3.2.3. Embeddings

🔧 Implementation Reference: Embeddings

Testable Pattern:

response = client.embeddings.create(model="text-embedding-ada-002", input="Search query")
vector = response.data[0].embedding  # List of 1536 floats

Error Handling Pattern:

from openai import AzureOpenAI, BadRequestError, RateLimitError

try:
    response = client.embeddings.create(model="text-embedding-ada-002", input=text)
    vector = response.data[0].embedding
except BadRequestError as e:
    if "maximum context length" in str(e):
        # Input exceeds token limit (8191 for ada-002)
        logging.error("Input text exceeds embedding model token limit")
        # Chunk the text and embed separately
except RateLimitError:
    # Implement exponential backoff for batch processing
    time.sleep(retry_delay)

CLI Equivalent (REST):

curl -X POST "https://{resource}.openai.azure.com/openai/deployments/{deployment}/embeddings?api-version=2024-08-01-preview" \
  -H "Content-Type: application/json" \
  -H "api-key: {key}" \
  -d '{"input": "Text to embed"}'