Fixes non UTF-8 surrogateescapes (#612)

Fixes non UTF-8 surrogateescapes 

Surrogate escapes in Unicode (non UTF-8 encoding) will be properly escaped with backslashes when encountered, versus breaking the transport layer.
Fixes #611

Author: tylerjharden

elasticsearch/transport.py

@@ -292,7 +292,7 @@ def perform_request(self, method, url, params=None, body=None):
 
         if body is not None:
             try:
-                body = body.encode('utf-8')
+                body = body.encode('utf-8', 'surrogatepass')            
             except (UnicodeDecodeError, AttributeError):
                 # bytes/str - no need to re-encode
                 pass

test_elasticsearch/test_transport.py

@@ -105,6 +105,13 @@ def test_body_bytes_get_passed_untouched(self):
         self.assertEquals(1, len(t.get_connection().calls))
         self.assertEquals(('GET', '/', None, body), t.get_connection().calls[0][0])
 
+    def test_body_surrogates_replaced_encoded_into_bytes(self):
+        t = Transport([{}], connection_class=DummyConnection)
+
+        t.perform_request('GET', '/', body='你好\uda6a')
+        self.assertEquals(1, len(t.get_connection().calls))
+        self.assertEquals(('GET', '/', None, b'\xe4\xbd\xa0\xe5\xa5\xbd\xed\xa9\xaa'), t.get_connection().calls[0][0])        
+                
     def test_kwargs_passed_on_to_connections(self):
         t = Transport([{'host': 'google.com'}], port=123)
         self.assertEquals(1, len(t.connection_pool.connections))