All notable changes to this Actor are documented here.
- Fix:
tests/fixtures/input.qa.json — the fixture shipped for cloud QA —
still pinned apifyProxyCountry: "AE", even though f383615e
(2026-09-25) had already corrected the model default and
input_schema.json prefill to "GB" after measuring AE as a guaranteed
Cloudflare challenge on this target (15/15 blocked) vs GB clearing (20
rows, same build). Because the QA fixture is the exact input Apify's
automated testing and our own actor-qa-engineer feed to the cloud, every
QA run kept re-triggering the already-fixed AE block and the Actor stayed
flagged UNDER_MAINTENANCE (QA run n1MU2y5Io4uhZauhq, 2026-09-29).
Fixture now matches the model default (RESIDENTIAL/GB); added
test_qa_fixture_proxy_matches_measured_good_default so the fixture and
the model default can't drift apart again silently.
- No scraping-logic change —
src/client.py, src/parser.py,
src/scraper.py are untouched. This release only re-aligns the QA input
with the already-shipped, already-measured-good default.
- Scaffolded the Actor skeleton:
ActorInput/ResultRow Pydantic
models, .actor/ schemas, PPE pricing placeholders, and store-listing
copy. The real Bayt scraping pipeline (src/client.py, src/parser.py,
src/scraper.py) is stubbed, not yet implemented — src/main.py
currently pushes a single placeholder: true record so the Actor
boots and exits cleanly.