lekss361/gendesign
1
0
Fork 0
Code Issues 57 Pull requests Projects Releases Packages Wiki Activity Actions

fix(tradein/estimator): remove dead tier-aware-ratio path — truncation artifact + footgun (#2002) #2015

Merged
lekss361 merged 1 commit from fix/tradein-remove-tier-ratio-footgun into main 2026-06-27 17:34:54 +00:00
Conversation 0 Commits 1 Files changed 8 +53 -779
8 changed files with 53 additions and 779 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

12
tradein-mvp/backend/app/core/config.py
View file

@ -323,18 +323,6 @@ class Settings(BaseSettings):
# ENV: ASKING_RATIO_PPM2_MAX.
asking_ratio_ppm2_max: int = 1_200_000
# ppm2-tier asking->sold lookup (#928) -- gate + empirical-Bayes shrink parameter.
# tier_aware_ratio_enabled=False (default) -> byte-identical flag-OFF path via
# asking_to_sold_ratios (old table). True -> tiered lookup from
# asking_to_sold_ratios_tiered (requires migration 098 applied).
# ENV: TIER_AWARE_RATIO_ENABLED.
tier_aware_ratio_enabled: bool = False
# Empirical-Bayes shrink weight denominator: w = n_deals / (n_deals + K).
# K=150 -> cell with 150 deals gets w=0.5 (50% own tier ratio, 50% rooms-'all').
# Higher K -> more shrinkage toward 'all' ratio (conservative); lower -> more tier.
# ENV: TIER_RATIO_SHRINK_K.
tier_ratio_shrink_k: int = 150
# SSRF-защита для admin scrape endpoints (#756).
# Список хостов которым разрешено передавать абсолютные URL в параметрах *_url.
# Относительные пути (без netloc) проходят без проверки — хост подставляется

220
tradein-mvp/backend/app/services/estimator.py
View file

@ -202,9 +202,18 @@ def _repair_coefficient(repair_state: str | None) -> float:
# Ratio дрейфует медленно (refresh-задача раз в сутки, Stage 4), поэтому 300с
# TTL более чем достаточно и снимает по SELECT'у с каждой оценки. Single-worker
# uvicorn/scheduler — GIL делает dict-доступ atomic enough (без явного lock).
#
# Tier-aware ppm²-путь (#928) УДАЛЁН (#2002): per-price-tier sold/asking — это
# артефакт ratio-of-truncated-medians. Бинирование SOLD-сделок по ASKING-перцентилям
# с делением within-tier медиан рождает мнимый «премиум продаётся ближе к asking»
# градиент (Monte-Carlo с КОНСТАНТНЫМ true sold/asking=0.84 точно воспроизводил
# прод-тиры 0.94/0.99/0.96 и перекошенный сплит сделок 57/27/15%; включение флага
# ухудшало прод — overall MAPE 14.6→17.2, эконом-bias +5.8→+18.1). Валидный
# within-price-tier sold/asking НЕВЫЧИСЛИМ из asking-less ДКП-сделок; корректное
# обусловливание дают per-rooms blend + захардженный hedonic (год+площадь).
_ASKING_SOLD_RATIO_CACHE_TTL_S = 300.0
# Cache key: (bucket, tier_or_None) — tier=None for flag-OFF / no-anchor path (#928).
_asking_sold_ratio_cache: dict[tuple[int, str | None], tuple[float | None, str | None, float]] = {}
# Cache key: rooms bucket (единственный legacy per-rooms путь после #2002).
_asking_sold_ratio_cache: dict[int, tuple[float | None, str | None, float]] = {}
def _get_asking_sold_ratio(
@ -216,210 +225,65 @@ def _get_asking_sold_ratio(
bucket = min(max(rooms or 0, 0), 4).
Flag-OFF (settings.tier_aware_ratio_enabled=False) или anchor_ppm2=None:
Запрос к asking_to_sold_ratios (старая таблица) — байт-в-байт как до #928.
Fallback: per-rooms строка → global -1.
Запрос к asking_to_sold_ratios (migration 080): per-rooms строка
(WHERE rooms_bucket = bucket AND district = '') → fallback на global -1
(basis='global_fallback'). Ничего нет → (None, None).
Flag-ON и anchor_ppm2 задан (#928 ppm2-tier path):
1. Читаем t33/t66 из asking_to_sold_tier_bounds для (bucket, '').
Нет bounds-строки → fallback к 'all' строке (шаг 3).
2. tier = 'low'/'mid'/'high' по anchor_ppm2 vs t33/t66.
Читаем asking_to_sold_ratios_tiered (bucket,'',tier) → tier_ratio, n_deals.
Shrink: w = n_deals/(n_deals + settings.tier_ratio_shrink_k).
Также читаем (bucket,'','all') как shrink-target.
shrunk = tier_ratio * w + all_ratio * (1 - w).
basis = f'per_rooms_tier:{tier}'.
3. Fallback (bucket,'','all') → basis='per_rooms_all'.
4. Fallback (-1,'','all') → basis='global_all'.
5. Ничего → (None, None).
anchor_ppm2 сохранён в сигнатуре для совместимости с call-site (резолвер
вызывается с ФИНАЛЬНЫМ headline ppm² после anchor/blend-мутаций — см.
_ratio_resolver), но БОЛЬШЕ НЕ используется: tier-aware путь удалён как
статистический артефакт (#2002, см. module-level комментарий выше).
Таблицы нет / любая ошибка → (None, None), НЕ raise (graceful).
Кэшируется на ключ (bucket, tier|None) с TTL _ASKING_SOLD_RATIO_CACHE_TTL_S.
Кэшируется на ключ bucket с TTL _ASKING_SOLD_RATIO_CACHE_TTL_S.
"""
bucket = min(max(rooms or 0, 0), 4)
# Determine tier key for cache:
# flag-OFF or no anchor_ppm2 → use None (legacy path, reads old table).
tier_key: str | None = None
use_tier_path = settings.tier_aware_ratio_enabled and anchor_ppm2 is not None
cached = _asking_sold_ratio_cache.get(bucket)
if cached is not None:
ratio, basis, fetched = cached
if (time.monotonic() - fetched) < _ASKING_SOLD_RATIO_CACHE_TTL_S:
return ratio, basis
cache_key: tuple[int, str | None] = (bucket, tier_key) # may be updated below
if not use_tier_path:
# ── Flag-OFF path: byte-identical behavior (legacy asking_to_sold_ratios) ──
# Use a distinct key "_legacy" to avoid collision with the flag-ON Step-3 'all'
# fallback which also caches under (bucket, None). Without this, a process that
# flips tier_aware_ratio_enabled ON without restart could serve the old-table ratio
# from the flag-OFF entry for up to the TTL — wrong table.
cache_key = (bucket, "_legacy")
cached = _asking_sold_ratio_cache.get(cache_key)
if cached is not None:
ratio, basis, fetched = cached
if (time.monotonic() - fetched) < _ASKING_SOLD_RATIO_CACHE_TTL_S:
return ratio, basis
ratio: float | None = None
basis: str | None = None
try:
row = db.execute(
text(
"""
SELECT ratio, basis FROM asking_to_sold_ratios
WHERE rooms_bucket = CAST(:b AS int) AND district = ''
"""
),
{"b": bucket},
).fetchone()
if row is None:
# Бакет тонкий (n<30 при seed'е) или отсутствует → global fallback (-1).
row = db.execute(
text(
"""
SELECT ratio, basis FROM asking_to_sold_ratios
WHERE rooms_bucket = -1 AND district = ''
"""
),
).fetchone()
if row is not None and row.ratio is not None:
ratio = float(row.ratio)
basis = row.basis
except Exception as exc:
# Таблицы может не быть на свежей/старой БД (миграция 080 не применена),
# либо транзакция в сбойном состоянии — graceful: без sold-коррекции.
# ОБЯЗАТЕЛЬНО rollback: неудачный SELECT помечает транзакцию
# InFailedSqlTransaction, и без отката следующий statement упал бы → 500.
logger.debug("asking_to_sold_ratio lookup skipped (graceful): %s", exc)
try:
db.rollback()
except Exception:
pass
ratio, basis = None, None
_asking_sold_ratio_cache[cache_key] = (ratio, basis, time.monotonic())
return ratio, basis
# ── Flag-ON path: ppm2-tier lookup (#928) ───────────────────────────────────
ratio: float | None = None
basis: str | None = None
try:
# Step 1: read bounds for this bucket.
bounds_row = db.execute(
row = db.execute(
text(
"""
SELECT t33, t66 FROM asking_to_sold_tier_bounds
SELECT ratio, basis FROM asking_to_sold_ratios
WHERE rooms_bucket = CAST(:b AS int) AND district = ''
"""
),
{"b": bucket},
).fetchone()
if bounds_row is not None:
t33 = float(bounds_row.t33)
t66 = float(bounds_row.t66)
tier_key = "low" if anchor_ppm2 < t33 else ("mid" if anchor_ppm2 < t66 else "high")
cache_key = (bucket, tier_key)
# Check cache for this (bucket, tier) key.
cached = _asking_sold_ratio_cache.get(cache_key)
if cached is not None:
c_ratio, c_basis, fetched = cached
if (time.monotonic() - fetched) < _ASKING_SOLD_RATIO_CACHE_TTL_S:
return c_ratio, c_basis
# Step 2: look up tier row + 'all' row for shrink.
tier_row = db.execute(
if row is None:
# Бакет тонкий (n<30 при seed'е) или отсутствует → global fallback (-1).
row = db.execute(
text(
"""
SELECT ratio, n_deals FROM asking_to_sold_ratios_tiered
WHERE rooms_bucket = CAST(:b AS int)
AND district = ''
AND ppm2_tier = CAST(:tier AS text)
SELECT ratio, basis FROM asking_to_sold_ratios
WHERE rooms_bucket = -1 AND district = ''
"""
),
{"b": bucket, "tier": tier_key},
).fetchone()
all_row = db.execute(
text(
"""
SELECT ratio FROM asking_to_sold_ratios_tiered
WHERE rooms_bucket = CAST(:b AS int)
AND district = ''
AND ppm2_tier = 'all'
"""
),
{"b": bucket},
).fetchone()
if tier_row is not None and all_row is not None:
tier_ratio = float(tier_row.ratio)
n_deals = int(tier_row.n_deals)
all_ratio = float(all_row.ratio)
k = settings.tier_ratio_shrink_k
w = n_deals / (n_deals + k)
ratio = tier_ratio * w + all_ratio * (1.0 - w)
basis = f"per_rooms_tier:{tier_key}"
_asking_sold_ratio_cache[cache_key] = (ratio, basis, time.monotonic())
return ratio, basis
# Step 3: fallback to 'all' for this bucket.
cache_key = (bucket, None)
cached = _asking_sold_ratio_cache.get(cache_key)
if cached is not None:
c_ratio, c_basis, fetched = cached
if (time.monotonic() - fetched) < _ASKING_SOLD_RATIO_CACHE_TTL_S:
return c_ratio, c_basis
all_row = db.execute(
text(
"""
SELECT ratio FROM asking_to_sold_ratios_tiered
WHERE rooms_bucket = CAST(:b AS int)
AND district = ''
AND ppm2_tier = 'all'
"""
),
{"b": bucket},
).fetchone()
if all_row is not None:
ratio = float(all_row.ratio)
basis = "per_rooms_all"
_asking_sold_ratio_cache[cache_key] = (ratio, basis, time.monotonic())
return ratio, basis
# Step 4: global fallback (-1, 'all').
global_key: tuple[int, str | None] = (-1, None)
cached = _asking_sold_ratio_cache.get(global_key)
if cached is not None:
c_ratio, c_basis, fetched = cached
if (time.monotonic() - fetched) < _ASKING_SOLD_RATIO_CACHE_TTL_S:
return c_ratio, c_basis
global_row = db.execute(
text(
"""
SELECT ratio FROM asking_to_sold_ratios_tiered
WHERE rooms_bucket = -1
AND district = ''
AND ppm2_tier = 'all'
"""
),
).fetchone()
if global_row is not None:
ratio = float(global_row.ratio)
basis = "global_all"
_asking_sold_ratio_cache[global_key] = (ratio, basis, time.monotonic())
return ratio, basis
if row is not None and row.ratio is not None:
ratio = float(row.ratio)
basis = row.basis
except Exception as exc:
# Таблицы нет / ошибка → graceful (migration 098 ещё не применена).
logger.debug("asking_to_sold_ratio tiered lookup skipped (graceful): %s", exc)
# Таблицы может не быть на свежей/старой БД (миграция 080 не применена),
# либо транзакция в сбойном состоянии — graceful: без sold-коррекции.
# ОБЯЗАТЕЛЬНО rollback: неудачный SELECT помечает транзакцию
# InFailedSqlTransaction, и без отката следующий statement упал бы → 500.
logger.debug("asking_to_sold_ratio lookup skipped (graceful): %s", exc)
try:
db.rollback()
except Exception:
pass
ratio, basis = None, None
# Step 5: nothing found.
return None, None
_asking_sold_ratio_cache[bucket] = (ratio, basis, time.monotonic())
return ratio, basis
# ── Avito IMV cache lookup (Stage 3) ────────────────────────────────────────

255
tradein-mvp/backend/app/tasks/asking_to_sold_ratio.py
View file

@ -189,237 +189,6 @@ _COUNTERS_SQL = text(
)
# ── Tiered refresh (#928): DELETE + re-derive for tier tables ────────────────
# атомарно DELETE WHERE district='' ПЕРЕД re-derive INSERT (те же семантика, что
# для asking_to_sold_ratios выше: stale строки не оставляем).
_DELETE_TIERED_ROWS_SQL = text(
"""
DELETE FROM asking_to_sold_ratios_tiered WHERE district = ''
"""
)
_DELETE_TIER_BOUNDS_SQL = text(
"""
DELETE FROM asking_to_sold_tier_bounds WHERE district = ''
"""
)
# ── Tiered derivation (BYTE-CONSISTENT с migration 098) ──────────────────────
# Те же CTE что в 098: bounds -> deal_tiered/ask_tiered -> tier_rows + all_rows + global_row.
# Bind params: :ppm2_min / :ppm2_max (как в _REDERIVE_SQL выше).
# CAST не нужен -- psycopg v3 передаёт int напрямую.
_REDERIVE_BOUNDS_SQL = text(
"""
WITH bounds AS (
SELECT
LEAST(GREATEST(rooms, 0), 4) AS rooms_bucket,
percentile_cont(0.3333) WITHIN GROUP (ORDER BY price_per_m2)::bigint AS t33,
percentile_cont(0.6667) WITHIN GROUP (ORDER BY price_per_m2)::bigint AS t66,
COUNT(*)::int AS n_listings
FROM listings
WHERE is_active
AND rooms IS NOT NULL
AND price_per_m2 BETWEEN :ppm2_min AND :ppm2_max
-- novostroyki guard (#1186): NULL = legacy вторичка до м.011
AND (listing_segment IS NULL OR listing_segment = 'vtorichka')
GROUP BY LEAST(GREATEST(rooms, 0), 4)
HAVING COUNT(*) > 0
)
INSERT INTO asking_to_sold_tier_bounds (rooms_bucket, district, t33, t66, n_listings)
SELECT rooms_bucket, ''::text, t33, t66, n_listings FROM bounds
"""
)
_REDERIVE_TIERED_SQL = text(
"""
WITH
bnd AS (
SELECT rooms_bucket, t33, t66 FROM asking_to_sold_tier_bounds WHERE district = ''
),
deal_tiered AS (
SELECT
LEAST(GREATEST(d.rooms, 0), 4) AS rooms_bucket,
CASE
WHEN d.price_per_m2 < b.t33 THEN 'low'
WHEN d.price_per_m2 < b.t66 THEN 'mid'
ELSE 'high'
END AS ppm2_tier,
d.price_per_m2
FROM deals d
JOIN bnd b ON b.rooms_bucket = LEAST(GREATEST(d.rooms, 0), 4)
WHERE d.source = 'rosreestr'
AND d.rooms IS NOT NULL
AND d.price_per_m2 BETWEEN :ppm2_min AND :ppm2_max
AND d.deal_date >= CURRENT_DATE - INTERVAL '12 months'
),
ask_tiered AS (
SELECT
LEAST(GREATEST(l.rooms, 0), 4) AS rooms_bucket,
CASE
WHEN l.price_per_m2 < b.t33 THEN 'low'
WHEN l.price_per_m2 < b.t66 THEN 'mid'
ELSE 'high'
END AS ppm2_tier,
l.price_per_m2
FROM listings l
JOIN bnd b ON b.rooms_bucket = LEAST(GREATEST(l.rooms, 0), 4)
WHERE l.is_active
AND l.rooms IS NOT NULL
AND l.price_per_m2 BETWEEN :ppm2_min AND :ppm2_max
-- novostroyki guard (#1186): NULL = legacy вторичка до м.011
AND (l.listing_segment IS NULL OR l.listing_segment = 'vtorichka')
),
deal_side_tier AS (
SELECT
rooms_bucket,
ppm2_tier,
percentile_cont(0.5) WITHIN GROUP (ORDER BY price_per_m2) AS sold_median,
COUNT(*)::int AS n_deals
FROM deal_tiered
GROUP BY rooms_bucket, ppm2_tier
),
ask_side_tier AS (
SELECT
rooms_bucket,
ppm2_tier,
percentile_cont(0.5) WITHIN GROUP (ORDER BY price_per_m2) AS ask_median,
COUNT(*)::int AS n_listings
FROM ask_tiered
GROUP BY rooms_bucket, ppm2_tier
),
tier_rows AS (
SELECT
d.rooms_bucket,
''::text AS district,
d.ppm2_tier,
(d.sold_median / a.ask_median)::numeric AS ratio,
round(d.sold_median)::bigint AS sold_median,
round(a.ask_median)::bigint AS ask_median,
d.n_deals,
a.n_listings,
12 AS window_months,
'per_rooms_tier'::text AS basis
FROM deal_side_tier d
JOIN ask_side_tier a USING (rooms_bucket, ppm2_tier)
WHERE d.n_deals >= 30
AND a.n_listings >= 30
AND a.ask_median IS NOT NULL
AND a.ask_median > 0
AND d.sold_median IS NOT NULL
AND d.sold_median > 0
),
deal_side_all AS (
SELECT
LEAST(GREATEST(rooms, 0), 4) AS rooms_bucket,
percentile_cont(0.5) WITHIN GROUP (ORDER BY price_per_m2) AS sold_median,
COUNT(*)::int AS n_deals
FROM deals
WHERE source = 'rosreestr'
AND rooms IS NOT NULL
AND price_per_m2 BETWEEN :ppm2_min AND :ppm2_max
AND deal_date >= CURRENT_DATE - INTERVAL '12 months'
GROUP BY LEAST(GREATEST(rooms, 0), 4)
),
ask_side_all AS (
SELECT
LEAST(GREATEST(rooms, 0), 4) AS rooms_bucket,
percentile_cont(0.5) WITHIN GROUP (ORDER BY price_per_m2) AS ask_median,
COUNT(*)::int AS n_listings
FROM listings
WHERE is_active
AND rooms IS NOT NULL
AND price_per_m2 BETWEEN :ppm2_min AND :ppm2_max
-- novostroyki guard (#1186): NULL = legacy вторичка до м.011
AND (listing_segment IS NULL OR listing_segment = 'vtorichka')
GROUP BY LEAST(GREATEST(rooms, 0), 4)
),
all_rows AS (
SELECT
d.rooms_bucket,
''::text AS district,
'all'::text AS ppm2_tier,
(d.sold_median / a.ask_median)::numeric AS ratio,
round(d.sold_median)::bigint AS sold_median,
round(a.ask_median)::bigint AS ask_median,
d.n_deals,
a.n_listings,
12 AS window_months,
'per_rooms_all'::text AS basis
FROM deal_side_all d
JOIN ask_side_all a USING (rooms_bucket)
WHERE d.n_deals >= 30
AND a.n_listings >= 30
AND a.ask_median IS NOT NULL
AND a.ask_median > 0
AND d.sold_median IS NOT NULL
AND d.sold_median > 0
),
deal_global AS (
SELECT
percentile_cont(0.5) WITHIN GROUP (ORDER BY price_per_m2) AS sold_median,
COUNT(*)::int AS n_deals
FROM deals
WHERE source = 'rosreestr'
AND rooms IS NOT NULL
AND price_per_m2 BETWEEN :ppm2_min AND :ppm2_max
AND deal_date >= CURRENT_DATE - INTERVAL '12 months'
),
ask_global AS (
SELECT
percentile_cont(0.5) WITHIN GROUP (ORDER BY price_per_m2) AS ask_median,
COUNT(*)::int AS n_listings
FROM listings
WHERE is_active
AND rooms IS NOT NULL
AND price_per_m2 BETWEEN :ppm2_min AND :ppm2_max
-- novostroyki guard (#1186): NULL = legacy вторичка до м.011
AND (listing_segment IS NULL OR listing_segment = 'vtorichka')
),
global_row AS (
SELECT
-1 AS rooms_bucket,
''::text AS district,
'all'::text AS ppm2_tier,
(d.sold_median / a.ask_median)::numeric AS ratio,
round(d.sold_median)::bigint AS sold_median,
round(a.ask_median)::bigint AS ask_median,
d.n_deals,
a.n_listings,
12 AS window_months,
'global_all'::text AS basis
FROM deal_global d
CROSS JOIN ask_global a
WHERE a.ask_median IS NOT NULL
AND a.ask_median > 0
AND d.sold_median IS NOT NULL
AND d.sold_median > 0
)
INSERT INTO asking_to_sold_ratios_tiered (
rooms_bucket, district, ppm2_tier, ratio, sold_median, ask_median,
n_deals, n_listings, window_months, basis
)
SELECT rooms_bucket, district, ppm2_tier, ratio, sold_median, ask_median,
n_deals, n_listings, window_months, basis FROM global_row
UNION ALL
SELECT rooms_bucket, district, ppm2_tier, ratio, sold_median, ask_median,
n_deals, n_listings, window_months, basis FROM all_rows
UNION ALL
SELECT rooms_bucket, district, ppm2_tier, ratio, sold_median, ask_median,
n_deals, n_listings, window_months, basis FROM tier_rows
"""
)
# ── Post-insert counters (tiered) ─────────────────────────────────────────────
_TIERED_COUNTERS_SQL = text(
"""
SELECT
(SELECT COUNT(*) FROM asking_to_sold_ratios_tiered WHERE district = '')
AS tiered_rows_written,
(SELECT COUNT(*) FROM asking_to_sold_tier_bounds WHERE district = '')
AS tier_bounds_written
"""
)
def recompute_asking_to_sold_ratios(db: Session, run_id: int) -> dict[str, int]:
"""Пересчитать asking_to_sold_ratios (TRUE-MIRROR refresh, #648 Stage 4).
@ -437,8 +206,6 @@ def recompute_asking_to_sold_ratios(db: Session, run_id: int) -> dict[str, int]:
"rows_written": 0,
"per_rooms_rows": 0,
"used_global_fallback": 0,
"tiered_rows_written": 0,
"tier_bounds_written": 0,
}
try:
# DELETE + re-derive INSERT в одной транзакции (НЕ коммитим между ними —
@ -455,35 +222,15 @@ def recompute_asking_to_sold_ratios(db: Session, run_id: int) -> dict[str, int]:
counters["per_rooms_rows"] = int(row["per_rooms_rows"] or 0)
counters["used_global_fallback"] = int(row["used_global_fallback"] or 0)
# Tiered refresh (#928): DELETE + re-derive bounds + tier rows в той же транзакции.
# Атомарно — таблицы никогда не пустые mid-refresh (всё в одном commit ниже).
db.execute(_DELETE_TIERED_ROWS_SQL)
db.execute(_DELETE_TIER_BOUNDS_SQL)
db.execute(
_REDERIVE_BOUNDS_SQL,
{"ppm2_min": _PPM2_MIN, "ppm2_max": settings.asking_ratio_ppm2_max},
)
db.execute(
_REDERIVE_TIERED_SQL,
{"ppm2_min": _PPM2_MIN, "ppm2_max": settings.asking_ratio_ppm2_max},
)
tier_row = db.execute(_TIERED_COUNTERS_SQL).mappings().first()
if tier_row is not None:
counters["tiered_rows_written"] = int(tier_row["tiered_rows_written"] or 0)
counters["tier_bounds_written"] = int(tier_row["tier_bounds_written"] or 0)
db.commit()
runs_mod.mark_done(db, run_id, counters)
logger.info(
"recompute_asking_to_sold_ratios run_id=%d done: "
"rows_written=%d per_rooms_rows=%d used_global_fallback=%d "
"tiered_rows_written=%d tier_bounds_written=%d",
"rows_written=%d per_rooms_rows=%d used_global_fallback=%d",
run_id,
counters["rows_written"],
counters["per_rooms_rows"],
counters["used_global_fallback"],
counters["tiered_rows_written"],
counters["tier_bounds_written"],
)
return counters
except Exception as exc:

291
tradein-mvp/backend/tests/services/test_asking_sold_tier_ratio.py
View file

@ -1,291 +0,0 @@
"""Unit tests for ppm2-tier asking->sold ratio lookup (#928).
Tests for the _get_asking_sold_ratio helper with tier_aware_ratio_enabled flag.
Five scenarios:
1. flag OFF -> uses old asking_to_sold_ratios table (byte-identical to pre-#928).
2. flag ON, anchor in low band -> returns shrunk tier ratio, basis startswith 'per_rooms_tier'.
3. flag ON, no tier row for that cell -> fallback to 'all' (basis 'per_rooms_all').
4. flag ON, no 'all' row either -> global (-1) fallback (basis 'global_all').
5. flag ON, no tables / empty -> (None, None), no raise.
Static callers audit: _get_asking_sold_ratio gained anchor_ppm2 param (default None),
so all existing call sites using positional (db, rooms) continue to work unchanged.
"""
from __future__ import annotations
import os
from typing import Any
from unittest.mock import MagicMock, patch
# Settings requires DATABASE_URL at init time. Set dummy DSN before any app import.
os.environ.setdefault("DATABASE_URL", "postgresql+psycopg://test:test@localhost/test_db")
class _FakeRow:
"""Stand-in for a SQLAlchemy Row."""
def __init__(self, **kw: Any) -> None:
for k, v in kw.items():
setattr(self, k, v)
def _clear_ratio_cache() -> None:
from app.services import estimator
estimator._asking_sold_ratio_cache.clear()
# ── helper: build a MagicMock db returning consecutive fetchone() results ─────
def _db_sequence(rows: list[Any]) -> MagicMock:
"""MagicMock db whose execute().fetchone() calls return `rows` in order."""
db = MagicMock()
db.execute.return_value.fetchone.side_effect = rows
return db
# ── 1. Flag OFF -> old table, ignores anchor_ppm2 ────────────────────────────
def test_flag_off_uses_old_table_ignores_anchor_ppm2() -> None:
"""Flag OFF: _get_asking_sold_ratio(db, 2, anchor_ppm2=90000) hits asking_to_sold_ratios."""
from app.core.config import settings
from app.services.estimator import _get_asking_sold_ratio
_clear_ratio_cache()
db = _db_sequence([_FakeRow(ratio=0.80, basis="per_rooms")])
with patch.object(settings, "tier_aware_ratio_enabled", False):
ratio, basis = _get_asking_sold_ratio(db, 2, anchor_ppm2=90_000.0)
assert ratio == 0.80
assert basis == "per_rooms"
# Only one DB query (old path, no tier tables touched)
assert db.execute.call_count == 1
# The query must target asking_to_sold_ratios (old table)
sql_str = str(db.execute.call_args_list[0].args[0])
assert "asking_to_sold_ratios" in sql_str
assert "tiered" not in sql_str
def test_flag_off_no_anchor_ppm2_still_works() -> None:
"""Flag OFF with no anchor_ppm2 (positional call) still works as before."""
from app.core.config import settings
from app.services.estimator import _get_asking_sold_ratio
_clear_ratio_cache()
db = _db_sequence([_FakeRow(ratio=0.82, basis="per_rooms")])
with patch.object(settings, "tier_aware_ratio_enabled", False):
ratio, basis = _get_asking_sold_ratio(db, 1)
assert ratio == 0.82
assert basis == "per_rooms"
def test_flag_off_cache_key_is_legacy_not_none() -> None:
"""Flag OFF must cache under (bucket, '_legacy'), NOT (bucket, None).
This prevents key collision with the flag-ON Step-3 'all' fallback which also
caches under (bucket, None). If the process flips tier_aware_ratio_enabled ON
without restart, the '_legacy' entry is not accidentally served as the 'all' row.
"""
from app.core.config import settings
from app.services import estimator
from app.services.estimator import _get_asking_sold_ratio
_clear_ratio_cache()
bucket = min(max(2, 0), 4)
db = _db_sequence([_FakeRow(ratio=0.77, basis="per_rooms")])
with patch.object(settings, "tier_aware_ratio_enabled", False):
_get_asking_sold_ratio(db, 2)
# Flag-OFF entry must be under "_legacy", not None.
assert (bucket, "_legacy") in estimator._asking_sold_ratio_cache
assert (bucket, None) not in estimator._asking_sold_ratio_cache
# ── 2. Flag ON, anchor in low band -> shrunk tier ratio ──────────────────────
def test_flag_on_low_band_returns_shrunk_tier_ratio() -> None:
"""Flag ON, anchor < t33 -> tier='low', shrunk = tier*w + all*(1-w)."""
from app.core.config import settings
from app.services.estimator import _get_asking_sold_ratio
_clear_ratio_cache()
# DB calls sequence:
# 1. bounds (t33=130000, t66=165000)
# 2. tier row (ratio=0.91, n_deals=200)
# 3. 'all' row (ratio=0.80)
db = _db_sequence(
[
_FakeRow(t33=130_000, t66=165_000), # bounds
_FakeRow(ratio=0.91, n_deals=200), # tier row
_FakeRow(ratio=0.80), # 'all' row
]
)
k = 150
w = 200 / (200 + k)
expected_ratio = 0.91 * w + 0.80 * (1.0 - w)
with (
patch.object(settings, "tier_aware_ratio_enabled", True),
patch.object(settings, "tier_ratio_shrink_k", k),
):
ratio, basis = _get_asking_sold_ratio(db, 2, anchor_ppm2=90_000.0)
assert ratio is not None
assert abs(ratio - expected_ratio) < 1e-9
assert basis is not None
assert basis.startswith("per_rooms_tier")
assert "low" in basis
# ── 3. Flag ON, no tier row -> fallback to 'all' ─────────────────────────────
def test_flag_on_no_tier_row_falls_back_to_all() -> None:
"""Flag ON, anchor in mid band, but no tier row for (bucket,mid) -> 'all' fallback."""
from app.core.config import settings
from app.services.estimator import _get_asking_sold_ratio
_clear_ratio_cache()
# DB calls:
# 1. bounds (t33=130000, t66=165000)
# 2. tier row -> None (cell not populated)
# 3. 'all' row (ratio=0.80) -- used as both shrink-target AND fallback
# Since tier_row is None, we skip shrink and fall through to step 3.
db = _db_sequence(
[
_FakeRow(t33=130_000, t66=165_000), # bounds
None, # tier row missing
_FakeRow(ratio=0.80), # 'all' row (shrink target query)
_FakeRow(ratio=0.80), # 'all' row (fallback step 3)
]
)
with (
patch.object(settings, "tier_aware_ratio_enabled", True),
patch.object(settings, "tier_ratio_shrink_k", 150),
):
ratio, basis = _get_asking_sold_ratio(db, 2, anchor_ppm2=145_000.0)
assert ratio == 0.80
assert basis == "per_rooms_all"
# ── 4. Flag ON, no 'all' row -> global fallback ───────────────────────────────
def test_flag_on_no_all_row_falls_back_to_global() -> None:
"""Flag ON, bounds present, tier+all rows missing -> global -1 fallback."""
from app.core.config import settings
from app.services.estimator import _get_asking_sold_ratio
_clear_ratio_cache()
db = _db_sequence(
[
_FakeRow(t33=130_000, t66=165_000), # bounds
None, # tier row -> None
None, # 'all' shrink-target -> None
None, # step-3 'all' fallback -> None
_FakeRow(ratio=0.79), # step-4 global (-1, all)
]
)
with patch.object(settings, "tier_aware_ratio_enabled", True):
ratio, basis = _get_asking_sold_ratio(db, 2, anchor_ppm2=145_000.0)
assert ratio == 0.79
assert basis == "global_all"
# ── 5. Flag ON, no tables / empty -> (None, None), no raise ──────────────────
def test_flag_on_no_tables_returns_none_none() -> None:
"""Flag ON, db.execute raises (tables not migrated) -> (None, None), no raise."""
from app.core.config import settings
from app.services.estimator import _get_asking_sold_ratio
_clear_ratio_cache()
db = MagicMock()
db.execute.side_effect = RuntimeError("relation asking_to_sold_tier_bounds does not exist")
with patch.object(settings, "tier_aware_ratio_enabled", True):
result = _get_asking_sold_ratio(db, 2, anchor_ppm2=90_000.0)
assert result == (None, None)
# rollback must be called on error
db.rollback.assert_called()
# ── 6. Flag ON, no bounds row -> step-3 'all' fallback (skip tier entirely) ──
def test_flag_on_no_bounds_row_falls_back_to_all() -> None:
"""Flag ON, bounds row None -> skip tier, go straight to 'all' step 3."""
from app.core.config import settings
from app.services.estimator import _get_asking_sold_ratio
_clear_ratio_cache()
db = _db_sequence(
[
None, # bounds -> None (bucket has no bound data)
_FakeRow(ratio=0.80), # step-3 'all' row
]
)
with patch.object(settings, "tier_aware_ratio_enabled", True):
ratio, basis = _get_asking_sold_ratio(db, 2, anchor_ppm2=90_000.0)
assert ratio == 0.80
assert basis == "per_rooms_all"
# ── 7. Cache: (bucket, tier) key memoises per tier ───────────────────────────
def test_tier_cache_key_is_bucket_tier_tuple() -> None:
"""Second call with same (bucket, anchor in same tier) -> served from cache after bounds.
Flow for second call:
- cache_key starts as (bucket, None) -- miss
- fetches bounds (1 DB query) -> determines tier_key="low"
- cache_key = (bucket, "low") -- HIT (populated by first call)
- returns cached ratio without further DB queries
Total DB queries: first call (3) + second call (1 bounds) = 4.
"""
from app.core.config import settings
from app.services.estimator import _get_asking_sold_ratio
_clear_ratio_cache()
db = _db_sequence(
[
# First call: bounds + tier_row + all_row
_FakeRow(t33=130_000, t66=165_000),
_FakeRow(ratio=0.91, n_deals=200),
_FakeRow(ratio=0.80),
# Second call: bounds only (tier cache hit after that)
_FakeRow(t33=130_000, t66=165_000),
]
)
with (
patch.object(settings, "tier_aware_ratio_enabled", True),
patch.object(settings, "tier_ratio_shrink_k", 150),
):
first = _get_asking_sold_ratio(db, 2, anchor_ppm2=90_000.0)
second = _get_asking_sold_ratio(db, 2, anchor_ppm2=95_000.0) # same 'low' tier
# Both calls return the same result (same tier).
assert first == second
# First call: 3 queries (bounds + tier_row + all_row).
# Second call: 1 query (bounds) + cache hit on (bucket, "low").
assert db.execute.call_count == 4

28
tradein-mvp/backend/tests/test_asking_to_sold_ratio.py
View file

@ -272,27 +272,17 @@ class _FakeMappingResult:
class _FakeDB:
"""Minimal stand-in for a SQLAlchemy Session — records execute() calls.
Execute call order (#928 extended task):
Execute call order:
1. DELETE asking_to_sold_ratios
2. REDERIVE INSERT asking_to_sold_ratios
3. COUNTERS SELECT (returns counters_row)
4. DELETE asking_to_sold_ratios_tiered (tiered rows)
5. DELETE asking_to_sold_tier_bounds (tier bounds)
6. REDERIVE_BOUNDS INSERT
7. REDERIVE_TIERED INSERT
8. TIERED_COUNTERS SELECT (returns tiered_counters_row)
"""
def __init__(
self,
counters_row: dict[str, Any],
tiered_counters_row: dict[str, Any] | None = None,
) -> None:
self._counters_row = counters_row
self._tiered_counters_row = tiered_counters_row or {
"tiered_rows_written": 0,
"tier_bounds_written": 0,
}
self.executed: list[Any] = []
self.committed = False
self.rolled_back = False
@ -302,9 +292,6 @@ class _FakeDB:
# 3rd call is the counters SELECT (asking_to_sold_ratios).
if len(self.executed) == 3:
return _FakeMappingResult(self._counters_row)
# 8th call is the tiered counters SELECT.
if len(self.executed) == 8:
return _FakeMappingResult(self._tiered_counters_row)
return _FakeMappingResult(None)
def commit(self) -> None:
@ -317,8 +304,8 @@ class _FakeDB:
def test_counter_logic_with_fake_db(monkeypatch: pytest.MonkeyPatch) -> None:
"""recompute_asking_to_sold_ratios maps the counters row into its return + marks done.
#928: counters now include tiered_rows_written + tier_bounds_written (5 keys total).
FakeDB has 8 execute() calls (old: 3, tiered: 5).
#2002: tiered refresh removed — counters are the 3 legacy keys, FakeDB has 3
execute() calls (DELETE + REDERIVE INSERT + COUNTERS SELECT).
"""
marked: dict[str, Any] = {}
monkeypatch.setattr(
@ -330,7 +317,6 @@ def test_counter_logic_with_fake_db(monkeypatch: pytest.MonkeyPatch) -> None:
db = _FakeDB(
counters_row={"rows_written": 4, "per_rooms_rows": 3, "used_global_fallback": 1},
tiered_counters_row={"tiered_rows_written": 15, "tier_bounds_written": 5},
)
out = ratio_mod.recompute_asking_to_sold_ratios(db, run_id=99) # type: ignore[arg-type]
@ -338,15 +324,11 @@ def test_counter_logic_with_fake_db(monkeypatch: pytest.MonkeyPatch) -> None:
"rows_written": 4,
"per_rooms_rows": 3,
"used_global_fallback": 1,
"tiered_rows_written": 15,
"tier_bounds_written": 5,
}
assert out == expected
assert db.committed is True
# Eight statements: DELETE(old) + REDERIVE(old) + COUNTERS(old)
# + DELETE_TIERED_ROWS + DELETE_TIER_BOUNDS + REDERIVE_BOUNDS + REDERIVE_TIERED
# + TIERED_COUNTERS (#928 extension).
assert len(db.executed) == 8
# Three statements: DELETE + REDERIVE INSERT + COUNTERS SELECT (legacy per-rooms).
assert len(db.executed) == 3
assert marked["run_id"] == 99
assert marked["counters"] == expected

2
tradein-mvp/backend/tests/test_backtest_fixture_roundtrip.py
View file

@ -164,7 +164,7 @@ def _build_fixture() -> dict[str, Any]:
{"price_per_m2": 308_000.0, "area_m2": 82.0, "rooms": 3, "floor": 6, "total_floors": 9},
],
anchor_tier_fetched=None,
ratio_calls=[[310_000.0, [0.92, "per_rooms_tier:high"]]],
ratio_calls=[[310_000.0, [0.92, "per_rooms"]]],
qi_calls=[],
qis_calls=[],
address="ул. Тестовая, 3",

2
tradein-mvp/backend/tests/test_estimator_ratio_tier_fix.py
View file

@ -163,7 +163,7 @@ def test_ratio_tier_uses_final_headline_after_anchor() -> None:
def _spy(db_inner: Any, rooms: Any, anchor_ppm2: Any = None) -> tuple:
captured_anchor_ppm2.append(anchor_ppm2)
return (0.78, "per_rooms_tier:high")
return (0.78, "per_rooms")
async def _run() -> Any:
with (

22
tradein-mvp/backend/tests/test_segment_guard_1186.py
View file

@ -74,9 +74,9 @@ def test_tier_a_anchor_gated_python_side_not_sql_guard() -> None:
# (a) Tier C SQL guard сохранён (поиск находит единственное вхождение — Tier C).
assert _GUARD_RE.search(src), "Tier C anchor SQL lost the canonical novostroyki guard"
# (b) Tier A теперь gated Python-side (#1774).
assert "estimate_sb_tier_a_allow_primary_if_secondary_present" in src, (
"Tier A no longer references the #1774 gating flag — gated relaxation missing"
)
assert (
"estimate_sb_tier_a_allow_primary_if_secondary_present" in src
), "Tier A no longer references the #1774 gating flag — gated relaxation missing"
# Hardened gate: secondary_count ≥ primary_count (заменил has_secondary).
assert "secondary_count" in src, "Tier A lacks secondary_count Python-side gate (#1774)"
assert "primary_count" in src, "Tier A lacks primary_count comparison in hardened gate (#1774)"
@ -94,8 +94,6 @@ def test_tier_c_anchor_has_canonical_guard_not_segment_param() -> None:
# ── asking_to_sold_ratio.py static checks ─────────────────────────────────────
_REDERIVE_SQL_TEXT = str(ratio_mod._REDERIVE_SQL.text)
_REDERIVE_BOUNDS_TEXT = str(ratio_mod._REDERIVE_BOUNDS_SQL.text)
_REDERIVE_TIERED_TEXT = str(ratio_mod._REDERIVE_TIERED_SQL.text)
def test_ask_side_cte_has_guard() -> None:
@ -114,20 +112,6 @@ def test_ask_global_cte_has_guard() -> None:
), f"Expected ≥2 guard occurrences in _REDERIVE_SQL (ask_side + ask_global), got {matches}"
def test_bounds_sql_has_guard() -> None:
"""_REDERIVE_BOUNDS_SQL (tier bounds percentiles) — guard обязателен."""
assert _GUARD_RE.search(_REDERIVE_BOUNDS_TEXT), "_REDERIVE_BOUNDS_SQL lacks novostroyki guard"
def test_tiered_ask_ctes_have_guard() -> None:
"""_REDERIVE_TIERED_SQL: ask_tiered + ask_side_all + ask_global — минимум 3 вхождения."""
matches = len(_GUARD_RE.findall(_REDERIVE_TIERED_TEXT))
assert matches >= 3, (
f"Expected ≥3 guard occurrences in _REDERIVE_TIERED_SQL "
f"(ask_tiered + ask_side_all + ask_global), got {matches}"
)
# ── Поведенческие тесты: _fetch_analogs (mock DB) ────────────────────────────