You get:
- code that works on happy path but fails on edge cases
- no tests for error conditions
- no mock examples for external dependencies
- tests that are hard to run or maintain
- low test coverage (untested code)
But unit tests are not optional.
They are the safety net for refactoring and deployment.
- Normal case: typical inputs, expected outputs
- Edge cases: empty lists, None values, boundary conditions
- Error cases: invalid inputs, exceptions raised
- Mock examples: external API calls, database queries
- Parametrized tests: test multiple inputs with one function
Without tests, you don’t know if your code works.
This framework forces AI to generate comprehensive unit tests.
Assume the role of a quality assurance engineer who writes thorough pytest unit tests. Your task is to generate unit tests for a Python function. Generate: 1. IMPORTS - pytest - The function to test - unittest.mock (if mocking needed) 2. NORMAL CASE TESTS (2-3) - Typical inputs - Expected outputs 3. EDGE CASE TESTS (2-3) - Empty inputs, None values, boundary conditions - Single-item collections, zero values 4. ERROR CASE TESTS (2-3) - Invalid input types - Exceptions raised 5. MOCK TESTS (if external dependencies) - Mock API calls - Mock database queries - Mock file I/O 6. PARAMETRIZED TESTS (optional) - Test multiple inputs with @pytest.mark.parametrize INPUTS: Function Code (paste the function to test): [PASTE FUNCTION CODE] External Dependencies (if any): [E.G., "Calls requests.get()", "Reads from file", "Queries database"] Edge Cases to Test (if known): [LIST OR "INFER"] Test Framework: [PYTEST / UNITTEST] Coverage Goal: [STATEMENT / BRANCH / PATH] RULES: - Test normal case first (happy path) - Test edge cases (boundaries, empty, None) - Test error cases (invalid inputs, exceptions) - Use pytest fixtures for setup/teardown - Use unittest.mock for external dependencies - Use parametrize for testing multiple inputs - Test names should describe what they test - Assert one thing per test (when possible)
- Paste the exact function code — imports and all.
- List external dependencies so the prompt knows what to mock.
- Run pytest after generation and fix any failures (some may need adjustment).
- Add tests for any custom logic not covered.
- Aim for statement coverage (not perfection) as a starting point.
Function Code:
def fetch_user_data(user_id, api_client):
response = api_client.get(f”/users/{user_id}”)
if response.status_code == 200:
return response.json()
elif response.status_code == 404:
return None
else:
raise Exception(f”API error: {response.status_code}”)
External Dependencies: Calls api_client.get() (needs mocking)
Edge Cases to Test: user_id not found (404), API returns 500 error
Test Framework: PYTEST
This framework improves outcomes by forcing:
- normal case testing (baseline)
- edge case testing (robustness)
- error case testing (resilience)
- mock examples (isolation)
- parametrized tests (efficiency)
Great unit tests don’t just prove code works — they prove it fails gracefully when it should.
Build Better AI Systems
Subscribe for advanced prompt engineering, AI coding tools, Python frameworks, and practical strategies for developers and engineers.