similarity search

Finding vectors closest to a query vector using distance metrics. The core operation in vector databases enabling semantic search and recommendation.

Syntax

vector-databases

results = collection.query(query_vector, top_k=10, metric="cosine")

Example

vector-databases

# Cosine similarity search with ChromaDB:
collection = client.get_collection("articles")

query_embedding = embed("best practices for React hooks")

results = collection.query(
    query_embeddings=[query_embedding],
    n_results=5,
    include=["documents", "distances", "metadatas"]
)

for doc, dist in zip(results["documents"][0], results["distances"][0]):
    print(f"Score: {1-dist:.3f} | {doc[:80]}...")