Build courses on automated testing in Python

Kristian Rother, Magdalena Rother, Daniel Szoska

The material here allows you to prepare training courses on automated testing in Python. It contains components that you can combine to tailor your course towards

• the testing framework used
• the experience of your participants
• the duration of the course

Build a course using the following formula:

course = prototype + challenges + tools

The prototype is the story you wrap your content in. It is the driving force that will motivate trainees to keep going. In this tutorial, Herman Melville's Moby Dick will be used as a prototype.

Challenges are problem descriptions relevant to the topic of the course. They are related to real-world problems. Challenges are more general than e.g. tasks for trainees. You can adjust challenges easily to fit to a given difficulty or time frame. They are the raw material to create both tasks and learning objectives.

Tools are what your trainees will use to solve the challenges. They include: Python programs, code examples, example data etc. Tools are interchangeable. The material given here works with different Python testing frameworks.

During the tutorial participants will implement automatic test functions that pass for the Moby Dick example. using the unittest module within 20'.

one hour for a group of experienced developers. For beginners, you may double or triple that time.

• Handout “Introduction to the unittest Framework in Python”
• Slide: Challenge
• Slide: Task description
• TextAnalyser.py
• TestTextAnalyzer.py

What are the 5 words that Melville used the most in Moby Dick? Extend the TextAnalyser class with a function that returns the five most frequent words from Melvilles “Moby Dick”. Add a corresponding test function to the TextAnalyserTests module.

Ensure that the test passes.

Testing adds precision to your work.

nosetests

EXPLAIN OUTPUT

• Write code
• debug
• refactor
• maintain software
• teamwork

• Unit Tests
• Acceptance Tests
• Integration Tests
• Regression Tests

• Coverate
• Mock Objects
• test suite
• TDD
• test data
• 1 assert per test
• border cases

• test functions
• TestCase subclasses
• asserts from nose.tools
• fixtures setUp() tearDown()
• test generators
• test detection
• test selection

• Unit Tests
• Integration Tests
• Acceptance Tests
• Regression tests

• Write tests first
• 1 assert per test
• Border cases
• Test data
• Mock objects

• Test coverage
• Test suite
• Setup script
• User Acceptance
• Functional testing

I. Writing automatic tests in Python

Warm-up exercise: run example test

1.1 nosetests

• nosetests
• -v
• explain output
• test functions (EXAMPLE)
• assert, nose.tools (EXAMPLE)

1.2 Unit Testing

• testing in isolation
• TestCase subclasses (EXAMPLE)
• self.assert statements

1.3 Integration Tests

• two together (EXAMPLE)

1.4 Acceptance Tests

• disambiguation: vs. Web tests, User acceptance tests, manual tests
• command line tests (EXAMPLE)

TASKS:

• TASK: write a Unit Test
• TASK: write a Integration Test
• TASK: write an Acceptance Test

II. Test-Driven Development

2.1 TDD

• What is TDD
• 1 assert per test
• when all tests pass, stop coding
• have test data
• what kind of tests to write?

2.2 Border cases

• empty, fail, smallest, biggest, nasty, sanity (EXAMPLE)

2.3 Fixtures

• in modules
• in classes (EXAMPLE)

2.4 Test generators

• generators (EXAMPLE)

III. Tests during refactoring

Warm-up: Puzzle: calculating coverage

3.1 Refactoring

• example strategy
• tests need structure, too
• mock objects (EXAMPLE)

3.2 Test detection

• test detection
• test selection

3.3 Building a Test Suite

• test suite
• setup script

3.4 Coverage

• run coverage

TASK: refactor TASK: run coverage

Feedback and comments are welcome at: krother@academis.eu

© 2013 Magdalena & Kristian Rother

Released under the conditions of a Creative Commons Attribution License 4.0.