Back to professional work

FreeWillResearch Analyst

Auto-update flow.

How I built itTry a test case

I built test websites to check whether our update flow detected changes and carried them into the data system.

  • GitHub
  • GitHub Pages
  • Git branches and tags
  • Automation
How I built the test sites and update checks

Websites change, and our data has to change with them. To check that our auto-update flow catches those changes and carries them into the data onboarding system, I built three controlled test sites in just three days. Each is an exact copy of a funder website, with 20 test-case branches. The sites and the flow are internal to my team, so this page walks through the method with generic examples instead.

  • Mirror a funder site as an exact static copy on GitHub Pages, with relative links.
  • List every file and its original URL in a manifest, including links that were already dead on the live site. Those stay on purpose.
  • Tag the untouched copy baseline-v1, and stop git from altering file bytes so content hashes match on every machine.
  • Make each test case a branch. Merging it to main switches it on.
  • Check the flow picked up the change, then reset to the baseline.
baseline-v1 test case 1 test case 2 test case 3 test case 4 test case 5 main +15 more reset to baseline-v1

One test site’s history. Five of the 20 test-case branches are drawn.

Walk through a test case

Generic examples to show the method. The real sites and flow stay internal.

main is on: baseline-v1, untouched

Pick a change, then merge it to main.

Illustrative output.