--- trunk/run.pl	2005/12/05 17:46:57	209
+++ trunk/run.pl	2006/08/01 13:59:47	606
@@ -4,111 +4,433 @@
 
 use Cwd qw/abs_path/;
 use File::Temp qw/tempdir/;
-use Data::Dumper;
 use lib './lib';
 
-use WebPAC::Lookup;
-use WebPAC::Input::ISIS;
+use WebPAC::Common 0.02;
+use WebPAC::Lookup 0.03;
+use WebPAC::Input 0.07;
 use WebPAC::Store 0.03;
-use WebPAC::Normalize::XML;
+use WebPAC::Normalize 0.11;
 use WebPAC::Output::TT;
-use WebPAC::Output::Estraier;
+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/;
 
-my $limit = shift @ARGV;
+use Proc::Queue size => 1;
+use POSIX ":sys_wait_h"; # imports WNOHANG
 
-my $config = LoadFile('conf/config.yml');
+=head1 NAME
 
-print "config = ",Dumper($config);
+run.pl - start WebPAC indexing
 
-my $type = lc($config->{input}->{type});
+B<this command will probably go away. Don't get used to it!>
 
-die "I know only how to handle input type isis, not '$type'!\n" unless ($type eq 'isis');
+Options:
 
-my $abs_path = abs_path($0);
-$abs_path =~ s#/[^/]*$#/#;
+=over 4
 
-my $lookup = new WebPAC::Lookup(
-	lookup_file => $config->{input}->{lookup},
-);
+=item --offset 42
 
+start loading (all) databases at offset 42
 
+=item --limit 100
 
-my $isis = new WebPAC::Input::ISIS(
-	code_page => $config->{webpac}->{webpac_encoding},
-	limit_mfn => $config->{input}->{limit},
-);
+limit loading to 100 records
 
-my $maxmfn = $isis->open(
-	filename => $config->{input}->{path},
-	code_page => $config->{input}->{encoding},	# database encoding
-);
+=item --clean
 
-my $path = './db/';
+remove database and Hyper Estraier index before indexing
 
-my $db = new WebPAC::Store(
-	path => $config->{webpac}->{db_path},
-);
+=item --only=database_name/input_filter
 
-my $n = new WebPAC::Normalize::XML(
-#	filter => { 'foo' => sub { shift } },
-	db => $db,
-	lookup_regex => $lookup->regex,
-	lookup => $lookup,
-);
+reindex just single database (legacy name is --one)
 
-$n->open(
-	tag => $config->{normalize}->{tag},
-	xml_file => $config->{normalize}->{path},
-);
+C</input_filter> is optional part which can be C<name>
+or C<type> from input
 
-my $out = new WebPAC::Output::TT(
-	include_path => $config->{webpac}->{template_path},
-	filters => { foo => sub { shift } },
-);
+=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
 
-my $est = new WebPAC::Output::Estraier(
-	%{ $config->{hyperestraier} }
+Force dump or input and marc record for debugging.
+
+=item --parallel 4
+
+Run databases in parallel (aproximatly same as number of processors in
+machine if you want to use full load)
+
+=back
+
+=cut
+
+my $offset;
+my $limit;
+
+my $clean = 0;
+my $config = 'conf/config.yml';
+my $debug = 0;
+my $only_filter;
+my $stats = 0;
+my $validate_path;
+my ($marc_normalize, $marc_output);
+my $marc_lint = 1;
+my $marc_dump = 0;
+
+my $parallel = 0;
+
+GetOptions(
+	"limit=i" => \$limit,
+	"offset=i" => \$offset,
+	"clean" => \$clean,
+	"one=s" => \$only_filter,
+	"only=s" => \$only_filter,
+	"config" => \$config,
+	"debug+" => \$debug,
+	"stats" => \$stats,
+	"validate=s" => \$validate_path,
+	"marc-normalize=s" => \$marc_normalize,
+	"marc-output=s" => \$marc_output,
+	"marc-lint!" => \$marc_lint,
+	"marc-dump!" => \$marc_dump,
+	"parallel=i" => \$parallel,
 );
 
-my $total_rows = 0;
+$config = LoadFile($config);
+
+print "config = ",dump($config) if ($debug);
 
-for ( 0 ... $isis->size ) {
+die "no databases in config file!\n" unless ($config->{databases});
 
-	my $row = $isis->fetch || next;
+my $log = _new WebPAC::Common()->_get_logger();
+$log->info( "-" x 79 );
 
-	my $mfn = $row->{'000'}->[0] || die "can't find MFN";
+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...");
+}
 
-	my $ds = $n->data_structure($row);
+# disable indexing when creating marc
+$use_indexer = undef if ($marc_normalize);
 
-#	print STDERR Dumper($row, $ds);
-
-#	my $html = $out->apply(
-#		template => 'html_ffzg.tt',
-#		data => $ds,
-#	);
-#
-#	# create test output
-#
-#	my $file = sprintf('out/%02d.html', $mfn );
-#	open(my $fh, '>', $file) or die "can't open $file: $!";
-#	print $fh $html;
-#	close($fh);
-#
-#	$html =~ s#\s*[\n\r]+\s*##gs;
-#
-#	print STDERR $html;
-
-	$est->add(
-		id => $mfn,
-		ds => $ds,
-		type => $config->{hyperestraier}->{type},
+my $total_rows = 0;
+my $start_t = time();
+
+my @links;
+my $indexer;
+
+if ($parallel) {
+	$log->info("Using $parallel processes for speedup");
+	Proc::Queue::size($parallel);
+}
+
+while (my ($database, $db_config) = each %{ $config->{databases} }) {
+
+	my ($only_database,$only_input) = split(m#/#, $only_filter) if ($only_filter);
+	next if ($only_database && $database !~ m/$only_database/i);
+
+	if ($parallel) {
+		my $f=fork;
+		if(defined ($f) and $f==0) {
+			$log->info("Created processes $$ for speedup");
+		} else {
+			next;
+		}
+	}
+
+	if ($use_indexer) {
+		my $indexer_config = $config->{$use_indexer} || $log->logdie("can't find '$use_indexer' part in confguration");
+		$indexer_config->{database} = $database;
+		$indexer_config->{clean} = $clean;
+		$indexer_config->{label} = $db_config->{name};
+
+		if ($use_indexer eq 'hyperestraier') {
+
+			# open Hyper Estraier database
+			use WebPAC::Output::Estraier '0.10';
+			$indexer = new WebPAC::Output::Estraier( %{ $indexer_config } );
+		
+		} elsif ($use_indexer eq 'kinosearch') {
+
+			# open KinoSearch
+			use WebPAC::Output::KinoSearch;
+			$indexer_config->{clean} = 1 unless (-e $indexer_config->{index_path});
+			$indexer = new WebPAC::Output::KinoSearch( %{ $indexer_config } );
+
+		} else {
+			$log->logdie("unknown use_indexer: $use_indexer");
+		}
+
+		$log->logide("can't continue without valid indexer") unless ($indexer);
+	}
+
+
+	#
+	# now WebPAC::Store
+	#
+	my $abs_path = abs_path($0);
+	$abs_path =~ s#/[^/]*$#/#;
+
+	my $db_path = $config->{webpac}->{db_path} . '/' . $database;
+
+	if ($clean) {
+		$log->info("creating new database '$database' in $db_path");
+		rmtree( $db_path ) || $log->warn("can't remove $db_path: $!");
+	} else {
+		$log->info("working on database '$database' in $db_path");
+	}
+
+	my $db = new WebPAC::Store(
+		path => $db_path,
+		database => $database,
+		debug => $debug,
 	);
 
-	$total_rows++;
 
-};
+	#
+	# 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)
+		)
+	);
 
-my $log = $lookup->_get_logger;
+	#
+	# add Hyper Estraier links to other databases
+	#
+	if (ref($db_config->{links}) eq 'ARRAY' && $use_indexer) {
+		foreach my $link (@{ $db_config->{links} }) {
+			if ($use_indexer eq 'hyperestraier') {
+				$log->info("saving link $database -> $link->{to} [$link->{credit}]");
+				push @links, {
+					from => $database,
+					to => $link->{to},
+					credit => $link->{credit},
+				};
+			} else {
+				$log->warn("NOT IMPLEMENTED WITH $use_indexer: adding link $database -> $link->{to} [$link->{credit}]");
+			}
+		}
+	}
+
+	# end forked process
+	if ($parallel) {
+		$log->info("parallel process $$ finished");
+		exit(0);
+	}
+
+}
+
+if ($parallel) {
+	# wait all children to finish
+	sleep(1) while wait != -1;
+	$log->info("all parallel processes finished");
+}
+
+foreach my $link (@links) {
+	$log->info("adding link $link->{from} -> $link->{to} [$link->{credit}]");
+	$indexer->add_link( %{ $link } );
+}
 
-$log->info("$total_rows records indexed");