X-Git-Url: http://git.equinoxoli.org/?p=migration-tools.git;a=blobdiff_plain;f=fingerprinter;h=e228c6edeb6751ac791c2bbb9d367bc0222d7694;hp=3d224a275b21bd4c97b3cdd868ad4f6b2e5f228b;hb=209c24bd96a2534843af84d1c0176e6dbe80b82d;hpb=08fc326b43e6d0e2ef25a3460c7d1876b22bf0cd diff --git a/fingerprinter b/fingerprinter index 3d224a2..e228c6e 100755 --- a/fingerprinter +++ b/fingerprinter @@ -16,17 +16,13 @@ $| = 1; initialize($conf); open OF, '>', $conf->{output}; -binmode(OF, ':utf8'); open XF, '>', $conf->{exception}; -binmode(XF, ':utf8'); for my $file (@ARGV) { print XF "Processing $file\n"; - #open my $records, '<:utf8', $file; # This hack dosn't play nice with the use MARC::File::XML ( BinaryEncoding => 'utf-8' ); hack my $batch = undef; my $record = undef; - #$batch = MARC::Batch->new('XML', $records); # The other part of the hack - $batch = MARC::Batch->new('XML', $file); + $batch = MARC::Batch->new($conf->{marctype}, $file); $batch->strict_off(); $batch->warnings_off(); @@ -43,10 +39,14 @@ for my $file (@ARGV) { next; } + # populate and normalize marc $marc = populate_marc($record, $id); - $marc = normalize_marc($marc); + normalize_marc($marc); unless (marc_isvalid($marc)) { dump_exception($marc); next; } + + # if everything looks good, score it and dump fingerprints + score_marc($marc, $record); dump_fingerprints($marc); $scount++; progress_ticker(); } @@ -74,17 +74,30 @@ sub populate_marc { # date1, date2 my $my_008 = $record->field('008'); - $my_008 = $my_008->as_string() if ($my_008); - unless (defined $my_008 and length $my_008 == 40) - { print XF ">> Bad 008 length in rec ",$marc{id},"\n"; return \%marc } - $marc{date1} = substr($my_008,7,4) if ($my_008); - $marc{date2} = substr($my_008,11,4) if ($my_008); # UNUSED + $marc{tag008} = $my_008->as_string() if ($my_008); + if (defined $marc{tag008}) { + unless (length $marc{tag008} == 40) { + $marc{tag008} = $marc{tag008} . ('|' x (40 - length($marc{tag008}))); + print XF ">> Short 008 padded to ",length($marc{tag008})," at rec $count\n"; + } + $marc{date1} = substr($marc{tag008},7,4) if ($marc{tag008}); + $marc{date2} = substr($marc{tag008},11,4) if ($marc{tag008}); # UNUSED + } + unless ($marc{date1} and $marc{date1} =~ /\d{4}/) { + my $my_260 = $record->field('260'); + my $date1 = $my_260->subfield('c') if $my_260; + if (defined $date1 and $date1 =~ /\d{4}/) { + $marc{date1} = $date1; + $marc{fudgedate} = 1; + print XF ">> using 260c as date1 at rec $count\n"; + } + } # item_form if ( $marc{record_type} =~ /[gkroef]/ ) { # MAP, VIS - $marc{item_form} = substr($my_008,29,1) if ($my_008); + $marc{item_form} = substr($marc{tag008},29,1) if ($marc{tag008}); } else { - $marc{item_form} = substr($my_008,23,1) if ($my_008); + $marc{item_form} = substr($marc{tag008},23,1) if ($marc{tag008}); } # isbns @@ -103,9 +116,22 @@ sub populate_marc { } } - # "Accompanying material" (300e) - $marc{accomp} = $record->field('300')->subfield('e') - if $record->field('300'); + # oclc + $marc{oclc} = []; + push @{ $marc{oclc} }, $record->field('001')->as_string() + if ($record->field('001') and $record->field('003') and + $record->field('003')->as_string() =~ /OCo{0,1}LC/); + for ($record->field('035')) { + my $oclc = $_->subfield('a'); + push @{ $marc{oclc} }, $oclc + if (defined $oclc and $oclc =~ /\(OCoLC\)/ and $oclc =~/([0-9]+)/); + } + + # "Accompanying material" and check for "copy" (300) + if ($record->field('300')) { + $marc{accomp} = $record->field('300')->subfield('e'); + $marc{tag300a} = $record->field('300')->subfield('a'); + } # issn, lccn, title, desc, pages, pub, pubyear, edition $marc{lccn} = $record->field('010')->subfield('a') if $record->field('010'); @@ -181,6 +207,91 @@ sub marc_isvalid { } +=head2 score_marc + +Assign a score to the record based on various criteria. + +Score is constructed by pushing elements onto a list, via a dispatch +table. This allows order of fingerprints in the output file to be +varied. + +=cut + +sub score_marc { + my ($marc, $record) = @_; + my @score = (); + my $json = '{'; + + #---------------------------------- + # static criteria scoring + #---------------------------------- + $marc->{misc_score} = 999; + $marc->{age_score} = 999999999999; + + # -1 if 008 has been padded, -2 if it doesn't exist + if ($marc->{tag008}) + { $marc->{misc_score}-- if ($marc->{tag008} =~ /\|$/) } + else + { $marc->{misc_score} -= 2 } + # -1 if date has been pulled from 260 + $marc->{misc_score}-- if $marc->{fudgedate}; + # -1 if this is a copy record + $marc->{misc_score}-- + if (defined $marc->{tag300a} and $marc->{tag300a} =~ /copy/i); + + # subtract record id if we want older records to win + #$marc->{age_score} -= $marc->{id} unless ($conf->{newwins}); + # handle arbitrary adjustments + $marc->{age_score} = 1; + if ($conf->{'arbitrarily-lose-above'}) { + $marc->{age_score} = 0 + if ($marc->{id} >= $conf->{'arbitrarily-lose-above'}); + } + if ($conf->{'arbitrarily-lose-below'}) { + $marc->{age_score} = 0 + if ($marc->{id} <= $conf->{'arbitrarily-lose-below'}); + } + + #---------------------------------- + # dynamic calculated scoring + #---------------------------------- + my %scores_code = ( + oclc => sub { return $marc->{oclc}[0] ? 1 : 0 }, + dlc => sub { + if ($record->field('040') and $record->field('040')->subfield('a')) + { return scalar($record->subfield( '040', 'a')) =~ /dlc/io ? 1 : 0 } + else { return 0 } + }, + num_650 => sub { + if ($record->field('650')) { + # can't say "scalar $record->field('650')"; MARC::Record + # behaves differently in list/scalar contexts + my @tags = $record->field('650'); + return sprintf("%04d", scalar @tags) + } else { return '0000' } + }, + num_tags=> sub { return sprintf( '%04d', scalar( $record->fields ) ) }, + enc_lvl => sub { + my $enc = substr($record->leader, 17, 1) || 'u'; + my %levels = ( ' ' => 9, 1 => 8, 2 => 7, 3 => 6, 4 => 5, 5 => 4, + 6 => 3, 7 => 2, 8 => 1, 'u' => 0, 'z' => 0 ); + return $levels{$enc} || 0; + } + ); + + #---------------------------------- + # assemble and store scores + #---------------------------------- + for ( @{ $conf->{dyn_scores} } ) { + push @score, $scores_code{$_}->($marc, $record); + $json .= $_ . ':' . $score[-1] . ','; + } + $json .= 'misc:' . $marc->{misc_score} . '}'; + + my $compact = join('-', $marc->{age_score}, $marc->{misc_score}, @score); + $marc->{score} = "$compact\t$json"; +} + =head2 dump_fingerprints =cut @@ -188,60 +299,69 @@ sub marc_isvalid { sub dump_fingerprints { my ($marc) = @_; - if ($conf->{runtype} eq "primary") { - print OF join("\t",$marc->{id}, $marc->{item_form}, - $marc->{date1}, $marc->{record_type}, - $marc->{bib_lvl}, $marc->{title}), "\n"; - } else { - if ((scalar @{ $marc->{isbns} } > 0) && $marc->{pages}) { - # case a : isbn and pages - foreach my $isbn ( @{ $marc->{isbns}} ) { - print OF join("\t", $marc->{id}, "case a", - $marc->{item_form}, $marc->{date1}, - $marc->{record_type}, - $marc->{bib_lvl}, $marc->{title}, - $isbn, $marc->{pages}), "\n"; - } - } + if ($conf->{fingerprints}{baseline}) { + print OF join("\t", $marc->{score}, $marc->{id}, 'baseline', + $marc->{item_form}, $marc->{date1}, $marc->{record_type}, + $marc->{bib_lvl}, $marc->{title}), "\n"; + } - if ($marc->{edition}) { # case b : edition - print OF join("\t", $marc->{id}, "case b", - $marc->{item_form}, $marc->{date1}, - $marc->{record_type}, $marc->{bib_lvl}, - $marc->{title}, $marc->{edition}), "\n"; + if ($conf->{fingerprints}{oclc} and scalar @{$marc->{oclc} }) { + for (@{$marc->{oclc} }) { + print OF join("\t", $marc->{score}, $marc->{id}, "oclc", + $marc->{item_form}, $marc->{date1}, + $marc->{record_type}, $marc->{bib_lvl}, + $marc->{title}, $_, "\n"); } + } - if ($marc->{issn}) { # case c : issn - print OF join("\t", $marc->{id}, "case c", + if ($conf->{fingerprints}{isbn}) { + if ((scalar @{ $marc->{isbns} } > 0) and $marc->{pages}) { + foreach my $isbn ( @{ $marc->{isbns}} ) { + print OF join("\t", $marc->{score}, $marc->{id}, "isbn", $marc->{item_form}, $marc->{date1}, - $marc->{record_type}, $marc->{bib_lvl}, - $marc->{title}, $marc->{issn}), "\n"; + $marc->{record_type}, + $marc->{bib_lvl}, $marc->{title}, + $isbn, $marc->{pages}), "\n"; + } } + } - if ($marc->{lccn}) { # case d : lccn - print OF join("\t", $marc->{id}, "case d", - $marc->{item_form}, $marc->{date1}, - $marc->{record_type}, $marc->{bib_lvl}, - $marc->{title}, $marc->{lccn}) ,"\n"; - } + if ($conf->{fingerprints}{edition} and $marc->{edition}) { + print OF join("\t", $marc->{score}, $marc->{id}, "edition", + $marc->{item_form}, $marc->{date1}, + $marc->{record_type}, $marc->{bib_lvl}, + $marc->{title}, $marc->{edition}), "\n"; + } - if ($marc->{accomp}) { # case e : accomp - print OF join("\t", $marc->{id}, "case d", - $marc->{item_form}, $marc->{date1}, - $marc->{record_type}, $marc->{bib_lvl}, - $marc->{title}, $marc->{accomp}) ,"\n"; - } + if ($conf->{fingerprints}{issn} and $marc->{issn}) { + print OF join("\t", $marc->{score}, $marc->{id}, "issn", + $marc->{item_form}, $marc->{date1}, + $marc->{record_type}, $marc->{bib_lvl}, + $marc->{title}, $marc->{issn}), "\n"; + } - # case z : author, publisher, pubyear, pages - if ($marc->{author} and $marc->{publisher} and $marc->{pubyear} - and $marc->{pages}) { - print OF join("\t", $marc->{id}, "case e", - $marc->{item_form}, $marc->{date1}, - $marc->{record_type}, $marc->{bib_lvl}, - $marc->{title}, $marc->{author}, - $marc->{publisher}, $marc->{pubyear}, - $marc->{pages}), "\n"; - } + if ($conf->{fingerprints}{lccn} and $marc->{lccn}) { + print OF join("\t", $marc->{score}, $marc->{id}, "lccn", + $marc->{item_form}, $marc->{date1}, + $marc->{record_type}, $marc->{bib_lvl}, + $marc->{title}, $marc->{lccn}) ,"\n"; + } + + if ($conf->{fingerprints}{accomp} and $marc->{accomp}) { + print OF join("\t", $marc->{score}, $marc->{id}, "accomp", + $marc->{item_form}, $marc->{date1}, + $marc->{record_type}, $marc->{bib_lvl}, + $marc->{title}, $marc->{accomp}) ,"\n"; + } + + if ($conf->{fingerprints}{authpub} and $marc->{author} and + $marc->{publisher} and $marc->{pubyear} and $marc->{pages}) { + print OF join("\t", $marc->{score}, $marc->{id}, "authpub", + $marc->{item_form}, $marc->{date1}, + $marc->{record_type}, $marc->{bib_lvl}, + $marc->{title}, $marc->{author}, + $marc->{publisher}, $marc->{pubyear}, + $marc->{pages}), "\n"; } } @@ -259,16 +379,16 @@ sub dump_exception { print XF "Undefined record at line $count; likely bad XML\n"; return; } - print XF "Record ", $marc->{id}, " did not make the cut: "; - print XF "Missing item_form. " unless ($marc->{item_form}); + print XF "Record ", $marc->{id}, " excluded: "; + print XF "missing item_form; " unless ($marc->{item_form}); unless (defined $marc->{date1}) - { print XF "Missing date1. " } + { print XF "missing date1; " } else - { print XF "Invalid date1: ", $marc->{date1}, " " + { print XF "invalid date1: '", $marc->{date1}, "'; " unless ($marc->{date1} =~ /\d{4}/); } - print XF "Missing record_type. " unless ($marc->{record_type}); - print XF "Missing bib_lvl. " unless ($marc->{bib_lvl}); - print XF "Missing title. " unless ($marc->{title}); + print XF "missing record_type; " unless ($marc->{record_type}); + print XF "missing bib_lvl; " unless ($marc->{bib_lvl}); + print XF "missing title " unless ($marc->{title}); print XF "\n"; } @@ -288,37 +408,70 @@ sub initialize { binmode(STDIN, ':utf8'); my $rc = GetOptions( $c, - 'incoming', - 'incumbent', 'exception|x=s', 'output|o=s', - 'runtype|r=s', + 'prefix|p=s', + 'marctype|m=s', 'subfield|s=s', 'tag|t=s', + 'fingerprints=s', + 'scores=s', + 'arbitrarily-lose-above=i', + 'arbitrarily-lose-below=i', + 'newwins', 'quiet|q', 'help|h', ); show_help() unless $rc; show_help() if ($c->{help}); - # set defaults if told to do so - if ($c->{incoming}) { - $c->{tag} = 903 unless defined $c->{tag}; - $c->{subfield} = 'a' unless defined $c->{subfield}; - $c->{output} = 'incoming.fp' unless defined $c->{output}; - $c->{exception} = 'incoming.ex' unless defined $c->{exception}; - $c->{runtype} = 'full' unless defined $c->{runtype}; - } elsif ($c->{incumbent}) { - $c->{tag} = 901 unless defined $c->{tag}; - $c->{subfield} = 'c' unless defined $c->{subfield}; - $c->{output} = 'incumbent.fp' unless defined $c->{output}; - $c->{exception} = 'incumbent.ex' unless defined $c->{exception}; - $c->{runtype} = 'full' unless defined $c->{runtype}; + # check fingerprints list for validity + if ($c->{fingerprints}) { + my %fps = (); + my %valid_fps = ( oclc => 1, isbn => 1, issn => 1, lccn => 1, + edition => 1, accomp => 1, authpub => 1, + baseline => 1, crap => 1, + ); + for (split /,/, $c->{fingerprints}) { + die "Invalid fingerprint '$_'\n" unless $valid_fps{$_}; + $fps{$_} = 1; + } + $c->{fingerprints} = \%fps + } else { + $c->{fingerprints} = {oclc => 1, isbn => 1, edition => 1, issn => 1, + lccn => 1, accomp => 1, authpub => 1}; + } + + # check scores list for validity + if ($c->{scores}) { + my %scores = (); + my %valid_scores = ( oclc => 1, dlc => 1, num_650 => 1, + num_tags => 1, enc_lvl => 1, + ); + for (split /,/, $c->{scores}) { + die "Invalid score mode '$_'\n" unless $valid_scores{$_}; + $scores{$_} = 1; + } + $c->{dyn_scores} = [split /,/, $c->{scores}]; + $c->{scores} = \%scores; + } else { + $c->{scores} = {oclc => 1, dlc => 1, num_650 => 1, + num_tags => 1, enc_lvl => 1}; + $c->{dyn_scores} = [ qw/oclc dlc num_650 num_tags enc_lvl/ ]; + } + + # set defaults + $c->{tag} = 903 unless defined $c->{tag}; + $c->{subfield} = 'a' unless defined $c->{subfield}; + $c->{marctype} = 'XML' unless defined $c->{marctype}; + if ($c->{prefix}) { + $c->{output} = join('.',$c->{prefix},'fp'); + $c->{exception} = join('.',$c->{prefix},'fp','ex'); } my @keys = keys %{$c}; show_help() unless (@ARGV and @keys); - for my $key ('runtype', 'tag', 'subfield', 'output', 'exception') + for my $key ('tag', 'subfield', 'output', 'exception') { push @missing, $key unless $c->{$key} } if (@missing) { print "Required option: ", join(', ', @missing), " missing!\n"; @@ -348,19 +501,30 @@ sub show_help { print < Req'd Arguments - --runtype=(primary|full) -r Do 'primary' or 'full' fingerprinting - --tag=N -t Which tag to use - --subfield=X -s Which subfield to use - --output= -o Output filename - --exceptions= -x Exception report filename + --output= -o Output filename + --exceptions= -x Exception report filename + or + --prefix=> -p Shared prefix for output/exception files. Will + produce PREFIX.fp and PREFIX.fp.ex Options - --incoming Set -r to 'full'; -t, -s, -o, -x to incoming defaults - --incumbent Set -r to 'full'; -t, -s, -o, -x to incumbent defaults - - Example: '$0 --incoming' is equivalent to - '$0 -r full -t 903 -s a -o incoming.fp -x incoming.ex' - - --quiet -q Don't write status messages to STDOUT + --tag=N -t Which tag to use (default 903) + --subfield=X -s Which subfield to use (default 'a') + --quiet -q Don't write status messages to STDOUT + + --fingerprints=LIST Fingerprints to generate, comma separated + Default: oclc,isbn,edition,issn,lccn,accomp,authpub + Others: baseline + + --scores=LIST Scores to calculate, comma separated + Default: oclc,dlc,num_650,num_tags,enc_level + --newwins New record IDs score higher (default is old wins) + --arbitrarily-lose-above + --arbitrarily-lose-below + --arbitrarily-decrease-score-by + Modify fingerprint scoring of records whose EG id is above or below a + given value, inclusive (so 5 is <= 5 or >= 5) such that they lose. + + --marctype=TYPE Defaults to 'XML' HELP exit 1; }