update 1

agent.py

@@ -1,18 +1,26 @@
+import warnings
+warnings.filterwarnings('ignore', category=FutureWarning)
+warnings.filterwarnings('ignore', category=DeprecationWarning)
+
 print("agent: > thu vien")
 print("-------------")
 from langchain.agents import AgentExecutor, create_tool_calling_agent
-from langchain.memory import ConversationBufferMemory
-from langchain.prompts import MessagesPlaceholder
-from langchain.prompts import ChatPromptTemplate
+from langchain_core.prompts import ChatPromptTemplate, MessagesPlaceholder
+from langchain_core.runnables.history import RunnableWithMessageHistory
+from langchain_community.chat_message_histories import ChatMessageHistory
 from tools import (model_gemini,
                    get_weather,
                    toeic_question)
-                
 
 tools = [get_weather, toeic_question]
 
-# Pull the prompt template from the hub
-# prompt = hub.pull("hwchase17/openai-tools-agent")
+# Store để lưu chat history
+store = {}
+
+def get_session_history(session_id: str):
+    if session_id not in store:
+        store[session_id] = ChatMessageHistory()
+    return store[session_id]
 
 system = """Bạn là một AI có thể sử dụng các công cụ (tool) sau [get_weather, toeic_question]:
 1. toeic_question tool: Trả lời các câu hỏi liên quan đến bài thi TOEIC (không hỏi thêm thông tin ngoài lề).
@@ -28,37 +36,40 @@ Quy tắc:
 - Không đoán thông tin nếu không chắc, hãy nói rõ nếu bạn không biết.
 - Khi dùng tool, chỉ chọn đúng 1 tool phù hợp nhất với câu hỏi.
 """
-prompt = ChatPromptTemplate.from_messages(
-    [
-        ("system", system),    
-        MessagesPlaceholder("chat_history"),  # thêm lịch sử trò chuyện      
-        ("human", "{input}"),
-        MessagesPlaceholder("agent_scratchpad")
-    ]
-)
-memory = ConversationBufferMemory(
-    memory_key="chat_history", return_messages=True)    #Tạo memory cho đoạn chat
 
-# Create the ReAct agent using the create_tool_calling_agent function
-# This function sets up an agent capable of calling tools based on the provided prompt.
+prompt = ChatPromptTemplate.from_messages([
+    ("system", system),    
+    MessagesPlaceholder("chat_history"),
+    ("human", "{input}"),
+    MessagesPlaceholder("agent_scratchpad")
+])
+
 agent = create_tool_calling_agent(
-    llm=model_gemini,  # Language model to use
-    tools=tools,  # List of tools available to the agent
-    prompt=prompt,  # Prompt template to guide the agent's responses
+    llm=model_gemini,
+    tools=tools,
+    prompt=prompt,
 )
 
-# Create the agent executor
-agent_executor = AgentExecutor.from_agent_and_tools(
-    agent=agent,  # The agent to execute
-    tools=tools,  # List of tools available to the agent
-    memory= memory,
-    verbose=True,  # Enable verbose logging
-    handle_parsing_errors=True,  # Handle parsing errors gracefully
+agent_executor = AgentExecutor(
+    agent=agent,
+    tools=tools,
+    verbose=True,
+    handle_parsing_errors=True,
 )
 
+agent_with_chat_history = RunnableWithMessageHistory(
+    agent_executor,
+    get_session_history,
+    input_messages_key="input",
+    history_messages_key="chat_history",
+)
 
-def chat_with_agent(user_input: str):
-    response = agent_executor.invoke({"input": user_input})
+def chat_with_agent(user_input: str, session_id: str = "default"):
+    response = agent_with_chat_history.invoke(
+        {"input": user_input},
+        config={"configurable": {"session_id": session_id}}
+    )
     return response["output"]
+
 print("agent: DONE")
 print("-------------")