Automating Tedious Unit Test Generation

What it does

The user leverages an AI chat agent to automatically create specific types of unit tests that would be time-consuming to write manually. They rely on the ease of verifying these generated tests for correctness, using the tool primarily for tedious or self-contained coding tasks rather than complex legacy systems.

Example output

import pytest

from your_module import calculate_discount  # Adjust import path as necessary

class TestCalculateDiscount:
    """Tests for the calculate_discount function."""

    def test_regular_price_no_discount(self):
        """Test that non-members pay full price."""
        assert calculate_discount(100, False) == 100.0
        assert calculate_discount(0, False) == 0.0
        assert calculate_discount(99.99, False) == 99.99

    def test_member_price_10_percent_off(self):
        """Test that members receive a 10% discount."""
        # Test standard integer price
        assert calculate_discount(100, True) == 90.0

        # Test floating point precision edge cases
        # 100 * 0.9 = 90.0 (exact in float)
        assert calculate_discount(100.0, True) == pytest.approx(90.0)

        # Test a price that results in non-terminating binary fraction
        # e.g., 10 * 0.9 = 9.0 (exact), but let's try something like 33.33
        assert calculate_discount(33.33, True) == pytest.approx(29.997, rel=1e-5)

    def test_zero_price(self):
        """Test behavior with zero price."""
        # Zero is not negative, so no error should be raised
        assert calculate_discount(0, False) == 0.0
        assert calculate_discount(0, True) == pytest.approx(0.0)
[... truncated, full run would continue]

## Example prompt

Act as an expert software engineer and generate unit tests for the following [paste code snippet or function here].
Focus on edge cases, boundary conditions, and error handling scenarios that are tedious to write manually.
Ensure the tests are self-contained and easy to verify for correctness without complex mocking dependencies.
Use a standard testing framework compatible with [specify language/framework, e.g., Jest/PyTest/JUnit].
Output only the test code blocks clearly separated from any explanations.

## How to build it

Open your chat agent (ChatGPT, Claude, or Copilot) and paste the example prompt. Adjust the inputs in the curly braces and run.