diff options
author | Chris Dyer <cdyer@cs.cmu.edu> | 2012-01-20 15:35:47 -0500 |
---|---|---|
committer | Chris Dyer <cdyer@cs.cmu.edu> | 2012-01-20 15:35:47 -0500 |
commit | f0bdd4de6455855d705d9056deb2e90c999dc740 (patch) | |
tree | ac4ebe783d4f613c6f0b01a28210b014f7ba7ed6 /word-aligner/stemmers | |
parent | d156a65ac638b574abfabfd78c949e122faada5d (diff) |
'pseudo model 2' that strictly favors a diagonal, with tunable parameters for p(null) and how sharp/flat the alignment distribution is around the diagonal
Diffstat (limited to 'word-aligner/stemmers')
-rwxr-xr-x | word-aligner/stemmers/ar.pl | 39 |
1 files changed, 39 insertions, 0 deletions
diff --git a/word-aligner/stemmers/ar.pl b/word-aligner/stemmers/ar.pl new file mode 100755 index 00000000..c85e883a --- /dev/null +++ b/word-aligner/stemmers/ar.pl @@ -0,0 +1,39 @@ +#!/usr/bin/perl -w + +use strict; +use utf8; + +binmode(STDIN, ":utf8"); +binmode(STDOUT,":utf8"); + +my $vocab = undef; +if (scalar @ARGV > 0) { + die "Only allow --vocab" unless ($ARGV[0] eq '--vocab' && scalar @ARGV == 1); + $vocab = 1; +} + +my %dict; +while(<STDIN>) { + chomp; + my @words = split /\s+/; + my @out = (); + for my $w (@words) { + my $tw = $dict{$w}; + if (!defined $tw) { + my $el = 4; + if ($w =~ /^(.st|.n|Al)/) { $el+=2; } + if ($w =~ /^(y|t|n)/) { $el++; } + if ($el > length($w)) { $el = length($w); } + $tw = substr $w, 0, $el; + $dict{$w} = $tw; + } + push @out, $tw; + } + if ($vocab) { + die "Expected exactly one word per line with --vocab: $_" unless scalar @out == 1; + print "$_ @out\n"; + } else { + print "@out\n"; + } +} + |