6 use Digest::SHA1 qw(sha1_base64);
9 my $conf = {}; # configuration hashref
12 my %fps = (); # records matching each fingerprint (and the lead)
13 my %recs = (); # fingerprints belonging to each record
15 open FP, '<', $ARGV[0] or die "Can't open input file: $!\n";
17 print "Loading and ranking fingerprints\n";
19 my @fields = split "\t", $_;
20 my $fp = populate_fingerprint(@fields);
21 rank_fingerprint($fp);
22 print "\r", ( int($i / $total) ), "% complete" unless ($i % 1000);
24 print "Writing matchset to disk\n";
29 sub populate_fingerprint {
31 my %fp = (); # zero fingerprint hash each time thru
33 # populate fp hash -- first the simple data
34 $fp{compact} = shift @fields;
35 $fp{json} = shift @fields;
36 $fp{id} = shift @fields;
37 # then smash everything else together, remove non-Roman characters, and
38 # generate a SHA1 hash to represent it
39 my $stripped = join('', @fields);
40 $stripped =~ s/[^A-Za-z0-9]//g;
41 $fp{sha1} = sha1_base64($stripped);
43 # populate records hash
44 $recs{ $fp{id} }{ $fp{sha1} } = {};
50 sub rank_fingerprint {
53 my $sha1 = $fp->{sha1};
55 my $islead = $recs{$id}{lead};
57 # only process records which haven't already been set as a sub
58 unless (defined $islead and $islead) {
59 unless ($fps{$sha1}) {
60 # haven't seen this fp before. create a new hashref with the current
62 $fps{$sha1} = { lead => { id => $id,
63 score => $fp->{compact} },
67 # have seen this fp. push record id onto matchlist
68 push @{ $fps{$sha1}{recs} }, $id;
69 if ($fp->{compact} > $fps{$sha1}{lead}{score}) {
70 # and set this record as lead if it scores higher than current lead
71 $recs{ $fps{$sha1}{lead}{id} }{lead} = 0; # unset current
72 $recs{ $id }{lead} = 1; # set new as lead
73 $fps{$sha1}{lead}{id} = $id;
74 $fps{$sha1}{lead}{score} = $fp->{compact};
83 Writes out a 2-column file of lead and subordinate records.
89 open OUT, '>', $conf->{output}
90 or die "Can't open ", $conf->{output}, "$!\n";
91 for my $id (keys %recs) {
92 next unless $recs{$id}{lead};
93 for my $sha1 ( keys %{$recs{$id}} ) {
94 for my $subid ( @{$fps{$sha1}{recs}} ) {
95 next if ($id == $subid);
96 next if defined $used{$subid};
98 print OUT "$id\t$subid\n";
108 # set mode on existing filehandles
109 binmode(STDIN, ':utf8');
111 my $rc = GetOptions( $c,
115 show_help() unless $rc;
116 show_help() if ($c->{help});
118 my @keys = keys %{$c};
119 show_help() unless (@ARGV and @keys);
120 for my $key ('output')
121 { push @missing, $key unless $c->{$key} }
123 print "Required option: ", join(', ', @missing), " missing!\n";
130 Usage is: compress_fingerprints -o OUTPUTFILE INPUTFILE