UI Automation Framework Overview and Evolution for Mobile App Testing

Overview

The increasing number of app features has caused regression testing to become time‑consuming and prone to missed test points; the framework aims to cover most regression cases, reduce manual effort, enable frequent UI testing, collect auxiliary data, and monitor main‑flow functionality in released versions.

Framework Introduction

Appium was chosen as the UI automation engine because it supports both Android and iOS with a single test script, handles native, WebView, and hybrid pages, and offers language‑agnostic extensibility.

UITest Structure

The UITest framework provides the foundation for stable and efficient UI automation.

Business Cases

Cases follow the PageObject pattern to separate concerns and simplify maintenance; they are stored within the UITest framework for core business and in separate projects for other business lines, allowing independent development while sharing common actions.

Test Results

Local results are saved in a dedicated folder under the report directory, containing device logs, failure screenshots, and an HTML summary report. Online results are uploaded to a result platform and persisted in a database for historical analysis.

UI Automation Evolution

Locator strategies have progressed from exclusive XPath to a combination of NAME and Text, reducing maintenance overhead. Image‑recognition handling was enhanced by separating platform‑specific image directories, matching by resolution, and falling back to name‑based lookup. WebView controls are now supported alongside native elements.

Control Operation Methods

Existence checks now consider screen visibility, filtering out off‑screen elements and performing scroll actions when needed. Click actions have been extended to WebElements, and additional data‑collection hooks were added. Swipe gestures compute precise coordinates based on device resolution, improving iOS swipe reliability.

Execution Strategies

Trigger parameters have been expanded for flexible CI integration. Multi‑platform execution runs Android and iOS tests concurrently on separate Appium ports using multiprocessing. Multi‑device execution distributes cases across devices, isolates result collection per device, and allows selective device targeting via the -d flag. Case distribution logic creates per‑device test suites and aggregates results per platform, controllable with the -g option.

Framework Execution Logic Comparison

Initial execution relied on manual script runs; the current approach uses a unified execution interface with configurable parameters, supporting parallel multi‑platform runs.

UI Case Selection Criteria

Early stages prioritize regression checklist items; later stages add smoke‑test‑derived UI cases before expanding coverage.

Future Directions

Planned enhancements include richer data extraction from UI cases, visual dashboards for test control, tighter CI integration, and more agile, automated execution pipelines.