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 0.11;
use WebPAC::Store 0.03;
use WebPAC::Normalize 0.11;
use WebPAC::Output::TT;
use WebPAC::Validate 0.06;
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, 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 --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;

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

my $hostname = `hostname`;
chomp($hostname);
$hostname =~ s/\..+$//;
if (-e "conf/$hostname.yml") {
        $config = "conf/$hostname.yml";
        $log->info("using host configuration file: $config");
}

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});

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

                my $indexer_config = $config->{$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} . '/' . $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}'" : ""
                );

                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,
                        lookup_coderef => sub {
                                my $rec = shift || return;
                                $lookup->add( $rec );
                        },
                        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);

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

                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 @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);

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

                        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) {
                                        if ( my $errors = $validate->validate_errors( $row, $input_db->dump ) ) {
                                                $log->error( "MFN $mfn validation error:\n",
                                                        $validate->report_error( $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 => $config->{$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::marc_leader(),
                                                        row => $row,
                                                );
                                                $i++;
                                        }

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

                                $total_rows++;
                        }

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

                        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)
                }

        }

        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	619	use WebPAC::Input 0.11;
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	652	use WebPAC::Validate 0.06;
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	638	disable indexing, modify_* in configuration and dump statistics about field
64			and subfield 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	611	my $log = _new WebPAC::Common()->_get_logger();
121
122			my $hostname = `hostname`;
123			chomp($hostname);
124			$hostname =~ s/\..+$//;
125			if (-e "conf/$hostname.yml") {
126			$config = "conf/$hostname.yml";
127			$log->info("using host configuration file: $config");
128			}
129
130	dpavlin	301	GetOptions(
131			"limit=i" => \$limit,
132			"offset=i" => \$offset,
133			"clean" => \$clean,
134	dpavlin	510	"one=s" => \$only_filter,
135			"only=s" => \$only_filter,
136	dpavlin	301	"config" => \$config,
137	dpavlin	560	"debug+" => \$debug,
138	dpavlin	507	"stats" => \$stats,
139	dpavlin	516	"validate=s" => \$validate_path,
140	dpavlin	552	"marc-normalize=s" => \$marc_normalize,
141			"marc-output=s" => \$marc_output,
142	dpavlin	556	"marc-lint!" => \$marc_lint,
143	dpavlin	559	"marc-dump!" => \$marc_dump,
144	dpavlin	606	"parallel=i" => \$parallel,
145	dpavlin	607	"only-links!" => \$only_links,
146	dpavlin	608	"merge" => \$merge,
147	dpavlin	301	);
148
149			$config = LoadFile($config);
150
151	dpavlin	611	#print "config = ",dump($config) if ($debug);
152	dpavlin	301
153	dpavlin	210	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	608
158			my $estcmd_fh;
159			my $estcmd_path = './estcmd-merge.sh';
160			if ($merge) {
161			open($estcmd_fh, '>', $estcmd_path) \|\| $log->logdie("can't open $estcmd_path: $!");
162			print $estcmd_fh 'cd /data/estraier/_node/ \|\| exit 1',$/;
163			print $estcmd_fh 'sudo /etc/init.d/hyperestraier stop',$/;
164			$log->info("created merge batch file $estcmd_path");
165			}
166
167
168	dpavlin	516	my $validate;
169			$validate = new WebPAC::Validate(
170			path => $validate_path,
171			) if ($validate_path);
172
173	dpavlin	608
174	dpavlin	430	my $use_indexer = $config->{use_indexer} \|\| 'hyperestraier';
175	dpavlin	509	if ($stats) {
176			$log->debug("option --stats disables update of indexing engine...");
177			$use_indexer = undef;
178			} else {
179			$log->info("using $use_indexer indexing engine...");
180			}
181	dpavlin	141
182	dpavlin	552	# disable indexing when creating marc
183			$use_indexer = undef if ($marc_normalize);
184
185	dpavlin	213	my $total_rows = 0;
186	dpavlin	389	my $start_t = time();
187	dpavlin	213
188	dpavlin	529	my @links;
189
190	dpavlin	606	if ($parallel) {
191			$log->info("Using $parallel processes for speedup");
192			Proc::Queue::size($parallel);
193			}
194
195	dpavlin	210	while (my ($database, $db_config) = each %{ $config->{databases} }) {
196	dpavlin	141
197	dpavlin	512	my ($only_database,$only_input) = split(m#/#, $only_filter) if ($only_filter);
198	dpavlin	510	next if ($only_database && $database !~ m/$only_database/i);
199	dpavlin	335
200	dpavlin	606	if ($parallel) {
201			my $f=fork;
202			if(defined ($f) and $f==0) {
203			$log->info("Created processes $$ for speedup");
204			} else {
205			next;
206			}
207			}
208
209	dpavlin	607	my $indexer;
210	dpavlin	509	if ($use_indexer) {
211	dpavlin	627
212			my $cfg_name = $use_indexer;
213			$cfg_name =~ s/\-.*$//;
214
215			my $indexer_config = $config->{$cfg_name} \|\| $log->logdie("can't find '$cfg_name' part in confguration");
216	dpavlin	509	$indexer_config->{database} = $database;
217			$indexer_config->{clean} = $clean;
218			$indexer_config->{label} = $db_config->{name};
219	dpavlin	431
220	dpavlin	608	# force clean if database has links
221			$indexer_config->{clean} = 1 if ($db_config->{links});
222
223	dpavlin	509	if ($use_indexer eq 'hyperestraier') {
224	dpavlin	255
225	dpavlin	509	# open Hyper Estraier database
226			use WebPAC::Output::Estraier '0.10';
227			$indexer = new WebPAC::Output::Estraier( %{ $indexer_config } );
228
229	dpavlin	627	} elsif ($use_indexer eq 'hyperestraier-native') {
230
231			# open Hyper Estraier database
232			use WebPAC::Output::EstraierNative;
233			$indexer = new WebPAC::Output::EstraierNative( %{ $indexer_config } );
234
235	dpavlin	509	} elsif ($use_indexer eq 'kinosearch') {
236	dpavlin	430
237	dpavlin	509	# open KinoSearch
238			use WebPAC::Output::KinoSearch;
239			$indexer_config->{clean} = 1 unless (-e $indexer_config->{index_path});
240			$indexer = new WebPAC::Output::KinoSearch( %{ $indexer_config } );
241	dpavlin	431
242	dpavlin	509	} else {
243			$log->logdie("unknown use_indexer: $use_indexer");
244			}
245
246			$log->logide("can't continue without valid indexer") unless ($indexer);
247	dpavlin	430	}
248
249
250	dpavlin	255	#
251	dpavlin	607	# store Hyper Estraier links to other databases
252			#
253			if (ref($db_config->{links}) eq 'ARRAY' && $use_indexer) {
254			foreach my $link (@{ $db_config->{links} }) {
255			if ($use_indexer eq 'hyperestraier') {
256	dpavlin	608	if ($merge) {
257			print $estcmd_fh 'sudo -u www-data estcmd merge ' . $database . ' ' . $link->{to},$/;
258			} else {
259			$log->info("saving link $database -> $link->{to} [$link->{credit}]");
260			push @links, sub {
261			$log->info("adding link $database -> $link->{to} [$link->{credit}]");
262			$indexer->add_link(
263			from => $database,
264			to => $link->{to},
265			credit => $link->{credit},
266			);
267			};
268			}
269	dpavlin	607	} else {
270			$log->warn("NOT IMPLEMENTED WITH $use_indexer: adding link $database -> $link->{to} [$link->{credit}]");
271			}
272			}
273			}
274			next if ($only_links);
275
276
277			#
278	dpavlin	255	# now WebPAC::Store
279			#
280	dpavlin	210	my $abs_path = abs_path($0);
281			$abs_path =~ s#/[^/]*$#/#;
282	dpavlin	141
283	dpavlin	210	my $db_path = $config->{webpac}->{db_path} . '/' . $database;
284	dpavlin	74
285	dpavlin	301	if ($clean) {
286	dpavlin	511	$log->info("creating new database '$database' in $db_path");
287	dpavlin	301	rmtree( $db_path ) \|\| $log->warn("can't remove $db_path: $!");
288			} else {
289	dpavlin	511	$log->info("working on database '$database' in $db_path");
290	dpavlin	301	}
291	dpavlin	255
292	dpavlin	210	my $db = new WebPAC::Store(
293			path => $db_path,
294	dpavlin	217	database => $database,
295	dpavlin	301	debug => $debug,
296	dpavlin	210	);
297	dpavlin	74
298	dpavlin	233
299	dpavlin	213	#
300			# now, iterate through input formats
301			#
302	dpavlin	74
303	dpavlin	213	my @inputs;
304			if (ref($db_config->{input}) eq 'ARRAY') {
305			@inputs = @{ $db_config->{input} };
306	dpavlin	255	} elsif ($db_config->{input}) {
307			push @inputs, $db_config->{input};
308	dpavlin	213	} else {
309	dpavlin	255	$log->info("database $database doesn't have inputs defined");
310	dpavlin	213	}
311	dpavlin	74
312	dpavlin	286	my @supported_inputs = keys %{ $config->{webpac}->{inputs} };
313
314	dpavlin	213	foreach my $input (@inputs) {
315	dpavlin	233
316	dpavlin	512	next if ($only_input && ($input->{name} !~ m#$only_input#i && $input->{type} !~ m#$only_input#i));
317	dpavlin	510
318	dpavlin	233	my $type = lc($input->{type});
319
320	dpavlin	286	die "I know only how to handle input types ", join(",", @supported_inputs), " not '$type'!\n" unless (grep(/$type/, @supported_inputs));
321	dpavlin	233
322	dpavlin	588	my $lookup;
323			if ($input->{lookup}) {
324			$lookup = new WebPAC::Lookup(
325			lookup_file => $input->{lookup},
326			);
327			delete( $input->{lookup} );
328			}
329	dpavlin	233
330	dpavlin	286	my $input_module = $config->{webpac}->{inputs}->{$type};
331	dpavlin	74
332	dpavlin	523	$log->info("working on input '$input->{name}' in $input->{path} [type: $input->{type}] using $input_module",
333			$input->{lookup} ? "lookup '$input->{lookup}'" : ""
334			);
335	dpavlin	286
336	dpavlin	638	if ($stats) {
337			# disable modification of records if --stats is in use
338			delete($input->{modify_records});
339			delete($input->{modify_file});
340			}
341
342	dpavlin	287	my $input_db = new WebPAC::Input(
343			module => $input_module,
344	dpavlin	585	encoding => $config->{webpac}->{webpac_encoding},
345	dpavlin	301	limit => $limit \|\| $input->{limit},
346			offset => $offset,
347	dpavlin	585	lookup_coderef => sub {
348			my $rec = shift \|\| return;
349			$lookup->add( $rec );
350			},
351	dpavlin	416	recode => $input->{recode},
352	dpavlin	507	stats => $stats,
353	dpavlin	600	modify_records => $input->{modify_records},
354	dpavlin	636	modify_file => $input->{modify_file},
355	dpavlin	287	);
356	dpavlin	286	$log->logdie("can't create input using $input_module") unless ($input);
357	dpavlin	113
358	dpavlin	287	my $maxmfn = $input_db->open(
359	dpavlin	285	path => $input->{path},
360	dpavlin	213	code_page => $input->{encoding}, # database encoding
361	dpavlin	523	%{ $input },
362	dpavlin	213	);
363	dpavlin	113
364	dpavlin	671	my $report_fh;
365			if ($stats \|\| $validate) {
366			my $path = "out/report/" . $database . '-' . $input->{name} . '.txt';
367			open($report_fh, '>', $path) \|\| $log->logdie("can't open $path: $!");
368
369			print $report_fh "Report for database '$database' input '$input->{name}' records ",
370			$offset \|\| 1, "-", $limit \|\| $input->{limit} \|\| $maxmfn, "\n\n";
371			$log->info("Generating report file $path");
372			}
373
374	dpavlin	539	my @norm_array = ref($input->{normalize}) eq 'ARRAY' ?
375			@{ $input->{normalize} } : ( $input->{normalize} );
376	dpavlin	210
377	dpavlin	552	if ($marc_normalize) {
378			@norm_array = ( {
379			path => $marc_normalize,
380			output => $marc_output \|\| 'out/marc/' . $database . '-' . $input->{name} . '.marc',
381			} );
382			}
383
384	dpavlin	539	foreach my $normalize (@norm_array) {
385	dpavlin	492
386	dpavlin	539	my $normalize_path = $normalize->{path} \|\| $log->logdie("can't find normalize path in config");
387	dpavlin	269
388	dpavlin	539	$log->logdie("Found '$normalize_path' as normalization file which isn't supported any more!") unless ( $normalize_path =~ m!\.pl$!i );
389	dpavlin	210
390	dpavlin	539	my $rules = read_file( $normalize_path ) or die "can't open $normalize_path: $!";
391	dpavlin	213
392	dpavlin	539	$log->info("Using $normalize_path for normalization...");
393	dpavlin	213
394	dpavlin	578	my $marc = new WebPAC::Output::MARC(
395			path => $normalize->{output},
396			lint => $marc_lint,
397			dump => $marc_dump,
398			) if ($normalize->{output});
399	dpavlin	540
400	dpavlin	539	# reset position in database
401			$input_db->seek(1);
402	dpavlin	290
403	dpavlin	674	# generate name of config key for indexer (strip everything after -)
404			my $indexer_config = $use_indexer;
405	dpavlin	675	$indexer_config =~ s/^(\w+)-?.*$/$1/g if ($indexer_config);
406	dpavlin	674
407	dpavlin	539	foreach my $pos ( 0 ... $input_db->size ) {
408	dpavlin	516
409	dpavlin	539	my $row = $input_db->fetch \|\| next;
410	dpavlin	516
411	dpavlin	539	my $mfn = $row->{'000'}->[0];
412	dpavlin	213
413	dpavlin	539	if (! $mfn \|\| $mfn !~ m#^\d+$#) {
414			$log->warn("record $pos doesn't have valid MFN but '$mfn', using $pos");
415			$mfn = $pos;
416			push @{ $row->{'000'} }, $pos;
417			}
418	dpavlin	508
419	dpavlin	213
420	dpavlin	539	if ($validate) {
421	dpavlin	653	if ( my $errors = $validate->validate_errors( $row, $input_db->dump ) ) {
422			$log->error( "MFN $mfn validation error:\n",
423	dpavlin	671	$validate->report_error( $errors )
424	dpavlin	652	);
425			}
426	dpavlin	539	}
427	dpavlin	213
428	dpavlin	595	my $ds_config = dclone($db_config);
429
430			# default values -> database key
431			$ds_config->{_} = $database;
432
433			# current mfn
434			$ds_config->{_mfn} = $mfn;
435
436			# attach current input
437			$ds_config->{input} = $input;
438
439	dpavlin	539	my $ds = WebPAC::Normalize::data_structure(
440			row => $row,
441			rules => $rules,
442			lookup => $lookup ? $lookup->lookup_hash : undef,
443	dpavlin	595	config => $ds_config,
444	dpavlin	541	marc_encoding => 'utf-8',
445	dpavlin	539	);
446	dpavlin	507
447	dpavlin	539	$db->save_ds(
448			id => $mfn,
449			ds => $ds,
450			prefix => $input->{name},
451			) if ($ds && !$stats);
452	dpavlin	210
453	dpavlin	539	$indexer->add(
454			id => $input->{name} . "/" . $mfn,
455			ds => $ds,
456	dpavlin	674	type => $config->{$indexer_config}->{type},
457	dpavlin	540	) if ($indexer && $ds);
458	dpavlin	539
459	dpavlin	580	if ($marc) {
460			my $i = 0;
461	dpavlin	559
462	dpavlin	580	while (my $fields = WebPAC::Normalize::_get_marc_fields( fetch_next => 1 ) ) {
463			$marc->add(
464			id => $mfn . ( $i ? "/$i" : '' ),
465			fields => $fields,
466			leader => WebPAC::Normalize::marc_leader(),
467			row => $row,
468			);
469			$i++;
470			}
471
472			$log->info("Created $i instances of MFN $mfn\n") if ($i > 1);
473			}
474
475	dpavlin	539	$total_rows++;
476			}
477
478	dpavlin	664	if ($validate) {
479			my $errors = $validate->report;
480	dpavlin	671	if ($errors) {
481			$log->info("validation errors:\n$errors\n" );
482			print $report_fh "$errors\n" if ($report_fh);
483			}
484	dpavlin	657	}
485
486	dpavlin	671	if ($stats) {
487			my $s = $input_db->stats;
488			$log->info("statistics of fields usage:\n$s");
489			print $report_fh "Statistics of fields usage:\n$s" if ($report_fh);
490			}
491	dpavlin	539
492	dpavlin	540	# close MARC file
493	dpavlin	578	$marc->finish if ($marc);
494	dpavlin	539
495	dpavlin	671	# close report
496	dpavlin	672	close($report_fh) if ($report_fh)
497	dpavlin	540	}
498
499	dpavlin	539	}
500
501	dpavlin	509	eval { $indexer->finish } if ($indexer && $indexer->can('finish'));
502	dpavlin	434
503	dpavlin	389	my $dt = time() - $start_t;
504	dpavlin	511	$log->info("$total_rows records ", $indexer ? "indexed " : "",
505			sprintf("in %.2f sec [%.2f rec/sec]",
506	dpavlin	389	$dt, ($total_rows / $dt)
507			)
508			);
509	dpavlin	255
510
511	dpavlin	606	# end forked process
512			if ($parallel) {
513			$log->info("parallel process $$ finished");
514			exit(0);
515			}
516
517	dpavlin	210	}
518
519	dpavlin	606	if ($parallel) {
520			# wait all children to finish
521			sleep(1) while wait != -1;
522			$log->info("all parallel processes finished");
523			}
524
525	dpavlin	608	#
526			# handle links or merge after indexing
527			#
528	dpavlin	607
529	dpavlin	608	if ($merge) {
530			print $estcmd_fh 'sudo /etc/init.d/hyperestraier start',$/;
531			close($estcmd_fh);
532			chmod 0700, $estcmd_path \|\| $log->warn("can't chmod 0700 $estcmd_path: $!");
533			system $estcmd_path;
534			} else {
535			foreach my $link (@links) {
536			$log->logdie("coderef in link ", Dumper($link), " is ", ref($link), " and not CODE") unless (ref($link) eq 'CODE');
537			$link->();
538			}
539	dpavlin	529	}