--- trunk/run.pl 2007/05/23 20:04:13 833 +++ trunk/run.pl 2009/05/19 11:55:30 1185 @@ -9,9 +9,9 @@ use WebPAC::Common 0.02; use WebPAC::Parser 0.08; use WebPAC::Input 0.16; -use WebPAC::Store 0.14; +use WebPAC::Store 0.15; use WebPAC::Normalize 0.22; -use WebPAC::Output::TT; +#use WebPAC::Output::TT; use WebPAC::Validate 0.11; use WebPAC::Output::MARC; use WebPAC::Config; @@ -22,8 +22,8 @@ use Data::Dump qw/dump/; use Storable qw/dclone/; use Pod::Usage qw/pod2usage/; +use LWP::Simple qw//; -use Proc::Queue size => 1; use POSIX ":sys_wait_h"; # imports WNOHANG =head1 NAME @@ -68,6 +68,14 @@ turn on extra validation of imput records, see L +You can use special variables C<$database> and $C<$input> in this parametar +like C<--validate 'conf/validate/$database-$input'> to construct filename + +=item --validate-delimiters path/to/validate_delimiters_file + +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. + =item --marc-generate Generate MARC file. This will automatically be on if file contains C directives. @@ -95,6 +103,10 @@ Create merged index of databases which have links +=item --mirror http://www.example.com + +Tries to download input path files from mirror URI + =back =cut @@ -108,12 +120,14 @@ my $only_filter; my $stats = 0; my $validate_path; +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 $mirror; my $help; my $log = _new WebPAC::Common()->_get_logger(); @@ -124,23 +138,29 @@ "clean" => \$clean, "one=s" => \$only_filter, "only=s" => \$only_filter, - "config" => \$config_path, + "config=s" => \$config_path, "debug+" => \$debug, "stats" => \$stats, "validate=s" => \$validate_path, + "validate-delimiters=s" => \$validate_delimiters_path, "marc-generate!" => \$marc_generate, "marc-lint!" => \$marc_lint, "marc-dump!" => \$marc_dump, "parallel=i" => \$parallel, "only-links!" => \$only_links, "merge" => \$merge, + "mirror=s" => \$mirror, "help" => \$help, ); +$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); @@ -163,26 +183,29 @@ $log->info("created merge batch file $estcmd_path"); } - my $validate; $validate = new WebPAC::Validate( - path => $validate_path, delimiters => $config->webpac('delimiters'), -) if ($validate_path); - +) if ($validate_path || $validate_delimiters_path); my $use_indexer = $config->use_indexer; $stats ||= $validate; if ($stats) { $log->debug("disabled indexing for stats collection"); $use_indexer = undef; -} else { +} elsif ( $use_indexer ) { $log->info("using $use_indexer indexing engine..."); } # 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 $start_t = time(); @@ -190,6 +213,8 @@ my @links; if ($parallel) { + eval 'use Proc::Queue size => 1;'; + die $@ if $@; $log->info("Using $parallel processes for speedup"); Proc::Queue::size($parallel); } @@ -203,9 +228,9 @@ 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 ($parallel) { @@ -234,27 +259,24 @@ if ($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); } @@ -288,52 +310,97 @@ # # now WebPAC::Store # - my $abs_path = abs_path($0); - $abs_path =~ s#/[^/]*$#/#; + my $store = new WebPAC::Store({ + debug => $debug, + }); - my $db_path = $config->webpac('db_path'); - if ($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 $store = new WebPAC::Store( - path => $db_path, - debug => $debug, - ); + 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 $module"); + } + } # # now, iterate through input formats # - my @inputs; - if (ref($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) { my $input_name = $input->{name} || $log->logdie("input without a name isn't valid: ",dump($input)); + if ( $input->{skip} ) { + $log->info("skip $input_name"); + next; + } + next if ($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 $input->{path}", @lookups ? " creating lookups: ".join(", ", @lookups) : "" ); @@ -343,15 +410,31 @@ delete($input->{modify_file}); } + if ( $mirror ) { + my $path = $input->{path} || die "no input path in ",dump( $input ); + + my $base = $path; + $base =~ s{/[^/]+$}{}; + mkpath $base unless -e $base; + + my $rc = LWP::Simple::mirror( "$mirror/$path", $path ); + if (LWP::Simple::is_error( $rc )) { + die "can't mirror $mirror/$path -> $path [$rc]"; + } else { + $log->info( "mirror ", $path, " [$rc] ", -s $path, " bytes" ); + } + + } + my $input_db = new WebPAC::Input( module => $input_module, - encoding => $config->webpac('webpac_encoding'), limit => $limit || $input->{limit}, offset => $offset, recode => $input->{recode}, 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); @@ -389,7 +472,7 @@ my $maxmfn = $input_db->open( path => $input->{path}, - code_page => $input->{encoding}, # database encoding + input_encoding => $input->{encoding}, # database encoding lookup_coderef => $lookup_coderef, lookup => $lookup_jar, %{ $input }, @@ -430,24 +513,37 @@ my $report_fh; if ($stats || $validate) { - my $path = "out/report/${database}-${input_name}.txt"; + my $out_report = 'out/report'; # FIXME move to config + mkpath $out_report unless -e $out_report; + 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"); + + 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 ); + } } my $marc; if ($marc_generate && $parser->have_rules( 'marc', $database, $input_name )) { + + my $out_marc = 'out/marc'; # FIXME move to config + mkpath $out_marc unless -e $out_marc; + $marc = new WebPAC::Output::MARC( - path => "out/marc/${database}-${input_name}.marc", + path => "$out_marc/${database}-${input_name}.marc", lint => $marc_lint, dump => $marc_dump, ); } - my $rules = $parser->normalize_rules($database,$input_name) || $log->logdie("no normalize rules found for $database/$input_name"); + 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"); # reset position in database @@ -483,6 +579,16 @@ } + # 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"); + } + } + + foreach my $pos ( 0 ... $input_db->size ) { my $row = $input_db->fetch || next; @@ -491,7 +597,7 @@ my $mfn = $row->{'000'}->[0]; - if (! $mfn || $mfn !~ m#^\d+$#) { + if (! $mfn || $mfn !~ m{^\d+$}) { $log->warn("record $pos doesn't have valid MFN but '$mfn', using $pos"); $mfn = $pos; push @{ $row->{'000'} }, $pos; @@ -507,45 +613,59 @@ next; # validation doesn't create any output } - my $ds = WebPAC::Normalize::data_structure( - row => $row, - rules => $rules, - lookup => $lookup_hash, - config => create_ds_config( $db_config, $database, $input, $mfn ), - marc_encoding => 'utf-8', - load_row_coderef => sub { - my ($database,$input,$mfn) = @_; - return $store->load_row( + if ($rules) { + + my $ds = WebPAC::Normalize::data_structure( + row => $row, + rules => $rules, + 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, + ); + }, + ); + + $log->debug("ds = ", sub { dump($ds) }); + + if ( $ds ) { + + $store->save_ds( database => $database, - input => $input, + input => $input_name, id => $mfn, - ); - }, - ); + ds => $ds, + ) if !$stats; - $log->debug("ds = ", sub { dump($ds) }) if ($ds); + $indexer->add( + id => "${input_name}/${mfn}", + ds => $ds, + type => $config->get($indexer_config)->{type}, + ) if $indexer; - $store->save_ds( - database => $database, - input => $input_name, - id => $mfn, - ds => $ds, - ) if ($ds && !$stats); - - $indexer->add( - id => "${input_name}/${mfn}", - ds => $ds, - type => $config->get($indexer_config)->{type}, - ) if ($indexer && $ds); + foreach my $out ( @output_modules ) { + $out->add( $mfn, $ds ) if $out->can('add'); + } + + } else { + $log->warn("record $pos didn't produce any output after normalization rules!") unless $marc; + } + } if ($marc) { my $i = 0; - while (my $fields = WebPAC::Normalize::_get_marc_fields( fetch_next => 1 ) ) { + while (my $fields = WebPAC::Normalize::MARC::_get_marc_fields( fetch_next => 1 ) ) { $marc->add( id => $mfn . ( $i ? "/$i" : '' ), fields => $fields, - leader => WebPAC::Normalize::_get_marc_leader(), + leader => WebPAC::Normalize::MARC::_get_marc_leader(), row => $row, ); $i++; @@ -553,6 +673,7 @@ $log->info("Created $i instances of MFN $mfn\n") if ($i > 1); } + } if ($validate) { @@ -562,7 +683,10 @@ print $report_fh "$errors\n" if ($report_fh); } - print $report_fh "\nAll possible subfields/delimiter templates:\n", $validate->delimiters_templates, "\n\n"; + print $report_fh "\nAll possible subfields/delimiter templates:\n", $validate->delimiters_templates( report => 1, current_input => 1 ), "\n\n"; + + # must be last thing that touches $validate for this input + $validate->reset; } if ($stats) { @@ -575,11 +699,14 @@ $marc->finish if ($marc); # close report - close($report_fh) if ($report_fh) - + close($report_fh) if ($report_fh); } - eval { $indexer->finish } if ($indexer && $indexer->can('finish')); + $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 " : "", @@ -603,6 +730,9 @@ $log->info("all parallel processes finished"); } +# save new delimiters if needed +$validate->save_delimiters_templates if ( $validate_delimiters_path ); + # # handle links or merge after indexing #