DSPy OpenTelemetry Integration

Overview

This integration provides support for using OpenTelemetry with the DSPy framework. It enables tracing and monitoring of applications built with DSPy.

Installation

1. Install traceAI DSPy

pip install traceAI-DSPy

2. Install DSPy

pip install dspy

Set Environment Variables

Set up your environment variables to authenticate with both FutureAGI.

import os

os.environ["FI_API_KEY"] = FI_API_KEY
os.environ["FI_SECRET_KEY"] = FI_SECRET_KEY

Quickstart

Register Tracer Provider

Set up the trace provider to establish the observability pipeline. The trace provider:

from fi_instrumentation import register
from fi_instrumentation.fi_types import ProjectType

trace_provider = register(
    project_type=ProjectType.OBSERVE,
    project_name="dspy_app",
    project_version_name="v1",
    session_name="chat-bot"
)

Configure DSPy Instrumentation

Instrument the DSPy client to enable telemetry collection. This step ensures that all interactions with the DSPy SDK are tracked and monitored.

from traceai_dspy import DSPyInstrumentor

DSPyInstrumentor().instrument(tracer_provider=trace_provider)

Create DSPy Components

Set up your DSPy client with built-in observability.

import dspy

class BasicQA(dspy.Signature):
    """Answer questions with short factoid answers."""

    question = dspy.InputField()
    answer = dspy.OutputField(desc="often between 1 and 5 words")

if __name__ == "__main__":
    turbo = dspy.LM(model="openai/gpt-4")

    dspy.settings.configure(lm=turbo)

    # Define the predictor.
    generate_answer = dspy.Predict(BasicQA)

    # Call the predictor on a particular input.
    pred = generate_answer(question="What is the capital of the united states?")
    print(f"Predicted Answer: {pred.answer}")