It's OpenTelemetry instrumentation (python) for OpenAI's library.

Project site: https://github.com/cartermp/opentelemetry-instrument-openai-py

Supported APIs:

• Chat
• Embeddings
• Moderation
• Image (generation, edit, variation)
• Audio (transcribe, translate)
• Completion (GPT-3)
• Edit (GPT-3)

Simple! First, install this package.

With autoinstrumentation agent:

poetry add opentelemetry-instrument-openai
poetry run opentelemetry-bootstrap -a install
poetry run opentelemetry-instrument python your_app.py

If you prefer to do it in code, you can do that too:

import openai
from dotenv import load_dotenv
from opentelemetry.instrument.openai import OpenAIInstrumentor

OpenAIInstrumentor().instrument()

load_dotenv()
openai.api_key = os.getenv("OPENAI_API_KEY")

openai.ChatCompletion.create(
    model="gpt-3.5-turbo",
    messages=[{"role":"user", "content":"Tell me a joke about opentelemetry"}],
)

You can then run your app normally with your own opentelemetry initialization.

Get poetry.

Now install and run tests:

poetry install
poetry run pytest

Now you can develop and run tests as you go!

Click the example folder and read the README.