Automation tests CI/CD integration

[anthonymarkQA]

Why is this change needed?

Integrate GUI automation tests in CI/CD via github actions

How was the change implemented?

separate e2e folder, to make sure it does not affect other files

New unit tests

Unit tests executed by the author

How to test manually

run local openspp via docker compose
cd into e2e, npx playwright test --headed

Related links

.github/workflows/e2e.yml

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 75.27%. Comparing base (f66e8a0) to head (fd14937).
⚠️ Report is 11 commits behind head on 19.0.

Additional details and impacted files

@@            Coverage Diff             @@
##             19.0     #442      +/-   ##
==========================================
+ Coverage   75.01%   75.27%   +0.26%     
==========================================
  Files         524      526       +2     
  Lines       35301    35423     +122     
==========================================
+ Hits        26480    26664     +184     
+ Misses       8821     8759      -62     

Flag	Coverage Δ
spp_base_common	91.07% <ø> (ø)
spp_programs	66.73% <ø> (+1.20%)	⬆️
spp_registry	87.79% <ø> (+0.64%)	⬆️
spp_security	69.56% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.
see 12 files with indirect coverage changes

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[gonzalesedwin1123]

Thanks for this — the suite itself is in good shape. I ran the full thing locally against this branch: 16/16 tests passed in 7.3 minutes (Playwright 1.60.0/Chromium 148, which is what the committed lockfile actually resolves to), including both full stack teardown/rebuilds. The journey coverage (install → registrants → groups → areas → approval workflow → vocabulary codes) is real and working. Review verdict is changes requested on one blocker plus a few workflow items.

Must fix before merge

The fixture PDFs in e2e/fixtures/ contain what appears to be real personal data. Three of the five look like internet-sourced sample documents of real, identifiable people: sample_BIR_2316.pdf is a filled and signed tax certificate (names, TINs, home address, birth date, salary figures), sample_authorization_letter.pdf names two specific individuals, and sample_cert_of_enrolment.pdf names a minor along with their real school and principal. For a social-protection platform whose core mission is protecting PII, we cannot commit these into the public repo's permanent history. (sample_valid_ID_parent.pdf — a cartoon mock-up — and sample_academic_calendar.pdf are fine.)

Note that no test currently uploads these files (there is no setInputFiles in the suite — tests 14–15 only create vocabulary codes named after them), so this costs nothing functionally:

1. Replace the three PDFs with self-made fabricated documents (obviously fictional names like Juan Dela Cruz, invalid TIN patterns, fictional school).
2. Rewrite the branch history (interactive rebase or fresh branch + force-push) so the original files never existed in any commit that lands on 19.0 — replacing them in a new commit is not enough, git history keeps the originals forever.

Should fix before merge

• The workflow's Start OpenSPP stack / Wait for healthy steps are wasted work: each spec's beforeAll calls resetStack(), which immediately does down -v and rebuilds. Drop the workflow steps (keep resetStack), or skip resetStack when CI already provides a fresh stack. (Also means ODOO_URL can't actually point anywhere remote, since helpers.ts shells out to local docker compose.)
• Use npm ci and cache on the lockfile so CI installs exactly what's locked.
• e2e/Dockerfile is referenced by nothing (the workflow installs Node directly, helpers use compose) and pins v1.44.0-jammy while the lockfile resolves 1.60.0. Either wire it up or delete it.
• Guard the Discord step so a missing/unset DISCORD_WEBHOOK secret doesn't fail the notify job (e.g. if: ${{ secrets.DISCORD_WEBHOOK != '' }} via an env indirection, or || true with a warning).
• Rename e2e/CLAUDE.md to e2e/README.md — its content is a README, and this repo doesn't commit CLAUDE.md files.

Non-blocking observations

• Heads-up that this workflow has never executed on an actual GitHub runner (it only triggers on push to 19.0, and workflow_dispatch needs the file on the default branch) — the first real run happens after merge. My local 7.3m run had a warm Docker build cache; a cold runner at ~5x slower compute will land somewhere in the 25–45 min range per matrix job, inside the 60-min timeout but worth watching on the shakedown run.
• See inline comments for the action-178 regex, the 30s hard sleep, and slowMo.

Happy to re-review as soon as the fixtures are replaced and the history is rewritten — that part is the only hard gate.

[anthonymarkQA]

Thanks for the thorough review @gonzalesedwin1123! Fixed everything — fixture PDFs replaced with fabricated ones and branch history rewritten (51b0ca8 → a2bb476, force-pushed), plus all the workflow/spec fixes below. Replied + resolved each item inline. Ready for re-review whenever you get a chance.