composo 0.2.16

Composo Python SDK

Composo Python SDK

A Python SDK for Composo evaluation services, providing both synchronous and asynchronous clients for evaluating LLM conversations with support for OpenAI and Anthropic formats.

Features

• Dual Client Support: Both synchronous and asynchronous clients
• Multiple LLM Provider Support: Native support for OpenAI and Anthropic formats
• Connection Pooling: Optimized HTTP client with connection reuse
• Retry Logic: Exponential backoff with jitter for robust API calls
• Type Safety: Full type hints and Pydantic models
• Context Managers: Proper resource management with context managers

Installation

pip install composo

Quick Start

Basic Usage

from composo import Composo, AsyncComposo

# Initialize client
client = Composo(api_key="your-api-key")

# Evaluate messages
messages = [
    {"role": "user", "content": "What is machine learning?"},
    {"role": "assistant", "content": "Machine learning is..."}
]

criteria = ["Reward responses that provide accurate technical explanations"]

result = client.evaluate(messages=messages, criteria=criteria)
print(f"Score: {result.score}")
print(f"Explanation: {result.explanation}")

Async Usage

import asyncio
from composo import AsyncComposo

async def main():
    async with AsyncComposo(api_key="your-api-key") as client:
        result = await client.evaluate(
            messages=messages,
            criteria=criteria
        )
        print(f"Score: {result.score}")

asyncio.run(main())

With LLM Results

import openai
from composo import Composo

# Get response from OpenAI
openai_client = openai.OpenAI(api_key="your-openai-key")
openai_result = openai_client.chat.completions.create(
    model="gpt-4",
    messages=[{"role": "user", "content": "What is machine learning?"}]
)

# Evaluate the response
composo_client = Composo(api_key="your-composo-key")
eval_result = composo_client.evaluate(
    messages=[{"role": "user", "content": "What is machine learning?"}],
    result=openai_result,
    criteria=["Reward accurate technical explanations"]
)

Configuration

Client Options

• api_key (required): Your Composo API key
• base_url (optional): Custom API endpoint (default: https://platform.composo.ai)
• num_retries (optional): Number of retry attempts (default: 1)
• model_core (optional): Specific model core for evaluation

Logging

The SDK uses Python's standard logging module. Configure logging level:

import logging
logging.getLogger("composo").setLevel(logging.INFO)

Error Handling

The SDK provides specific exception types:

from composo import (
    ComposoError,
    RateLimitError,
    MalformedError,
    APIError,
    AuthenticationError
)

try:
    result = client.evaluate(messages=messages, criteria=criteria)
except RateLimitError:
    print("Rate limit exceeded")
except AuthenticationError:
    print("Invalid API key")
except ComposoError as e:
    print(f"Composo error: {e}")

Performance Optimization

• Connection Pooling: HTTP clients reuse connections for better performance
• Context Managers: Use context managers to properly close connections
• Async Support: Use async client for high-throughput scenarios