Test Generator with One-Assertion Rule

Name: Test Generator with One-Assertion Rule
Rating: 92
Author: svo

Overview

This skill generates tests following the project's mandatory testing standards:

•One assertion per test (critical rule - violations will break the build)
•Descriptive test names using test_should_[expected]_when_[condition] pattern
•Arrange-Act-Assert (AAA) structure for clarity
•Proper mocking to isolate behavior under test
•100% coverage requirement

When to Use This Skill

Use this skill when:

•Creating tests for new code
•Adding missing test coverage
•Refactoring tests that violate the one-assertion rule
•Converting multi-assertion tests into separate test functions

Critical Rule: ONE ASSERTION PER TEST

Each test function MUST contain exactly ONE assertion. If you need multiple assertions, create separate test functions.

Wrong (Multiple Assertions)

python

def test_user_creation(self):
    user = create_user()
    assert_that(user.name).is_equal_to("John")  # ❌ Multiple assertions
    assert_that(user.email).is_equal_to("john@example.com")  # ❌

Correct (One Assertion Each)

python

def test_should_set_user_name_when_user_is_created(self):
    user = create_user()
    assert_that(user.name).is_equal_to("John")  # ✓ Single assertion

def test_should_set_user_email_when_user_is_created(self):
    user = create_user()
    assert_that(user.email).is_equal_to("john@example.com")  # ✓ Single assertion

Test Naming Convention

Test names must be descriptive sentences following this pattern:

code

test_should_[expected_behavior]_when_[condition]

Examples

•test_should_return_404_when_resource_is_not_found()
•test_should_create_user_when_valid_data_is_provided()
•test_should_raise_validation_error_when_email_is_invalid()
•test_should_call_repository_read_method_when_use_case_executes()
•test_should_increment_counter_when_event_is_processed()

Naming Tips

•Start with test_should_
•Describe the expected outcome (what should happen)
•End with _when_ followed by the condition or trigger
•Be specific and descriptive
•Use full words, not abbreviations

Test Structure: Arrange-Act-Assert (AAA)

Every test should follow this three-part structure:

python

def test_should_return_coconut_when_id_exists(self):
    # Arrange - Set up test data and mocks
    repository = Mock()
    use_case = GetCoconutUseCase(repository)
    coconut_id = uuid.uuid4()
    expected_coconut = Coconut(id=coconut_id)
    repository.read.return_value = expected_coconut

    # Act - Execute the behavior being tested
    result = use_case.execute(coconut_id)

    # Assert - Verify the expected outcome (ONE assertion only)
    assert_that(result).is_equal_to(expected_coconut)

Section Guidelines

Arrange:

•Create test data
•Set up mocks and their return values
•Initialize the object under test with dependencies
•Configure the test scenario

Act:

•Execute the single method or behavior being tested
•Store the result if needed for assertion
•Should typically be one line

Assert:

•Verify ONE expected outcome
•MUST use assertpy library (assert_that) - bare assert statements are FORBIDDEN
•Use pytest.raises for exception testing

Testing by Layer

Domain Model Tests

Test Pydantic models and domain entities:

python

from python_sprint_zero.domain.model.coconut import Coconut

def test_should_create_coconut_with_id(self):
    # Arrange
    coconut_id = uuid.uuid4()

    # Act
    coconut = Coconut(id=coconut_id)

    # Assert
    assert_that(coconut.id).is_equal_to(coconut_id)

def test_should_support_none_id(self):
    # Arrange & Act
    coconut = Coconut(id=None)

    # Assert
    assert_that(coconut.id).is_none()

Repository Interface Tests (Domain)

Test the interface definition exists and has correct signatures:

python

from python_sprint_zero.domain.repository.coconut_repository import CoconutQueryRepository

def test_should_define_read_method(self):
    # Arrange & Act
    method = getattr(CoconutQueryRepository, 'read', None)

    # Assert
    assert_that(method).is_not_none()

Repository Implementation Tests (Infrastructure)

Test concrete repository implementations with proper setup:

python

from python_sprint_zero.infrastructure.persistence.in_memory.in_memory_coconut_query_repository import (
    InMemoryCoconutQueryRepository
)

class TestInMemoryCoconutQueryRepository:
    @pytest.fixture
    def storage(self):
        return {}

    @pytest.fixture
    def repository(self, storage):
        return InMemoryCoconutQueryRepository(storage)

    def test_should_return_coconut_when_id_exists(self, repository, storage):
        # Arrange
        coconut_id = uuid.uuid4()
        coconut = Coconut(id=coconut_id)
        storage[coconut_id] = coconut

        # Act
        result = repository.read(coconut_id)

        # Assert
        assert_that(result).is_equal_to(coconut)

    def test_should_raise_exception_when_id_not_found(self, repository):
        # Arrange
        coconut_id = uuid.uuid4()

        # Act & Assert
        with pytest.raises(Exception) as excinfo:
            repository.read(coconut_id)

        assert_that(str(excinfo.value)).contains("not found")

Use Case Tests (Application)

Test use cases with mocked dependencies:

python

from unittest.mock import Mock
from python_sprint_zero.application.use_case.coconut_use_case import GetCoconutUseCase

class TestGetCoconutUseCase:
    @pytest.fixture
    def mock_repository(self):
        return Mock()

    @pytest.fixture
    def use_case(self, mock_repository):
        return GetCoconutUseCase(mock_repository)

    def test_should_call_repository_read_method(self, use_case, mock_repository):
        # Arrange
        coconut_id = uuid.uuid4()
        mock_repository.read.return_value = Coconut(id=coconut_id)

        # Act
        use_case.execute(coconut_id)

        # Assert
        mock_repository.read.assert_called_once_with(coconut_id)

    def test_should_return_coconut_from_repository(self, use_case, mock_repository):
        # Arrange
        coconut_id = uuid.uuid4()
        expected_coconut = Coconut(id=coconut_id)
        mock_repository.read.return_value = expected_coconut

        # Act
        result = use_case.execute(coconut_id)

        # Assert
        assert_that(result).is_equal_to(expected_coconut)

Controller Tests (Interface)

Test API controllers with TestClient:

python

from fastapi.testclient import TestClient
from unittest.mock import Mock

class TestCoconutController:
    @pytest.fixture
    def mock_use_case(self):
        return Mock()

    @pytest.fixture
    def controller(self, mock_use_case):
        return CoconutController(
            get_use_case=mock_use_case,
            authentication_dependency=lambda x: None
        )

    @pytest.fixture
    def client(self, controller):
        app = FastAPI()
        app.include_router(controller.router)
        return TestClient(app)

    def test_should_return_200_when_coconut_exists(self, client, mock_use_case):
        # Arrange
        coconut_id = uuid.uuid4()
        mock_use_case.execute.return_value = Coconut(id=coconut_id)

        # Act
        response = client.get(f"/coconut/{coconut_id}")

        # Assert
        assert_that(response.status_code).is_equal_to(200)

    def test_should_return_coconut_data_when_requested(self, client, mock_use_case):
        # Arrange
        coconut_id = uuid.uuid4()
        mock_use_case.execute.return_value = Coconut(id=coconut_id)

        # Act
        response = client.get(f"/coconut/{coconut_id}")

        # Assert
        assert_that(response.json()["id"]).is_equal_to(str(coconut_id))

Mocking Guidelines

When to Mock

•External dependencies (databases, APIs, file systems)
•Repository implementations in use case tests
•Use cases in controller tests
•Authentication dependencies

What NOT to Mock

•The object under test itself
•Simple domain models (Pydantic models)
•Data transfer objects (DTOs)

Mock Setup

python

from unittest.mock import Mock

# Create mock
mock_repository = Mock()

# Configure return value
mock_repository.read.return_value = Coconut(id=uuid.uuid4())

# Configure exception
mock_repository.read.side_effect = Exception("Not found")

# Verify call
mock_repository.read.assert_called_once_with(coconut_id)

# Verify call with any arguments
mock_repository.read.assert_called_once()

Common Testing Patterns

Testing Exceptions

python

def test_should_raise_exception_when_not_found(self, use_case, mock_repository):
    # Arrange
    coconut_id = uuid.uuid4()
    mock_repository.read.side_effect = Exception("Not found")

    # Act & Assert
    with pytest.raises(Exception) as excinfo:
        use_case.execute(coconut_id)

    assert_that(str(excinfo.value)).contains("Not found")

Testing with Fixtures

python

class TestMyFeature:
    @pytest.fixture
    def sample_id(self):
        return uuid.uuid4()

    @pytest.fixture
    def repository(self):
        return InMemoryRepository({})

    def test_should_use_fixture(self, repository, sample_id):
        # Use fixtures in test
        pass

Testing HTTP Status Codes

python

def test_should_return_404_when_not_found(self, client, mock_use_case):
    # Arrange
    coconut_id = uuid.uuid4()
    mock_use_case.execute.side_effect = Exception("Not found")

    # Act
    response = client.get(f"/coconut/{coconut_id}")

    # Assert
    assert_that(response.status_code).is_equal_to(404)

Test File Organization

File Naming

•Test files mirror source structure exactly
•Prefix with test_: test_coconut_use_case.py

Class Organization

python

class TestGetCoconutUseCase:
    # Group related tests in classes
    # Use fixtures for setup
    # One behavior = one class

    def test_should_call_repository(self):
        pass

    def test_should_return_result(self):
        pass

class TestCreateCoconutUseCase:
    # Separate class for separate use case
    pass

Splitting Multi-Assertion Tests

When you encounter tests with multiple assertions, split them:

Before (Violates Rule)

python

def test_user_creation(self):
    user = create_user("John", "john@example.com")
    assert_that(user.name).is_equal_to("John")
    assert_that(user.email).is_equal_to("john@example.com")
    assert_that(user.is_active).is_true()

After (Follows Rule)

python

def test_should_set_name_when_user_created(self):
    # Arrange & Act
    user = create_user("John", "john@example.com")

    # Assert
    assert_that(user.name).is_equal_to("John")

def test_should_set_email_when_user_created(self):
    # Arrange & Act
    user = create_user("John", "john@example.com")

    # Assert
    assert_that(user.email).is_equal_to("john@example.com")

def test_should_activate_user_when_created(self):
    # Arrange & Act
    user = create_user("John", "john@example.com")

    # Assert
    assert_that(user.is_active).is_true()

Note: Duplicating the Arrange-Act code is acceptable and encouraged. Each test should be independently readable.

Assertion Libraries

Using assertpy (Preferred)

python

from assertpy import assert_that

# Equality
assert_that(result).is_equal_to(expected)

# None checks
assert_that(value).is_none()
assert_that(value).is_not_none()

# Boolean
assert_that(value).is_true()
assert_that(value).is_false()

# Contains
assert_that(string).contains("substring")
assert_that(list).contains(item)

# Type checks
assert_that(value).is_instance_of(MyClass)

# Comparisons
assert_that(value).is_greater_than(5)
assert_that(value).is_less_than(10)

Using pytest.raises

python

import pytest

with pytest.raises(ValueError) as excinfo:
    function_that_raises()

assert_that(str(excinfo.value)).contains("expected message")

Coverage Requirements

•100% test coverage is mandatory
•Every function, class, and method must have tests
•Tests must be meaningful, not just coverage-seeking
•Use tox to verify coverage

bash

# Run all tests with coverage
tox

# Run specific tests during TDD iteration
tox -- tests/specific_test.py

# View coverage report
coverage report

Test Checklist

Before completing test generation, verify:

• Each test has exactly ONE assertion
• Test names follow test_should_X_when_Y pattern
• Tests use Arrange-Act-Assert structure
• Appropriate use of mocks for dependencies
• Tests are isolated and independent
• All edge cases covered (happy path, errors, edge cases)
• Tests verified with tox (NOT just pytest - must pass all quality gates)
• Coverage is 100%: confirmed by tox output

Common Mistakes to Avoid

•Multiple assertions - Split into separate tests
•Using bare assert statements - ALWAYS use assert_that from assertpy
•Vague test names - Use descriptive names with should/when
•Testing implementation details - Test behavior, not internals
•Not using mocks - Mock external dependencies
•Shared state between tests - Each test should be independent
•Missing edge cases - Test happy path, errors, and boundaries
•No AAA structure - Always use Arrange-Act-Assert
•Testing multiple behaviors - One test = one behavior
•Using pytest for final verification - Use tox to verify all quality gates pass

References

See the references/ directory for:

•Complete test examples from the codebase
•Test patterns by layer
•Common testing scenarios