From 36faf01602d28d5bb5f030e0e03c8e7dd2078445 Mon Sep 17 00:00:00 2001
From: redpony <redpony@ec762483-ff6d-05da-a07a-a48fb63a330f>
Date: Wed, 20 Oct 2010 17:19:17 +0000
Subject: cz norm

git-svn-id: https://ws10smt.googlecode.com/svn/trunk@685 ec762483-ff6d-05da-a07a-a48fb63a330f
---
 word-aligner/ortho-norm/cz.pl | 17 +++++++++++++++++
 1 file changed, 17 insertions(+)
 create mode 100755 word-aligner/ortho-norm/cz.pl

(limited to 'word-aligner')

diff --git a/word-aligner/ortho-norm/cz.pl b/word-aligner/ortho-norm/cz.pl
new file mode 100755
index 00000000..9d8bb07b
--- /dev/null
+++ b/word-aligner/ortho-norm/cz.pl
@@ -0,0 +1,17 @@
+#!/usr/bin/perl -w
+use strict;
+use utf8;
+
+while(<STDIN>) {
+  $_ = lc $_;
+  s/ al-/ al/g;
+  s/^al-/al/;
+  s/j/i/g;
+  s/dž/j/g;
+  s/š/sh/g;
+  s/á/a/g;
+  s/č/ch/g;
+  s/í/i/g;
+  print;
+}
+
-- 
cgit v1.2.3