README.md

This repo explores a wide range of testing techniques — mocking, fixtures, test-class inheritance, a simulated HTTP server and continuous integration — using simple multiple-choice quizzes stored as JSON, and compares the automated tests through unittest and pytest.

It was also my first steps with Markdown documentation.

The test suite is the heart of the project: it stacks several techniques on a single use case.

Clone the repository :

git clone https://github.com/DevCTx/unit_tests_practice

Into the cloned folder :

python -m venv venv
source venv/bin/activate              # Linux / macOS  (venv\Scripts\activate on Windows)
pip install -r requirements.txt       # to use questionnaire.py only
pip install -r requirements-dev.txt   # to use questionnaire_unit_tests.py

python3 ./questionnaire_import.py 

More info about the original conversion : questionnaire_import.md

python3 ./questionnaire.py ./json_questionnaires/<questionnaire.json>

More info about the JSON format expected and Error messages : questionnaire.md

python3 -m unittest questionnaire_unit_test

pytest -v questionnaire_unit_test.py -s

or use ./json_test_files folder to check 25 intentionally broken quizzes targeting error cases

python3 questionnaire.py json_test_files/json_00_empty.json

More info about the 25 tests : questionnaire_unit_test.md

Explanation

In questionnaire_unit_test.py, load_json_argv() is the entry point.

It receives sys.argv, validates that there is exactly one .json argument. detects whether it is a URL or a local file, and delegates to load_json_data_from_URL() or load_json_data_from_file().

It always returns a tuple (json_data, file_path).

Testing highlights

The objective was to test these various characteristics and methods:

• Test-class inheritance to reuse base cases by changing a single parameter.
• Mocking with unittest.mock.patch to simulate sys.argv.
• Threaded local HTTP server (ThreadingMixIn + SimpleHTTPRequestHandler) to test remote loading without relying on the internet.
• setUp/tearDown (unittest) vs @pytest.fixture (pytest) for the same need, to compare both frameworks directly.

Continuous integration

GitHub Actions runs on every push and pull request to master: dependency install (Python 3.10), flake8 lint, then the test suite under both unittest and pytest.

• Python 3.10+
• jsonschema==4.17.0
• parameterized==0.8.1
• requests==2.28.1
• validators==0.20.0

• pytest==8.0.0
• flake8==6.0.0