llmeval-sdk 0.1.4

Python SDK for the evaluAte LLM evaluation framework

llmeval - Python SDK for evaluAte

A Python client library for the evaluAte LLM evaluation framework.

Installation

pip install -e .

For development with all extras:

pip install -e ".[dev]"

Quick Start

from llmeval import EvalClient

# Initialize the client
client = EvalClient(base_url="http://127.0.0.1:8080")

# Check server health
status = client.health_check()
print(status)

# Get available models
models = client.get_models()
print(f"Available models: {models}")

# Run a single evaluation
result = client.run_eval(
    model="anthropic:claude-sonnet-4",
    prompt="What is the capital of France?",
    expected="Paris",
    judge_model="gemini:gemini-1.5-pro"
)

print(f"Model output: {result.model_output}")
print(f"Judge verdict: {result.judge_verdict}")
print(f"Passed: {result.passed}")

Features

• ✅ Simple, intuitive API
• ✅ Type-safe with Pydantic models
• ✅ Batch evaluation support
• ✅ Real-time WebSocket streaming
• ✅ Jupyter notebook integration
• ✅ pandas DataFrame utilities
• ✅ Comprehensive error handling
• ✅ Context manager support

Documentation

See the examples/ directory for more usage examples:

• basic_usage.py - Simple examples
• advanced_usage.py - Advanced patterns
• streaming_example.py - WebSocket streaming
• jupyter_example.ipynb - Jupyter notebook

Requirements

• Python 3.8+
• requests
• pydantic
• websockets
• pandas

License

MIT License