From 782fb27af98ed98256cc25c832131c59c8e9ce9c Mon Sep 17 00:00:00 2001 From: Chris Dyer Date: Mon, 5 Nov 2012 21:34:14 -0500 Subject: script to add sos/eos --- corpus/add-sos-eos.pl | 24 ++++++++++++++++++++++++ 1 file changed, 24 insertions(+) create mode 100755 corpus/add-sos-eos.pl (limited to 'corpus') diff --git a/corpus/add-sos-eos.pl b/corpus/add-sos-eos.pl new file mode 100755 index 00000000..5e2d44cb --- /dev/null +++ b/corpus/add-sos-eos.pl @@ -0,0 +1,24 @@ +#!/usr/bin/perl -w +use strict; + +while(<>) { + chomp; + my @fields = split / \|\|\| /; + my ($ff, $ee, $aa) = @fields; + die "Expected: foreign ||| target ||| alignments" unless scalar @fields == 3; + my @fs = split /\s+/, $ff; + my @es = split /\s+/, $ee; + my @as = split /\s+/, $aa; + my @oas = (); + push @oas, '0-0'; + my $flen = scalar @fs; + my $elen = scalar @es; + for my $ap (@as) { + my ($a, $b) = split /-/, $ap; + die "Bad format in: @as" unless defined $a && defined $b; + push @oas, ($a + 1) . '-' . ($b + 1); + } + push @oas, ($flen + 1) . '-' . ($elen + 1); + print " $ff ||| $ee ||| @oas\n"; +} + -- cgit v1.2.3