Merged PR 2490: Fix chat vectorisation which was using the original words list and not the

Fix chat vectorisation which was using the original words list and not the tokenized one.

Related work items: #4549

Author: Pedro Teixeira

src/embedding/chat_process.py

@@ -55,8 +55,9 @@ async def chat_request(self, msg: ait_c.ChatRequestMessage):
                 EmbeddingChatProcessWorker.__spacy_wrapper.tokenizeSpacy(s)
                 for s in list
             ]
+
             try:
-                vecs = await self.w2v_client.get_vectors_for_words(words)
+                vecs = await self.w2v_client.get_vectors_for_words(x_tokens_testset[0])
             except aiohttp.client_exceptions.ClientConnectorError as exc:
                 self.logger.warn(
                     "Could not receive response from w2v service - {}".format(