Tips and Tricks #93: Use Embeddings for Semantic Search

Implement semantic search using text embeddings for more relevant results than keyword matching.

Code Snippet

from openai import OpenAI
import numpy as np

client = OpenAI()

def get_embedding(text: str) -> list[float]:
    """Generate embedding for text using OpenAI."""
    response = client.embeddings.create(
        model="text-embedding-3-small",
        input=text
    )
    return response.data[0].embedding

def cosine_similarity(a: list[float], b: list[float]) -> float:
    """Calculate cosine similarity between two vectors."""
    return np.dot(a, b) / (np.linalg.norm(a) * np.linalg.norm(b))

# Index documents
documents = ["Python is great for data science", "JavaScript powers the web"]
doc_embeddings = [get_embedding(doc) for doc in documents]

# Search
query = "machine learning programming"
query_embedding = get_embedding(query)

# Find most similar
similarities = [cosine_similarity(query_embedding, de) for de in doc_embeddings]
best_match_idx = np.argmax(similarities)
print(f"Best match: {documents[best_match_idx]}")

Why This Helps

• Finds semantically similar content, not just keyword matches
• Works across languages and synonyms
• Foundation for RAG applications

How to Test

• Test with synonyms and paraphrases
• Compare results to keyword search

When to Use

Search functionality, recommendation systems, document similarity, RAG pipelines.

Performance/Security Notes

Use vector databases (Pinecone, Weaviate) for production. Cache embeddings for performance.

References

• https://platform.openai.com/docs/guides/embeddings

Try this tip in your next project and share your results in the comments!