src/cm/utils/string_utils.py
changeset 155 f436e2fef076
parent 149 0f2c5744b39b
child 175 4f072edc51a1
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/src/cm/utils/string_utils.py	Tue Feb 09 22:20:08 2010 +0100
@@ -0,0 +1,15 @@
+import chardet
+
+def to_unicode(input):
+    if type(input) == str:
+        res = None
+        for encoding in [chardet.detect(input)['encoding'], 'utf8', 'latin1']:
+            try:
+                res = unicode(input, encoding)
+                break;
+            except UnicodeDecodeError:
+                pass
+        if not res:
+            raise Exception('UnicodeDecodeError: could not decode')
+        return res
+    return input
\ No newline at end of file