Maintain the historical chat conversation per user;

chatbot-core/llm.py

@@ -22,22 +22,28 @@
 class LLMBase:
     """LLM Base containing common methods"""
 
+    mem_key: str = field(init=False)
+    conversation_max_window: int = field(init=False)
     qdrant_client: QdrantClient = field(init=False)
     llm_model: Any = field(init=False)
+    user_memory_mapping: dict = field(init=False)
     memory: Any = field(init=False)
     embedding_model: CustomEmbeddingsWrapper = field(init=False)
     rag_chain: Optional[Any] = None
 
     def __post_init__(self, mem_key: str = "chat_history", conversation_max_window: int = 3):
         self.llm_model = None
         self.qdrant_client = None
-        self.memory = None
+
+        self.mem_key = mem_key
+        self.conversation_max_window = conversation_max_window
 
         try:
             self.qdrant_client = QdrantClient(host=settings.QDRANT_DB_HOST, port=settings.QDRANT_DB_PORT)
         except Exception as e:
             raise Exception(f"Qdrant client is not properly setup. {str(e)}")
-        self.memory = ConversationBufferWindowMemory(k=conversation_max_window, memory_key=mem_key, return_messages=True)
+
+        self.user_memory_mapping = {}
 
         self.embedding_model = CustomEmbeddingsWrapper(
             url=settings.EMBEDDING_MODEL_URL,
@@ -113,23 +119,44 @@ def create_chain(self, db_collection_name: str):
         rag_chain = create_retrieval_chain(history_aware_retriever, chat_response_chain)
         return rag_chain
 
-    def execute_chain(self, query: str, db_collection_name: str = settings.QDRANT_DB_COLLECTION_NAME):
+    def execute_chain(self, user_id: str, query: str, db_collection_name: str = settings.QDRANT_DB_COLLECTION_NAME):
         """
         Executes the chain
         """
         if not self.rag_chain:
             self.rag_chain = self.create_chain(db_collection_name=db_collection_name)
 
-        response = self.rag_chain.invoke({"input": query, "chat_history": self.get_message_history()["chat_history"]})
-        self.memory.chat_memory.add_message(HumanMessage(content=query))
-        self.memory.chat_memory.add_message(AIMessage(content=response["answer"]))
+        if "user_id" not in self.user_memory_mapping:
+            self.user_memory_mapping[user_id] = ConversationBufferWindowMemory(
+                k=self.conversation_max_window, memory_key=self.mem_key, return_messages=True
+            )
+
+        memory = self.user_memory_mapping[user_id]
+
+        response = self.rag_chain.invoke(
+            {"input": query, "chat_history": self.get_message_history(user_id=user_id)["chat_history"]}
+        )
+        memory.chat_memory.add_message(HumanMessage(content=query))
+        memory.chat_memory.add_message(AIMessage(content=response["answer"]))
+        self.user_memory_mapping[user_id] = memory
+
         return response["answer"] if "answer" in response else ""
 
-    def get_message_history(self):
+    def get_message_history(self, user_id: str):
         """
         Returns the historical conversational data
         """
-        return self.memory.load_memory_variables({})
+        if "user_id" in self.user_memory_mapping:
+            return self.user_memory_mapping[user_id].load_memory_variables({})
+        return {}
+
+    def delete_message_history_by_user(self, user_id: str) -> bool:
+        """Deletes the message history based on user id"""
+        if "user_id" in self.user_memory_mapping:
+            del self.user_memory_mapping[user_id]
+            logger.info(f"Successfully delete the {user_id} conversational history.")
+            return True
+        return False
 
 
 @dataclass