From eed3ff80e00b4b4fd0f0c7d5758ba5638f3bc7fe Mon Sep 17 00:00:00 2001
From: Josip Pisoj <josip@pisoj.com>
Date: Mon, 8 Mar 2010 18:12:13 +0000
Subject: [PATCH] bug #3662: SoundEx works only for ascii characters

svn: r14700
---
 src/soundex.py | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/src/soundex.py b/src/soundex.py
index c9c00d35f..19ee3f988 100644
--- a/src/soundex.py
+++ b/src/soundex.py
@@ -28,6 +28,7 @@ Provide soundex calculation
 #
 #-------------------------------------------------------------------------
 import string
+import unicodedata
 
 #-------------------------------------------------------------------------
 #
@@ -46,7 +47,8 @@ TABLE  = string.maketrans('ABCDEFGIJKLMNOPQRSTUVXYZ',
 def soundex(strval):
     "Return the soundex value to a string argument."
 
-    strval = strval.upper().strip()
+    strval = unicodedata.normalize('NFKD', 
+                    strval.upper().strip()).encode('ASCII', 'ignore')
     if not strval:
         return "Z000"
     strval = strval.encode('iso-8859-1')