Moving "src-tokenized" under "translation"

No need to repeat it every time in n-best lists

API.md

@@ -66,7 +66,7 @@ The response structure includes:
 
 -   *errorCode* (number): error code (see below), returned always
 -   *errorMessage* (string): a description of the error, returned always
--   **translation** (list of structures): contains the translated data as a list of 
+-   *translation* (list of structures): contains the translated data as a list of 
     sentences, each of which contains the following:
 
     -   *translated* (list of structures): list of translations of one sentence
@@ -79,11 +79,15 @@ The response structure includes:
         -   *rank* (integer): rank of the translation option (ranked according
             to the scoring, counting from 0; this may be omitted as the rank
             is given by the order in the *translated* list)
-        -   *src-tokenized* (string): tokenization of the source sentence (for
-            the translation of multiple sentences and/or alignment information)
         -   Further items if alignment information or multiple translation
             options are requested (see below).
 
+    -   *src-tokenized* (string): tokenization of the source sentence (for
+        the translation of multiple sentences and/or alignment information)
+
+    -   *src* (string): source sentence in its original form (optional)
+
+
 -   *translationId*: string, globally unique ID of the transaction 
     (may be omitted)