Commit graph

109 commits

Author SHA1 Message Date
bot-backend
67a852caf1 feat(tradein/matching): fias_exact tier + estimate fias persist + suggest fias_id contract
All checks were successful
CI / backend-tests (pull_request) Has been skipped
CI / frontend-tests (pull_request) Has been skipped
CI / openapi-codegen-check (pull_request) Has been skipped
CI / changes (pull_request) Successful in 10s
2026-07-02 20:49:04 +03:00
bot-backend
3909db745b fix(tradein/estimator): не отравлять asking→sold ratio-кэш транзиентной ошибкой
All checks were successful
CI / frontend-tests (pull_request) Has been skipped
CI / openapi-codegen-check (pull_request) Has been skipped
CI / changes (pull_request) Successful in 6s
CI / backend-tests (pull_request) Has been skipped
Блок «ОЖИДАЕМАЯ ЦЕНА СДЕЛКИ» (expected_sold, «с учётом торга») молча гас в «—» у
части оценок с полноценным headline (прод: high-confidence, 17 аналогов, цена
есть — expected_sold NULL). Причина: `_get_asking_sold_ratio` gracefully ловит
ЛЮБУЮ ошибку БД → (None, None), но запись в `_asking_sold_ratio_cache` шла
БЕЗУСЛОВНО (после try/except) → один транзиентный сбой (poisoned tx от
вышестоящего graceful-except, коннект-хиккап) отравлял кэш бакета `(None, None)`
на весь TTL (300с) и гасил expected_sold для ВСЕХ оценок этого rooms-бакета на
воркере до истечения TTL.

Фикс: ранний `return None, None` из except БЕЗ записи в кэш → следующая оценка
ретраит. Кэшируем только успешный lookup (ratio может быть None, если строки
реально нет — стабильный факт БД, безопасно кэшировать).

Прод-диагностика (30 дней, 304 оценки): 39 пустых expected_sold, из них 37 —
легитимно n_analogs=0 (нет аналогов), а 2 — этот баг (headline есть, ratio-кэш
отравлён). Таблица asking_to_sold_ratios полна по всем бакетам + global fallback
0.843 → при исправном lookup ratio всегда резолвится.

Юнит-тесты: транзиентная ошибка → None БЕЗ кэша; error→retry не залипает;
успех кэшируется.
2026-07-02 20:34:10 +03:00
ecc3ab5aab fix(tradein/estimator): честный asking_to_sold_ratio — бейдж «−N% к рынку» больше не врёт
All checks were successful
CI / changes (pull_request) Successful in 8s
CI / frontend-tests (pull_request) Has been skipped
CI / backend-tests (pull_request) Has been skipped
CI / openapi-codegen-check (pull_request) Has been skipped
Оценщик клиента жаловался на «большой интервал между рекомендованной ценой
и оценкой». Разбор: бейдж «−23% к рынку» (web HeroSummary + PDF, формула
round((1−ratio)×100)) систематически завышал скидку.

Root cause: сохранённый asking_to_sold_ratio — это СЫРОЙ per-rooms/tier дисконт
из ratio_resolver, но фактический expected_sold сдвинут относительно median×ratio
последующими корректировками: hedonic year+area (#2002, factor ∈ [0.75, 1.30], ON
by default), le_asking-clamp и corridor-clamp. Пример с прода (451de30b): median
7.75M × raw 0.771 = 5.97M, hedonic ×1.226 → expected_sold 7.32M — но stored ratio
остался 0.771, тогда как фактическое expected_sold/median = 0.945. Бейдж показывал
«−23%» вместо честных «−5%».

Fix: после финализации expected_sold пересчитываем сохранённый asking_to_sold_ratio
как реальное expected_sold_price/median_price (честный дескриптор). Сам expected_sold
(выкуп) НЕ трогаем — hedonic-uplift остаётся прибит к sale-модели, buyout не падает
до наивного median×raw. Порог _RATIO_DESCRIPTOR_EPS=1e-4 отсекает шум округления:
без сдвига (hedonic OFF, нет клампа) табличный ratio сохраняется байт-в-байт →
регрессия на не-зажатых оценках отсутствует.

Стор asking_to_sold_ratio — чисто ДЕСКРИПТОР (web/PDF/history badge), НЕ калибровочный
вход: калибровочный ratio живёт в таблице asking_to_sold_ratios (refresh-task, читает
resolver) — не тронута. Backtest #1966 скорит expected_sold_per_m2 (не stored ratio) —
не затронут (expected_sold без изменений).

Tests: 3 новых в test_estimator_price_spine.py (инвариант при hedonic-uplift +
corridor-clamp; byte-identical регрессия без сдвига); поправлен
test_global_fallback_basis_carried_through (hedonic OFF для сырого ratio).
Full suite: 2749 passed (кроме pre-existing test_search_cache_hit).

Refs #2141
2026-07-02 18:26:19 +03:00
18b22c9df8 fix(tradein/v2): canonical sources_used — kill «ИСТОЧНИКОВ X/7» POST↔GET desync
All checks were successful
CI / changes (pull_request) Successful in 7s
CI / backend-tests (pull_request) Has been skipped
CI / frontend-tests (pull_request) Has been skipped
CI / openapi-codegen-check (pull_request) Has been skipped
Единый helper estimator._canonical_sources(analogs, valuation_flags) — источник
правды для sources_used, зовётся идентично на POST (estimate_quality) и
GET-rehydrate (get_estimate) из ТЕХ ЖЕ persisted analogs + оценочных флагов.

Root cause: три расходящиеся деривации — POST брал sources_used из top-N
analogs_lots, GET возвращал persisted-колонку (радиусный набор + quarter_index),
source_counts на GET считался из persisted analogs. Источник (напр. cian) мог
попасть в source_counts, но не в sources_used → счётчик «X/7» прыгал 4→5 между
POST-ответом и reload(GET).

- sources_used = {листинговые из persisted analogs} ∪ {avito_imv/yandex_valuation/
  cian_valuation}. Детерминированно отсортирован.
- source_counts на POST теперь тоже из analogs_lots (не полной metadata-выборки)
  → инвариант source_counts.keys() ⊆ sources_used на POST и GET.
- POST персистит канонический sources_used в колонку (history/PDF консистентны для
  новых строк); GET рехайдрейтит его же helper'ом — чинит и СТАРЫЕ строки
  (листинговая часть пересобирается из analogs, quarter_index/радиусный шум
  отбрасывается фильтром оценочных флагов).

Оценочные флаги персистятся в колонке sources_used и читаются оттуда на GET —
реконструкция не требуется.

Repro 451de30b: до — sources_used=[avito,avito_imv,domklik,yandex], source_counts
имеет cian (не в sources_used); после — sources_used=[avito,avito_imv,cian,
domklik,yandex] (5/7), counts.keys() ⊆ sources_used.

Part of #2087 (M1).
2026-07-02 17:22:20 +03:00
24b107ddb5 feat(tradein): #2043 BE-1 — surface cv/source_counts/created_at + /history.sources_used + cache-KPI
Данные уже считаются в estimator — отдаём наружу для /trade-in/v2 (снимает
approximate-флаги CV / счётчиков источников / даты отчёта / «ИСТОЧНИКОВ N/7»).

- AggregatedEstimate += cv (float|None), source_counts (dict[str,int]), created_at.
- estimator: _cv_from_ppm2 / _source_counts helpers. cv прокинут через
  PricingResult — anchor-путь берёт CV комплов (anchor["cv"]), radius-путь — CV
  радиусной ₽/м²-выборки. source_counts считается по ПОЛНОЙ выборке (metadata_lots)
  до top-N отсечки. created_at = момент создания.
- POST /estimate возвращает все три; GET /estimate/{id} пересчитывает cv/
  source_counts из сохранённых analogs (best-effort) + created_at из колонки.
- /history: += sources_used (jsonb) в проекции → колонка «ИСТОЧНИКОВ N/7».
- /cache-stats: += avg_median_price (по median_price>0) + repeat_address_pct
  (доля строк с неуникальным address). Честный best-effort по persisted-оценкам.

Refs #2043
2026-07-02 16:50:18 +03:00
e23dabe4f5 fix(tradein/v2): wire РАДИУС (backend radius_m) + map +/- zoom + building image + deglue word+word
All checks were successful
CI / changes (pull_request) Successful in 7s
CI / backend-tests (pull_request) Has been skipped
CI / frontend-tests (pull_request) Has been skipped
CI / openapi-codegen-check (pull_request) Has been skipped
Live-browser audit (round 2): РАДИУС/+- dead, building photo missing.
- РАДИУС: backend BE-2 — optional radius_m (int 100–5000) on TradeInEstimateInput,
  threaded into estimate_quality comp search (base + fallback). None preserves the
  exact two-tier default (1000m primary / 2000m fallback) byte-identically. FE: РАДИУС
  dropdown with **Авто** default (sends no radius_m → legacy behaviour, NOT a silent
  500m narrowing) + 300/500/1000/2000 overrides; included as radius_m on submit; outer
  map ring tracks the value.
- Map +/− buttons: now a real zoom (transform: scale() on the map content layer, step
  .25, clamp .6–2.5); controls/bracket stay unscaled.
- building.png: next/image didn't apply basePath → 400; switched to plain <img> with
  NEXT_PUBLIC_BASE_PATH-prefixed src (→ 200) + onError fallback. HERO АДРЕС card шоу.
- deglueAddr: also splits word+word glue ('БольшаковаГеологическая'→'Большакова
  Геологическая'), still preserves house letters (14А, д.5К).

next build green (/v2 34.3 kB); estimator None-default byte-identical (ge=100 rules out
falsy-0); code-reviewer  after the Авто-default fix (avoided a default-radius regression).
2026-06-28 18:03:07 +03:00
44b9305d89 feat(tradein/estimator): elite-segment manual_review trigger by asking ppm² (#2002)
All checks were successful
Deploy Trade-In / changes (push) Successful in 10s
Deploy Trade-In / build-frontend (push) Has been skipped
Deploy Trade-In / build-browser (push) Has been skipped
Deploy Trade-In / test (push) Successful in 1m25s
Deploy Trade-In / build-backend (push) Successful in 51s
Deploy Trade-In / deploy (push) Successful in 48s
2026-06-27 20:15:38 +00:00
e39fb86a9b feat(tradein): AI-curated premium-building overlay — class + false-positive demotion (#2002)
All checks were successful
Deploy Trade-In / changes (push) Successful in 10s
Deploy Trade-In / build-browser (push) Has been skipped
Deploy Trade-In / test (push) Successful in 1m29s
Deploy Trade-In / deploy (push) Successful in 52s
Deploy Trade-In / build-frontend (push) Has been skipped
Deploy Trade-In / build-backend (push) Successful in 56s
2026-06-27 18:43:05 +00:00
275e7e9618 feat(tradein/estimator): manual_review recommendation flag (#2002)
All checks were successful
Deploy Trade-In / changes (push) Successful in 11s
Deploy Trade-In / build-frontend (push) Has been skipped
Deploy Trade-In / build-browser (push) Has been skipped
Deploy Trade-In / test (push) Successful in 1m30s
Deploy Trade-In / build-backend (push) Successful in 1m2s
Deploy Trade-In / deploy (push) Successful in 49s
2026-06-27 18:17:24 +00:00
50684ce87e fix(tradein/estimator): remove dead tier-aware-ratio path — truncation artifact + footgun (#2002)
All checks were successful
Deploy Trade-In / changes (push) Successful in 10s
Deploy Trade-In / build-frontend (push) Has been skipped
Deploy Trade-In / build-browser (push) Has been skipped
Deploy Trade-In / test (push) Successful in 1m25s
Deploy Trade-In / deploy (push) Successful in 49s
Deploy Trade-In / build-backend (push) Successful in 53s
2026-06-27 17:34:54 +00:00
840d64a4a0 feat(tradein): premium_houses MV + premium_building flag (#2002)
All checks were successful
Deploy Trade-In / changes (push) Successful in 10s
Deploy Trade-In / build-browser (push) Has been skipped
Deploy Trade-In / build-frontend (push) Has been skipped
Deploy Trade-In / test (push) Successful in 1m33s
Deploy Trade-In / build-backend (push) Successful in 56s
Deploy Trade-In / deploy (push) Successful in 51s
2026-06-27 16:40:39 +00:00
c45445b726 feat(tradein/estimator): hedonic year+area correction on expected_sold (#2002)
All checks were successful
Deploy Trade-In / changes (push) Successful in 11s
Deploy Trade-In / build-frontend (push) Has been skipped
Deploy Trade-In / build-browser (push) Has been skipped
Deploy Trade-In / build-backend (push) Successful in 54s
Deploy Trade-In / deploy (push) Successful in 47s
Deploy Trade-In / test (push) Successful in 1m25s
2026-06-27 16:06:08 +00:00
9448a945d4 feat(tradein/estimator): calibrated 80% prediction interval for expected_sold range (#1966)
All checks were successful
Deploy Trade-In / test (push) Successful in 1m26s
Deploy Trade-In / deploy (push) Successful in 47s
Deploy Trade-In / changes (push) Successful in 10s
Deploy Trade-In / build-frontend (push) Has been skipped
Deploy Trade-In / build-browser (push) Has been skipped
Deploy Trade-In / build-backend (push) Successful in 51s
2026-06-27 14:01:24 +00:00
24a80ee0cf wip: extract pricing spine (pre-test) 2026-06-27 12:51:14 +03:00
bcdc7ccb04 fix(estimator): deterministic comp tie-breaker + n_analogs counts priced analogs
All checks were successful
CI / backend-tests (pull_request) Has been skipped
CI / changes (pull_request) Successful in 6s
CI / frontend-tests (pull_request) Has been skipped
CI / openapi-codegen-check (pull_request) Has been skipped
Found by adversarial valuation audit (2 confirmed, bot-safe).

FIX A (#5): both radius comp queries (Tier H ~3990, Tier W ~4135) ended with
a bare ORDER BY relevance_score; on score ties Postgres returned rows in
undefined order, so the same /analyze could pick different comps across runs.
Append deterministic tiebreaker: relevance_score ASC, distance_m ASC,
scraped_at DESC NULLS LAST, id ASC (id = listings PK → total order). Added id
to each base CTE; outer projection unchanged (no leak downstream).

FIX B (#2): _filter_outliers keeps rows with price_per_m2 IS NULL, but the
median is built from prices_ppm2 (drops them) while n_analogs counted all of
listings_clean — overstating contributing comps ("Найдено N аналогов"
misleading; all-price-less -> median=0 but n_analogs>0). Count n_analogs from
prices_ppm2 in the radius path. #698 anchor overwrite + #691 zero-analog->low
guard unaffected; listings_clean itself not filtered.

Adds tests/test_estimator_n_analogs_priced.py (verified to fail on old code).
Audit also flagged velocity fan-out (false-positive: 0 duplicate domrf_obj_id
on prod) and >/>= disclosure tweaks (cosmetic) — deliberately not changed.

Refs #1871
2026-06-27 05:57:04 +05:00
5d5b2981fc Merge pull request 'fix(estimator): флаг + тесты для ghost-anchor guard и radius-dedup (#1871)' (#1919) from fix/estimator-trust-ghost-dedup into main
Some checks failed
Deploy Trade-In / test (push) Blocked by required conditions
Deploy Trade-In / build-backend (push) Blocked by required conditions
Deploy Trade-In / build-frontend (push) Blocked by required conditions
Deploy Trade-In / build-browser (push) Blocked by required conditions
Deploy Trade-In / deploy (push) Blocked by required conditions
Deploy Trade-In / changes (push) Has been cancelled
Reviewed-on: #1919
2026-06-26 16:18:58 +00:00
2fe34ac785 fix(estimator): ghost-anchor — confidence='low' when radius analogs empty (#1871) (#1920)
All checks were successful
Deploy Trade-In / changes (push) Successful in 10s
Deploy Trade-In / build-frontend (push) Has been skipped
Deploy Trade-In / build-browser (push) Has been skipped
Deploy Trade-In / test (push) Successful in 1m26s
Deploy Trade-In / deploy (push) Successful in 48s
Deploy Trade-In / build-backend (push) Successful in 58s
2026-06-26 16:07:12 +00:00
68f7c0e015 fix(estimator): запрет confidence=high при n_analogs=0 (#1871 ghost-anchor)
Добавляет settings-флаг estimate_confidence_floor_no_analogs (дефолт True),
гейтящий вызов _enforce_zero_analog_low перед сборкой AggregatedEstimate.
При n_analogs==0 и confidence!='low' форсит 'low' + добавляет caveat в
explanation («без сопоставимых аналогов рядом»), предотвращая показ выдуманного
«высокого» доверия когда оценка построена только на внешних оценщиках
(yandex_valuation/cian_valuation) без реальных рыночных аналогов.
2026-06-26 18:50:19 +03:00
77a3d3f75a fix(estimator): enable split-home disclosure at validated threshold 1.2 (audit #1871)
All checks were successful
CI / backend-tests (pull_request) Has been skipped
CI / frontend-tests (pull_request) Has been skipped
CI / changes (pull_request) Successful in 6s
CI / openapi-codegen-check (pull_request) Has been skipped
Включает wide-corridor disclosure (PR #1880, был default OFF) — но с
ИСПРАВЛЕННЫМ порогом по prod-данным.

Проверка на trade_in_estimates (60д): corridor_pct = (range_high-range_low)/
median_price имеет median≈0.48, p90≈0.93, p99≈1.38; **30.7% оценок > 0.6**.
Старый порог 0.6 прилепил бы caveat «дом разбит на секции» к ~трети оценок —
большинство НЕ split-дома (широкий коридор бывает от разных причин). Это была
бы ложная атрибуция (дезинформация).

- estimate_wide_corridor_threshold 0.6 → 1.2: genuine split-дома из аудита =
  148-170% (1.48-1.70) → 1.2 ловит только экстремальный хвост (>p99), не трогая
  нормальную оценочную неопределённость.
- estimate_wide_corridor_disclosure_enabled False → True (включаем после
  валидации). ENV-override сохранён (kill-switch).
- Формулировка смягчена: «дом разбит на секции» → «вероятно, дом разбит на
  секции разной этажности ИЛИ разнородный фонд» — мы ИНФЕРИМ split по ширине
  коридора, не доказываем структурно.
- Логика не тронута: фаерит только Tier A + corridor>threshold, только понижает
  confidence + дописывает explanation; point/median/range не трогает.

Тесты: обновлены default-ассерты (флаг ON, порог 1.2) + 2 behavioral boundary-
теста (фаерит при pct>threshold, НЕ фаерит при умеренном 1.045<1.2 — доказывает
что raise убрал false-positive flood). 13 passed.

Refs #1871
2026-06-24 01:54:58 +05:00
1b23fc5d95 fix(estimator): radius (source,source_id) dedup + split-house corridor disclosure (audit #1871 P2)
All checks were successful
CI / changes (pull_request) Successful in 6s
CI / backend-tests (pull_request) Has been skipped
CI / frontend-tests (pull_request) Has been skipped
CI / openapi-codegen-check (pull_request) Has been skipped
Два P2-фикса в чувствительном estimator.py, оба за флагами.

DEDUP (estimate_radius_dedup_enabled, default ON):
radius-путь _fetch_analogs (Tier S-canonical/S-fallback/H/W) кэпил только
per-address (rn_addr ≤ MAX_ANALOGS_PER_ADDRESS=5), но (source,source_id)-дубли
делят один address и выживали на разных rn_addr рангах → раздували n_analogs
(prod 2026-06-23: yandex 48, cian 9, n1 5 excess). Добавлен rn_dup-window
PARTITION BY source, CASE(id:source_id|url:source_url|ctid) ORDER BY scraped_at
DESC + AND rn_dup=1 во все 4 тира — зеркалит anchor-dedup (:1556). CASE/ctid-guard
не схлопывает legacy NULL-source_id. ctid доступен (все тиры FROM listings
single-table). psycopg3: column::type (легально), bind-param :: отсутствует.

SPLIT DISCLOSURE (estimate_wide_corridor_disclosure_enabled, default OFF):
Tier A матчит по address-regex (намеренно — дом дробится на house_id), на
split-доме разной этажности comp_min..max растягивает коридор до 148-170%.
При corridor_pct > threshold (0.6) → понизить confidence на ступень + caveat
«дом разбит на секции разной этажности — оценка ориентировочная». НЕ трогает
point/median/range — только confidence+explanation. Helper _downgrade_confidence.
Default OFF (no-regress, включим осознанно). Merge house_id отклонён (recon:
вреден — Tier A намеренно обходит house_id-группировку).

Тесты: +22 (dedup 4 тира + семантика + psycopg3-regex guard; split disclosure
через реальный _run_estimate harness + does-not-touch-median/range + helper).
399 estimator passed, ruff clean, mypy pre-existing-only.

Refs #1871
2026-06-23 14:48:02 +05:00
41e9c820c9 fix(tradein/estimator): ghost-anchor invariant guard — n_analogs=0 forces low (audit #1871 P1.2)
All checks were successful
CI / changes (pull_request) Successful in 7s
CI / backend-tests (pull_request) Has been skipped
CI / frontend-tests (pull_request) Has been skipped
CI / openapi-codegen-check (pull_request) Has been skipped
Audit #1871 P1.2: estimate 5fcc1e99 показывал confidence='high' при
n_analogs=0 (median 38.45M ₽, sources=['yandex_valuation']) при реальной
истории листингов дома 4.2-6.5M. По prod-DB это ЕДИНСТВЕННАЯ legacy-строка
от 2026-05-30, давно expired (TTL оценок = 1 день, все 66 high/medium-c-n<5
строк expired). Текущий mainline уже честен. Фикс — defensive invariant,
чтобы ghost не мог родиться снова из будущих external-valuation-blend /
rehydrate / migration путей.

Backend (estimator.py):
- helper _enforce_zero_analog_low(confidence, n_analogs, explanation, ...)
  форсит 'low' + caveat когда n_analogs==0 and confidence!='low'.
  Идемпотентен (low+0 → low без дубля caveat), не трогает high+N>0, реальный
  median внешнего сервиса сохраняется (fake-коэффициенты не вводятся),
  logger.warning для observability.
- Вызов перед сборкой AggregatedEstimate в estimate_quality (единственный
  уязвимый return-путь; _empty_estimate хардкодит low — структурно безопасен).
- +5 unit-тестов (high+0→low+caveat, medium+0→low, low+0 идемпотент,
  high+N unchanged, None-explanation).

Frontend (HeroSummary, HeroTransparency, history):
- effectiveConfidence = n_analogs===0 ? 'low' : confidence — defensive depth
  на случай legacy-rehydrate / контракт-drift. Caveat-текст из backend
  confidence_explanation (не хардкод). Type-safe (ConfidenceLevel).

Refs #1871
2026-06-23 11:05:34 +05:00
92725c6122 fix(tradein): Mera-audit fix-4 — верифицировать DEAL_MAX_PPM2 vs prod данных
All checks were successful
CI / changes (pull_request) Successful in 7s
CI / backend-tests (pull_request) Has been skipped
CI / frontend-tests (pull_request) Has been skipped
CI / openapi-codegen-check (pull_request) Has been skipped
Данные на 2026-06-21 (prod deals table, 49 791 записей):
  p90=172 419 ₽/м², p95=199 583 ₽/м², p99=278 681 ₽/м², p99.9=500 000 ₽/м²
  deals >800k = 13 штук (0.026%) — нерыночные outlier'ы, не легитимный премиум ЕКБ.
  listings p99 (is_active, <800k фильтр) = 392 795 ₽/м².

Вывод: DEAL_MAX_PPM2=800_000 НЕ срезает легитимный ЕКБ-премиум (p99.9 сделок=500k).
Значение NOT изменено — задокументировано в комментарии рядом с константой
для обоснования порога и последующей ревалидации при росте рынка.
2026-06-21 09:06:36 +03:00
af1d882745 fix(tradein): Mera-audit fix-3 — cross-source dedup в _fetch_price_trend
Один объект на avito_imv + yandex_valuation с разными ext_item_id создаёт
дубликаты в house_placement_history и double-count'ится в помесячной медиане.
Решение: CTE с DISTINCT ON (round(area_m2), floor, price, price_date) с
приоритетом avito_imv через CASE sort. Легитимные разные квартиры (разные
площадь/этаж/цена) не затрагиваются.

За флагом estimate_price_trend_dedup_enabled (дефолт True в settings).
False → точно старое поведение (backward-compat). Graceful: без изменений
в логике min_points и возврата None при недостатке данных.

Тесты: 6 новых тест-кейсов в test_estimator_price_trend_dedup.py (SQL содержит
DISTINCT ON, flag OFF не содержит, Source1 preferred, данные возвращаются корректно).
2026-06-21 09:05:57 +03:00
81ee864d80 fix(tradein): Mera-audit fix-2 — _is_plausible_deal floor edge + area/price guards
Ужесточает валидацию ДКП-сделок в _is_plausible_deal:
- floor < 1 или floor=0 → drop (floor=-5/0/999 из битого парсера)
- area_m2 задана и <= 0 → drop (нулевая/отрицательная площадь из битого парсера)
- price_rub задана и <= 0 → drop (нерыночная/техническая сделка)
- floor=None → допустимо (graceful, нечем судить)

Все новые параметры area_m2/price_rub keyword-only с дефолтом None →
backward-compatible для существующих вызовов. Caller _fetch_deals обновлён
для передачи area_m2/price_rub из данных сделки.

Тесты: 16 новых тест-кейсов в test_deals_sanitize.py (floor=-5, floor=999,
floor=None ok, area_m2=0 drop, price_rub=-1 drop и пр.).
2026-06-21 09:04:54 +03:00
b2cd2925fd fix(estimator): clamp expected_sold <= asking (ratio cap 1.0) -- smoke 3к exp_sold>asking
All checks were successful
CI / frontend-tests (pull_request) Has been skipped
CI / openapi-codegen-check (pull_request) Has been skipped
CI / changes (pull_request) Successful in 8s
CI / backend-tests (pull_request) Has been skipped
2026-06-21 00:35:23 +03:00
f2cfe66b02 fix(scrapers): yandex history skip area<=0 items
_save_yandex_history_items теперь пропускает items с area_m2 is None или
area_m2 <= 0 (битый парс «0,5 м²» и пр.) перед сохранением в
house_placement_history. Грязь копилась в БД и искажала price_trend.

Estimator защищён NULLIF на уровне SQL, но фильтрация на входе надёжнее.
Лог кол-ва отброшенных (info). Батч не падает из-за одного битого item.
2026-06-21 00:25:43 +03:00
b6ea986134 fix(estimator): radius-path нижний floor от DKP-коридора (анти-undershoot)
Anchor-путь имеет hard floor от comp_min_ppm2×(1-tol). Radius-путь аналогичной
защиты снизу не имел — quarter-index-вниз или corridor-clamp могли занизить
median неоправданно (asymmetry).

Добавлен floor: если итоговый median_ppm2 < dkp_low_ppm2 × factor → поднять
до floor + лог. Только radius-путь (anchor_tier is None); без dkp_raw → no-op.
За флагами estimate_radius_floor_enabled (default True) и
estimate_radius_floor_factor (default 0.8).
2026-06-21 00:25:05 +03:00
352dfc8053 fix(estimator): asking_to_sold ratio tier по финальному headline
_get_asking_sold_ratio теперь вызывается ПОСЛЕ anchor/IMV-blend/quarter-index/
corridor-clamp, получая ФИНАЛЬНЫЙ median_ppm2 для tier-placement (#928 audit).

До фикса: tier резолвился по pre-anchor radius-медиане (~105k), но ratio
применялся к post-anchor headline (~300k) — несовпадение tier'а в premium-
сегменте. После: ratio берётся из того же tier, что и финальный headline.

Graceful: нет ratio → expected_sold_* = None, headline не меняется.
2026-06-21 00:24:02 +03:00
daa2edfac3 test(estimator): unit tests для audit fixes #audit-1..5 + empty-series guard
All checks were successful
CI / changes (pull_request) Successful in 6s
CI / backend-tests (pull_request) Has been skipped
CI / frontend-tests (pull_request) Has been skipped
CI / openapi-codegen-check (pull_request) Has been skipped
18 тестов покрывают все 5 фиксов: low-conf gate, analog_tier enum,
price_trend freshness, sigma=0 guard, sber staleness, thin-market flag.
Попутный fix: _load_sber_index_series — guard 'if not series: continue'
предотвращает max() на пустом dict при MagicMock rows (truthy=True).
2026-06-20 15:06:59 +03:00
a6ef2c73e1 fix(estimator): data-age guards — sber staleness + IMV thin-market (#audit-5)
5a: _load_sber_index_series логирует warning если latest месяц серии
старее sber_index_max_age_days (дефолт 35) — расчёт не блокируется.
5b: AvitoImvSummary.thin_market (bool, дефолт False) — True когда
market_count < avito_imv_thin_market_threshold (дефолт 10). Все три
точки конструирования AvitoImvSummary обновлены. Warning при thin.
2026-06-20 15:03:19 +03:00
5b2a39cd5e fix(estimator): MAD-clip после similarity-weighting + sigma=0 guard (#audit-4)
sigma=0 guard: safe_area_sigma2/safe_floor_sigma2 предотвращает div/0 при
нулевом sigma из конфига. Post-weight MAD-clip (estimate_sb_clip_after_weight,
дефолт True) выполняется ПОСЛЕ Gaussian weighting — видовые/топ-юниты
с высоким ppm² больше не выкидываются до учёта similarity. За флагом.
2026-06-20 15:01:36 +03:00
17edac532b fix(estimator): Yandex history freshness filter в price_trend (#audit-3)
_fetch_price_trend(freshness_months) — новый параметр, дефолт из
settings.estimate_price_trend_max_age_months (6 мес).
WHERE scraped_at > now() - N months исключает устаревшие items
(yandex_valuation 2024 и т.д.) из house_placement_history fallback.
houses_price_dynamics (Source 1) не затронут. psycopg v3 CAST syntax.
2026-06-20 15:00:55 +03:00
6b99f7d238 fix(estimator): структурный analog_tier + address_precision в API (#audit-2)
Добавляет AggregatedEstimate.analog_tier — стабильный enum для фронта:
same_building (Tier A), micro_radius (Tier C), district (S/H/0), city (W),
null (нет данных). Фронт не парсит tier из explanation.
address_precision уже был в схеме (подтверждено: line 208). Explanation не
удаляется — фронт fallback'ает на него.
2026-06-20 15:00:12 +03:00
55c7bc2c72 fix(estimator): anchor low-confidence gate (#audit-1)
Якорь с confidence=low ИЛИ (n<gate_min_n И FSD>gate_max_fsd) не заменяет
headline — fallback на radius-median. Дефолты (min_n=3, max_fsd=0.20) не
режут здоровые якоря (n≥4, FSD<0.15 проходят). За флагом
estimate_sb_low_conf_gate_enabled (дефолт True). Логируется с причиной.
2026-06-20 14:59:05 +03:00
494f2e00e0 fix(tradein/estimator): clamp premium headline к коридору сделок + робастный anchor (#1795)
All checks were successful
CI / changes (pull_request) Successful in 6s
CI / backend-tests (pull_request) Has been skipped
CI / frontend-tests (pull_request) Has been skipped
CI / openapi-codegen-check (pull_request) Has been skipped
2026-06-19 20:34:48 +03:00
99e2a5d75c feat(tradein): estimate accepts client-provided coords, skips geocode (Variant A)
All checks were successful
CI / changes (pull_request) Successful in 6s
CI / backend-tests (pull_request) Has been skipped
CI / frontend-tests (pull_request) Has been skipped
CI / openapi-codegen-check (pull_request) Has been skipped
2026-06-19 14:15:02 +03:00
80380dc32a fix(tradein): budget Cian valuation call to stop 25s estimate stalls
All checks were successful
CI / openapi-codegen-check (pull_request) Has been skipped
CI / changes (pull_request) Successful in 5s
CI / backend-tests (pull_request) Has been skipped
CI / frontend-tests (pull_request) Has been skipped
The Cian Valuation enrichment was the only ungated external call in
estimate_quality() — its internal curl timeout (~25s) blocks the whole
/estimate request on every cache-miss, intermittently pushing latency
past the gateway/patience window (perceived as 'сервис не считает').

Wrap it in _with_budget() (mirrors the existing Yandex-valuation guard):
on timeout it degrades to None — the same graceful path as a network
error — instead of stalling. New setting estimate_cian_valuation_timeout_s
(default 8.0s, env ESTIMATE_CIAN_VALUATION_TIMEOUT_S).
2026-06-19 12:43:28 +03:00
07d9cc8e8b fix(estimator): #1774 Tier A admits same-building novostroyki resales when secondary-dominant
All checks were successful
CI / changes (pull_request) Successful in 6s
CI / backend-tests (pull_request) Has been skipped
CI / frontend-tests (pull_request) Has been skipped
CI / openapi-codegen-check (pull_request) Has been skipped
Tier A (same building) dropped cian listings tagged listing_segment='novostroyki'
even in delivered buildings where that tag marks owner resales/переуступки
(sale_type=free). Gated relaxation of the #1186 guard, Tier A ONLY: include
novostroyki rows iff the same house has secondary (vtorichka/NULL) listings AND
secondary_count >= primary_count — secondary-dominant signal of a delivered house.
Prevents a lone vtorichka + clustered developer-priced novostroyki from anchoring
on застройщик prices after MAD-clip drops the rare secondary (~27% overvaluation).
Pure-primary / primary-dominated houses keep the #1186 guard. Behind
estimate_sb_tier_a_allow_primary_if_secondary_present (default True).

Dedup duplicate active rows by (source, source_id) primary key (codebase canon,
namespaced id/url tags, source_url fallback) — collapses rows whose source_url
differs only by trailing slash / query-param (e.g. cian source_id=330047129 ×2).

Tier C / radius / asking_to_sold_ratio paths untouched. 84 unit tests pass.
2026-06-19 14:00:13 +05:00
6eda4a6d0a fix(estimator): DKP corridor uses P10/P90 not min/max (#1520)
All checks were successful
CI / changes (push) Successful in 8s
CI / changes (pull_request) Successful in 7s
CI / backend-tests (push) Has been skipped
CI / frontend-tests (push) Has been skipped
CI / openapi-codegen-check (push) Has been skipped
CI / backend-tests (pull_request) Has been skipped
CI / frontend-tests (pull_request) Has been skipped
CI / openapi-codegen-check (pull_request) Has been skipped
Replace absolute min/max with robust percentiles so a single outlier ДКП
deal cannot shift the soft-bound corridor boundaries. Small samples fall
back gracefully via linear interpolation (n=3: P10≈index 0.2, P90≈index 2.8).
2026-06-17 20:50:15 +03:00
86e9ea2937 fix(week-review): автофиксы код-ревью — 169 issue (label «week ревью 1»)
Многоагентный аудит + имплементация: один воркер на файл, точечные правки.
Верификация: py_compile (47/47 .py) + tsc --noEmit (0 ошибок). Unit-тесты
не прогонялись (окружение не поднято: rollup native dep / нет pytest-venv).

Полностью исправлено (169): #1336, #1337, #1339, #1340, #1341, #1342, #1343, #1345, #1346, #1348, #1349, #1350, #1351, #1354, #1356, #1358, #1359, #1360, #1362, #1364, #1365, #1366, #1367, #1368, #1369, #1370, #1371, #1372, #1373, #1374, #1375, #1376, #1377, #1378, #1379, #1380, #1381, #1382, #1384, #1385, #1386, #1387, #1388, #1389, #1390, #1391, #1392, #1394, #1395, #1396, #1397, #1399, #1400, #1401, #1402, #1403, #1404, #1408, #1409, #1410, #1411, #1412, #1413, #1414, #1415, #1416, #1417, #1418, #1420, #1423, #1425, #1426, #1427, #1428, #1429, #1430, #1431, #1432, #1433, #1434, #1435, #1437, #1438, #1439, #1440, #1441, #1442, #1443, #1444, #1445, #1446, #1447, #1448, #1449, #1450, #1451, #1452, #1453, #1454, #1455, #1456, #1457, #1458, #1459, #1460, #1461, #1462, #1463, #1464, #1465, #1466, #1467, #1468, #1469, #1471, #1472, #1473, #1474, #1476, #1478, #1479, #1481, #1482, #1483, #1484, #1485, #1487, #1488, #1489, #1490, #1491, #1492, #1493, #1494, #1495, #1496, #1497, #1499, #1500, #1501, #1502, #1504, #1505, #1506, #1507, #1510, #1514, #1515, #1516, #1517, #1518, #1519, #1521, #1522, #1523, #1524, #1525, #1526, #1527, #1528, #1529, #1531, #1532, #1533, #1534, #1535, #1536, #1537, #1538

Частично (9, in-file часть, остаток cross-file): #1361, #1419, #1422, #1424, #1470, #1475, #1477, #1480, #1498
Требуют cross-file (3, не тронуты): #1338, #1363, #1421
Пропущено (1): #1539

Не входило в партию: 22 needs-Leha issue (нужны решения владельца).

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-15 20:21:11 +05:00
a9a489bba5 refactor(tradein): убрать dead listing_segment param из _fetch_anchor_comps + негативные guard-asserts (review fixup, #1186)
All checks were successful
CI / frontend-tests (push) Has been skipped
CI / changes (pull_request) Successful in 6s
CI / backend-tests (pull_request) Has been skipped
CI / frontend-tests (pull_request) Has been skipped
CI / changes (push) Successful in 6s
CI / backend-tests (push) Has been skipped
- Удалён параметр listing_segment: str | None = None из сигнатуры _fetch_anchor_comps
  (guard хардкожен в SQL-блоках Tier A/C, параметр не использовался)
- Убраны seg_counts / target_segment в caller-блоке SAME-BUILDING ANCHOR (~L2174)
- Docstring Tier C обновлён: «вторичка-канон guard (#1186): NULL = legacy вторичка»
- test_segment_guard_1186.py: добавлены 4 негативных assert'а:
  · no bare = 'vtorichka' без IS NULL OR
  · no deprecated <> / != 'novostroyki' form
  · _fetch_anchor_comps не содержит listing_segment в сигнатуре
  · Tier C не использует CAST(:segment AS text) (параметрическую форму)
2026-06-12 12:01:00 +03:00
b70ae46885 fix(tradein): guard listing_segment во всех читателях listings — novostroyki вне comps (#1186)
Канон-предикат: (listing_segment IS NULL OR listing_segment = 'vtorichka')
NULL = legacy вторичка до миграции 011 (sweep'ы были secondary-only) — отбрасывать нельзя.

Затронуты все SELECT-пути, влияющие на оценку и корректирующие коэффициенты:
- estimator.py: _COMMON_WHERE (Tier S/H), Tier W inline, Tier A anchor (добавлен впервые),
  Tier C anchor (заменён с параметрического CAST(:segment) на канон — прежний вариант
  отбрасывал NULL-строки при segment='vtorichka').
- asking_to_sold_ratio.py: ask_side, ask_global (_REDERIVE_SQL); bounds (_REDERIVE_BOUNDS_SQL);
  ask_tiered, ask_side_all, ask_global (_REDERIVE_TIERED_SQL) — 6 мест.
- sber_index.py: reconciliation-запрос (диагностический asking-median).

Не затронуты: скрейперы/scrape_pipeline, search matview (catalog display, не оценка),
admin stats, geocoding, cian price history, matching — обосновано в инвентаризации.

Тесты: новый test_segment_guard_1186.py (статика SQL + поведение mock); обновлён
test_asking_to_sold_ratio.py (subset-тест нормализует guard, добавленный в refresh но
отсутствующий в 080-seed).
2026-06-12 11:52:09 +03:00
3294bc2021 fix(tradein): exclude novostroyki from radius-median comp pool (#935)
Some checks failed
Deploy Trade-In / test (push) Successful in 31s
Deploy Trade-In / deploy (push) Blocked by required conditions
Deploy Trade-In / changes (push) Successful in 5s
Deploy Trade-In / build-frontend (push) Has been skipped
Deploy Trade-In / build-browser (push) Has been skipped
Deploy Trade-In / build-backend (push) Has been cancelled
Co-authored-by: bot-backend <bot-backend@gendsgn.local>
Co-committed-by: bot-backend <bot-backend@gendsgn.local>
2026-05-31 21:33:23 +00:00
571980c56e feat(tradein): ppm²-tier segmentation of asking→sold ratio (#928, flag OFF) (#934)
All checks were successful
Deploy Trade-In / changes (push) Successful in 5s
Deploy Trade-In / build-frontend (push) Has been skipped
Deploy Trade-In / build-browser (push) Has been skipped
Deploy Trade-In / test (push) Successful in 31s
Deploy Trade-In / build-backend (push) Successful in 44s
Deploy Trade-In / deploy (push) Successful in 36s
Co-authored-by: lekss361 <lekss361@gendsgn.local>
Co-committed-by: lekss361 <lekss361@gendsgn.local>
2026-05-31 21:22:39 +00:00
4999c877ae feat(tradein): СберИндекс time-adjust for frozen Rosreestr ДКП + per-dashboard filter fix (#794) (#919)
Some checks failed
Deploy Trade-In / build-frontend (push) Has been skipped
Deploy Trade-In / build-backend (push) Has been skipped
Deploy Trade-In / deploy (push) Has been skipped
Deploy Trade-In / changes (push) Successful in 5s
Deploy Trade-In / build-browser (push) Has been skipped
Deploy Trade-In / test (push) Failing after 33s
Co-authored-by: lekss361 <lekss361@gendsgn.local>
Co-committed-by: lekss361 <lekss361@gendsgn.local>
2026-05-31 18:48:22 +00:00
42de30cd5c refactor(tradein): unify kadastr_num -> cadastral_number (#732)
p3 naming-consistency (cadastral cols 0% filled). Migration 094_cadastral_unify.sql
(idempotent DO-block IF EXISTS): drop+recreate listings_search_mv (MV) +
v_data_quality (view), RENAME deals/house_metadata kadastr_num->cadastral_number,
DROP listings.kadastr_num + listings_kadastr_idx. Code SQL-string renames in
search_query/estimator/base (Lot dataclass field NOT renamed). 240 tests pass.

WARN: destructive DDL auto-applies to prod on deploy -> MERGE-GATE post-demo.
Reviewer: verify recreated MV/view DDL vs LIVE pg_get_viewdef before apply
(reconstructed from migration history 050/046, not live capture).

Refs #732
2026-05-31 15:41:52 +03:00
ec82742456 feat(tradein): same-building anchor safe-guard (#755)
CLIENT-VISIBLE pricing (killer-factor, demo 2026-06-01). The same-building
anchor replaced headline median/range/confidence with an over-confident
premium on as few as 2 same-building asking comps. Human-chosen (Anton)
3-part safe-guard, all implemented:

1. min_comps 2 -> 4: anchor fires only with >=4 same-building comps.
2. confidence cap: at <5 anchor comps, confidence cannot be 'high' -> medium.
3. MAD-clip: drop price_per_m2 outliers among anchor comps before aggregation
   (k=estimate_sb_mad_k=3.5); if <min_comps survive, anchor does NOT fire
   (fallback to radius-median). Guards against price typos inflating anchor.

Anchor still REPLACES the headline (uplift-no-replace variant NOT chosen).
quarter-index/radius-tier/asking-sold untouched. 8 new unit tests; 61 anchor
tests pass, ruff clean.

MERGE-GATE (client-visible, like #764/#858): do NOT merge without a green
backtest #763 (held-out MAPE <= 18.7% + canonical A/B Малышева 125 /
Заводская 44а). Backtest needs live DB — pre-merge human/qa step.

Refs #755, #694
2026-05-31 12:19:50 +03:00
3f31d1f409 fix(estimator): sanity-clamp quarter-index factor (#859) (#863)
All checks were successful
Deploy Trade-In / changes (push) Successful in 5s
Deploy Trade-In / test (push) Successful in 28s
Deploy Trade-In / build-frontend (push) Has been skipped
Deploy Trade-In / build-backend (push) Successful in 39s
Deploy Trade-In / deploy (push) Successful in 34s
Co-authored-by: bot-backend <bot-backend@gendsgn.local>
Co-committed-by: bot-backend <bot-backend@gendsgn.local>
2026-05-31 07:20:59 +00:00
b16a45ef87 feat(tradein): per-cadastral-quarter price index in estimator (#764) (#858)
All checks were successful
Deploy Trade-In / changes (push) Successful in 5s
Deploy Trade-In / build-frontend (push) Has been skipped
Deploy Trade-In / test (push) Successful in 28s
Deploy Trade-In / build-backend (push) Successful in 40s
Deploy Trade-In / deploy (push) Successful in 34s
Co-authored-by: bot-backend <bot-backend@gendsgn.local>
Co-committed-by: bot-backend <bot-backend@gendsgn.local>
2026-05-31 06:53:21 +00:00
fd75f4d47a perf(estimator): executemany batch insert in _save_yandex_history_items (Refs #826) (#832)
Some checks failed
Deploy Trade-In / changes (push) Successful in 26s
Deploy Trade-In / build-frontend (push) Has been skipped
Deploy Trade-In / test (push) Failing after 27s
Deploy Trade-In / build-backend (push) Has been skipped
Deploy Trade-In / deploy (push) Has been skipped
Co-authored-by: bot-backend <bot-backend@gendsgn.local>
Co-committed-by: bot-backend <bot-backend@gendsgn.local>
2026-05-30 20:07:06 +00:00