isnt-that-odd 0.1.0

A Python library that determines if a number is even using LLM APIs

isn't that odd

A Python library that determines if a given number is even or not by sending prompts to LLM APIs. Built with LiteLLM for universal LLM support and structured output for reliable responses.

Features

• 🤖 LLM-Powered: Uses Large Language Models to determine if numbers are even or odd
• 🔌 Universal Support: Built on LiteLLM to support any LLM provider (OpenAI, Anthropic, local models, etc.)
• 📊 Structured Output: Forces LLM responses to be structured JSON for reliability
• 🧪 Comprehensive Testing: Full test suite with pytest and promptfoo integration
• 🚀 Modern Python: Type hints, Pydantic models, and modern Python practices
• 📦 Easy Installation: Simple setup with uv package manager

Installation

Using uv (Recommended)

# Install uv if you haven't already
pip install uv

# Clone the repository
git clone https://github.com/yourusername/isnt_that_odd.git
cd isnt_that_odd

# Install dependencies
uv sync

# Install in development mode
uv pip install -e .

Quick Start

from isnt_that_odd import is_even

# Check if a number is even
result = is_even(42)
print(result)  # True

result = is_even(43)
print(result)  # False

Usage

Command Line Interface

The library provides a command-line interface with two main commands:

Check a single number

# Basic usage
isnt-that-odd check 42

# With custom model
isnt-that-odd check --model gpt-4 42

# With verbose output
isnt-that-odd check --verbose 42

# With custom API key
isnt-that-odd check --api-key YOUR_KEY 42

Run benchmark mode

# Run benchmark with 20 random numbers
isnt-that-odd benchmark --count 20

# Run benchmark with custom range and verbose output
isnt-that-odd benchmark --count 50 --min -5000 --max 5000 --verbose

# Run benchmark with specific model
isnt-that-odd benchmark --count 100 --model claude-3-sonnet-20240229

The benchmark mode:

• Generates random numbers within a specified range
• Tests the LLM's accuracy in determining even/odd numbers
• Provides detailed statistics including accuracy percentage and timing
• Shows breakdown by even vs odd numbers
• Reports examples of incorrect predictions (with verbose mode)

Basic Usage

from isnt_that_odd import is_even, EvenChecker

# Simple function call
is_even(10)      # True
is_even(11)      # False
is_even(0)       # True
is_even(-4)      # True
is_even(-7)      # False
is_even(10.5)    # True (integer part 10 is even)

Advanced Usage

from isnt_that_odd import EvenChecker

# Create a custom checker
checker = EvenChecker(
    model="gpt-4",
    api_key="your-api-key-here"
)

# Check multiple numbers
numbers = [2, 3, 4, 5, 6, 7, 8, 9, 10]
results = [checker.check(num) for num in numbers]
print(results)  # [True, False, True, False, True, False, True, False, True]

Using Different LLM Providers

from isnt_that_odd import EvenChecker

# OpenAI
checker = EvenChecker(
    model="gpt-4",
    api_key="your-openai-key"
)

# Anthropic
checker = EvenChecker(
    model="claude-3-sonnet-20240229",
    api_key="your-anthropic-key"
)

# Local models (e.g., Ollama)
checker = EvenChecker(
    model="llama2",
    base_url="http://localhost:11434/v1"
)

# Azure OpenAI
checker = EvenChecker(
    model="gpt-4",
    api_key="your-azure-key",
    base_url="https://your-resource.openai.azure.com/openai/deployments/your-deployment"
)

API Reference

is_even(number, model="gpt-3.5-turbo", api_key=None, base_url=None)

Convenience function to check if a number is even.

Parameters:

• number: The number to check (int, float, or string)
• model: LLM model to use (default: "gpt-3.5-turbo")
• api_key: API key for the LLM service
• base_url: Base URL for the LLM service (for open-source models)

Returns:

• bool: True if the number is even, False if odd

EvenChecker(model="gpt-3.5-turbo", api_key=None, base_url=None)

Main class for checking if numbers are even.

Methods:

• check(number): Check if a number is even
• _create_prompt(number): Create the prompt for the LLM

Benchmark Functions

The CLI module also provides programmatic access to benchmark functionality:

from isnt_that_odd.cli import run_benchmark, generate_random_numbers

# Generate random numbers for testing
numbers = generate_random_numbers(count=100, min_val=-1000, max_val=1000)

# Run benchmark programmatically
run_benchmark(
    count=50,
    model="gpt-3.5-turbo",
    api_key="your-key",
    verbose=True,
    min_val=-500,
    max_val=500
)

How It Works

1. Prompt Engineering: The library creates a carefully crafted prompt that instructs the LLM to determine if a number is even or odd
2. Structured Output: Uses JSON response format to ensure the LLM returns structured data
3. Fallback Parsing: If JSON parsing fails, falls back to text analysis
4. Error Handling: Comprehensive error handling for API failures and parsing issues

Example Prompt

You are a mathematical assistant. Your task is to determine if the given number is even or odd.

Number: 42

Instructions:
1. A number is even if it is divisible by 2 with no remainder
2. A number is odd if it is not divisible by 2
3. For decimal numbers, only the integer part matters
4. For negative numbers, the same rules apply
5. Zero (0) is considered even

Please respond with ONLY a JSON object containing a boolean field "is_even":
- Set "is_even" to true if the number is even
- Set "is_even" to false if the number is odd

Example response format:
{"is_even": true}

Your response:

Testing

Running Tests

# Run all tests
pytest

# Run with coverage
pytest --cov=src/isnt_that_odd

# Run specific test file
pytest tests/test_core.py

Prompt Testing with promptfoo

# Install promptfoo
npm install -g promptfoo

# Run prompt tests
promptfoo eval -c promptfoo.yaml

Development

Setting up Development Environment

# Install development dependencies
uv sync --group dev

# Install pre-commit hooks
pre-commit install

# Run code formatting
black src/ tests/
isort src/ tests/

# Run linting
ruff check src/ tests/

# Run type checking
mypy src/

Project Structure

isnt_that_odd/
├── src/
│   └── isnt_that_odd/
│       ├── __init__.py
│       ├── core.py
│       └── cli.py
├── examples/
│   ├── basic_usage.py
│   └── benchmark_example.py
├── tests/
│   ├── __init__.py
│   ├── test_core.py
│   └── test_cli.py
├── pyproject.toml
├── promptfoo.yaml
└── README.md

Contributing

1. Fork the repository
2. Create a feature branch (git checkout -b feature/amazing-feature)
3. Commit your changes (git commit -m 'Add some amazing feature')
4. Push to the branch (git push origin feature/amazing-feature)
5. Open a Pull Request

License

This project is licensed under the MIT License - see the LICENSE file for details.

Acknowledgments

• Inspired by vibesort - GPT-powered array sorting using structured output
• Built with LiteLLM for universal LLM support
• Uses Pydantic for data validation
• Tested with pytest and promptfoo

Why "isn't that odd"?

The name is a playful reference to the library's purpose of determining whether numbers are even or odd. It's also a bit of a pun - when you ask if something "isn't that odd," you're questioning whether it's unusual, just like the library questions whether a number is odd!