summaryrefslogtreecommitdiff
path: root/word-aligner/stemmers/mg.pl
diff options
context:
space:
mode:
authorPatrick Simianer <p@simianer.de>2011-11-03 00:38:49 +0100
committerPatrick Simianer <p@simianer.de>2011-11-03 00:38:49 +0100
commitbcfb20c1e1e9434ae04a30240859db6a353457d0 (patch)
tree4611721111f4fef5bd132ed3236b820c4e16fa9c /word-aligner/stemmers/mg.pl
parent4df1aea9e6edce1c79059d3310d55947be0a302e (diff)
parent6de8f58cd13813bf33af4903bf386439683c0fd6 (diff)
Merge remote-tracking branch 'upstream/master'
Diffstat (limited to 'word-aligner/stemmers/mg.pl')
-rwxr-xr-xword-aligner/stemmers/mg.pl39
1 files changed, 39 insertions, 0 deletions
diff --git a/word-aligner/stemmers/mg.pl b/word-aligner/stemmers/mg.pl
new file mode 100755
index 00000000..2f79a94e
--- /dev/null
+++ b/word-aligner/stemmers/mg.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 $el = 5;
+ if ($w =~ /(ndz|ndr|nts|ntr)/) { $el++; }
+ if ($w =~ /^(mp|mb|nd)/) { $el++; }
+ if ($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";
+ }
+}
+