FAISS vector store setup in LangChain - Practice Problems & Coding Challenges

Challenge - 5 Problems

❓ component_behavior

intermediate

FAISS Vector Store Initialization Output

What will be the output of this code snippet when initializing a FAISS vector store with LangChain?

LangChain

from langchain.vectorstores import FAISS
from langchain.embeddings import OpenAIEmbeddings

embeddings = OpenAIEmbeddings()
texts = ["Hello world", "FAISS vector store", "LangChain example"]

vector_store = FAISS.from_texts(texts, embeddings)
print(len(vector_store.index))

A Raises TypeError

B 3

C 0

D Length of the embedding dimension

💡 Hint

Check the attributes of the FAISS vector store object and what 'index' represents.

The FAISS vector store's 'index' attribute is a FAISS index object, which does not support len(). Trying to get its length raises a TypeError.

❓ state_output

intermediate

Number of Vectors Stored in FAISS

After adding three documents to a FAISS vector store, what is the number of vectors stored?

LangChain

from langchain.vectorstores import FAISS
from langchain.embeddings import OpenAIEmbeddings

embeddings = OpenAIEmbeddings()
texts = ["Doc1", "Doc2", "Doc3"]

vector_store = FAISS.from_texts(texts, embeddings)

print(vector_store.index.ntotal)

A 0

B 3

C Raises AttributeError

D Number of embedding dimensions

💡 Hint

The 'ntotal' attribute of the FAISS index shows how many vectors are stored.

The 'ntotal' attribute returns the total number of vectors stored in the FAISS index, which matches the number of texts added.

📝 Syntax

advanced

Correct Syntax for Creating FAISS Vector Store

Which option correctly creates a FAISS vector store from texts using LangChain embeddings?

A FAISS.from_texts(texts, embeddings=OpenAIEmbeddings())()

B FAISS.from_texts(texts, OpenAIEmbeddings)

C FAISS.from_texts(texts, embeddings=OpenAIEmbeddings())

D FAISS.from_texts(texts, embeddings=OpenAIEmbeddings)

💡 Hint

Remember to instantiate the embeddings class before passing it.

The method expects an instance of the embeddings class, not the class itself or a call of a call.

🔧 Debug

advanced

Error Raised When Using FAISS Without Embeddings

What error will this code raise?

LangChain

from langchain.vectorstores import FAISS

texts = ["sample text"]
vector_store = FAISS.from_texts(texts, embeddings=None)

A AttributeError: 'NoneType' has no attribute 'embed_documents'

B TypeError: 'NoneType' object is not callable

C ValueError: embeddings must be provided

D No error, creates empty vector store

💡 Hint

Check what happens when embeddings is None and the method tries to call embed_documents.

The code tries to call embed_documents on None, causing an AttributeError.

🧠 Conceptual

expert

FAISS Vector Store Persistence Behavior

Which statement about saving and loading a FAISS vector store in LangChain is correct?

A Loading a FAISS vector store automatically downloads embeddings from the cloud.

B Saving the FAISS index alone is enough to restore the vector store with all metadata.

C FAISS vector stores cannot be saved or loaded; they must be rebuilt every time.

D You must save the FAISS index and the associated metadata separately to reload correctly.

💡 Hint

Think about what data FAISS stores and what LangChain manages separately.

FAISS stores only vectors and indexes; LangChain manages metadata separately, so both must be saved and loaded.