cz norm

git-svn-id: https://ws10smt.googlecode.com/svn/trunk@685 ec762483-ff6d-05da-a07a-a48fb63a330f

1 files changed, 17 insertions, 0 deletions

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;
+}
+