Machine Learning powered regression
summary
Proposes automated full-regression by storing baseline screenshots for every sitemap URL and comparing new screenshots to detect unexpected changes, claiming orders-of-magnitude savings versus manual or brittle Selenium suites; implements a crawler that runs screenshots via BrowserStack and stitches multi-viewport captures, generates Selenium tests for hover/click states, and exposes APIs; handles 2,000 pages in ~4 hours, produces 50+ GB for huge sites, and is open-source work in progress with ML attempts to separate template vs content.
AI-written summary · about
chapters (10)
- 0:00Introduction and welcome
- 2:44Problem with regression testing
- 8:06Snapshot testing approach
- 13:03Challenges with large projects
- 19:14Goal of the open-source tool
- 25:23Competitive advantages
- 28:16QA efficiency with automation
- 30:05Machine learning applications
- 37:11Future roadmap and open-source plans
- 45:22Q&A and closing
related talks
39mSpot the difference: automating visual regression testing
37mSpot the Difference; Automating Visual Regression Testing
37mThe Colorful World of Visual Test Automation
20mVisual testing with Visual Regression Tracker
27mWhat To Expect When You're Expecting: A guide to Regression Testing
21mVisual Regression Testing: From a Tool to a Process
18mImplementing a Visual CSS Testing Framework
46mSW Testing: Can ML save us?
