my-project-public

repository

loading code, commits, and activity

repositories

loading repo index

#1	---
#2	title: LangGraph
#3	---
#4
#5	Build a personalized Customer Support AI Agent using LangGraph for conversation flow and Mem0 for memory retention. This integration enables context-aware and efficient support experiences.
#6
#7	## Overview
#8
#9	In this guide, we'll create a Customer Support AI Agent that:
#10	1. Uses LangGraph to manage conversation flow
#11	2. Leverages Mem0 to store and retrieve relevant information from past interactions
#12	3. Provides personalized responses based on user history
#13
#14	## Setup and Configuration
#15
#16	Install necessary libraries:
#17
#18	```bash
#19	pip install langgraph langchain-openai mem0ai python-dotenv
#20	```
#21
#22
#23	Import required modules and set up configurations:
#24
#25	<Note>Remember to get the Mem0 API key from [Mem0 Platform](https://app.mem0.ai).</Note>
#26
#27	```python
#28	from typing import Annotated, TypedDict, List
#29	from langgraph.graph import StateGraph, START
#30	from langgraph.graph.message import add_messages
#31	from langchain_openai import ChatOpenAI
#32	from mem0 import MemoryClient
#33	from langchain_core.messages import SystemMessage, HumanMessage, AIMessage
#34	from dotenv import load_dotenv
#35
#36	load_dotenv()
#37
#38	# Configuration
#39	# OPENAI_API_KEY = 'sk-xxx' # Replace with your actual OpenAI API key
#40	# MEM0_API_KEY = 'your-mem0-key' # Replace with your actual Mem0 API key
#41
#42	# Initialize LangChain and Mem0
#43	llm = ChatOpenAI(model="gpt-4")
#44	mem0 = MemoryClient()
#45	```
#46
#47	## Define State and Graph
#48
#49	Set up the conversation state and LangGraph structure:
#50
#51	```python
#52	class State(TypedDict):
#53	messages: Annotated[List[HumanMessage \| AIMessage], add_messages]
#54	mem0_user_id: str
#55
#56	graph = StateGraph(State)
#57	```
#58
#59	## Create Chatbot Function
#60
#61	Define the core logic for the Customer Support AI Agent:
#62
#63	```python
#64	def chatbot(state: State):
#65	messages = state["messages"]
#66	user_id = state["mem0_user_id"]
#67
#68	try:
#69	# Retrieve relevant memories
#70	memories = mem0.search(messages[-1].content, user_id=user_id)
#71
#72	# Handle dict response format
#73	memory_list = memories['results']
#74
#75	context = "Relevant information from previous conversations:\n"
#76	for memory in memory_list:
#77	context += f"- {memory['memory']}\n"
#78
#79	system_message = SystemMessage(content=f"""You are a helpful customer support assistant. Use the provided context to personalize your responses and remember user preferences and past interactions.
#80	{context}""")
#81
#82	full_messages = [system_message] + messages
#83	response = llm.invoke(full_messages)
#84
#85	# Store the interaction in Mem0
#86	try:
#87	interaction = [
#88	{
#89	"role": "user",
#90	"content": messages[-1].content
#91	},
#92	{
#93	"role": "assistant",
#94	"content": response.content
#95	}
#96	]
#97	result = mem0.add(interaction, user_id=user_id)
#98	print(f"Memory saved: {len(result.get('results', []))} memories added")
#99	except Exception as e:
#100	print(f"Error saving memory: {e}")
#101
#102	return {"messages": [response]}
#103
#104	except Exception as e:
#105	print(f"Error in chatbot: {e}")
#106	# Fallback response without memory context
#107	response = llm.invoke(messages)
#108	return {"messages": [response]}
#109	```
#110
#111	## Set Up Graph Structure
#112
#113	Configure the LangGraph with appropriate nodes and edges:
#114
#115	```python
#116	graph.add_node("chatbot", chatbot)
#117	graph.add_edge(START, "chatbot")
#118	graph.add_edge("chatbot", "chatbot")
#119
#120	compiled_graph = graph.compile()
#121	```
#122
#123	## Create Conversation Runner
#124
#125	Implement a function to manage the conversation flow:
#126
#127	```python
#128	def run_conversation(user_input: str, mem0_user_id: str):
#129	config = {"configurable": {"thread_id": mem0_user_id}}
#130	state = {"messages": [HumanMessage(content=user_input)], "mem0_user_id": mem0_user_id}
#131
#132	for event in compiled_graph.stream(state, config):
#133	for value in event.values():
#134	if value.get("messages"):
#135	print("Customer Support:", value["messages"][-1].content)
#136	return
#137	```
#138
#139	## Main Interaction Loop
#140
#141	Set up the main program loop for user interaction:
#142
#143	```python
#144	if __name__ == "__main__":
#145	print("Welcome to Customer Support! How can I assist you today?")
#146	mem0_user_id = "alice" # You can generate or retrieve this based on your user management system
#147	while True:
#148	user_input = input("You: ")
#149	if user_input.lower() in ['quit', 'exit', 'bye']:
#150	print("Customer Support: Thank you for contacting us. Have a great day!")
#151	break
#152	run_conversation(user_input, mem0_user_id)
#153	```
#154
#155	## Key Features
#156
#157	1. Memory Integration: Uses Mem0 to store and retrieve relevant information from past interactions.
#158	2. Personalization: Provides context-aware responses based on user history.
#159	3. Flexible Architecture: LangGraph structure allows for easy expansion of the conversation flow.
#160	4. Continuous Learning: Each interaction is stored, improving future responses.
#161
#162	## Conclusion
#163
#164	By integrating LangGraph with Mem0, you can build a personalized Customer Support AI Agent that can maintain context across interactions and provide personalized assistance.
#165
#166	<CardGroup cols={2}>
#167	<Card title="LangChain Integration" icon="link" href="/integrations/langchain">
#168	Build conversational agents with LangChain and Mem0
#169	</Card>
#170	<Card title="CrewAI Integration" icon="users" href="/integrations/crewai">
#171	Create multi-agent systems with CrewAI
#172	</Card>
#173	</CardGroup>
#174
#175

z6Mkq5mY3JWtxoxUobWcfNHm7AkRubgSWEZTkBVqZXJviFZ5/my-project-public