--- trunk/run.pl 2006/05/15 13:32:18 508 +++ trunk/run.pl 2006/09/11 20:26:01 672 @@ -4,21 +4,26 @@ use Cwd qw/abs_path/; use File::Temp qw/tempdir/; -use Data::Dumper; use lib './lib'; use WebPAC::Common 0.02; -use WebPAC::Lookup; -use WebPAC::Input 0.03; +use WebPAC::Lookup 0.03; +use WebPAC::Input 0.11; use WebPAC::Store 0.03; -use WebPAC::Normalize::XML; -use WebPAC::Normalize::Set; +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 @@ -42,22 +47,55 @@ remove database and Hyper Estraier index before indexing -=item --only=database_name +=item --only=database_name/input_filter reindex just single database (legacy name is --one) +C is optional part which can be C +or C from input + =item --config conf/config.yml path to YAML configuration file -=item --force-set +=item --stats -force conversion C<< normalize->path >> in C from -C<.xml> to C<.pl> +disable indexing, modify_* in configuration and dump statistics about field +and subfield usage for each input -=item --stats +=item --validate path/to/validation_file + +turn on extra validation of imput records, see L + +=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) -dump statistics about used fields and subfields in each input +=item --only-links + +Create just links + +=item --merge + +Create merged index of databases which have links =back @@ -69,65 +107,174 @@ my $clean = 0; my $config = 'conf/config.yml'; my $debug = 0; -my $only_db_name; -my $force_set = 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_db_name, - "only=s" => \$only_db_name, + "one=s" => \$only_filter, + "only=s" => \$only_filter, "config" => \$config, - "debug" => \$debug, - "force-set" => \$force_set, + "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 = ",Dumper($config) if ($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 = $config->{use_indexer} || 'hyperestraier'; -$log->info("using $use_indexer indexing engine..."); +if ($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 %{ $config->{databases} }) { - next if ($only_db_name && $database !~ m/$only_db_name/i); + 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}; - - 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 } ); + my $cfg_name = $use_indexer; + $cfg_name =~ s/\-.*$//; - } else { - $log->logdie("unknown use_indexer: $use_indexer"); + 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 } ); + + $use_indexer = 'hyperestraier'; + + } 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); - $log->logide("can't continue without valid indexer") unless ($indexer); # # now WebPAC::Store @@ -138,10 +285,10 @@ my $db_path = $config->{webpac}->{db_path} . '/' . $database; if ($clean) { - $log->info("creating new database $database in $db_path"); + $log->info("creating new database '$database' in $db_path"); rmtree( $db_path ) || $log->warn("can't remove $db_path: $!"); } else { - $log->debug("working on $database in $db_path"); + $log->info("working on database '$database' in $db_path"); } my $db = new WebPAC::Store( @@ -168,141 +315,223 @@ 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 = new WebPAC::Lookup( - lookup_file => $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}; - $log->info("working on input '$input->{path}' [$input->{type}] using $input_module lookup '$input->{lookup}'"); + $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, - 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}, + 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 $n = new WebPAC::Normalize::XML( - # filter => { 'foo' => sub { shift } }, - db => $db, - lookup_regex => $lookup->regex, - lookup => $lookup, - prefix => $input->{name}, - ); + 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 $rules; - my $normalize_path = $input->{normalize}->{path}; + my @norm_array = ref($input->{normalize}) eq 'ARRAY' ? + @{ $input->{normalize} } : ( $input->{normalize} ); - if ($force_set) { - my $new_norm_path = $normalize_path; - $new_norm_path =~ s/\.xml$/.pl/; - if (-e $new_norm_path) { - $log->debug("--force-set replaced $normalize_path with $new_norm_path"); - $normalize_path = $new_norm_path; - } else { - $log->debug("--force-set failed on $new_norm_path, fallback to $normalize_path"); - } + if ($marc_normalize) { + @norm_array = ( { + path => $marc_normalize, + output => $marc_output || 'out/marc/' . $database . '-' . $input->{name} . '.marc', + } ); } - if ($normalize_path =~ m/\.xml$/i) { - $n->open( - tag => $input->{normalize}->{tag}, - xml_file => $normalize_path, - ); - } elsif ($normalize_path =~ m/\.(?:yml|yaml)$/i) { - $n->open_yaml( - path => $normalize_path, - tag => $input->{normalize}->{tag}, - ); - } elsif ($normalize_path =~ m/\.(?:pl)$/i) { - $n = undef; - $log->info("using WebPAC::Normalize::Set to process $normalize_path"); - $rules = read_file( $normalize_path ) or die "can't open $normalize_path: $!"; - } + foreach my $normalize (@norm_array) { - foreach my $pos ( 0 ... $input_db->size ) { + my $normalize_path = $normalize->{path} || $log->logdie("can't find normalize path in config"); - my $row = $input_db->fetch || next; + $log->logdie("Found '$normalize_path' as normalization file which isn't supported any more!") unless ( $normalize_path =~ m!\.pl$!i ); - my $mfn = $row->{'000'}->[0]; + my $rules = read_file( $normalize_path ) or die "can't open $normalize_path: $!"; - if (! $mfn || $mfn !~ m#^\d+$#) { - $log->warn("record $pos doesn't have valid MFN but '$mfn', using $pos"); - $mfn = $pos; - push @{ $row->{'000'} }, $pos; - } + $log->info("Using $normalize_path for normalization..."); - - my $ds; - if ($n) { - $ds = $n->data_structure($row); - } else { - $ds = WebPAC::Normalize::Set::data_structure( + 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) { + 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_hash, + 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); + ) 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++; } - $indexer->add( - id => $input->{name} . "/" . $mfn, - ds => $ds, - type => $config->{$use_indexer}->{type}, - ); + if ($validate) { + my $errors = $validate->report; + if ($errors) { + $log->info("validation errors:\n$errors\n" ); + print $report_fh "$errors\n" if ($report_fh); + } + } - $total_rows++; - } + 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); + } - $log->info("statistics of fields usage:\n", $input_db->stats) if ($stats); + # close MARC file + $marc->finish if ($marc); - }; + # close report + close($report_fh) if ($report_fh) + } + + } - eval { $indexer->finish } if ($indexer->can('finish')); + eval { $indexer->finish } if ($indexer && $indexer->can('finish')); my $dt = time() - $start_t; - $log->info("$total_rows records indexed in " . - sprintf("%.2f sec [%.2f rec/sec]", + $log->info("$total_rows records ", $indexer ? "indexed " : "", + sprintf("in %.2f sec [%.2f rec/sec]", $dt, ($total_rows / $dt) ) ); - # - # add Hyper Estraier links to other databases - # - if (ref($db_config->{links}) eq 'ARRAY') { - foreach my $link (@{ $db_config->{links} }) { - if ($use_indexer eq 'hyperestraier') { - $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}]"); - } - } + + # 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->(); + } +}