fix(pdf): brand-spoof + median=0 empty-state + SSRF + дисклеймер + sources (Refs #772) (#789)
Some checks failed
Deploy Trade-In / changes (push) Successful in 5s
Deploy Trade-In / build-frontend (push) Has been skipped
Deploy Trade-In / deploy (push) Has been cancelled
Deploy Trade-In / test (push) Successful in 25s
Deploy Trade-In / build-backend (push) Successful in 39s

Co-authored-by: bot-backend <bot-backend@gendsgn.local>
Co-committed-by: bot-backend <bot-backend@gendsgn.local>
This commit is contained in:
bot-backend 2026-05-30 16:55:56 +00:00 committed by bot-reviewer
parent 96027357b5
commit 38d44f46b7
3 changed files with 628 additions and 25 deletions

View file

@ -342,11 +342,13 @@ def get_estimate(
def estimate_pdf( def estimate_pdf(
estimate_id: UUID, estimate_id: UUID,
db: Annotated[Session, Depends(get_db)], db: Annotated[Session, Depends(get_db)],
brand: str | None = None,
x_authenticated_user: Annotated[str | None, Header(alias="X-Authenticated-User")] = None, x_authenticated_user: Annotated[str | None, Header(alias="X-Authenticated-User")] = None,
) -> Response: ) -> Response:
"""Скачать 4-страничный PDF-отчёт для оценки trade-in. """Скачать 4-страничный PDF-отчёт для оценки trade-in.
Бренд PDF определяется по владельцу оценки (estimate.created_by brand slug),
а НЕ из ?brand= query param (#7 brand-spoofing fix).
Возвращает application/pdf attachment. Возвращает application/pdf attachment.
404 оценка не найдена. 404 оценка не найдена.
410 оценка просрочена (TTL 24ч). 410 оценка просрочена (TTL 24ч).
@ -430,9 +432,13 @@ def estimate_pdf(
"has_balcony": row.has_balcony, "has_balcony": row.has_balcony,
} }
from app.core.auth import get_brand_for_user as _brand_for_user
from app.services.brand import get_brand as _resolve_brand from app.services.brand import get_brand as _resolve_brand
brand_obj = _resolve_brand(brand, db) # #7 brand-spoofing fix: бренд определяется по владельцу оценки,
# а не по ?brand= query param (который был удалён из сигнатуры).
owner_brand_slug = _brand_for_user(row.created_by) if row.created_by else None
brand_obj = _resolve_brand(owner_brand_slug, db)
pdf_bytes = generate_trade_in_pdf(estimate, input_snapshot, brand=brand_obj) pdf_bytes = generate_trade_in_pdf(estimate, input_snapshot, brand=brand_obj)
filename = f"trade-in-{brand_obj.slug}-{estimate_id}.pdf" filename = f"trade-in-{brand_obj.slug}-{estimate_id}.pdf"
logger.info( logger.info(

View file

@ -19,6 +19,7 @@ import datetime as dt
import html as _html import html as _html
import io import io
import logging import logging
import re
from urllib.parse import urlparse from urllib.parse import urlparse
from uuid import UUID from uuid import UUID
@ -95,6 +96,85 @@ def _safe_url(url: str | None, *, cdn: bool = False) -> str | None:
return url return url
# ── Logo URL и color-sanitation (#13 SSRF / CSS-injection guard) ─────────────
# Допустимые домены для логотипов брендов (white-label).
_ALLOWED_LOGO_CDN: frozenset[str] = frozenset(
{
"gk-praktika.ru",
"www.gk-praktika.ru",
"prinzip.ru",
"www.prinzip.ru",
"cdn.gk-praktika.ru",
}
)
_COLOR_RE = re.compile(r"^#[0-9A-Fa-f]{6}$")
def _safe_logo_url(url: str | None) -> str | None:
"""Валидирует logo_url бренда: только https:// + allowlist домен + не data:/file:.
data: URI допускаются для инлайн-логотипов (нет сетевого запроса).
Всё остальное отбрасывается. Если передан file:// отбрасывается.
"""
if not url:
return None
try:
p = urlparse(url)
except Exception:
return None
if p.scheme == "data":
# data: URI безопасны — нет SSRF, только инлайн
return url
if p.scheme != "https":
# file://, http://, javascript: — все отбрасываем
return None
if p.netloc not in _ALLOWED_LOGO_CDN:
logger.warning("logo_url domain not in allowlist: %s", p.netloc)
return None
return url
def _safe_color(color: str | None, fallback: str) -> str:
"""Принимает только #RRGGBB hex. Всё остальное → fallback.
Защита от CSS-injection через primary_color/accent_color бренда
(WeasyPrint вставляет значение напрямую в style=).
"""
if color and _COLOR_RE.match(color):
return color
return fallback
# ── WeasyPrint custom url_fetcher (SSRF guard) ───────────────────────────────
def _make_safe_url_fetcher(): # type: ignore[no-untyped-def]
"""Возвращает url_fetcher для WeasyPrint, блокирующий file:// и нелоговые схемы.
Разрешены: https:// (логотипы + шрифты CDN) и data: (инлайн SVG/PNG).
Блокированы: file://, http://, ftp:// и любые другие.
"""
def fetcher(url: str, timeout: int = 10, ssl_context=None): # type: ignore[no-untyped-def]
# Импорт отложен до момента первого реального fetch, чтобы тесты могли
# подменять weasyprint.urls через sys.modules без ModuleNotFoundError.
from weasyprint.urls import default_url_fetcher
try:
p = urlparse(url)
except Exception:
raise ValueError(f"Invalid URL in PDF resource: {url!r}") from None
if p.scheme in ("data", "https"):
return default_url_fetcher(url, timeout=timeout, ssl_context=ssl_context)
raise ValueError(
f"PDF resource scheme '{p.scheme}' is not allowed (only data: and https:). "
f"URL: {url!r}"
)
return fetcher
# ── Source pseudo-logos (текстовые pill-badges как у Брусники с логотипами) ── # ── Source pseudo-logos (текстовые pill-badges как у Брусники с логотипами) ──
_SOURCE_LOGO_COLORS: dict[str, tuple[str, str]] = { _SOURCE_LOGO_COLORS: dict[str, tuple[str, str]] = {
"avito": ("#00aaff", "#fff"), # Avito brand blue (упрощённо) "avito": ("#00aaff", "#fff"), # Avito brand blue (упрощённо)
@ -433,13 +513,28 @@ def _build_cover(estimate: AggregatedEstimate, input_snapshot: dict, brand) -> s
sub_label="Диапазон цен по фактическим сделкам", sub_label="Диапазон цен по фактическим сделкам",
) )
logo_url = _safe_logo_url(brand.logo_url)
logo_html = (
f"<img src='{_html.escape(logo_url)}' alt='{_html.escape(brand.name)}' "
f"style='max-height:28pt;max-width:120pt;object-fit:contain;vertical-align:middle;'/>"
if logo_url
else f"<span style='font-size:12pt;font-weight:900;letter-spacing:0.06em;'>"
f"{_html.escape(brand.name).upper()}</span>"
)
disclaimer_html = ""
if brand.pdf_disclaimer:
disclaimer_html = (
f"<p style='margin-top:10pt;font-size:7.5pt;color:#6b7280;line-height:1.4;"
f"border-top:1pt solid #e6e8ec;padding-top:6pt;'>"
f"{_html.escape(brand.pdf_disclaimer)}</p>"
)
return f""" return f"""
<div style="page-break-after:always;"> <div style="page-break-after:always;">
<!-- Top brand bar компактный --> <!-- Top brand bar компактный -->
<div style="background:{brand.primary_color};color:#fff;padding:8pt 14pt;margin-bottom:8pt; <div style="background:{brand.primary_color};color:#fff;padding:8pt 14pt;margin-bottom:8pt;">
font-size:12pt;font-weight:900;letter-spacing:0.06em;"> {logo_html}
{_html.escape(brand.name).upper()}
</div> </div>
<h1 style="font-size:12pt;font-weight:700;margin:4pt 0 8pt 0;text-transform:uppercase; <h1 style="font-size:12pt;font-weight:700;margin:4pt 0 8pt 0;text-transform:uppercase;
@ -502,6 +597,7 @@ def _build_cover(estimate: AggregatedEstimate, input_snapshot: dict, brand) -> s
padding-top:8pt;"> padding-top:8pt;">
<strong>Этот отчёт онлайн:</strong> {settings.public_url}?id={estimate.estimate_id} <strong>Этот отчёт онлайн:</strong> {settings.public_url}?id={estimate.estimate_id}
</p> </p>
{disclaimer_html}
</div> </div>
""" """
@ -528,10 +624,12 @@ def _deals_range(deals: list[AnalogLot], fallback: tuple[int, int]) -> tuple[int
def _build_listings_page(estimate: AggregatedEstimate, input_snapshot: dict, brand) -> str: # type: ignore[no-untyped-def,type-arg] def _build_listings_page(estimate: AggregatedEstimate, input_snapshot: dict, brand) -> str: # type: ignore[no-untyped-def,type-arg]
n_total = estimate.n_analogs n_total = estimate.n_analogs
n_with_repair = max(3, int(n_total * 0.4)) # упрощение: ~40% с указанным ремонтом # Используем n_analogs — estimator уже фильтрует по repair_state при поиске.
# Fabricated "~40%" удалён (#33): показываем тот же count, что прошёл фильтр.
n_with_repair = n_total
# Source logos (pseudo) — 5 максимальных # Source logos (pseudo) — берём из estimate.sources_used (не захардкоженный список).
sources_to_show = estimate.sources_used or ["avito", "cian", "domklik", "yandex", "rosreestr"] sources_to_show = estimate.sources_used or []
sources_html = "".join(_source_logo_pill(s) for s in sources_to_show[:5]) sources_html = "".join(_source_logo_pill(s) for s in sources_to_show[:5])
# Params правой колонки — параметры поиска (НЕ конкретной квартиры) # Params правой колонки — параметры поиска (НЕ конкретной квартиры)
@ -674,15 +772,17 @@ def _examples_rows(lots: list[AnalogLot]) -> str:
) )
else: else:
addr_cell = addr addr_cell = addr
ppm2_cell = _fmt_ppm2(lot.price_per_m2) if lot.price_per_m2 is not None else ""
price_cell = _fmt_ppm2(lot.price_rub) if lot.price_rub is not None else ""
rows.append( rows.append(
"<tr>" "<tr>"
f"<td style='padding:7pt 4pt;border-bottom:1px solid #f3f4f6;'>{addr_cell}</td>" f"<td style='padding:7pt 4pt;border-bottom:1px solid #f3f4f6;'>{addr_cell}</td>"
f"<td style='padding:7pt 4pt;border-bottom:1px solid #f3f4f6;'>" f"<td style='padding:7pt 4pt;border-bottom:1px solid #f3f4f6;'>"
f"{_source_badge_inline(lot.source)}</td>" f"{_source_badge_inline(lot.source)}</td>"
f"<td style='padding:7pt 4pt;border-bottom:1px solid #f3f4f6;text-align:right;" f"<td style='padding:7pt 4pt;border-bottom:1px solid #f3f4f6;text-align:right;"
f"font-variant-numeric:tabular-nums;'>{_fmt_ppm2(lot.price_per_m2)}</td>" f"font-variant-numeric:tabular-nums;'>{ppm2_cell}</td>"
f"<td style='padding:7pt 4pt;border-bottom:1px solid #f3f4f6;text-align:right;" f"<td style='padding:7pt 4pt;border-bottom:1px solid #f3f4f6;text-align:right;"
f"font-variant-numeric:tabular-nums;'>{_fmt_ppm2(lot.price_rub)}</td>" f"font-variant-numeric:tabular-nums;'>{price_cell}</td>"
f"<td style='padding:7pt 4pt;border-bottom:1px solid #f3f4f6;text-align:right;" f"<td style='padding:7pt 4pt;border-bottom:1px solid #f3f4f6;text-align:right;"
f"font-variant-numeric:tabular-nums;'>{lot.days_on_market or ''}</td>" f"font-variant-numeric:tabular-nums;'>{lot.days_on_market or ''}</td>"
"</tr>" "</tr>"
@ -698,9 +798,13 @@ def _build_deals_page(estimate: AggregatedEstimate, input_snapshot: dict, brand)
today = dt.date.today() today = dt.date.today()
period_start = today - dt.timedelta(days=estimate.period_months * 30) period_start = today - dt.timedelta(days=estimate.period_months * 30)
# Источники для сделок — Этажи / Домклик / Росреестр # Источники для сделок — берём из estimate.sources_used (не захардкоженный список).
deal_sources = ["etazhi", "domklik", "rosreestr"] # Фильтруем по известным источникам сделок; fallback к пустому (не fabricate).
sources_html = "".join(_source_logo_pill(s) for s in deal_sources) _deal_source_keys = {"etazhi", "domklik", "rosreestr"}
deal_sources = [s for s in (estimate.sources_used or []) if s in _deal_source_keys]
if not deal_sources:
deal_sources = [s for s in (estimate.sources_used or [])]
sources_html = "".join(_source_logo_pill(s) for s in deal_sources[:5])
area = float(input_snapshot.get("area_m2", 0) or 0) area = float(input_snapshot.get("area_m2", 0) or 0)
area_min = round(area * 0.85, 1) area_min = round(area * 0.85, 1)
@ -1146,15 +1250,51 @@ h1, h2, h3 {{ margin: 0; }}
# ── Public API ─────────────────────────────────────────────────────────────── # ── Public API ───────────────────────────────────────────────────────────────
def _build_insufficient_data_page(estimate: AggregatedEstimate, input_snapshot: dict, brand) -> str: # type: ignore[no-untyped-def,type-arg]
"""Страница-заглушка для случаев, когда данных недостаточно (#697/#9).
Рендерится вместо числового отчёта при estimate.insufficient_data=True.
Предотвращает публикацию «0,0 млн» и fabricated таблиц потерь.
"""
address = _html.escape(input_snapshot.get("address", ""))
report_num = _report_number(estimate.estimate_id)
today = dt.date.today()
return f"""
<div>
<div style="background:{brand.primary_color};color:#fff;padding:8pt 14pt;margin-bottom:8pt;
font-size:12pt;font-weight:900;letter-spacing:0.06em;">
{_html.escape(brand.name).upper()}
</div>
<p style="font-size:9pt;color:#6b7280;margin:0 0 18pt 0;">
{report_num} · {today.strftime("%d.%m.%Y")} · {address}
</p>
<div style="margin:40pt auto;text-align:center;max-width:320pt;">
<div style="font-size:32pt;margin-bottom:12pt;"></div>
<h2 style="font-size:14pt;font-weight:700;color:#1a1d23;margin-bottom:8pt;">
Недостаточно данных для оценки
</h2>
<p style="font-size:10pt;color:#6b7280;line-height:1.5;">
По данному объекту не найдено достаточного количества аналогов
для формирования надёжной рыночной оценки.
Пожалуйста, уточните параметры или обратитесь к специалисту.
</p>
</div>
</div>
"""
def generate_trade_in_pdf( def generate_trade_in_pdf(
estimate: AggregatedEstimate, estimate: AggregatedEstimate,
input_snapshot: dict, # type: ignore[type-arg] input_snapshot: dict, # type: ignore[type-arg]
*, *,
brand=None, # type: ignore[no-untyped-def] brand=None, # type: ignore[no-untyped-def]
) -> bytes: ) -> bytes:
"""Генерирует 4-страничный WeasyPrint PDF в стиле Брусника.Обмен. """Генерирует WeasyPrint PDF в стиле Брусника.Обмен.
Pages: Если estimate.insufficient_data=True рендерит одностраничный empty-state
вместо числового отчёта («Недостаточно данных», #697/#9).
Pages (normal):
1. Cover отчёта + параметры + 2 диапазона + 3 блока «Что важно» 1. Cover отчёта + параметры + 2 диапазона + 3 блока «Что важно»
2. Listings рынок аналогов + источники + таблица примеров 2. Listings рынок аналогов + источники + таблица примеров
3. Deals фактические сделки + источники + таблица 3. Deals фактические сделки + источники + таблица
@ -1186,25 +1326,45 @@ def generate_trade_in_pdf(
pdf_disclaimer=None, pdf_disclaimer=None,
) )
# Sanitize brand colors before inserting into CSS/style attributes (#13).
brand = Brand(
slug=brand.slug,
name=brand.name,
logo_url=brand.logo_url,
primary_color=_safe_color(brand.primary_color, "#1d4ed8"),
accent_color=_safe_color(brand.accent_color, "#f59e0b"),
footer_text=brand.footer_text,
pdf_disclaimer=brand.pdf_disclaimer,
)
# #697/#9: insufficient data → empty-state page, NOT fabricated 0-median report.
if estimate.insufficient_data:
body_html = _build_insufficient_data_page(estimate, input_snapshot, brand)
else:
body_html = (
_build_cover(estimate, input_snapshot, brand)
+ _build_listings_page(estimate, input_snapshot, brand)
+ _build_deals_page(estimate, input_snapshot, brand)
+ _build_offer_page(estimate, brand)
)
html_str = ( html_str = (
'<!DOCTYPE html><html lang="ru"><head><meta charset="UTF-8">' '<!DOCTYPE html><html lang="ru"><head><meta charset="UTF-8">'
f"<title>Trade-In — {_html.escape(input_snapshot.get('address', ''))}</title>" f"<title>Trade-In — {_html.escape(input_snapshot.get('address', ''))}</title>"
"</head><body>" f"</head><body>{body_html}</body></html>"
+ _build_cover(estimate, input_snapshot, brand)
+ _build_listings_page(estimate, input_snapshot, brand)
+ _build_deals_page(estimate, input_snapshot, brand)
+ _build_offer_page(estimate, brand)
+ "</body></html>"
) )
css_str = _build_css(brand=brand) css_str = _build_css(brand=brand)
# base_url=file:/// чтобы WeasyPrint мог resolve file:// URLs для PNG # SSRF guard (#13): кастомный url_fetcher блокирует file:// и нелоговые схемы.
pdf_bytes = HTML(string=html_str, base_url="file:///").write_pdf( # base_url=None — не resolve никакие относительные URL (все ресурсы data: или https:).
stylesheets=[CSS(string=css_str)] pdf_bytes = HTML(string=html_str, base_url=None).write_pdf(
stylesheets=[CSS(string=css_str)],
url_fetcher=_make_safe_url_fetcher(),
) )
logger.info( logger.info(
"PDF generated estimate_id=%s brand=%s size=%d (Brusnika-style)", "PDF generated estimate_id=%s brand=%s size=%d insufficient_data=%s",
estimate.estimate_id, estimate.estimate_id,
brand.slug, brand.slug,
len(pdf_bytes), len(pdf_bytes),
estimate.insufficient_data,
) )
return pdf_bytes return pdf_bytes

View file

@ -0,0 +1,437 @@
"""PDF security + correctness tests (#7/#9/#13/#33).
Tests cover:
- Part A (#7): brand derived from owner, not query param (tested at service level via
generate_trade_in_pdf API-layer change verified separately via grep).
- Part B (#9): insufficient_data=True → empty-state page, NOT numeric report / loss table.
- Part C (#8): pdf_disclaimer / logo_url from brand object rendered on cover.
- Part D (#13): _safe_logo_url rejects file:// / non-allowlist; _safe_color rejects
non-hex; _make_safe_url_fetcher blocks file:// scheme.
- Part E (#33): n_with_repair real count; sources_used respected; None fields no crash.
WeasyPrint is stubbed tests exercise HTML builders only (consistent with
tests/services/test_trade_in_pdf_dual_price.py pattern).
"""
from __future__ import annotations
import os
import sys
from datetime import UTC, datetime, timedelta
from unittest.mock import MagicMock, patch
from uuid import uuid4
os.environ.setdefault("DATABASE_URL", "postgresql+psycopg://test:test@localhost:5432/test")
_wp_mock = MagicMock()
sys.modules.setdefault("weasyprint", _wp_mock)
import pytest # noqa: E402
from app.schemas.trade_in import AggregatedEstimate, AnalogLot # noqa: E402
from app.services.brand import Brand # noqa: E402
from app.services.exporters import trade_in_pdf as mod # noqa: E402
# ── Fixtures ─────────────────────────────────────────────────────────────────
_GENERIC = Brand(
slug="generic",
name="Trade-In",
logo_url=None,
primary_color="#1d4ed8",
accent_color="#f59e0b",
footer_text=None,
pdf_disclaimer=None,
)
_PRAKTIKA = Brand(
slug="praktika",
name="Практика",
logo_url="https://gk-praktika.ru/images/logo/logo-pink.svg",
primary_color="#235F49",
accent_color="#1FCECB",
footer_text="ГК «Практика» · Екатеринбург",
pdf_disclaimer="Отчёт носит ориентировочный характер. ГК Практика, 2026.",
)
_SNAPSHOT = {
"address": "Екатеринбург, ул. Ленина, 1",
"area_m2": 50.0,
"rooms": 2,
"floor": 3,
"total_floors": 9,
"year_built": 2010,
"house_type": "panel",
"repair_state": "standard",
"has_balcony": True,
}
def _estimate(**overrides) -> AggregatedEstimate:
base = dict(
estimate_id=uuid4(),
median_price_rub=10_000_000,
range_low_rub=9_000_000,
range_high_rub=11_000_000,
median_price_per_m2=200_000,
confidence="high",
n_analogs=15,
period_months=24,
analogs=[],
actual_deals=[],
expires_at=datetime.now(UTC) + timedelta(days=30),
)
base.update(overrides)
return AggregatedEstimate(**base)
def _zero_estimate(**overrides) -> AggregatedEstimate:
"""Оценка с median=0 — insufficient_data=True."""
base = dict(
estimate_id=uuid4(),
median_price_rub=0,
range_low_rub=0,
range_high_rub=0,
median_price_per_m2=0,
confidence="low",
n_analogs=0,
period_months=24,
analogs=[],
actual_deals=[],
expires_at=datetime.now(UTC) + timedelta(days=30),
)
base.update(overrides)
return AggregatedEstimate(**base)
# ── Part B (#9): insufficient_data guard ─────────────────────────────────────
def test_insufficient_data_renders_empty_state_not_price() -> None:
"""median=0 → _build_insufficient_data_page called, no numeric headline."""
est = _zero_estimate()
assert est.insufficient_data is True
html = mod._build_insufficient_data_page(est, _SNAPSHOT, _GENERIC)
assert "Недостаточно данных" in html
# no fabricated price / loss table
assert "млн. руб." not in html
assert "Общие финансовые потери" not in html
def test_insufficient_data_page_contains_address_and_report_num() -> None:
est = _zero_estimate()
html = mod._build_insufficient_data_page(est, _SNAPSHOT, _GENERIC)
assert "Ленина" in html # address
assert "EKБ-" in html # report number prefix
def test_generate_pdf_insufficient_skips_offer_page() -> None:
"""generate_trade_in_pdf с insufficient_data → offer page (loss table) не рендерится."""
est = _zero_estimate()
# Stub WeasyPrint HTML/CSS so we can inspect the html_str fed to it.
captured: list[str] = []
class _FakeHTML:
def __init__(self, string=None, base_url=None, **kw):
captured.append(string or "")
def write_pdf(self, stylesheets=None, url_fetcher=None):
return b"%PDF-fake"
class _FakeCSS:
def __init__(self, string=None):
pass
with patch.dict(sys.modules, {"weasyprint": MagicMock(HTML=_FakeHTML, CSS=_FakeCSS)}):
import app.services.exporters.trade_in_pdf as _mod
_mod.generate_trade_in_pdf(est, _SNAPSHOT, brand=_GENERIC)
assert captured, "HTML was never passed to WeasyPrint mock"
html = captured[0]
assert "Недостаточно данных" in html
assert "Общие финансовые потери" not in html
assert "ФОРМИРОВАНИЕ ВЫКУПНОЙ СТОИМОСТИ" not in html
def test_generate_pdf_normal_estimate_has_offer_page() -> None:
"""Полная оценка → offer page присутствует."""
est = _estimate()
captured: list[str] = []
class _FakeHTML:
def __init__(self, string=None, base_url=None, **kw):
captured.append(string or "")
def write_pdf(self, stylesheets=None, url_fetcher=None):
return b"%PDF-fake"
class _FakeCSS:
def __init__(self, string=None):
pass
with patch.dict(sys.modules, {"weasyprint": MagicMock(HTML=_FakeHTML, CSS=_FakeCSS)}):
import app.services.exporters.trade_in_pdf as _mod
_mod.generate_trade_in_pdf(est, _SNAPSHOT, brand=_GENERIC)
assert captured
html = captured[0]
assert "ФОРМИРОВАНИЕ ВЫКУПНОЙ СТОИМОСТИ" in html
# ── Part C (#8): disclaimer + logo_url in cover ──────────────────────────────
def test_cover_renders_pdf_disclaimer() -> None:
est = _estimate()
html = mod._build_cover(est, _SNAPSHOT, _PRAKTIKA)
assert "Отчёт носит ориентировочный характер" in html
def test_cover_disclaimer_absent_when_none() -> None:
est = _estimate()
html = mod._build_cover(est, _SNAPSHOT, _GENERIC)
# No stray disclaimer block
assert "ориентировочный характер" not in html
def test_cover_renders_logo_img_when_valid_url() -> None:
est = _estimate()
html = mod._build_cover(est, _SNAPSHOT, _PRAKTIKA)
assert "<img" in html
assert "gk-praktika.ru" in html
def test_cover_falls_back_to_name_when_logo_none() -> None:
est = _estimate()
html = mod._build_cover(est, _SNAPSHOT, _GENERIC)
assert "<img" not in html
assert "TRADE-IN" in html.upper()
# ── Part D (#13): _safe_logo_url ─────────────────────────────────────────────
def test_safe_logo_url_accepts_allowlisted_https() -> None:
url = "https://gk-praktika.ru/images/logo.svg"
assert mod._safe_logo_url(url) == url
def test_safe_logo_url_rejects_file_scheme() -> None:
assert mod._safe_logo_url("file:///etc/passwd") is None
def test_safe_logo_url_rejects_http() -> None:
assert mod._safe_logo_url("http://gk-praktika.ru/logo.svg") is None
def test_safe_logo_url_rejects_unknown_domain() -> None:
assert mod._safe_logo_url("https://evil.com/logo.svg") is None
def test_safe_logo_url_accepts_data_uri() -> None:
data = "data:image/svg+xml;base64,PHN2Zy8+"
assert mod._safe_logo_url(data) == data
def test_safe_logo_url_none_input() -> None:
assert mod._safe_logo_url(None) is None
# ── Part D (#13): _safe_color ─────────────────────────────────────────────────
def test_safe_color_accepts_valid_hex() -> None:
assert mod._safe_color("#1d4ed8", "#000000") == "#1d4ed8"
assert mod._safe_color("#AbCdEf", "#000000") == "#AbCdEf"
def test_safe_color_rejects_css_expression() -> None:
assert mod._safe_color("expression(alert(1))", "#1d4ed8") == "#1d4ed8"
def test_safe_color_rejects_short_hex() -> None:
# 3-digit hex not accepted (require strict 6-digit)
assert mod._safe_color("#abc", "#1d4ed8") == "#1d4ed8"
def test_safe_color_rejects_none() -> None:
assert mod._safe_color(None, "#1d4ed8") == "#1d4ed8"
# ── Part D (#13): url_fetcher blocks file:// ─────────────────────────────────
def test_url_fetcher_rejects_file_scheme() -> None:
"""_make_safe_url_fetcher returns a fetcher that raises on file:// URLs."""
fake_default = MagicMock(return_value={"content": b"fake"})
urls_mod = MagicMock()
urls_mod.default_url_fetcher = fake_default
fetcher = mod._make_safe_url_fetcher()
with patch.dict(sys.modules, {"weasyprint.urls": urls_mod}):
with pytest.raises(ValueError, match="not allowed"):
fetcher("file:///etc/passwd")
def test_url_fetcher_allows_data_uri() -> None:
fake_default = MagicMock(return_value={"content": b"fake"})
urls_mod = MagicMock()
urls_mod.default_url_fetcher = fake_default
fetcher = mod._make_safe_url_fetcher()
with patch.dict(sys.modules, {"weasyprint.urls": urls_mod}):
result = fetcher("data:image/svg+xml;base64,PHN2Zy8+")
fake_default.assert_called_once()
assert result == {"content": b"fake"}
def test_url_fetcher_allows_https() -> None:
fake_default = MagicMock(return_value={"content": b"fake"})
urls_mod = MagicMock()
urls_mod.default_url_fetcher = fake_default
fetcher = mod._make_safe_url_fetcher()
with patch.dict(sys.modules, {"weasyprint.urls": urls_mod}):
fetcher("https://gk-praktika.ru/logo.svg")
fake_default.assert_called_once()
def test_url_fetcher_rejects_http() -> None:
fake_default = MagicMock(return_value={"content": b"fake"})
urls_mod = MagicMock()
urls_mod.default_url_fetcher = fake_default
fetcher = mod._make_safe_url_fetcher()
with patch.dict(sys.modules, {"weasyprint.urls": urls_mod}):
with pytest.raises(ValueError, match="not allowed"):
fetcher("http://example.com/style.css")
# ── Part D (#13): base_url no longer "file:///" ───────────────────────────────
def test_generate_pdf_no_file_base_url() -> None:
"""generate_trade_in_pdf must NOT pass base_url='file:///' to WeasyPrint."""
est = _estimate()
html_calls: list[dict] = []
class _FakeHTML:
def __init__(self, string=None, base_url=None, **kw):
html_calls.append({"base_url": base_url})
def write_pdf(self, stylesheets=None, url_fetcher=None):
return b"%PDF-fake"
class _FakeCSS:
def __init__(self, string=None):
pass
with patch.dict(sys.modules, {"weasyprint": MagicMock(HTML=_FakeHTML, CSS=_FakeCSS)}):
import app.services.exporters.trade_in_pdf as _mod
_mod.generate_trade_in_pdf(est, _SNAPSHOT, brand=_GENERIC)
assert html_calls, "HTML() was never called"
base_url = html_calls[0]["base_url"]
assert base_url != "file:///", f"base_url should NOT be 'file:///' (got {base_url!r})"
# ── Part E (#33): n_with_repair + sources + None-guards ──────────────────────
def _analog(**overrides) -> AnalogLot:
base = dict(
address="ул. Тест, 1",
source="avito",
source_url="https://avito.ru/1",
price_rub=5_000_000,
price_per_m2=100_000,
area_m2=50.0,
rooms=2,
floor=3,
total_floors=9,
listing_date=None,
days_on_market=None,
distance_m=200,
)
base.update(overrides)
return AnalogLot(**base)
def test_n_with_repair_equals_n_analogs_not_fabricated() -> None:
"""n_with_repair должен совпадать с n_analogs (не fabricated ~40%)."""
est = _estimate(n_analogs=12, sources_used=["avito"])
html = mod._build_listings_page(est, _SNAPSHOT, _GENERIC)
# both counts should show 12 (no fabricated sub-count)
assert "12 шт." in html
def test_listings_page_no_crash_with_no_sources() -> None:
est = _estimate(sources_used=[])
html = mod._build_listings_page(est, _SNAPSHOT, _GENERIC)
assert "РЫНОК КВАРТИР" in html
def test_listings_sources_from_estimate_not_hardcoded() -> None:
"""sources_used=['yandex'] → только yandex pill, не avito/cian/domklik/..."""
est = _estimate(sources_used=["yandex"])
html = mod._build_listings_page(est, _SNAPSHOT, _GENERIC)
assert "Я.Недвижимость" in html
# avito and cian should NOT appear if not in sources_used
assert "Avito" not in html
assert "Циан" not in html
def test_deals_sources_from_estimate_not_hardcoded() -> None:
"""sources_used=['rosreestr'] → только rosreestr badges на deals-странице."""
est = _estimate(sources_used=["rosreestr"])
html = mod._build_deals_page(est, _SNAPSHOT, _GENERIC)
assert "Росреестр" in html
# Этажи/Домклик — не должны появиться
assert "Этажи" not in html
assert "Домклик" not in html
def test_examples_rows_none_price_per_m2_no_crash() -> None:
"""AnalogLot с price_per_m2=None — _examples_rows не падает, рендерит «—»."""
# price_per_m2 is typed int (not Optional) in AnalogLot, but defensively
# the renderer must guard it. We pass a valid lot and verify no crash.
lot = _analog(price_per_m2=100_000, days_on_market=None)
html = mod._examples_rows([lot])
assert "100 000" in html
assert "5 000 000" in html
def test_examples_rows_empty_list_no_crash() -> None:
html = mod._examples_rows([])
assert "Нет данных" in html
def test_build_listings_page_none_year_built_no_crash() -> None:
snap = dict(_SNAPSHOT)
snap["year_built"] = None
est = _estimate()
html = mod._build_listings_page(est, snap, _GENERIC)
assert "РЫНОК КВАРТИР" in html
def test_build_deals_page_none_year_built_no_crash() -> None:
snap = dict(_SNAPSHOT)
snap["year_built"] = None
est = _estimate()
html = mod._build_deals_page(est, snap, _GENERIC)
assert "СДЕЛКИ" in html.upper()
# ── Part A (#7): brand derived from owner (service-level check) ───────────────
def test_brand_not_taken_from_query_param_docstring() -> None:
"""Verify the PDF endpoint no longer has a 'brand' query parameter."""
import inspect
from app.api.v1.trade_in import estimate_pdf
sig = inspect.signature(estimate_pdf)
param_names = list(sig.parameters.keys())
assert (
"brand" not in param_names
), "estimate_pdf should NOT have a 'brand' query param after #7 fix"