Batch Processing High-Dimensional Vectors | Dataford Interview Questions

Optimize a Python function that performs batch processing on high-dimensional feature vectors.

Implement batch_cosine_similarity(vectors, query) to return one cosine similarity score for each vector. All vectors have the same dimension as query. A zero vector, or a zero query vector, must produce a score of 0.0. Optimize the implementation by avoiding repeated calculations that are shared across the batch.

Input: a list of numeric vectors and one numeric query vector.

Output: a list of floating-point similarity scores in input order.

Constraints

vectors may be empty.
Every vector has the same dimension as query.
Vector components are finite numeric values.
The output preserves the order of vectors.
A zero vector or zero query vector produces a score of 0.0.

Function Signature

def batch_cosine_similarity(vectors, query):

Your question is Batch Processing High-Dimensional Vectors. Start with the requirements in the Question tab.

Run and submit as often as you like. When you're ready, talk me through your approach or go straight to the code.

Problem