Similarity search vs MMR retrieval in LangChain - Interactive Practice

Practice - 5 Tasks

1fill in blank

easy

Complete the code to perform a similarity search using LangChain.

LangChain

results = vectorstore.[1](query="What is AI?", k=3)

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

Asearch_documents

Bsimilarity_search

Cmmr_search

Dget_relevant_documents

Attempts:

3 left

💡 Hint

Common Mistakes

Using MMR search method instead of similarity search.

The similarity_search method finds documents most similar to the query.

2fill in blank

medium

Complete the code to perform an MMR retrieval with LangChain.

LangChain

results = vectorstore.[1](query="Explain machine learning", k=4, lambda_mult=0.5)

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

Amax_marginal_relevance_search

Bsimilarity_search

Cretrieve_mmr

Dsearch_mmr

Attempts:

3 left

💡 Hint

Common Mistakes

Using similarity search instead of MMR search.

The max_marginal_relevance_search method performs Maximal Marginal Relevance retrieval balancing relevance and diversity.

3fill in blank

hard

Fix the error in the MMR search call by completing the missing method name.

LangChain

docs = vectorstore.[1](query="Benefits of AI", k=5, lambda_mult=0.7)

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

Asearch

Bsimilarity_search

Cmax_marginal_relevance_search

Dretrieve

Attempts:

3 left

💡 Hint

Common Mistakes

Using generic search methods without MMR support.

The correct method for MMR retrieval is max_marginal_relevance_search.

4fill in blank

hard

Fill both blanks to create a dictionary comprehension that maps documents to their similarity scores greater than 0.8.

LangChain

high_scores = {doc: score for doc, score in vectorstore.similarity_search_with_score(query) if score [1] 0.8 and doc.[2] is not None}

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

A>

B<

Cmetadata

Dcontent

Attempts:

3 left

💡 Hint

Common Mistakes

Using '<' instead of '>' for score comparison.

Checking wrong document attribute.

The comprehension filters scores greater than 0.8 and checks if document metadata exists.

5fill in blank

hard

Fill all three blanks to create a dictionary comprehension that maps document titles to their content length for documents retrieved by MMR with score above 0.5.

LangChain

doc_info = {doc.[1]: len(doc.[2]) for doc, score in vectorstore.[3](query, k=6, lambda_mult=0.6) if score > 0.5}

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

Ametadata['title']

Bpage_content

Cmax_marginal_relevance_search_with_score

Dsimilarity_search

Attempts:

3 left

💡 Hint

Common Mistakes

Using similarity_search instead of max_marginal_relevance_search_with_score.

Accessing wrong document attributes.

This comprehension extracts titles from metadata, counts content length, and uses max_marginal_relevance_search_with_score for MMR retrieval.