Memory-augmented retrieval in LangChain - Mini Project: Build & Apply

Memory-augmented retrieval with LangChain

📖 Scenario: You are building a simple chatbot that remembers previous conversations to answer questions better. You will use LangChain's memory feature to store chat history and retrieve it during new queries.

🎯 Goal: Create a LangChain chatbot that uses memory to recall past messages and improve responses.

📋 What You'll Learn

Create a list called chat_history with two example messages

Create a variable called memory using ConversationBufferMemory

Create a ChatOpenAI instance called chat with model name gpt-3.5-turbo

Create a ConversationChain called conversation using chat and memory

💡 Why This Matters

🌍 Real World

Chatbots that remember past conversations provide better, more personalized answers. This is useful in customer support, virtual assistants, and interactive apps.

💼 Career

Understanding memory-augmented retrieval is key for building advanced conversational AI systems used in many tech companies.

1

Set up initial chat history

Create a list called chat_history with these two dictionaries exactly: {'role': 'user', 'content': 'Hello!'} and {'role': 'assistant', 'content': 'Hi! How can I help you?'}.

LangChain

# Create chat_history list with two messages
# Your code here

Need a hint?

Use a Python list with two dictionaries inside. Each dictionary has keys 'role' and 'content'.

2

Create ConversationBufferMemory

Create a variable called memory by instantiating ConversationBufferMemory with memory_key='chat_history' and return_messages=True.

LangChain

chat_history = [
    {'role': 'user', 'content': 'Hello!'},
    {'role': 'assistant', 'content': 'Hi! How can I help you?'}
]
# Create memory using ConversationBufferMemory
# Your code here

Need a hint?

Import ConversationBufferMemory from langchain.memory and create memory with the specified parameters.

3

Create ChatOpenAI instance

Create a variable called chat by instantiating ChatOpenAI with model_name='gpt-3.5-turbo'.

LangChain

from langchain.memory import ConversationBufferMemory
from langchain.chat_models import ChatOpenAI

chat_history = [
    {'role': 'user', 'content': 'Hello!'},
    {'role': 'assistant', 'content': 'Hi! How can I help you?'}
]
memory = ConversationBufferMemory(memory_key='chat_history', return_messages=True)
# Create chat instance with model_name 'gpt-3.5-turbo'
# Your code here

Need a hint?

Import ChatOpenAI from langchain.chat_models and create chat with the given model name.

4

Create ConversationChain with memory

Create a variable called conversation by instantiating ConversationChain with llm=chat and memory=memory.

LangChain

from langchain.memory import ConversationBufferMemory
from langchain.chat_models import ChatOpenAI
from langchain.chains import ConversationChain

chat_history = [
    {'role': 'user', 'content': 'Hello!'},
    {'role': 'assistant', 'content': 'Hi! How can I help you?'}
]
memory = ConversationBufferMemory(memory_key='chat_history', return_messages=True)
chat = ChatOpenAI(model_name='gpt-3.5-turbo')
# Create ConversationChain with llm=chat and memory=memory
# Your code here

Need a hint?

Import ConversationChain from langchain.chains and create conversation with the given parameters.