A small, generic end-to-end Test Automation Framework (TAF) for web UI testing using Python, pytest, and Playwright (sync API).
The project implements the Page Object Model (POM) and is designed to be easy to use, extend, run locally and in CI.

KEY POINTS / FEATURES

• Language: Python
• Test runner: pytest
• Browser automation: Playwright (Chromium only, sync API)
• Design: Page Object Model (pages/), simple dataclasses in model/
• Headed mode by default (can be toggled to headless)
• Per-test isolation using fresh browser contexts (fast and independent tests)
• Video recording and Playwright tracing support (kept on failures by default)
• HTML test report generated (2 types of reports are available)
• No credentials stored in source — uses environment variables / .env for local convenience
• Running tests in parallel
• Integration with CI/CD (GitHub Actions)

Activate venv:

source .venv/bin/activate

Run the full suite (headed by default):

pytest

Run headless:

export HEADED=false
pytest

Run a single test file:

pytest tests/test_checkout.py

Enable slow-motion for debugging:

export SLOW_MO=50
pytest tests/test_login.py

Running tests in parallel

The pytest-xdist package is used to run tests in parallel (multiple worker processes).

pytest -n 2

Reports:

A simple HTML report is generated at artifacts/report.html.

Running tests with report generation:

pytest -q --html=artifacts/report.html --self-contained-html

Note: this html-report can be generated only if tests are not run in parallel.

Reporting with Allure

The Allure is used for rich HTML reports (attachments, history, etc).

Running tests with report generation:

# [optionally ] remove previous artifacts:
rm -rf artifacts/allure-results artifacts/allure-report artifacts/videos artifacts/traces

# initial attempt
ALLURE_AUTO_GENERATE=1 RUN_ATTEMPT=1 ALLURE_RESULTS_DIR=artifacts/allure-results/attempt_1 VIDEO_DIR=artifacts/videos pytest --alluredir=artifacts/allure-results/attempt_1

# [optionally] re-run failed tests
RUN_ATTEMPT=2 ALLURE_RESULTS_DIR=artifacts/allure-results/attempt_2 VIDEO_DIR=artifacts/videos pytest --last-failed --alluredir=artifacts/allure-results/attempt_2

# [optionally, if used both attempts] merge Allure results from both attempts into one report
mkdir -p artifacts/allure-results/merged
cp -a artifacts/allure-results/attempt_1/. artifacts/allure-results/merged/ || true
cp -a artifacts/allure-results/attempt_2/. artifacts/allure-results/merged/ || true
allure generate artifacts/allure-results/merged -o artifacts/allure-report --clean

Opening the generated report:

allure open artifacts/allure-report

Record video:

Video is generated at artifacts/videos if the test is failed (by default). Set env KEEP_VIDEOS=true to keep videos for all tests.

Trace tests

Tracing is generated at artifacts/traces if the test is failed (by default).

To view the trace run:

playwright show-trace artifacts/traces/<your-trace-file>.zip
# or
python -m playwright show-trace artifacts/traces/<your-trace-file>.zip

Continuous Integration (GitHub Actions)

This project uses GitHub Actions to run automated tests on every push and pull request.

• Workflow file: .github/workflows/ci.yml
• Runs on Ubuntu latest with Python 3.11
• Installs dependencies, Playwright browsers, and executes the full pytest suite
• Uploads Allure report, videos, and traces as build artifacts
• Fails the job if the tests are failed

CI Behavior

• All tests are executed in CI (no fail-fast by default).
• If a test fails, artifacts (videos, traces) are available in the GitHub Actions run for debugging.
• Secrets (like SAUCE_USERNAME and SAUCE_PASSWORD) can be added to the repository’s Settings > Secrets and variables > Actions for secure usage in tests.
• The Allure report is generates as an artifact.
• The last Allure report is available here: https://somatori.github.io/web_taf_python_pytest_playwright/

Local vs CI

• Locally: run pytest (optionally pytest --headed to see the browser).
• In CI: tests run in headless mode by default, artifacts are stored automatically.

Test markers

Pytest markers are used to group and run subsets of tests quickly.

Registered markers:

• smoke — fast, high-level smoke tests that exercise critical paths.
• sanity — slightly broader sanity checks that validate core functionality.

Marking tests:

import pytest

@pytest.mark.smoke
def test_example(page, credentials):
    ...

Running tests by a marker:

pytest -q -m smoke