From da222df300e4f87ad185a7decbf119ad56aa34e0 Mon Sep 17 00:00:00 2001
From: Chris Dyer <redpony@gmail.com>
Date: Fri, 29 Jan 2010 15:56:59 +0000
Subject: word aligner checkin

---
 word-aligner/ortho-norm/ur.pl | 34 ++++++++++++++++++++++++++++++++++
 1 file changed, 34 insertions(+)
 create mode 100755 word-aligner/ortho-norm/ur.pl

(limited to 'word-aligner/ortho-norm/ur.pl')

diff --git a/word-aligner/ortho-norm/ur.pl b/word-aligner/ortho-norm/ur.pl
new file mode 100755
index 00000000..d125b744
--- /dev/null
+++ b/word-aligner/ortho-norm/ur.pl
@@ -0,0 +1,34 @@
+#!/usr/bin/perl -w
+use strict;
+use utf8;
+
+binmode(STDIN, ":utf8");
+binmode(STDOUT, ":utf8");
+
+my @out = ();
+while(<STDIN>) {
+  chomp;
+  my @words = split /\s+/;
+  for my $of (@words) {
+    if (length($of) > 1 && !($of =~ /\d/)) {
+      $of =~ s/\$/sh/g;
+    }
+    $of =~ s/([a-z])\~/$1$1/g;
+    $of =~ s/E/'/g;
+    $of =~ s/^Aw/o/g;
+    $of =~ s/\|/a/g;
+    $of =~ s/@/h/g;
+    $of =~ s/c/ch/g;
+    $of =~ s/x/kh/g;
+    $of =~ s/\*/dh/g;
+    $of =~ s/w/o/g;
+    $of =~ s/Z/dh/g;
+    $of =~ s/y/i/g;
+    $of =~ s/Y/a/g;
+    $of = lc $of;
+    $of = lc $of;
+    push @out, $of;
+  }
+  print "@out\n";
+}
+
-- 
cgit v1.2.3