summaryrefslogtreecommitdiff
path: root/word-aligner/ortho-norm
diff options
context:
space:
mode:
authorAvneesh Saluja <asaluja@gmail.com>2013-03-28 18:28:16 -0700
committerAvneesh Saluja <asaluja@gmail.com>2013-03-28 18:28:16 -0700
commit5b8253e0e1f1393a509fb9975ba8c1347af758ed (patch)
tree1790470b1d07a0b4973ebce19192e896566ea60b /word-aligner/ortho-norm
parent2389a5a8a43dda87c355579838559515b0428421 (diff)
parentb203f8c5dc8cff1b9c9c2073832b248fcad0765a (diff)
fixed conflicts
Diffstat (limited to 'word-aligner/ortho-norm')
-rwxr-xr-xword-aligner/ortho-norm/ru.pl44
1 files changed, 44 insertions, 0 deletions
diff --git a/word-aligner/ortho-norm/ru.pl b/word-aligner/ortho-norm/ru.pl
new file mode 100755
index 00000000..34452d06
--- /dev/null
+++ b/word-aligner/ortho-norm/ru.pl
@@ -0,0 +1,44 @@
+#!/usr/bin/perl -w
+use strict;
+use utf8;
+binmode(STDIN,":utf8");
+binmode(STDOUT,":utf8");
+while(<STDIN>) {
+ $_ = uc $_;
+ s/А/a/g;
+ s/І/i/g;
+ s/Б/b/g;
+ s/В/v/g;
+ s/Г/g/g;
+ s/Д/d/g;
+ s/Е/e/g;
+ s/Ж/zh/g;
+ s/З/z/g;
+ s/И/i/g;
+ s/Й/i/g;
+ s/К/k/g;
+ s/Л/l/g;
+ s/М/m/g;
+ s/Н/n/g;
+ s/О/o/g;
+ s/П/p/g;
+ s/Р/r/g;
+ s/С/s/g;
+ s/Т/t/g;
+ s/У/u/g;
+ s/Ф/f/g;
+ s/Х/kh/g;
+ s/Ц/c/g;
+ s/Ч/ch/g;
+ s/Ш/sh/g;
+ s/Щ/shch/g;
+ s/Ъ//g;
+ s/Ы//g;
+ s/Ь//g;
+ s/Э/e/g;
+ s/Ю/yo/g;
+ s/Я/ya/g;
+ $_ = lc $_;
+ print;
+}
+