Hybrid search (keyword + semantic) in LangChain - Interactive Code Practice

Practice - 5 Tasks

1fill in blank

easy

Complete the code to import the LangChain vector store for hybrid search.

LangChain

from langchain.vectorstores import [1]

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

AChroma

BPinecone

CFAISS

DWeaviate

Attempts:

3 left

💡 Hint

Common Mistakes

Choosing a vector store not typically used for hybrid search.

Confusing the import path or name.

The FAISS vector store is commonly used for hybrid search combining keyword and semantic search in LangChain.

2fill in blank

medium

Complete the code to initialize the FAISS vector store with embeddings.

LangChain

vector_store = FAISS.from_texts(texts, [1])

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

Aembeddings

Bembedding_model

Cembedding_function

Dembedding

Attempts:

3 left

💡 Hint

Common Mistakes

Using a singular form like 'embedding' instead of 'embeddings'.

Passing an incorrect variable name.

The parameter 'embeddings' is used to pass the embedding model or function to FAISS for vectorizing texts.

3fill in blank

hard

Fix the error in the hybrid search query combining keyword and semantic search.

LangChain

results = vector_store.hybrid_search(query=[1], keyword_weight=0.5, semantic_weight=0.5)

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

Atext_query

Bquery_text

Csearch_query

Dquery

Attempts:

3 left

💡 Hint

Common Mistakes

Using incorrect parameter names like 'query_text' or 'search_query'.

Passing the query as a positional argument instead of keyword argument.

The method expects the parameter named 'query' to receive the search string.

4fill in blank

hard

Fill both blanks to set weights for keyword and semantic parts in hybrid search.

LangChain

results = vector_store.hybrid_search(query='example', keyword_weight=[1], semantic_weight=[2])

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

A0.7

B0.3

C1.0

D0.5

Attempts:

3 left

💡 Hint

Common Mistakes

Setting weights that do not sum up or are out of range.

Confusing which weight corresponds to keyword or semantic.

Commonly, keyword_weight is set to 0.5 and semantic_weight to 0.3 or similar to balance the search.

5fill in blank

hard

Fill all three blanks to create a hybrid search function combining keyword and semantic queries.

LangChain

def hybrid_search_function(query):
    keyword_results = vector_store.keyword_search(query=[1])
    semantic_results = vector_store.semantic_search(query=[2])
    combined = keyword_results[:[3]] + semantic_results[:5]
    return combined

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

Aquery

C3

D5

Attempts:

3 left

💡 Hint

Common Mistakes

Using different variable names for query in keyword and semantic search.

Not slicing the results to limit combined output.

Both keyword and semantic search use the same 'query' parameter. Limiting keyword results to 3 and semantic to 5 balances the combined output.