Session management for multi-user RAG in LangChain - Interactive Code Practice

Practice - 5 Tasks

1fill in blank

easy

Complete the code to create a new session ID for each user.

LangChain

from uuid import [1]

session_id = [1]()

Drag options to blanks, or click blank then click option'

Agetuuid

Brandom

Cuuid4

Duuid1

Attempts:

3 left

💡 Hint

Common Mistakes

Using random instead of uuid4 for unique session IDs.

Using uuid1 which includes MAC address and timestamp, less privacy.

We use uuid4() to generate a unique session ID for each user session.

2fill in blank

medium

Complete the code to initialize a LangChain ConversationBufferMemory for a user session.

LangChain

from langchain.memory import [1]

memory = [1]()

Drag options to blanks, or click blank then click option'

AConversationBufferMemory

BConversationSummaryMemory

CChatMessageHistory

DConversationBufferWindowMemory

Attempts:

3 left

💡 Hint

Common Mistakes

Using ConversationSummaryMemory which summarizes instead of storing full history.

Using ChatMessageHistory which is a message container, not memory.

ConversationBufferMemory stores the full conversation history in memory for the session.

3fill in blank

hard

Fix the error in the code to retrieve the correct session memory from a sessions dictionary.

LangChain

sessions = {}

user_id = 'user123'
memory = sessions.get([1], None)
if memory is None:
    memory = ConversationBufferMemory()
    sessions[user_id] = memory

Drag options to blanks, or click blank then click option'

Asessions

Buser_id

C'user_id'

Dmemory

Attempts:

3 left

💡 Hint

Common Mistakes

Using the string 'user_id' instead of the variable user_id.

Trying to get memory with an undefined variable.

We must use the variable user_id as the key to get the session memory, not the string 'user_id'.

4fill in blank

hard

Fill both blanks to update the retriever and memory for a new user session.

LangChain

from langchain.vectorstores import FAISS

retriever = FAISS.load_local('index_dir', [1])
memory = [2]()

Drag options to blanks, or click blank then click option'

Aembedding_function

BConversationBufferMemory

COpenAIEmbeddings()

DChatMessageHistory

Attempts:

3 left

💡 Hint

Common Mistakes

Passing embedding_function string instead of an embeddings object.

Using ChatMessageHistory instead of ConversationBufferMemory.

FAISS.load_local requires the embedding function like OpenAIEmbeddings(), and memory should be ConversationBufferMemory.

5fill in blank

hard

Fill the blanks to create a multi-user session manager class with memory and retriever.

LangChain

class SessionManager:
    def __init__(self, retriever):
        self.sessions = {}
        self.retriever = [1]

    def get_session(self, user_id):
        if user_id not in self.sessions:
            self.sessions[user_id] = [2]()
        return self.sessions[user_id]

    def query(self, user_id, question):
        memory = self.get_session(user_id)
        # Use retriever and memory to answer question
        results = self.retriever.similarity_search(question)
        # Imagine combining results with memory here
        return results

Drag options to blanks, or click blank then click option'

Aretriever

BConversationBufferMemory

CConversationSummaryMemory

DChatMessageHistory

Attempts:

3 left

💡 Hint

Common Mistakes

Using ChatMessageHistory instead of ConversationBufferMemory for session memory.

Not storing retriever correctly in the class.

The retriever is passed in init, memory is ConversationBufferMemory for sessions, and ConversationSummaryMemory is a distractor here.