--- trunk/run.pl	2006/09/25 15:26:12	707
+++ trunk/run.pl	2008/08/02 15:11:06	1098
@@ -7,12 +7,12 @@
 use lib './lib';
 
 use WebPAC::Common 0.02;
-use WebPAC::Parser 0.04;
-use WebPAC::Input 0.13;
-use WebPAC::Store 0.03;
-use WebPAC::Normalize 0.11;
+use WebPAC::Parser 0.08;
+use WebPAC::Input 0.16;
+use WebPAC::Store 0.15;
+use WebPAC::Normalize 0.22;
 use WebPAC::Output::TT;
-use WebPAC::Validate 0.06;
+use WebPAC::Validate 0.11;
 use WebPAC::Output::MARC;
 use WebPAC::Config;
 use Getopt::Long;
@@ -21,6 +21,7 @@
 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
@@ -31,7 +32,7 @@
 
 B<this command will probably go away. Don't get used to it!>
 
-Options:
+=head1 OPTIONS
 
 =over 4
 
@@ -67,17 +68,22 @@
 
 turn on extra validation of imput records, see L<WebPAC::Validation>
 
-=item --marc-normalize conf/normalize/mapping.pl
+You can use special variables C<$database> and $C<$input> in this parametar
+like C<--validate 'conf/validate/$database-$input'> to construct filename
 
-This option specifies normalisation file for MARC creation
+=item --validate-delimiters path/to/validate_delimiters_file
 
-=item --marc-output out/marc/test.marc
+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.
 
-Optional path to output file
+=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 C<--marc-normalize> is used. You can disable 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
@@ -110,12 +116,14 @@
 my $only_filter;
 my $stats = 0;
 my $validate_path;
-my ($marc_normalize, $marc_output);
+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();
 
@@ -125,27 +133,40 @@
 	"clean" => \$clean,
 	"one=s" => \$only_filter,
 	"only=s" => \$only_filter,
-	"config" => \$config_path,
+	"config=s" => \$config_path,
 	"debug+" => \$debug,
 	"stats" => \$stats,
 	"validate=s" => \$validate_path,
-	"marc-normalize=s" => \$marc_normalize,
-	"marc-output=s" => \$marc_output,
+	"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,
 );
 
+$marc_generate = 0 if ( $validate_delimiters_path );
+
+pod2usage(-verbose => 2) if ($help);
+
 my $config = new WebPAC::Config( path => $config_path );
 
+WebPAC::Normalize::_debug( $debug - 1 ) if $debug > 1;
+
 #print "config = ",dump($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';
@@ -156,27 +177,29 @@
 	$log->info("created merge batch file $estcmd_path");
 }
 
-
 my $validate;
 $validate = new WebPAC::Validate(
-	path => $validate_path,
-) if ($validate_path);
-
+	delimiters => $config->webpac('delimiters'),
+) if ($validate_path || $validate_delimiters_path);
 
 my $use_indexer = $config->use_indexer;
+$stats ||= $validate;
 if ($stats) {
-	$log->debug("option --stats disables update of indexing engine...");
+	$log->debug("disabled indexing for stats collection");
 	$use_indexer = undef;
-} else {
+} elsif ( $use_indexer ) {
 	$log->info("using $use_indexer indexing engine...");
 }
 
-# disable indexing when creating marc
-$use_indexer = undef if ($marc_normalize);
-
 # parse normalize files and create source files for lookup and normalization
 
-my $parser = new WebPAC::Parser( config => $config );
+my ($only_database,$only_input) = split(m#/#, $only_filter) if $only_filter;
+
+my $parser = new WebPAC::Parser(
+	config => $config,
+	only_database => $only_database,
+	only_input => $only_input,
+);
 
 my $total_rows = 0;
 my $start_t = time();
@@ -197,9 +220,9 @@
 	return $c;
 }
 
-while (my ($database, $db_config) = each %{ $config->databases }) {
+foreach my $database ( sort keys %{ $config->databases } ) {
+	my $db_config = $config->databases->{$database};
 
-	my ($only_database,$only_input) = split(m#/#, $only_filter) if ($only_filter);
 	next if ($only_database && $database !~ m/$only_database/i);
 
 	if ($parallel) {
@@ -212,7 +235,7 @@
 	}
 
 	my $indexer;
-	if ($use_indexer) {
+	if ($use_indexer && $parser->have_rules( 'search', $database )) {
 
 		my $cfg_name = $use_indexer;
 		$cfg_name =~ s/\-.*$//;
@@ -228,27 +251,24 @@
 		if ($use_indexer eq 'hyperestraier') {
 
 			# open Hyper Estraier database
-			use WebPAC::Output::Estraier '0.10';
+			require WebPAC::Output::Estraier;
 			$indexer = new WebPAC::Output::Estraier( %{ $indexer_config } );
 		
 		} elsif ($use_indexer eq 'hyperestraier-native') {
 
 			# open Hyper Estraier database
-			use WebPAC::Output::EstraierNative;
+			require 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 } );
+			die "no longer supported";
 
 		} else {
 			$log->logdie("unknown use_indexer: $use_indexer");
 		}
 
-		$log->logide("can't continue without valid indexer") unless ($indexer);
+		$log->logdie("can't continue without valid indexer") unless ($indexer);
 	}
 
 
@@ -282,51 +302,92 @@
 	#
 	# now WebPAC::Store
 	#
-	my $abs_path = abs_path($0);
-	$abs_path =~ s#/[^/]*$#/#;
+	my $store = new WebPAC::Store({
+		debug => $debug,
+	});
 
-	my $db_path = $config->get('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");
+	#
+	# prepare output
+	#
+	my @outputs = force_array( $db_config->{output}, sub {
+		$log->error("Database $database doesn't have any outputs defined. Do you want to remove it from configuration?" );
+	} );
+
+	my @output_modules;
+
+	foreach my $output ( @outputs ) {
+
+#warn '## output = ',dump( $output );
+
+		my $module = $output->{module} || $log->logdie("need module in output section of $database");
+		$module = 'WebPAC::Output::' . $module unless $module =~ m/::/;
+	
+		$log->debug("loading output module $module");
+		eval "require $module";
+
+		# add database to arugemnts for output filter
+		$output->{database} = $database;
+		$output->{clean} = $clean;
+
+		$log->debug("calling $module->new(",dump( $output ),")");
+		my $out = new $module->new( $output );
+		if ( $out->init ) {
+			push @output_modules, $out;
+		} else {
+			$log->warn("SKIPPED $module");
+		}
 	}
 
-	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 {
+
+	my @inputs = force_array( $db_config->{input}, sub {
 		$log->info("database $database doesn't have inputs defined");
-	}
+	} );
+
+	if ( -e 'out/debug' ) {	# fixme flag?
+		my $out;
+		foreach my $i ( @inputs ) {
+			warn dump( $i );
+			next unless defined $i->{normalize};
+			warn dump( $i->{normalize} );
+			foreach my $normalize ( @{ $i->{normalize} } ) {
+				my $path = $normalize->{path};
+				$out .= qq/\n##\n## $path\n##\n\n/;
+				$out .= read_file( $path );
+			}
+		}
+		my $all = "out/debug/all-normalize.pl";
+		write_file( $all, $out );
+		warn "### all normalize for this input saved to: $all";
+	};
 
 	foreach my $input (@inputs) {
 
-		next if ($only_input && ($input->{name} !~ m#$only_input#i && $input->{type} !~ m#$only_input#i));
+		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')));
+		# FIXME check if input module exists
+		my $input_module = $input->{module};
 
-		my $input_module = $config->webpac('inputs')->{$type};
+		if ( ! $input_module ) {
+			if ( grep(/$type/, $config->webpac('inputs')) ) {
+				$input_module = $config->webpac('inputs')->{$type};
+			} else {
+				$log->logdie("I know only how to handle input types ", join(",", $config->webpac('inputs') ), " not '$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",
+		$log->info("working on $database/$input_name with $input_module on $input->{path}",
 			@lookups ? " creating lookups: ".join(", ", @lookups) : ""
 		);
 
@@ -345,15 +406,15 @@
 			stats => $stats,
 			modify_records => $input->{modify_records},
 			modify_file => $input->{modify_file},
+			input_config => $input,
 		);
 		$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...");
+			$log->warn("$database/$input_name has depriciated lookup definition, removing it...");
 			delete( $input->{lookup} );
 		}
 
-		my $lookup;
 		my $lookup_coderef;
 
 		if (@lookups) {
@@ -367,11 +428,10 @@
 				WebPAC::Normalize::data_structure(
 					row => $rec,
 					rules => $rules,
-					lookup => $lookup,
 					config => create_ds_config( $db_config, $database, $input, $mfn ),
 				);
 
-				warn "current lookup = ", dump($lookup) if ($lookup);
+				#warn "current lookup: ", dump(WebPAC::Normalize::_get_lookup());
 			};
 
 			WebPAC::Normalize::_set_lookup( undef );
@@ -380,143 +440,240 @@
 
 		};
 
+		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},
+				);
+			},
+
 		);
 
-		$log->debug("created following lookups: ", dump( WebPAC::Normalize::_get_lookup() ) );
+		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';
+			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 ",
+			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 ( $validate ) {
+				$validate->read_validate_file( $validate->fill_in( $validate_path, database => $database, input => $input_name ) ) if ( $validate_path );
+				$validate->read_validate_delimiters_file( $validate->fill_in( $validate_delimiters_path, database => $database, input => $input_name ) ) if ( $validate_delimiters_path );
+			}
+		}
 
-		if ($marc_normalize) {
-			@norm_array = ( {
-				path => $marc_normalize,
-				output => $marc_output || 'out/marc/' . $database . '-' . $input->{name} . '.marc',
-			} );
+		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,
+			);
 		}
 
-		foreach my $normalize (@norm_array) {
+		my $rules = $parser->normalize_rules($database,$input_name);
+		$log->logwarn("no normalize rules for $database/$input_name") unless $rules;
+
+		$log->debug("parsed normalize rules:\n$rules");
 
-			my $normalize_path = $normalize->{path} || $log->logdie("can't find normalize path in config");
+		# reset position in database
+		$input_db->seek(1);
 
-			$log->logdie("Found '$normalize_path' as normalization file which isn't supported any more!") unless ( $normalize_path =~ m!\.pl$!i );
+		# 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));
+					}
+				}
+			}
 
-			my $rules = read_file( $normalize_path ) or die "can't open $normalize_path: $!";
+			$log->debug("lookup_hash = ", sub { dump( $lookup_hash ) });
+		}
 
-			$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});
+		# setup input name for all output filters
+		foreach my $out ( @output_modules ) {
+			if ( $out->can('input') ) {
+				$out->input( $input_name );
+			} else {
+				$log->warn("output filter ",ref($out)," doesn't support input name");
+			}
+		}
 
-			# 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 ) {
 
-			foreach my $pos ( 0 ... $input_db->size ) {
+			my $row = $input_db->fetch || next;
 
-				my $row = $input_db->fetch || next;
+			$total_rows++;
 
-				my $mfn = $row->{'000'}->[0];
+			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 (! $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( $errors )
-						);
-					}
+			if ($validate) {
+				if ( my $errors = $validate->validate_rec( $row, $input_db->dump_ascii ) ) {
+					$log->error( "MFN $mfn validation error:\n",
+						$validate->report_error( $errors )
+					);
 				}
+				next;	# validation doesn't create any output
+			}
+
+			if ($rules) {
 
 				my $ds = WebPAC::Normalize::data_structure(
 					row => $row,
 					rules => $rules,
-					lookup => $lookup ? $lookup->lookup_hash : undef,
+					lookup => $lookup_hash,
 					config => create_ds_config( $db_config, $database, $input, $mfn ),
 					marc_encoding => 'utf-8',
+					load_row_coderef => sub {
+						my ($database,$input,$mfn) = @_;
+#warn "### load_row($database,$input,$mfn) from data_structure\n";
+						return $store->load_row(
+							database => $database,
+							input => $input,
+							id => $mfn,
+						);
+					},
 				);
 
-				$db->save_ds(
-					id => $mfn,
-					ds => $ds,
-					prefix => $input->{name},
-				) if ($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::marc_leader(),
-							row => $row,
-						);
-						$i++;
+				$log->debug("ds = ", sub { dump($ds) });
+
+				if ( $ds ) {
+
+					$store->save_ds(
+						database => $database,
+						input => $input_name,
+						id => $mfn,
+						ds => $ds,
+					) if !$stats;
+
+					$indexer->add(
+						id => "${input_name}/${mfn}",
+						ds => $ds,
+						type => $config->get($indexer_config)->{type},
+					) if $indexer;
+
+					foreach my $out ( @output_modules ) {
+						$out->add( $mfn, $ds ) if $out->can('add');
 					}
 
-					$log->info("Created $i instances of MFN $mfn\n") if ($i > 1);
+				} else {
+					$log->warn("record $pos didn't produce any output after normalization rules!") unless $marc;
 				}
-
-				$total_rows++;
 			}
 
-			if ($validate) {
-				my $errors = $validate->report;
-				if ($errors) {
-					$log->info("validation errors:\n$errors\n" );
-					print $report_fh "$errors\n" if ($report_fh);
+			if ($marc) {
+				my $i = 0;
+
+				while (my $fields = WebPAC::Normalize::MARC::_get_marc_fields( fetch_next => 1 ) ) {
+					$marc->add(
+						id => $mfn . ( $i ? "/$i" : '' ),
+						fields => $fields,
+						leader => WebPAC::Normalize::MARC::_get_marc_leader(),
+						row => $row,
+					);
+					$i++;
 				}
+
+				$log->info("Created $i instances of MFN $mfn\n") if ($i > 1);
 			}
 
-			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 ($report_fh);
+		}
+
+		if ($validate) {
+			my $errors = $validate->report;
+			if ($errors) {
+				$log->info("validation errors:\n$errors\n" );
+				print $report_fh "$errors\n" if ($report_fh);
 			}
 
-			# close MARC file
-			$marc->finish if ($marc);
+			print $report_fh "\nAll possible subfields/delimiter templates:\n", $validate->delimiters_templates( report => 1, current_input => 1 ), "\n\n";
 
-			# close report
-			close($report_fh) if ($report_fh)
+			# must be last thing that touches $validate for this input
+			$validate->reset;
+		}
+
+		if ($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'));
 
+	foreach my $out ( @output_modules ) {
+		$out->finish if $out->can('finish');
+	}
+
 	my $dt = time() - $start_t;
 	$log->info("$total_rows records ", $indexer ? "indexed " : "",
 		sprintf("in %.2f sec [%.2f rec/sec]",
@@ -539,6 +696,9 @@
 	$log->info("all parallel processes finished");
 }
 
+# save new delimiters if needed
+$validate->save_delimiters_templates if ( $validate_delimiters_path );
+
 #
 # handle links or merge after indexing
 #