fix(tradein): flag coarse-geo estimates as low-confidence (#693) #707
No reviewers
Labels
No labels
admin
analytics
auth
automation
bug
business
chore
ci
compliance
data
data-moat
docs
duplicate
dx
enhancement
Fable 5 ревью
feedback/max
generative
GG-форсайт
needs-discussion
needs-human
observability
pause-bots
performance
priority/p0
priority/p1
priority/p2
priority/p3
scope/backend
scope/db
scope/devops
scope/frontend
scope/qa
scrapers
security
site-finder
stage/1
stage/2
status/blocked
status/done
status/needs-analysis
status/needs-fix
status/qa
status/ready
status/review
status/wip
tech-debt
tradein
ux
week ревью 1
wontfix
вторичка
ИРД
No milestone
No project
No assignees
2 participants
Notifications
Due date
No due date set.
Dependencies
No dependencies set.
Reference: lekss361/gendesign#707
Loading…
Add table
Reference in a new issue
No description provided.
Delete branch "fix/693-nongeocodable-confidence"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Summary
qc_geo>=2— центроид, не дом/улица), радиусный поиск брал аналоги по случайной окрестности → ложно-уверенная medium. Теперь —low+ флаг «оценка ориентировочная».Conservative gate (без ложных downgrade)
Понижение ТОЛЬКО при ВСЕХ условиях:
dadata.qc_geo>=2(позитивный coarse-сигнал) Иanchor_tier is None(same-building путь #691 с реальными комплами дома НЕ трогаем) Иmedian_price>0. Без dadata/qc_geo — не понижаем.Test plan
pytest tests/test_same_building_anchor.py→ 41 passed (+4: coarse+no-anchor→low+флаг; coarse+anchor→не понижен; precise→без изменений; dadata=None→без ложного downgrade).ruff@0.7.4 checkclean.Closes #693
Flag coarse-geo estimates as low-confidence (#693). Downgrade to low + honest explanation when DaData qc_geo>=2 (coarse) AND same-building anchor did NOT fire AND median_price>0. Verified: qc_geo>=2=approximate matches the codebase convention (dadata.py:49 + schemas/trade_in.py:184, label map {2:settlement,3:city,4:region} consistent). Conservative guards: only when relying on radius-around-centroid (anchor_tier None) ? if address-resolved anchor fired (#691 path), no downgrade. Changes ONLY confidence label + explanation string (null-safe append) ? NO median/expected_sold mutation, no #674/#682 regression. Only estimator.py PR this pass, 0 conflicts. APPROVE.