fix(backend): address PR review — drop debug print, mirror intent onto Chat

Removes a leftover debug print in process_message that flooded logs with
every Redis stream payload (potentially exposing user prompts).

Also restores the pre-PR behavior where Chat.chat_intent tracks the latest
message's intent. _resolve_chat_intent now mirrors the persisted intent
onto chat_message.chat as well, so the PastConversations badge (which
reads chat.chat_intent_name from ChatSerializer) populates correctly for
chats created after this PR.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

Author: tahierhussain

backend/backend/application/context/llm_context.py

@@ -78,6 +78,10 @@ def _resolve_chat_intent(self, chat_message_id: str, data: dict, content: Any) -
             chat_message = ChatMessage.objects.get(chat_message_id=chat_message_id)
             chat_message.chat_intent = chat_intent
             chat_message.save(update_fields=["chat_intent"])
+            # Mirror the pre-PR behavior where Chat.chat_intent tracked the
+            # latest message's intent (used by PastConversations badge).
+            chat_message.chat.chat_intent = chat_intent
+            chat_message.chat.save(update_fields=["chat_intent"])
         except Exception as e:
             logging.error(f"Failed to persist chat_intent={intent_name}: {e}")
 
@@ -93,7 +97,6 @@ def process_message(
         discussion_status: str
     ):
         data = json.loads(payload["data"])
-        print("\n=============\n", data, "\n==========\n")
         if payload.get("type") == "status" and payload.get("status") == "failed":
             payload = json.loads(payload["data"])
             if payload and "error_message" in payload: