint,
361			dump => $marc_dump,
362			) if ($normalize->{output});
363	dpavlin	540
364	dpavlin	539	# reset position in database
365			$input_db->seek(1);
366	dpavlin	290
367	dpavlin	539	foreach my $pos ( 0 ... $input_db->size ) {
368	dpavlin	516
369	dpavlin	539	my $row = $input_db->fetch \|\| next;
370	dpavlin	516
371	dpavlin	539	my $mfn = $row->{'000'}->[0];
372	dpavlin	213
373	dpavlin	539	if (! $mfn \|\| $mfn !~ m#^\d+$#) {
374			$log->warn("record $pos doesn't have valid MFN but '$mfn', using $pos");
375			$mfn = $pos;
376			push @{ $row->{'000'} }, $pos;
377			}
378	dpavlin	508
379	dpavlin	213
380	dpavlin	539	if ($validate) {
381			my @errors = $validate->validate_errors( $row );
382			$log->error( "MFN $mfn validation errors:\n", join("\n", @errors) ) if (@errors);
383			}
384	dpavlin	213
385	dpavlin	595	my $ds_config = dclone($db_config);
386
387			# default values -> database key
388			$ds_config->{_} = $database;
389
390			# current mfn
391			$ds_config->{_mfn} = $mfn;
392
393			# attach current input
394			$ds_config->{input} = $input;
395
396	dpavlin	539	my $ds = WebPAC::Normalize::data_structure(
397			row => $row,
398			rules => $rules,
399			lookup => $lookup ? $lookup->lookup_hash : undef,
400	dpavlin	595	config => $ds_config,
401	dpavlin	541	marc_encoding => 'utf-8',
402	dpavlin	539	);
403	dpavlin	507
404	dpavlin	539	$db->save_ds(
405			id => $mfn,
406			ds => $ds,
407			prefix => $input->{name},
408			) if ($ds && !$stats);
409	dpavlin	210
410	dpavlin	539	$indexer->add(
411			id => $input->{name} . "/" . $mfn,
412			ds => $ds,
413			type => $config->{$use_indexer}->{type},
414	dpavlin	540	) if ($indexer && $ds);
415	dpavlin	539
416	dpavlin	580	if ($marc) {
417			my $i = 0;
418	dpavlin	559
419	dpavlin	580	while (my $fields = WebPAC::Normalize::_get_marc_fields( fetch_next => 1 ) ) {
420			$marc->add(
421			id => $mfn . ( $i ? "/$i" : '' ),
422			fields => $fields,
423			leader => WebPAC::Normalize::marc_leader(),
424			row => $row,
425			);
426			$i++;
427			}
428
429			$log->info("Created $i instances of MFN $mfn\n") if ($i > 1);
430			}
431
432	dpavlin	539	$total_rows++;
433			}
434
435			$log->info("statistics of fields usage:\n", $input_db->stats) if ($stats);
436
437	dpavlin	540	# close MARC file
438	dpavlin	578	$marc->finish if ($marc);
439	dpavlin	539
440	dpavlin	540	}
441
442	dpavlin	539	}
443
444	dpavlin	509	eval { $indexer->finish } if ($indexer && $indexer->can('finish'));
445	dpavlin	434
446	dpavlin	389	my $dt = time() - $start_t;
447	dpavlin	511	$log->info("$total_rows records ", $indexer ? "indexed " : "",
448			sprintf("in %.2f sec [%.2f rec/sec]",
449	dpavlin	389	$dt, ($total_rows / $dt)
450			)
451			);
452	dpavlin	255
453
454	dpavlin	606	# end forked process
455			if ($parallel) {
456			$log->info("parallel process $$ finished");
457			exit(0);
458			}
459
460	dpavlin	210	}
461
462	dpavlin	606	if ($parallel) {
463			# wait all children to finish
464			sleep(1) while wait != -1;
465			$log->info("all parallel processes finished");
466			}
467
468	dpavlin	608	#
469			# handle links or merge after indexing
470			#
471	dpavlin	607
472	dpavlin	608	if ($merge) {
473			print $estcmd_fh 'sudo /etc/init.d/hyperestraier start',$/;
474			close($estcmd_fh);
475			chmod 0700, $estcmd_path \|\| $log->warn("can't chmod 0700 $estcmd_path: $!");
476			system $estcmd_path;
477			} else {
478			foreach my $link (@links) {
479			$log->logdie("coderef in link ", Dumper($link), " is ", ref($link), " and not CODE") unless (ref($link) eq 'CODE');
480			$link->();
481			}
482	dpavlin	529	}