Demo Performing Similarity Search

pip install openai numpy

import openai
import numpy as np

def text_embedding(text: str) -> list[float]:
    response = openai.Embedding.create(
        model="text-embedding-ada-002",
        input=text
    )
    return response["data"][0]["embedding"]

phrases = [
    "Most of the websites provide the users with the choice of accepting or denying cookies",
    "Olivia went to the bank to open a savings account",
    "Sam sat under a tree that was on the bank of a river",
    "John's cookies were only half-baked but he still carries them for Mary"
]

embeddings = [text_embedding(p) for p in phrases]
print(f"Embedding dimension: {len(embeddings[0])}")  # Expect 1536

def vector_similarity(vec1: list[float], vec2: list[float]) -> float:
    a, b = np.array(vec1), np.array(vec2)
    return float(np.dot(a, b) / (np.linalg.norm(a) * np.linalg.norm(b)))

def find_most_similar(query: str) -> str:
    q_emb = text_embedding(query)
    scores = [vector_similarity(q_emb, emb) for emb in embeddings]
    ranked = sorted(zip(scores, phrases), reverse=True, key=lambda x: x[0])
    best_score, best_phrase = ranked[0]
    print(f"Query: {query!r}\nBest match ({best_score:.2f}): {best_phrase}\n")
    return best_phrase

find_most_similar("Sam sat under a tree that was on the bank of a river")
find_most_similar("Mary got the biscuits from John that were not fully baked")
find_most_similar("It's recommended to put your savings in a financial institution")
find_most_similar("You get refreshed when you spend time with nature")
find_most_similar("Cookies are covered by GDPR if they collect information about users that could be used to identify them")