Featured project 5 · Python · Keyword-driven

Robot Framework UI + API

UI and API test automation with Robot Framework 7, SeleniumLibrary and RequestsLibrary on the Page Object Model: 30 tests in about 70 seconds with 4 parallel pabot processes.

View source code View documentation View latest CI runs View test report

Purpose

Shows keyword-driven automation built like maintainable software: the same tool I use at work (Robot Framework + Python), applied to a public demo shop (Sauce Demo) and a public API (DummyJSON).

Technology stack

Area	Tools
Framework	Robot Framework 7
UI	SeleniumLibrary 6.9
API	RequestsLibrary 0.9
Parallel runs	pabot
Code quality	Robocop linter (documentation required on every test and keyword)
CI	GitHub Actions: lint gate, Chrome and Firefox matrix, nightly run

Architecture overview

 tests/*.robot           business language, no locators, no HTTP calls
        │
 pages / components      api service objects
 (locators live here)    (URLs and bodies live here)
        │                        │
 common.resource + ShopLibrary.py    ApiSession.resource
        │                        │
   SeleniumLibrary            RequestsLibrary

Key engineering decisions

Tests never touch a locator, URL path or HTTP verb; a UI or API change is one edit in one resource file
One page object per screen; shared screen parts (the header) are components
Action keywords do not assert; check keywords say "Should" (Submit Login Form vs Login Error Message Should Be)
Python only where Robot syntax is awkward: money with Decimal, sort checks, random test data

Implemented features

15 UI tests: login, product list and sorting, cart, checkout with the tax maths
15 API tests: auth, products, carts
Argument files for common, CI and smoke runs; tags for selection
Keyword names that read like a manual test script, with documentation on every test and keyword

Setup and execution

python -m venv .venv && .venv\Scripts\activate
pip install -r requirements.txt
pabot --processes 4 --testlevelsplit --artifacts png --artifactsinsubfolders \
      --outputdir results -A args/common.args tests
robocop check

Sample test results

Robot Framework report: 30 tests passed — Robot Framework report of a full parallel run.

CI workflow

Lint first (Robocop and a dry run), then the suite on Chrome and Firefox in parallel with pabot, a pass/fail table in the run summary, and the logs as artifacts; also a nightly run.