--- trunk/run.pl	2006/07/02 10:27:06	558
+++ trunk/run.pl	2006/08/23 11:04:32	613
@@ -7,20 +7,23 @@
 use lib './lib';
 
 use WebPAC::Common 0.02;
-use WebPAC::Lookup;
-use WebPAC::Input 0.03;
+use WebPAC::Lookup 0.03;
+use WebPAC::Input 0.10;
 use WebPAC::Store 0.03;
-use WebPAC::Normalize;
+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 MARC::Record 2.0;	# need 2.0 for utf-8 encoding see marcpm.sf.net
-use MARC::Lint;
 use Data::Dump qw/dump/;
+use Storable qw/dclone/;
+
+use Proc::Queue size => 1;
+use POSIX ":sys_wait_h"; # imports WNOHANG
 
 =head1 NAME
 
@@ -77,6 +80,23 @@
 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
@@ -92,6 +112,20 @@
 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,
@@ -100,28 +134,43 @@
 	"one=s" => \$only_filter,
 	"only=s" => \$only_filter,
 	"config" => \$config,
-	"debug" => \$debug,
+	"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);
+#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...");
@@ -137,21 +186,36 @@
 my $start_t = time();
 
 my @links;
-my $indexer;
 
-my $lint = new MARC::Lint if ($marc_lint);
+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
@@ -174,6 +238,33 @@
 
 
 	#
+	# 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);
@@ -218,9 +309,13 @@
 
 		die "I know only how to handle input types ", join(",", @supported_inputs), " not '$type'!\n" unless (grep(/$type/, @supported_inputs));
 
-		my $lookup = new WebPAC::Lookup(
-			lookup_file => $input->{lookup},
-		) if ($input->{lookup});
+		my $lookup;
+		if ($input->{lookup}) {
+			$lookup = new WebPAC::Lookup(
+				lookup_file => $input->{lookup},
+			);
+			delete( $input->{lookup} );
+		}
 
 		my $input_module = $config->{webpac}->{inputs}->{$type};
 
@@ -230,12 +325,16 @@
 
 		my $input_db = new WebPAC::Input(
 			module => $input_module,
-			code_page => $config->{webpac}->{webpac_encoding},
+			encoding => $config->{webpac}->{webpac_encoding},
 			limit => $limit || $input->{limit},
 			offset => $offset,
-			lookup => $lookup,
+			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);
 
@@ -265,13 +364,11 @@
 
 			$log->info("Using $normalize_path for normalization...");
 
-			my $marc_fh;
-			if (my $path = $normalize->{output}) {
-				open($marc_fh, '>', $path) ||
-					$log->logdie("can't open MARC output $path: $!");
-
-				$log->info("Creating MARC export file $path", $marc_lint ? ' (with lint)' : '', "\n");
-			}
+			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);
@@ -294,11 +391,22 @@
 					$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',
 				);
 
@@ -314,25 +422,20 @@
 					type => $config->{$use_indexer}->{type},
 				) if ($indexer && $ds);
 
-				if ($marc_fh) {
-					my $marc = new MARC::Record;
-					$marc->encoding( 'utf-8' );
-					my @marc_fields = WebPAC::Normalize::_get_marc_fields();
-					if (! @marc_fields) {
-						$log->warn("MARC record $mfn is empty, skipping");
-					} else {
-						$marc->add_fields( @marc_fields );
-						if ($marc_lint) {
-							$lint->check_record( $marc );
-							my $err = join( "\n", $lint->warnings );
-							$log->error("MARC lint detected warning on MFN $mfn\n",
-								"Original imput row: ",dump($row), "\n",
-								"Normalized MARC row: ",dump(@marc_fields), "\n",
-								"MARC lint warnings: ",$err,"\n"
-							) if ($err);
-						}
-						print $marc_fh $marc->as_usmarc;
+				if ($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++;
@@ -341,7 +444,7 @@
 			$log->info("statistics of fields usage:\n", $input_db->stats) if ($stats);
 
 			# close MARC file
-			close($marc_fh) if ($marc_fh);
+			$marc->finish if ($marc);
 
 		}
 
@@ -356,27 +459,33 @@
 		)
 	);
 
-	#
-	# 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);
 	}
 
 }
 
-foreach my $link (@links) {
-	$log->info("adding link $link->{from} -> $link->{to} [$link->{credit}]");
-	$indexer->add_link( %{ $link } );
+if ($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->();
+	}
 }