qananasikq/iaai-parser
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Initial project version

Browse Source
This commit is contained in:
qananasikq
2026-04-08 23:36:19 +03:00
commit 50d107cfba
54 changed files with 4592 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

16
.dockerignore Normal file
View File

@@ -0,0 +1,16 @@
__pycache__/
.pytest_cache/
.coverage
coverage.xml
.venv/
*.pyc
*.pyo
*.pyd
*.log
*.db
.env
.git/
.vscode/
*.egg-info/
dist/
build/

64
.env.example Normal file
View File

@@ -0,0 +1,64 @@
IAAI_HEADLESS=true
IAAI_LOG_LEVEL=INFO
# IAAI_LOG_FILE=iaai_scraper.log
# Network capture settings.
IAAI_CAPTURE_SAME_ORIGIN_ONLY=true
IAAI_MAX_CAPTURED_REQUESTS=40
IAAI_MAX_CAPTURED_JSON_RESPONSES=30
# Sequential listing-first mode.
IAAI_CARS_LISTING_URL=https://www.iaai.com/Vehiclelisting/Cars
IAAI_MAX_PAGES_PER_RUN=10
IAAI_MAX_VEHICLES_PER_RUN=30
IAAI_PAGE_LINK_LIMIT=100
IAAI_INCLUDE_PAGINATION=true
IAAI_COLLECT_CURRENT_PAGE_ONLY=false
# Human-like pacing.
IAAI_HUMAN_PACE_ENABLED=true
IAAI_AFTER_LISTING_OPEN_MIN_S=2.5
IAAI_AFTER_LISTING_OPEN_MAX_S=4.5
IAAI_AFTER_FILTER_ACTION_MIN_S=2.0
IAAI_AFTER_FILTER_ACTION_MAX_S=4.0
IAAI_BEFORE_VEHICLE_OPEN_MIN_S=1.5
IAAI_BEFORE_VEHICLE_OPEN_MAX_S=3.0
IAAI_AFTER_VEHICLE_OPEN_MIN_S=1.0
IAAI_AFTER_VEHICLE_OPEN_MAX_S=2.5
IAAI_BETWEEN_VEHICLES_MIN_S=3.0
IAAI_BETWEEN_VEHICLES_MAX_S=6.0
IAAI_AFTER_PAGE_CHANGE_MIN_S=3.0
IAAI_AFTER_PAGE_CHANGE_MAX_S=6.0
# Sync settings
IAAI_SYNC_ONLY_NEW=true
# Retry / backoff
IAAI_RETRY_DELAY_SECONDS=2.5
IAAI_RETRY_BACKOFF_MULTIPLIER=2.0
IAAI_RETRY_JITTER_SECONDS=0.25
# Proxy (HTTP/HTTPS preferred for Playwright)
# Chromium does not support SOCKS5 proxy authentication directly.
# Use residential or mobile USA proxy.
# IAAI_PROXY_SERVER=http://proxy.example.com:8080
# IAAI_PROXY_USERNAME=
# IAAI_PROXY_PASSWORD=
# Database (PostgreSQL).
IAAI_DATABASE_URL=postgresql+psycopg2://iaai:iaai@postgres:5432/iaai_scraper
IAAI_DATABASE_ECHO=false
IAAI_DATABASE_POOL_SIZE=5
IAAI_DATABASE_MAX_OVERFLOW=10
# ─── Redis (Celery broker) ─────────────────────────────────
IAAI_REDIS_URL=redis://redis:6379/0
# ─── Celery ────────────────────────────────────────────────
CELERY_BROKER_URL=redis://redis:6379/0
CELERY_RESULT_BACKEND=redis://redis:6379/0
CELERY_TASK_SOFT_TIME_LIMIT=600
CELERY_TASK_TIME_LIMIT=900
CELERY_WORKER_CONCURRENCY=1
CELERY_BEAT_SYNC_INTERVAL_MINUTES=60
CELERY_BEAT_SYNC_LIMIT=26

6
.gitattributes vendored Normal file
View File

@@ -0,0 +1,6 @@
# Force LF for shell scripts (critical for Docker on Windows)
*.sh text eol=lf
entrypoint.sh text eol=lf
# Default for other files
* text=auto

17
.gitignore vendored Normal file
View File

@@ -0,0 +1,17 @@
__pycache__/
.pytest_cache/
.coverage
coverage.xml
.venv/
*.pyc
*.pyo
*.pyd
*.log
*.db
.env
.vscode/
*.egg-info/
dist/
build/
artifacts/
celerybeat-schedule*

22
Dockerfile Normal file
View File

@@ -0,0 +1,22 @@
FROM mcr.microsoft.com/playwright/python:v1.58.0-noble
ENV PYTHONDONTWRITEBYTECODE=1 \
PYTHONUNBUFFERED=1
WORKDIR /app
# Xvfb for headed Chromium in container (IAAI blocks headless)
RUN apt-get update -qq && apt-get install -y --no-install-recommends xvfb \
&& rm -rf /var/lib/apt/lists/*
COPY requirements.txt ./
RUN pip install --no-cache-dir -r requirements.txt
COPY . .
RUN chmod +x entrypoint.sh
STOPSIGNAL SIGINT
# По умолчанию — API, но worker/beat переопределяют CMD в docker-compose
ENTRYPOINT ["./entrypoint.sh"]
CMD ["uvicorn", "iaai_scraper.api.app:app", "--host", "0.0.0.0", "--port", "8000"]

135
README.md Normal file
View File

@@ -0,0 +1,135 @@
# IAAI Scraper
Небольшой сервис для сбора автомобилей с [iaai.com](https://www.iaai.com).
Собирает ссылки из листинга, открывает карточки машин, достаёт данные из HTML и XHR, сохраняет всё в PostgreSQL.
## Что важно по сайту
У IAAI стоит **Imperva Incapsula** — внешний anti-bot и WAF.
- headless Chromium без прокси часто режется
- часть данных грузится через XHR, часть есть сразу в HTML
- есть cookie banner
Поэтому в проекте используется Playwright, паузы между действиями и прокси.
## Запуск
```bash
cp .env.example .env # подправить под себя
docker compose up -d
```
Поднимутся 5 контейнеров: postgres, redis, api, worker, beat. API на `http://localhost:8000`.
Swagger-документация: `http://localhost:8000/docs`
## API
**Здоровье и статистика:**
- `GET /health` — статус сервиса и подключения к БД
- `GET /api/v1/stats` — сколько машин/картинок в базе, топ брендов
**Машины:**
- `GET /api/v1/cars` — список с пагинацией
- `GET /api/v1/cars/{id}` — карточка с картинками
- `GET /api/v1/cars/by-origin/{origin_id}` — поиск по IAAI stock number
**Задачи:**
- `POST /api/v1/tasks/sync-vehicle` — скрапнуть одну машину по URL
- `POST /api/v1/tasks/sync-listing` — запустить полный обход листинга
- `GET /api/v1/tasks/{task_id}` — статус задачи
- `GET /api/v1/tasks` — все задачи
- `GET /api/v1/sync-runs` — история запусков
Пример:
```bash
curl -X POST http://localhost:8000/api/v1/tasks/sync-vehicle
-H "Content-Type: application/json" \
-d '{"vehicle_url": "https://www.iaai.com/VehicleDetail/45089484~US"}'
```
## CLI
Для локального запуска без API и Celery:
```bash
python main.py init-db
python main.py collect-listing --make Toyota
python main.py sync-vehicle "https://www.iaai.com/VehicleDetail/45089484~US"
python main.py sync-listing --limit 10
```
## Структура
```
iaai_scraper/
scraper.py — основной код скрапера
cli.py — CLI-команды
proxy_bridge.py — HTTP→SOCKS5 мост
browser/
factory.py — запуск браузера и контекста
listing.py — сбор ссылок из листинга
network.py — перехват XHR/fetch
pace.py — паузы между действиями
parsing/
parser.py — разбор карточки машины
mapper.py — нормализация данных
storage/
models.py — SQLAlchemy модели
schemas.py — Pydantic схемы
enums.py — справочники значений
db.py — работа с БД
api/
app.py — FastAPI приложение
deps.py — зависимости
routes/
health.py — GET /health
cars.py — машины и статистика
tasks.py — задачи и история запусков
worker/
celery_app.py — конфиг Celery
tasks.py — фоновые задачи
core/
config.py — настройки
logs.py — логирование
retry.py — retry-логика
utils.py — утилиты
alembic/ — миграции БД
```
## Конфигурация
Основные переменные лежат в `.env.example`:
**БД:** `IAAI_DATABASE_URL`, `IAAI_DATABASE_POOL_SIZE`
**Redis:** `IAAI_REDIS_URL`
**Celery:** `CELERY_BROKER_URL`, `CELERY_BEAT_SYNC_INTERVAL_MINUTES`
**Скрапер:** `IAAI_HEADLESS`, `IAAI_SYNC_ONLY_NEW`, `IAAI_MAX_PAGES_PER_RUN`
**Прокси:** `IAAI_PROXY_SERVER`, `IAAI_PROXY_USERNAME`, `IAAI_PROXY_PASSWORD`
**Паузы:** `IAAI_BETWEEN_VEHICLES_MIN_S`, `IAAI_AFTER_PAGE_CHANGE_MAX_S` и т.д.
## Миграции
В Docker миграции запускаются автоматически при старте API-контейнера.
Вручную:
```bash
alembic upgrade head
alembic revision --autogenerate -m "add_column_x"
```
## Тесты
```bash
pytest -q
```
Тесты идут на SQLite in-memory, без внешних сервисов.

40
alembic.ini Normal file
View File

@@ -0,0 +1,40 @@
# Alembic Configuration File
[alembic]
script_location = alembic
prepend_sys_path = .
sqlalchemy.url = postgresql+psycopg2://iaai:iaai@localhost:5432/iaai_scraper
[loggers]
keys = root,sqlalchemy,alembic
[handlers]
keys = console
[formatters]
keys = generic
[logger_root]
level = WARN
handlers = console
qualname =
[logger_sqlalchemy]
level = WARN
handlers =
qualname = sqlalchemy.engine
[logger_alembic]
level = INFO
handlers =
qualname = alembic
[handler_console]
class = StreamHandler
args = (sys.stderr,)
level = NOTSET
formatter = generic
[formatter_generic]
format = %(levelname)-5.5s [%(name)s] %(message)s
datefmt = %H:%M:%S

58
alembic/env.py Normal file
View File

@@ -0,0 +1,58 @@
"""Alembic env.py — подключение к БД через Settings."""
import os
import sys
from logging.config import fileConfig
from alembic import context
from sqlalchemy import engine_from_config, pool
# Добавляем корень проекта в sys.path
sys.path.insert(0, os.path.abspath(os.path.join(os.path.dirname(__file__), "..")))
from iaai_scraper.core.config import Settings
from iaai_scraper.storage.models import Base
config = context.config
# Logging
if config.config_file_name is not None:
fileConfig(config.config_file_name)
# Подставляем URL из Settings (env vars имеют приоритет)
settings = Settings()
config.set_main_option("sqlalchemy.url", settings.database.url)
target_metadata = Base.metadata
def run_migrations_offline() -> None:
"""Run migrations in 'offline' mode."""
url = config.get_main_option("sqlalchemy.url")
context.configure(
url=url,
target_metadata=target_metadata,
literal_binds=True,
dialect_opts={"paramstyle": "named"},
)
with context.begin_transaction():
context.run_migrations()
def run_migrations_online() -> None:
"""Run migrations in 'online' mode."""
connectable = engine_from_config(
config.get_section(config.config_ini_section, {}),
prefix="sqlalchemy.",
poolclass=pool.NullPool,
)
with connectable.connect() as connection:
context.configure(connection=connection, target_metadata=target_metadata)
with context.begin_transaction():
context.run_migrations()
if context.is_offline_mode():
run_migrations_offline()
else:
run_migrations_online()

25
alembic/script.py.mako Normal file
View File

@@ -0,0 +1,25 @@
"""${message}
Revision ID: ${up_revision}
Revises: ${down_revision | comma,n}
Create Date: ${create_date}
"""
from typing import Sequence, Union
from alembic import op
import sqlalchemy as sa
${imports if imports else ""}
# revision identifiers, used by Alembic.
revision: str = ${repr(up_revision)}
down_revision: Union[str, None] = ${repr(down_revision)}
branch_labels: Union[str, Sequence[str], None] = ${repr(branch_labels)}
depends_on: Union[str, Sequence[str], None] = ${repr(depends_on)}
def upgrade() -> None:
${upgrades if upgrades else "pass"}
def downgrade() -> None:
${downgrades if downgrades else "pass"}

105
alembic/versions/001_initial.py Normal file
View File

@@ -0,0 +1,105 @@
"""Initial schema — cars, images, sync_runs, scrape_tasks
Revision ID: 001_initial
Revises:
Create Date: 2026-04-08
"""
from typing import Sequence, Union
from alembic import op
import sqlalchemy as sa
revision: str = "001_initial"
down_revision: Union[str, None] = None
branch_labels: Union[str, Sequence[str], None] = None
depends_on: Union[str, Sequence[str], None] = None
def upgrade() -> None:
# --- cars ---
op.create_table(
"cars",
sa.Column("id", sa.BigInteger, primary_key=True, autoincrement=True),
sa.Column("parser_id", sa.String(50), nullable=False, unique=True),
sa.Column("brand", sa.String(50), nullable=False),
sa.Column("model", sa.String(50), nullable=False),
sa.Column("year", sa.Integer, nullable=True),
sa.Column("price", sa.BigInteger, nullable=True),
sa.Column("currency", sa.String(10), nullable=False, server_default="USD"),
sa.Column("mileage", sa.Integer, nullable=False, server_default="0"),
sa.Column("country", sa.String(10), nullable=False, server_default="NA"),
sa.Column("is_sold", sa.Boolean, nullable=False, server_default=sa.text("false")),
sa.Column("color", sa.String, nullable=False, server_default="other"),
sa.Column("drive", sa.String(10), nullable=True),
sa.Column("gearbox", sa.String(10), nullable=True),
sa.Column("steering_wheel", sa.String(10), nullable=True),
sa.Column("body_type", sa.String(20), nullable=False, server_default="OTHER"),
sa.Column("engine_volume", sa.Integer, nullable=True),
sa.Column("selling_type", sa.String(20), nullable=False, server_default="NA"),
sa.Column("one_owner", sa.Boolean, nullable=False, server_default=sa.text("false")),
sa.Column("new_car", sa.Boolean, nullable=False, server_default=sa.text("false")),
sa.Column("is_hidden", sa.Boolean, nullable=False, server_default=sa.text("false")),
sa.Column("origin", sa.String(20), nullable=False, server_default="NA"),
sa.Column("origin_url", sa.String, nullable=False),
sa.Column("origin_id", sa.String, nullable=False, unique=True),
sa.Column("is_damaged", sa.Boolean, nullable=False, server_default=sa.text("false")),
sa.Column("evaluation", sa.String, nullable=True),
sa.Column("non_smoking", sa.Boolean, nullable=False, server_default=sa.text("true")),
sa.Column("rental", sa.Boolean, nullable=False, server_default=sa.text("false")),
sa.Column("repair_history", sa.Boolean, nullable=False, server_default=sa.text("false")),
sa.Column("slug", sa.String, nullable=False),
sa.Column("last_seen_at", sa.DateTime(timezone=True), nullable=False, server_default=sa.func.now()),
sa.Column("content_hash", sa.String(64), nullable=False, server_default=""),
sa.Column("raw_attributes", sa.Text, nullable=True),
)
op.create_index("ix_cars_origin_url", "cars", ["origin_url"])
op.create_index("ix_cars_origin_id", "cars", ["origin_id"], unique=True)
op.create_index("ix_cars_content_hash", "cars", ["content_hash"])
# --- images ---
op.create_table(
"images",
sa.Column("id", sa.BigInteger, primary_key=True, autoincrement=True),
sa.Column("fullres_image", sa.String, nullable=False),
sa.Column("preview_image", sa.String, nullable=False),
sa.Column("order_index", sa.Integer, nullable=False),
sa.Column("car_id", sa.BigInteger, sa.ForeignKey("cars.id", ondelete="CASCADE"), nullable=False),
)
# --- sync_runs ---
op.create_table(
"sync_runs",
sa.Column("id", sa.BigInteger, primary_key=True, autoincrement=True),
sa.Column("started_at", sa.DateTime(timezone=True), nullable=False, server_default=sa.func.now()),
sa.Column("finished_at", sa.DateTime(timezone=True), nullable=True),
sa.Column("status", sa.Text, nullable=False),
sa.Column("lane", sa.Text, nullable=False),
sa.Column("ids_fetched", sa.Integer, nullable=False, server_default="0"),
sa.Column("cars_upserted", sa.Integer, nullable=False, server_default="0"),
sa.Column("cars_failed", sa.Integer, nullable=False, server_default="0"),
sa.Column("images_upserted", sa.Integer, nullable=False, server_default="0"),
sa.Column("error_summary", sa.Text, nullable=True),
)
# --- scrape_tasks ---
op.create_table(
"scrape_tasks",
sa.Column("id", sa.BigInteger, primary_key=True, autoincrement=True),
sa.Column("celery_task_id", sa.String(255), nullable=False, unique=True),
sa.Column("task_type", sa.String(50), nullable=False),
sa.Column("status", sa.String(20), nullable=False, server_default="pending"),
sa.Column("vehicle_url", sa.Text, nullable=True),
sa.Column("created_at", sa.DateTime(timezone=True), nullable=False, server_default=sa.func.now()),
sa.Column("started_at", sa.DateTime(timezone=True), nullable=True),
sa.Column("finished_at", sa.DateTime(timezone=True), nullable=True),
sa.Column("result_summary", sa.Text, nullable=True),
sa.Column("error_message", sa.Text, nullable=True),
)
op.create_index("ix_scrape_tasks_celery_task_id", "scrape_tasks", ["celery_task_id"], unique=True)
def downgrade() -> None:
op.drop_table("scrape_tasks")
op.drop_table("sync_runs")
op.drop_table("images")
op.drop_table("cars")

105
docker-compose.yml Normal file
View File

@@ -0,0 +1,105 @@
services:
# ─── PostgreSQL ───────────────────────────────────────────
postgres:
image: postgres:16-alpine
container_name: iaai-postgres
restart: unless-stopped
environment:
POSTGRES_USER: iaai
POSTGRES_PASSWORD: iaai
POSTGRES_DB: iaai_scraper
ports:
- "5432:5432"
volumes:
- pgdata:/var/lib/postgresql/data
healthcheck:
test: ["CMD-SHELL", "pg_isready -U iaai -d iaai_scraper"]
interval: 5s
timeout: 3s
retries: 5
# ─── Redis (Celery broker) ────────────────────────────────
redis:
image: redis:7-alpine
container_name: iaai-redis
restart: unless-stopped
ports:
- "6379:6379"
healthcheck:
test: ["CMD", "redis-cli", "ping"]
interval: 5s
timeout: 3s
retries: 5
# ─── FastAPI ──────────────────────────────────────────────
api:
build: .
container_name: iaai-api
restart: unless-stopped
env_file:
- path: .env
required: false
environment:
IAAI_DATABASE_URL: postgresql+psycopg2://iaai:iaai@postgres:5432/iaai_scraper
IAAI_REDIS_URL: redis://redis:6379/0
CELERY_BROKER_URL: redis://redis:6379/0
CELERY_RESULT_BACKEND: redis://redis:6379/0
ports:
- "8000:8000"
depends_on:
postgres:
condition: service_healthy
redis:
condition: service_healthy
command: >
uvicorn iaai_scraper.api.app:app
--host 0.0.0.0 --port 8000 --workers 2
# ─── Celery Worker ────────────────────────────────────────
worker:
build: .
container_name: iaai-worker
restart: unless-stopped
env_file:
- path: .env
required: false
environment:
IAAI_DATABASE_URL: postgresql+psycopg2://iaai:iaai@postgres:5432/iaai_scraper
IAAI_REDIS_URL: redis://redis:6379/0
CELERY_BROKER_URL: redis://redis:6379/0
CELERY_RESULT_BACKEND: redis://redis:6379/0
depends_on:
postgres:
condition: service_healthy
redis:
condition: service_healthy
stop_grace_period: 60s
command: >
celery -A iaai_scraper.worker.celery_app worker
--loglevel=info --concurrency=1 --pool=prefork
-Q scraping --without-heartbeat
# ─── Celery Beat (периодический планировщик) ──────────────
beat:
build: .
container_name: iaai-beat
restart: unless-stopped
env_file:
- path: .env
required: false
environment:
IAAI_DATABASE_URL: postgresql+psycopg2://iaai:iaai@postgres:5432/iaai_scraper
IAAI_REDIS_URL: redis://redis:6379/0
CELERY_BROKER_URL: redis://redis:6379/0
CELERY_RESULT_BACKEND: redis://redis:6379/0
depends_on:
postgres:
condition: service_healthy
redis:
condition: service_healthy
command: >
celery -A iaai_scraper.worker.celery_app beat
--loglevel=info
volumes:
pgdata:

57
entrypoint.sh Normal file
View File

@@ -0,0 +1,57 @@
#!/bin/bash
set -e
# Start HTTP→SOCKS5 proxy bridge if SOCKS5 upstream is configured
if [ -n "$SOCKS5_PROXY_HOST" ]; then
echo "[entrypoint] Starting proxy bridge (HTTP :8899 → SOCKS5 $SOCKS5_PROXY_HOST:${SOCKS5_PROXY_PORT:-1002})..."
if [ -z "$IAAI_PROXY_SERVER" ]; then
export IAAI_PROXY_SERVER="http://127.0.0.1:8899"
elif [ "$IAAI_PROXY_SERVER" = "http://localhost:8899" ]; then
export IAAI_PROXY_SERVER="http://127.0.0.1:8899"
fi
echo "[entrypoint] Using browser proxy: $IAAI_PROXY_SERVER"
python -m iaai_scraper.proxy_bridge &
BRIDGE_PID=$!
sleep 1
if ! kill -0 $BRIDGE_PID 2>/dev/null; then
echo "[entrypoint] ERROR: iaai_scraper.proxy_bridge failed to start"
exit 1
fi
echo "[entrypoint] Proxy bridge started (PID $BRIDGE_PID)"
fi
# Xvfb needed only for worker (Playwright) — not for API or beat
NEEDS_XVFB=false
case "$1" in
celery*|python*main*)
NEEDS_XVFB=true
;;
*)
# Check if any arg contains "worker"
for arg in "$@"; do
case "$arg" in
*worker*) NEEDS_XVFB=true; break ;;
esac
done
;;
esac
if [ "$NEEDS_XVFB" = "true" ]; then
# IAAI blocks headless Chromium on Linux; run headed via Xvfb virtual display
export DISPLAY=:99
export IAAI_HEADLESS=false
Xvfb :99 -screen 0 1920x1080x24 -nolisten tcp &
XVFB_PID=$!
sleep 0.5
echo "[entrypoint] Xvfb started (PID $XVFB_PID, DISPLAY=$DISPLAY)"
fi
# Run Alembic migrations (only for API service, skip for worker/beat)
case "$1" in
uvicorn*)
echo "[entrypoint] Running Alembic migrations..."
alembic upgrade head || echo "[entrypoint] WARNING: Alembic migration failed, continuing..."
;;
esac
exec "$@"

1
iaai_scraper/__init__.py Normal file
View File

@@ -0,0 +1 @@
__all__: list[str] = []

3
iaai_scraper/api/__init__.py Normal file
View File

@@ -0,0 +1,3 @@
from .app import create_app
__all__ = ["create_app"]

38
iaai_scraper/api/app.py Normal file
View File

@@ -0,0 +1,38 @@
from contextlib import asynccontextmanager
from fastapi import FastAPI
from ..core.config import Settings
from ..storage.db import PersistenceService
from .routes import cars, health, tasks
@asynccontextmanager
async def lifespan(app: FastAPI):
persistence: PersistenceService = app.state.persistence
persistence.create_tables()
yield
def create_app(settings: Settings | None = None) -> FastAPI:
_settings = settings or Settings()
app = FastAPI(
title="IAAI Scraper API",
description="REST API для управления задачами скрапинга IAAI и просмотра данных",
version="1.0.0",
lifespan=lifespan,
)
app.state.settings = _settings
app.state.persistence = PersistenceService(_settings)
app.include_router(health.router, tags=["health"])
app.include_router(cars.router, prefix="/api/v1", tags=["cars"])
app.include_router(tasks.router, prefix="/api/v1", tags=["tasks"])
return app
# uvicorn entrypoint
app = create_app()

7
iaai_scraper/api/deps.py Normal file
View File

@@ -0,0 +1,7 @@
from fastapi import Request
from ..storage.db import PersistenceService
def get_persistence(request: Request) -> PersistenceService:
return request.app.state.persistence

0
iaai_scraper/api/routes/__init__.py Normal file
View File

137
iaai_scraper/api/routes/cars.py Normal file
View File

@@ -0,0 +1,137 @@
from fastapi import APIRouter, Depends, HTTPException, Query
from sqlalchemy import func, select
from ..deps import get_persistence
from ...storage.db import PersistenceService
from ...storage.models import Car, Image
router = APIRouter()
@router.get("/cars")
def list_cars(
page: int = Query(1, ge=1),
per_page: int = Query(20, ge=1, le=100),
brand: str | None = None,
model: str | None = None,
year_min: int | None = None,
year_max: int | None = None,
is_sold: bool | None = None,
persistence: PersistenceService = Depends(get_persistence),
):
with persistence.session_scope() as session:
query = select(Car)
if brand:
query = query.where(Car.brand.ilike(f"%{brand}%"))
if model:
query = query.where(Car.model.ilike(f"%{model}%"))
if year_min is not None:
query = query.where(Car.year >= year_min)
if year_max is not None:
query = query.where(Car.year <= year_max)
if is_sold is not None:
query = query.where(Car.is_sold == is_sold)
# Общее количество.
count_query = select(func.count()).select_from(query.subquery())
total = session.execute(count_query).scalar() or 0
# пагинация
offset = (page - 1) * per_page
cars = session.execute(
query.order_by(Car.last_seen_at.desc()).offset(offset).limit(per_page)
).scalars().all()
return {
"total": total,
"page": page,
"per_page": per_page,
"pages": (total + per_page - 1) // per_page if per_page else 0,
"items": [_car_to_dict(car) for car in cars],
}
@router.get("/cars/{car_id}")
def get_car(
car_id: int,
persistence: PersistenceService = Depends(get_persistence),
):
with persistence.session_scope() as session:
car = session.get(Car, car_id)
if car is None:
raise HTTPException(status_code=404, detail="Car not found")
return _car_to_dict(car, include_images=True)
@router.get("/cars/by-origin/{origin_id}")
def get_car_by_origin(
origin_id: str,
persistence: PersistenceService = Depends(get_persistence),
):
with persistence.session_scope() as session:
car = session.execute(
select(Car).where(Car.origin_id == origin_id)
).scalars().first()
if car is None:
raise HTTPException(status_code=404, detail="Car not found")
return _car_to_dict(car, include_images=True)
@router.get("/stats")
def get_stats(persistence: PersistenceService = Depends(get_persistence)):
with persistence.session_scope() as session:
total_cars = session.execute(select(func.count(Car.id))).scalar() or 0
total_images = session.execute(select(func.count(Image.id))).scalar() or 0
brands = session.execute(
select(Car.brand, func.count(Car.id))
.group_by(Car.brand)
.order_by(func.count(Car.id).desc())
.limit(20)
).all()
return {
"total_cars": total_cars,
"total_images": total_images,
"top_brands": [{"brand": b, "count": c} for b, c in brands],
}
def _car_to_dict(car: Car, include_images: bool = False) -> dict:
result = {
"id": car.id,
"parser_id": car.parser_id,
"brand": car.brand,
"model": car.model,
"year": car.year,
"price": car.price,
"currency": car.currency,
"mileage": car.mileage,
"country": car.country,
"is_sold": car.is_sold,
"color": car.color,
"drive": car.drive,
"gearbox": car.gearbox,
"steering_wheel": car.steering_wheel,
"body_type": car.body_type,
"engine_volume": car.engine_volume,
"selling_type": car.selling_type,
"origin": car.origin,
"origin_url": car.origin_url,
"origin_id": car.origin_id,
"is_damaged": car.is_damaged,
"slug": car.slug,
"last_seen_at": car.last_seen_at.isoformat() if car.last_seen_at else None,
"content_hash": car.content_hash,
}
if include_images:
result["images"] = [
{
"id": img.id,
"fullres_image": img.fullres_image,
"preview_image": img.preview_image,
"order_index": img.order_index,
}
for img in sorted(car.images, key=lambda i: i.order_index)
]
return result

24
iaai_scraper/api/routes/health.py Normal file
View File

@@ -0,0 +1,24 @@
from fastapi import APIRouter, Depends
from sqlalchemy import text
from ..deps import get_persistence
from ...storage.db import PersistenceService
router = APIRouter()
@router.get("/health")
def health_check(persistence: PersistenceService = Depends(get_persistence)):
db_ok = False
try:
with persistence.session_scope() as session:
session.execute(text("SELECT 1"))
db_ok = True
except Exception:
pass
return {
"status": "ok" if db_ok else "degraded",
"service": "iaai-scraper-api",
"database": "connected" if db_ok else "unavailable",
}

174
iaai_scraper/api/routes/tasks.py Normal file
View File

@@ -0,0 +1,174 @@
from datetime import datetime, timezone
from fastapi import APIRouter, Depends, HTTPException, Query
from pydantic import BaseModel
from sqlalchemy import select, func
from ..deps import get_persistence
from ...storage.db import PersistenceService
from ...storage.models import ScrapeTask, SyncRun
from ...worker.tasks import sync_vehicle_task, sync_listing_task
router = APIRouter()
# --- схемы
class SyncVehicleRequest(BaseModel):
vehicle_url: str
lane: str = "iaai"
class SyncListingRequest(BaseModel):
make: str | None = None
model: str | None = None
lane: str = "iaai_cars"
limit: int | None = None
only_new: bool | None = None
# --- эндпоинты
@router.post("/tasks/sync-vehicle")
def start_sync_vehicle(
body: SyncVehicleRequest,
persistence: PersistenceService = Depends(get_persistence),
):
result = sync_vehicle_task.apply_async(
args=(body.vehicle_url,),
kwargs={"lane": body.lane},
queue="scraping",
)
# регистрируем в бд
persistence.create_scrape_task(
celery_task_id=result.id,
task_type="sync_vehicle",
vehicle_url=body.vehicle_url,
)
return {
"task_id": result.id,
"status": "queued",
"vehicle_url": body.vehicle_url,
}
@router.post("/tasks/sync-listing")
def start_sync_listing(
body: SyncListingRequest,
persistence: PersistenceService = Depends(get_persistence),
):
result = sync_listing_task.apply_async(
kwargs={
"make": body.make,
"model": body.model,
"lane": body.lane,
"limit": body.limit,
"only_new": body.only_new,
},
queue="scraping",
)
persistence.create_scrape_task(
celery_task_id=result.id,
task_type="sync_listing",
)
return {
"task_id": result.id,
"status": "queued",
}
@router.get("/tasks/{task_id}")
def get_task_status(
task_id: str,
persistence: PersistenceService = Depends(get_persistence),
):
task_info = persistence.get_scrape_task(task_id)
if task_info is None:
raise HTTPException(status_code=404, detail="Task not found")
return task_info
@router.get("/tasks")
def list_tasks(
page: int = Query(1, ge=1),
per_page: int = Query(20, ge=1, le=100),
status: str | None = None,
task_type: str | None = None,
persistence: PersistenceService = Depends(get_persistence),
):
with persistence.session_scope() as session:
query = select(ScrapeTask)
if status:
query = query.where(ScrapeTask.status == status)
if task_type:
query = query.where(ScrapeTask.task_type == task_type)
total = session.execute(
select(func.count()).select_from(query.subquery())
).scalar() or 0
offset = (page - 1) * per_page
tasks = session.execute(
query.order_by(ScrapeTask.created_at.desc()).offset(offset).limit(per_page)
).scalars().all()
return {
"total": total,
"page": page,
"per_page": per_page,
"items": [
{
"id": t.id,
"celery_task_id": t.celery_task_id,
"task_type": t.task_type,
"status": t.status,
"vehicle_url": t.vehicle_url,
"created_at": t.created_at.isoformat() if t.created_at else None,
"started_at": t.started_at.isoformat() if t.started_at else None,
"finished_at": t.finished_at.isoformat() if t.finished_at else None,
"result_summary": t.result_summary,
"error_message": t.error_message,
}
for t in tasks
],
}
@router.get("/sync-runs")
def list_sync_runs(
page: int = Query(1, ge=1),
per_page: int = Query(20, ge=1, le=100),
persistence: PersistenceService = Depends(get_persistence),
):
with persistence.session_scope() as session:
total = session.execute(select(func.count(SyncRun.id))).scalar() or 0
offset = (page - 1) * per_page
runs = session.execute(
select(SyncRun).order_by(SyncRun.started_at.desc()).offset(offset).limit(per_page)
).scalars().all()
return {
"total": total,
"page": page,
"per_page": per_page,
"items": [
{
"id": r.id,
"started_at": r.started_at.isoformat() if r.started_at else None,
"finished_at": r.finished_at.isoformat() if r.finished_at else None,
"status": r.status,
"lane": r.lane,
"ids_fetched": r.ids_fetched,
"cars_upserted": r.cars_upserted,
"cars_failed": r.cars_failed,
"images_upserted": r.images_upserted,
"error_summary": r.error_summary,
}
for r in runs
],
}

6
iaai_scraper/browser/__init__.py Normal file
View File

@@ -0,0 +1,6 @@
from .factory import BrowserFactory
from .listing import ListingCollector
from .network import NetworkCapture
from .pace import HumanPacer
__all__ = ["BrowserFactory", "ListingCollector", "NetworkCapture", "HumanPacer"]

117
iaai_scraper/browser/factory.py Normal file
View File

@@ -0,0 +1,117 @@
import json
import logging
import random
from playwright.sync_api import Browser, BrowserContext, Playwright
from ..core.config import Settings
logger = logging.getLogger("iaai_scraper.browser")
def _build_init_script() -> str:
"""JS-патч признаков автоматизации."""
hardware_concurrency = random.choice([4, 8, 12, 16])
device_memory = random.choice([4, 8, 16])
languages = ["en-US", "en"]
return f"""
(() => {{
const define = (obj, prop, value) => {{
try {{
Object.defineProperty(obj, prop, {{ get: () => value, configurable: true }});
}} catch (e) {{}}
}};
define(navigator, 'webdriver', undefined);
define(navigator, 'platform', 'Win32');
define(navigator, 'vendor', 'Google Inc.');
define(navigator, 'language', '{languages[0]}');
define(navigator, 'languages', {json.dumps(languages)});
define(navigator, 'hardwareConcurrency', {hardware_concurrency});
define(navigator, 'deviceMemory', {device_memory});
define(navigator, 'maxTouchPoints', 0);
if (!window.chrome) {{
Object.defineProperty(window, 'chrome', {{
value: {{ runtime: {{}}, app: {{}}, csi: () => ({{}}), loadTimes: () => ({{}}) }},
configurable: true
}});
}}
const originalQuery = navigator.permissions && navigator.permissions.query;
if (originalQuery) {{
navigator.permissions.query = (params) => (
params && params.name === 'notifications'
? Promise.resolve({{ state: Notification.permission }})
: originalQuery(params)
);
}}
const originalGetParameter = WebGLRenderingContext.prototype.getParameter;
WebGLRenderingContext.prototype.getParameter = function(parameter) {{
if (parameter === 37445) return 'Intel Inc.';
if (parameter === 37446) return 'Intel Iris OpenGL Engine';
return originalGetParameter.call(this, parameter);
}};
}})();
"""
class BrowserFactory:
def __init__(self, settings: Settings) -> None:
self.settings = settings
def create_browser(self, playwright: Playwright) -> Browser:
launch_kwargs: dict = {
"headless": self.settings.headless,
"args": [
"--disable-blink-features=AutomationControlled",
"--no-default-browser-check",
"--disable-dev-shm-usage",
"--disable-features=IsolateOrigins,site-per-process",
],
}
proxy_dict = self.settings.proxy.to_playwright_dict()
if proxy_dict:
launch_kwargs["proxy"] = proxy_dict
logger.info("Using proxy: %s", self.settings.proxy.server)
try:
logger.info("Trying to launch real Chrome channel")
return playwright.chromium.launch(channel="chrome", **launch_kwargs)
except Exception:
logger.warning("Chrome channel launch failed, falling back to Chromium")
return playwright.chromium.launch(**launch_kwargs)
def create_context(self, browser: Browser, storage_state: str | None = None) -> BrowserContext:
viewport = random.choice(self.settings.fingerprint.viewport_presets)
timezone_id = random.choice(self.settings.fingerprint.timezone_candidates)
color_scheme = random.choice(["light", "dark"])
context = browser.new_context(
storage_state=storage_state or None,
user_agent=self.settings.fingerprint.user_agent,
viewport=viewport,
screen=viewport,
device_scale_factor=random.choice([1, 1.25]),
is_mobile=False,
has_touch=False,
locale=self.settings.fingerprint.locale,
timezone_id=timezone_id,
color_scheme=color_scheme,
java_script_enabled=True,
ignore_https_errors=False,
extra_http_headers={
"Accept-Language": "en-US,en;q=0.9",
"DNT": "1",
"Upgrade-Insecure-Requests": "1",
"Sec-CH-UA": self.settings.fingerprint.sec_ch_ua,
"Sec-CH-UA-Mobile": "?0",
"Sec-CH-UA-Platform": '"Windows"',
},
)
context.set_default_timeout(self.settings.default_timeout_ms)
context.set_default_navigation_timeout(self.settings.default_timeout_ms)
context.add_init_script(_build_init_script())
return context

157
iaai_scraper/browser/listing.py Normal file
View File

@@ -0,0 +1,157 @@
import logging
import re
from dataclasses import asdict, dataclass, field
from typing import Any
from urllib.parse import urljoin
from playwright.sync_api import Page
from .pace import HumanPacer
from ..core.config import Settings
from ..core.utils import first_non_empty
logger = logging.getLogger("iaai_scraper.listing")
VEHICLE_HREF_RE = re.compile(r"/VehicleDetail/\d+(?:~[A-Z]{2})?", re.IGNORECASE)
@dataclass(slots=True)
class ListingVehicleLink:
# Ссылка на карточку
href: str
title: str = ""
lot_number: str | None = None
@dataclass(slots=True)
class ListingPageResult:
# Результат страницы листинга
source_url: str
page_number: int
vehicle_links: list[ListingVehicleLink] = field(default_factory=list)
pagination_available: bool = False
next_page_detected: bool = False
class ListingCollector:
def __init__(self, settings: Settings, pacer: HumanPacer) -> None:
# Сборщик ссылок листинга
self.settings = settings
self.pacer = pacer
def open_cars_listing(self, page: Page) -> None:
# Открываем листинг и ждём ссылки
logger.info("Opening cars listing page: %s", self.settings.listing.cars_url)
page.goto(self.settings.listing.cars_url, wait_until="domcontentloaded")
try:
page.wait_for_load_state("networkidle", timeout=15000)
except Exception:
logger.debug("networkidle timeout on listing page, continuing with current state")
try:
page.wait_for_selector("a[href*='/VehicleDetail/']", timeout=20000)
except Exception:
logger.warning("Vehicle links did not appear within timeout; page may not have rendered fully")
self.pacer.after_listing_open()
def apply_filters(self, page: Page, make: str | None = None, model: str | None = None) -> dict[str, str | None]:
# Пробуем make/model фильтры
applied = {"make": None, "model": None}
if make and self._try_fill_filter_input(page, ["input[placeholder*='Make']", "input[aria-label*='Make']"], make):
applied["make"] = make
self.pacer.after_filter_action()
if model and self._try_fill_filter_input(page, ["input[placeholder*='Model']", "input[aria-label*='Model']"], model):
applied["model"] = model
self.pacer.after_filter_action()
return applied
def collect_current_page(self, page: Page, page_number: int = 1) -> ListingPageResult:
# Собираем ссылки текущей страницы
anchors = page.locator("a[href*='/VehicleDetail/']")
total = min(anchors.count(), self.settings.listing.page_link_limit)
links: list[ListingVehicleLink] = []
seen: set[str] = set()
for idx in range(total):
anchor = anchors.nth(idx)
href = anchor.get_attribute("href") or ""
match = VEHICLE_HREF_RE.search(href)
if not match:
continue
absolute = urljoin(self.settings.home_url, match.group(0))
if absolute in seen:
continue
seen.add(absolute)
title = first_non_empty([anchor.get_attribute("title"), anchor.text_content(), ""]) or ""
links.append(ListingVehicleLink(href=absolute, title=str(title).strip()))
if len(links) >= self.settings.listing.max_vehicles_per_run:
break
next_page_detected = self._has_next_page(page)
return ListingPageResult(source_url=page.url, page_number=page_number, vehicle_links=links, pagination_available=next_page_detected, next_page_detected=next_page_detected)
def go_to_next_page(self, page: Page) -> bool:
# Переход на следующую страницу
selectors = ["a[aria-label*='Next']", "button[aria-label*='Next']", "a.pagination-next", "button.pagination-next", "a:has-text('Next')", "button:has-text('Next')"]
for selector in selectors:
locator = page.locator(selector).first
if locator.count() == 0:
continue
disabled = (locator.get_attribute("disabled") or "").lower()
aria_disabled = (locator.get_attribute("aria-disabled") or "").lower()
classes = (locator.get_attribute("class") or "").lower()
if disabled or aria_disabled == "true" or "disabled" in classes:
continue
self.pacer.move_mouse_to(page, locator)
locator.click()
try:
page.wait_for_load_state("networkidle", timeout=15000)
except Exception:
pass
self.pacer.after_page_change()
return True
return False
def collect_listing_links(self, page: Page, *, make: str | None = None, model: str | None = None) -> dict[str, Any]:
# Последовательный сбор с лимитами
self.open_cars_listing(page)
applied_filters = self.apply_filters(page, make=make, model=model)
pages: list[dict[str, object]] = []
all_links: list[str] = []
for page_number in range(1, max(1, self.settings.listing.max_pages_per_run) + 1):
page_result = self.collect_current_page(page, page_number=page_number)
pages.append({"page_number": page_result.page_number, "source_url": page_result.source_url, "links_found": len(page_result.vehicle_links), "vehicle_links": [asdict(item) for item in page_result.vehicle_links], "next_page_detected": page_result.next_page_detected})
for item in page_result.vehicle_links:
if item.href not in all_links:
all_links.append(item.href)
if len(all_links) >= self.settings.listing.max_vehicles_per_run:
break
if len(all_links) >= self.settings.listing.max_vehicles_per_run or self.settings.listing.collect_current_page_only or not self.settings.listing.include_pagination or not page_result.next_page_detected:
break
if not self.go_to_next_page(page):
break
return {"listing_url": self.settings.listing.cars_url, "applied_filters": applied_filters, "pages_collected": len(pages), "vehicles_collected": len(all_links), "vehicle_urls": all_links, "pages": pages, "strategy": {"sequential": True, "collect_current_page_only": self.settings.listing.collect_current_page_only, "include_pagination": self.settings.listing.include_pagination, "max_pages_per_run": self.settings.listing.max_pages_per_run, "max_vehicles_per_run": self.settings.listing.max_vehicles_per_run}}
@staticmethod
def _try_fill_filter_input(page: Page, selectors: list[str], value: str) -> bool:
# Несколько селекторов для фильтра
for selector in selectors:
locator = page.locator(selector).first
if locator.count() == 0:
continue
try:
locator.click()
locator.fill(value)
page.keyboard.press("Enter")
try:
page.wait_for_load_state("networkidle", timeout=15000)
except Exception:
pass
return True
except Exception:
continue
return False
@staticmethod
def _has_next_page(page: Page) -> bool:
# Проверка кнопки next
for selector in ["a[aria-label*='Next']", "button[aria-label*='Next']", "a.pagination-next", "button.pagination-next", "a:has-text('Next')", "button:has-text('Next')"]:
if page.locator(selector).count() > 0:
return True
return False

122
iaai_scraper/browser/network.py Normal file
View File

@@ -0,0 +1,122 @@
import json
import logging
from dataclasses import dataclass, field
from typing import Any
from urllib.parse import urlparse
from playwright.sync_api import Page, Request, Response
from ..core.config import Settings
logger = logging.getLogger("iaai_scraper.network")
@dataclass
class NetworkCapture:
"""XHR/fetch перехватчик."""
settings: Settings
requests: list[dict[str, Any]] = field(default_factory=list)
json_responses: list[dict[str, Any]] = field(default_factory=list)
_seen_req: set[str] = field(default_factory=set)
_seen_resp: set[str] = field(default_factory=set)
_origin: str | None = None
def attach(self, page: Page, origin_url: str | None = None) -> None:
# Подписка на сетевые события
try:
self._origin = urlparse(origin_url or page.url).netloc.lower() or None
except Exception:
self._origin = None
page.on("request", self._on_request)
page.on("response", self._on_response)
def _is_same_origin(self, url: str) -> bool:
# Фильтр по домену
if not self.settings.capture.capture_same_origin_only or not self._origin:
return True
netloc = urlparse(url).netloc.lower()
return netloc == self._origin or netloc.endswith(".iaai.com")
def _on_request(self, request: Request) -> None:
# Берём только xhr/fetch
if request.resource_type not in {"xhr", "fetch"}:
return
if not self._is_same_origin(request.url):
return
if len(self.requests) >= self.settings.capture.max_requests:
return
key = f"{request.method}:{request.url}:{request.post_data or ''}"
# Убираем дубли
if key in self._seen_req:
return
self._seen_req.add(key)
self.requests.append({
"url": request.url, "method": request.method,
"resource_type": request.resource_type, "post_data": request.post_data,
})
def _on_response(self, response: Response) -> None:
# Сохраняем JSON
request = response.request
if request.resource_type not in {"xhr", "fetch"}:
return
if not self._is_same_origin(response.url):
return
if len(self.json_responses) >= self.settings.capture.max_json_responses:
return
if not self._is_json(response):
return
key = f"{request.method}:{response.url}:{response.status}"
if key in self._seen_resp:
return
self._seen_resp.add(key)
try:
payload = response.json()
except Exception:
try:
payload = json.loads(response.text())
except Exception:
return
self.json_responses.append({
"url": response.url, "status": response.status,
"request_method": request.method, "post_data": request.post_data,
"resource_type": request.resource_type, "payload": payload,
"category": self._categorize(response.url),
})
@staticmethod
def _is_json(response: Response) -> bool:
ct = (response.headers.get("content-type") or "").lower()
if "application/json" in ct or "+json" in ct:
return True
url = response.url.lower()
return any(m in url for m in ["/api/", "/graphql", "vehicledetail", "vehicle", "auction", "bid", "images", "media"])
@staticmethod
def _categorize(url: str) -> str:
# Простая категория URL
low = url.lower()
mapping = {
"images": ["image", "media", "photos", "gallery"],
"bids": ["bid", "offer", "buy-now", "buynow"],
"auction": ["auction", "sale", "lane", "branch"],
"vehicle": ["vehicle", "detail", "vin", "damage", "runanddrive"],
"documents": ["title", "document", "report"],
}
for cat, markers in mapping.items():
if any(m in low for m in markers):
return cat
return "other"
def export(self) -> dict[str, Any]:
responses = sorted(self.json_responses, key=lambda i: (i["category"] == "other", i["url"]))
return {
"requests": self.requests,
"json_responses": responses,
"capture_limits": {
"same_origin_only": self.settings.capture.capture_same_origin_only,
"max_requests": self.settings.capture.max_requests,
"max_json_responses": self.settings.capture.max_json_responses,
},
}

46
iaai_scraper/browser/pace.py Normal file
View File

@@ -0,0 +1,46 @@
import random
import time
from playwright.sync_api import Locator, Page
from ..core.config import Settings
class HumanPacer:
"""Random паузы между действиями."""
def __init__(self, settings: Settings) -> None:
self.settings = settings
def pause(self, min_s: float, max_s: float) -> None:
if self.settings.pace.enabled:
time.sleep(random.uniform(min_s, max_s))
def after_listing_open(self) -> None:
self.pause(self.settings.pace.after_listing_open_min_s, self.settings.pace.after_listing_open_max_s)
def after_filter_action(self) -> None:
self.pause(self.settings.pace.after_filter_action_min_s, self.settings.pace.after_filter_action_max_s)
def before_vehicle_open(self) -> None:
self.pause(self.settings.pace.before_vehicle_open_min_s, self.settings.pace.before_vehicle_open_max_s)
def after_vehicle_open(self) -> None:
self.pause(self.settings.pace.after_vehicle_open_min_s, self.settings.pace.after_vehicle_open_max_s)
def between_vehicles(self) -> None:
self.pause(self.settings.pace.between_vehicles_min_s, self.settings.pace.between_vehicles_max_s)
def after_page_change(self) -> None:
self.pause(self.settings.pace.after_page_change_min_s, self.settings.pace.after_page_change_max_s)
def move_mouse_to(self, page: Page, locator: Locator) -> None:
try:
box = locator.bounding_box()
except Exception:
box = None
if not box:
return
x = box["x"] + box["width"] * random.uniform(0.2, 0.8)
y = box["y"] + box["height"] * random.uniform(0.2, 0.8)
page.mouse.move(x, y, steps=random.randint(8, 18))

82
iaai_scraper/cli.py Normal file
View File

@@ -0,0 +1,82 @@
import argparse
from pathlib import Path
from .core.config import Settings
from .core.utils import save_to_json
from .scraper import IAAIScraper
def build_parser() -> argparse.ArgumentParser:
parser = argparse.ArgumentParser(description="IAAI scraper CLI")
parser.add_argument("--headless", choices=["true", "false"], default=None, help="Override headless mode")
parser.add_argument("--debug", action="store_true", help="Enable DEBUG logging")
subparsers = parser.add_subparsers(dest="command", required=True)
default_output_dir = Path("artifacts/json")
subparsers.add_parser("init-db", help="Create DB tables")
listing_parser = subparsers.add_parser("collect-listing", help="Collect vehicle URLs from listing page")
listing_parser.add_argument("--make", default=None)
listing_parser.add_argument("--model", default=None)
listing_parser.add_argument("--output", default=str(default_output_dir / "iaai_listing_links.json"))
scrape_parser = subparsers.add_parser("scrape-vehicle", help="Scrape a vehicle detail page")
scrape_parser.add_argument("vehicle_url")
scrape_parser.add_argument("--output", default=str(default_output_dir / "iaai_vehicle_detail.json"))
sync_vehicle_parser = subparsers.add_parser("sync-vehicle", help="Scrape + upsert one vehicle")
sync_vehicle_parser.add_argument("vehicle_url")
sync_vehicle_parser.add_argument("--lane", default="iaai")
sync_vehicle_parser.add_argument("--output", default=str(default_output_dir / "iaai_sync_vehicle.json"))
sync_listing_parser = subparsers.add_parser("sync-listing", help="Collect listing + sync all vehicles")
sync_listing_parser.add_argument("--make", default=None)
sync_listing_parser.add_argument("--model", default=None)
sync_listing_parser.add_argument("--lane", default="iaai_cars")
sync_listing_parser.add_argument("--limit", type=int, default=None)
sync_listing_parser.add_argument("--only-new", choices=["true", "false"], default=None)
sync_listing_parser.add_argument("--output", default=str(default_output_dir / "iaai_sync_listing.json"))
return parser
def main() -> None:
parser = build_parser()
args = parser.parse_args()
runtime_settings: Settings | None = None
if args.headless is not None or args.debug:
runtime_settings = Settings()
if args.headless is not None:
runtime_settings.headless = args.headless == "true"
if args.debug:
runtime_settings.log_level = "DEBUG"
with IAAIScraper(runtime_settings) as scraper:
if args.command == "init-db":
data = scraper.init_db()
print(f"DB initialized: {data}")
return
elif args.command == "collect-listing":
data = scraper.collect_listing(make=args.make, model=args.model)
elif args.command == "scrape-vehicle":
data = scraper.scrape_vehicle_detail(args.vehicle_url)
elif args.command == "sync-vehicle":
data = scraper.sync_vehicle(args.vehicle_url, lane=args.lane)
else:
only_new = None if args.only_new is None else args.only_new == "true"
data = scraper.sync_listing(
make=args.make,
model=args.model,
lane=args.lane,
limit=args.limit,
only_new=only_new,
)
save_to_json(data, Path(args.output))
print(f"Saved to {Path(args.output).resolve()}")
if __name__ == "__main__":
main()

1
iaai_scraper/core/__init__.py Normal file
View File

@@ -0,0 +1 @@
__all__: list[str] = []

137
iaai_scraper/core/config.py Normal file
View File

@@ -0,0 +1,137 @@
import os
from dataclasses import dataclass, field
from dotenv import load_dotenv
load_dotenv()
@dataclass(slots=True)
class FingerprintConfig:
user_agent: str = (
"Mozilla/5.0 (Windows NT 10.0; Win64; x64) "
"AppleWebKit/537.36 (KHTML, like Gecko) "
"Chrome/135.0.0.0 Safari/537.36"
)
viewport_presets: tuple[dict[str, int], ...] = (
{"width": 1920, "height": 1080},
{"width": 1600, "height": 900},
{"width": 1536, "height": 864},
{"width": 1440, "height": 900},
{"width": 1366, "height": 768},
)
timezone_candidates: tuple[str, ...] = (
"America/New_York",
"America/Chicago",
"America/Los_Angeles",
)
locale: str = "en-US"
sec_ch_ua: str = '"Google Chrome";v="135", "Chromium";v="135", "Not.A/Brand";v="24"'
@dataclass(slots=True)
class CaptureConfig:
capture_same_origin_only: bool = os.getenv("IAAI_CAPTURE_SAME_ORIGIN_ONLY", "true").strip().lower() in {"1", "true", "yes", "on"}
max_requests: int = int(os.getenv("IAAI_MAX_CAPTURED_REQUESTS", "40"))
max_json_responses: int = int(os.getenv("IAAI_MAX_CAPTURED_JSON_RESPONSES", "30"))
@dataclass(slots=True)
class HumanPaceConfig:
enabled: bool = os.getenv("IAAI_HUMAN_PACE_ENABLED", "true").strip().lower() in {"1", "true", "yes", "on"}
after_listing_open_min_s: float = float(os.getenv("IAAI_AFTER_LISTING_OPEN_MIN_S", "2.5"))
after_listing_open_max_s: float = float(os.getenv("IAAI_AFTER_LISTING_OPEN_MAX_S", "4.5"))
after_filter_action_min_s: float = float(os.getenv("IAAI_AFTER_FILTER_ACTION_MIN_S", "2.0"))
after_filter_action_max_s: float = float(os.getenv("IAAI_AFTER_FILTER_ACTION_MAX_S", "4.0"))
before_vehicle_open_min_s: float = float(os.getenv("IAAI_BEFORE_VEHICLE_OPEN_MIN_S", "0.5"))
before_vehicle_open_max_s: float = float(os.getenv("IAAI_BEFORE_VEHICLE_OPEN_MAX_S", "1.5"))
after_vehicle_open_min_s: float = float(os.getenv("IAAI_AFTER_VEHICLE_OPEN_MIN_S", "0.3"))
after_vehicle_open_max_s: float = float(os.getenv("IAAI_AFTER_VEHICLE_OPEN_MAX_S", "1.0"))
between_vehicles_min_s: float = float(os.getenv("IAAI_BETWEEN_VEHICLES_MIN_S", "0.5"))
between_vehicles_max_s: float = float(os.getenv("IAAI_BETWEEN_VEHICLES_MAX_S", "1.5"))
after_page_change_min_s: float = float(os.getenv("IAAI_AFTER_PAGE_CHANGE_MIN_S", "3.0"))
after_page_change_max_s: float = float(os.getenv("IAAI_AFTER_PAGE_CHANGE_MAX_S", "6.0"))
@dataclass(slots=True)
class ListingConfig:
cars_url: str = os.getenv("IAAI_CARS_LISTING_URL", "https://www.iaai.com/Vehiclelisting/Cars")
max_pages_per_run: int = int(os.getenv("IAAI_MAX_PAGES_PER_RUN", "5"))
max_vehicles_per_run: int = int(os.getenv("IAAI_MAX_VEHICLES_PER_RUN", "100"))
page_link_limit: int = int(os.getenv("IAAI_PAGE_LINK_LIMIT", "200"))
include_pagination: bool = os.getenv("IAAI_INCLUDE_PAGINATION", "true").strip().lower() in {"1", "true", "yes", "on"}
collect_current_page_only: bool = os.getenv("IAAI_COLLECT_CURRENT_PAGE_ONLY", "false").strip().lower() in {"1", "true", "yes", "on"}
@dataclass(slots=True)
class DatabaseConfig:
url: str = os.getenv("IAAI_DATABASE_URL", "postgresql+psycopg2://iaai:iaai@localhost:5432/iaai_scraper")
echo: bool = os.getenv("IAAI_DATABASE_ECHO", "false").strip().lower() in {"1", "true", "yes", "on"}
pool_size: int = int(os.getenv("IAAI_DATABASE_POOL_SIZE", "5"))
max_overflow: int = int(os.getenv("IAAI_DATABASE_MAX_OVERFLOW", "10"))
@dataclass(slots=True)
class RedisConfig:
url: str = os.getenv("IAAI_REDIS_URL", "redis://localhost:6379/0")
@dataclass(slots=True)
class CeleryConfig:
broker_url: str = os.getenv("CELERY_BROKER_URL", "") or ""
result_backend: str = os.getenv("CELERY_RESULT_BACKEND", "") or ""
task_soft_time_limit: int = int(os.getenv("CELERY_TASK_SOFT_TIME_LIMIT", "600"))
task_time_limit: int = int(os.getenv("CELERY_TASK_TIME_LIMIT", "900"))
worker_concurrency: int = int(os.getenv("CELERY_WORKER_CONCURRENCY", "1"))
beat_sync_interval_minutes: int = int(os.getenv("CELERY_BEAT_SYNC_INTERVAL_MINUTES", "60"))
beat_sync_limit: int = int(os.getenv("CELERY_BEAT_SYNC_LIMIT", "26"))
@dataclass(slots=True)
class ProxyConfig:
server: str | None = (os.getenv("IAAI_PROXY_SERVER") or "").strip() or None
username: str | None = (os.getenv("IAAI_PROXY_USERNAME") or "").strip() or None
password: str | None = (os.getenv("IAAI_PROXY_PASSWORD") or "").strip() or None
@property
def enabled(self) -> bool:
return bool(self.server)
def to_playwright_dict(self) -> dict[str, str] | None:
if not self.server:
return None
result: dict[str, str] = {"server": self.server}
if self.username:
result["username"] = self.username
if self.password:
result["password"] = self.password
return result
@dataclass(slots=True)
class Settings:
home_url: str = "https://www.iaai.com/"
default_timeout_ms: int = int(os.getenv("IAAI_TIMEOUT_MS", "45000"))
network_settle_ms: int = int(os.getenv("IAAI_NETWORK_SETTLE_MS", "800"))
max_retries: int = int(os.getenv("IAAI_MAX_RETRIES", "3"))
retry_delay_seconds: float = float(os.getenv("IAAI_RETRY_DELAY_SECONDS", "2.5"))
retry_backoff_multiplier: float = float(os.getenv("IAAI_RETRY_BACKOFF_MULTIPLIER", "2.0"))
retry_jitter_seconds: float = float(os.getenv("IAAI_RETRY_JITTER_SECONDS", "0.25"))
headless: bool = os.getenv("IAAI_HEADLESS", "true").strip().lower() in {"1", "true", "yes", "on"}
log_level: str = os.getenv("IAAI_LOG_LEVEL", "INFO")
log_file: str | None = os.getenv("IAAI_LOG_FILE") or None
enable_trace_id_logs: bool = os.getenv("IAAI_ENABLE_TRACE_ID_LOGS", "true").strip().lower() in {"1", "true", "yes", "on"}
sync_only_new: bool = os.getenv("IAAI_SYNC_ONLY_NEW", "true").strip().lower() in {"1", "true", "yes", "on"}
raw_output_json: str | None = os.getenv("IAAI_RAW_OUTPUT_JSON") or None
scheduler_interval_minutes: int = int(os.getenv("IAAI_SCHEDULER_INTERVAL_MINUTES", "60"))
fingerprint: FingerprintConfig = field(default_factory=FingerprintConfig)
capture: CaptureConfig = field(default_factory=CaptureConfig)
pace: HumanPaceConfig = field(default_factory=HumanPaceConfig)
listing: ListingConfig = field(default_factory=ListingConfig)
database: DatabaseConfig = field(default_factory=DatabaseConfig)
redis: RedisConfig = field(default_factory=RedisConfig)
celery: CeleryConfig = field(default_factory=CeleryConfig)
proxy: ProxyConfig = field(default_factory=ProxyConfig)
settings = Settings()

31
iaai_scraper/core/logs.py Normal file
View File

@@ -0,0 +1,31 @@
import logging
import sys
from contextvars import ContextVar
TRACE_ID: ContextVar[str] = ContextVar("trace_id", default="-")
class TraceIdFilter(logging.Filter):
def filter(self, record: logging.LogRecord) -> bool:
record.trace_id = TRACE_ID.get()
return True
def set_trace_id(trace_id: str) -> None:
TRACE_ID.set(trace_id)
def setup_logging(level: str = "INFO", log_file: str | None = None) -> None:
handlers: list[logging.Handler] = [logging.StreamHandler(sys.stdout)]
if log_file:
handlers.append(logging.FileHandler(log_file, encoding="utf-8"))
trace_filter = TraceIdFilter()
for handler in handlers:
handler.addFilter(trace_filter)
logging.basicConfig(
level=getattr(logging, level.upper(), logging.INFO),
format="%(asctime)s | %(levelname)s | %(name)s | trace=%(trace_id)s | %(message)s",
handlers=handlers,
force=True,
)

49
iaai_scraper/core/retry.py Normal file
View File

@@ -0,0 +1,49 @@
import logging
import random
import time
from collections.abc import Callable
from functools import wraps
from typing import Any
from playwright.sync_api import Error, TimeoutError as PlaywrightTimeoutError
logger = logging.getLogger("iaai_scraper.retry")
RETRYABLE_EXCEPTIONS = (
PlaywrightTimeoutError,
Error,
ConnectionError,
OSError,
TimeoutError,
)
def retryable(
max_attempts: int,
delay_seconds: float = 2.5,
backoff_multiplier: float = 2.0,
jitter_seconds: float = 0.0,
) -> Callable[[Callable[..., Any]], Callable[..., Any]]:
def decorator(func: Callable[..., Any]) -> Callable[..., Any]:
@wraps(func)
def wrapper(*args: Any, **kwargs: Any) -> Any:
last_error: Exception | None = None
for attempt in range(1, max_attempts + 1):
try:
return func(*args, **kwargs)
except RETRYABLE_EXCEPTIONS as exc:
last_error = exc
logger.warning("%s failed on attempt %s/%s: %s", func.__name__, attempt, max_attempts, exc)
if attempt < max_attempts:
sleep_for = delay_seconds * (backoff_multiplier ** (attempt - 1))
if jitter_seconds > 0:
sleep_for += random.uniform(0, jitter_seconds)
logger.debug("Retrying %s in %.2fs", func.__name__, sleep_for)
time.sleep(sleep_for)
if last_error is not None:
raise last_error
raise RuntimeError("Retry wrapper failed without a captured exception")
return wrapper
return decorator

57
iaai_scraper/core/utils.py Normal file
View File

@@ -0,0 +1,57 @@
import json
import random
import re
import time
from pathlib import Path
from typing import Any, Iterable
# Файловые утилиты
def save_to_json(data: Any, filename: str | Path) -> None:
path = Path(filename)
path.parent.mkdir(parents=True, exist_ok=True)
path.write_text(json.dumps(data, ensure_ascii=False, indent=2), encoding="utf-8")
# Небольшая случайная пауза между запросами
def short_sleep(a: float = 0.10, b: float = 0.35) -> None:
time.sleep(random.uniform(a, b))
# Маскирование персональных данных
def mask_email(email: str) -> str:
if "@" not in email:
return "***"
local, domain = email.split("@", 1)
safe_local = local[:2] + "***" if len(local) > 2 else local[:1] + "*"
return f"{safe_local}@{domain}"
# Возвращает первое непустое значение
def first_non_empty(values: Iterable[Any]) -> Any | None:
for value in values:
if value not in (None, "", [], {}, ()):
return value
return None
# Регулярные выражения для VIN, lot и price
VIN_RE = re.compile(r"\b([A-HJ-NPR-Z0-9]{17})\b", re.IGNORECASE)
LOT_RE = re.compile(r"\b(\d{7,10})\b")
PRICE_RE = re.compile(r"\$\s?([\d,]+(?:\.\d{1,2})?)")
# Глубокий рекурсивный поиск значений по ключам
def deep_find_key(obj, target_keys: set[str], max_depth: int = 64, _depth: int = 0) -> list:
found = []
if _depth >= max_depth:
return found
if isinstance(obj, dict):
for key, value in obj.items():
if key.lower() in target_keys:
found.append(value)
found.extend(deep_find_key(value, target_keys, max_depth=max_depth, _depth=_depth + 1))
elif isinstance(obj, list):
for item in obj:
found.extend(deep_find_key(item, target_keys, max_depth=max_depth, _depth=_depth + 1))
return found

1
iaai_scraper/parsing/__init__.py Normal file
View File

@@ -0,0 +1 @@
__all__: list[str] = []

412
iaai_scraper/parsing/mapper.py Normal file
View File

@@ -0,0 +1,412 @@
import hashlib
import json
import re
from datetime import datetime, timezone
from typing import Any
from urllib.parse import urlparse
from ..core.utils import first_non_empty
from ..storage.enums import (
BODY_TYPE_ENUM_VALUES,
COUNTRY_ENUM_VALUES,
CURRENCY_ENUM_VALUES,
DRIVE_ENUM_VALUES,
GEARBOX_ENUM_VALUES,
ORIGIN_ENUM_VALUES,
SELLING_TYPE_ENUM_VALUES,
STEERING_WHEEL_ENUM_VALUES,
)
from ..storage.schemas import CarRecord, ImageRecord
class CarMapper:
"""IAAI data → CarRecord."""
BODY_MAP = {
"sedan": "SEDAN", "coupe": "COUPE", "hatchback": "HATCHBACK", "sport utility": "SUV",
"suv": "SUV", "wagon": "STATION_WAGON", "station wagon": "STATION_WAGON", "pickup": "PICKUP",
"pickup truck": "PICKUP", "crew cab": "PICKUP", "extended cab": "PICKUP", "regular cab": "PICKUP",
"quad cab": "PICKUP", "double cab": "PICKUP", "king cab": "PICKUP", "mega cab": "PICKUP",
"supercab": "PICKUP", "supercrew": "PICKUP", "truck": "TRUCK", "van": "MINIVAN",
"minivan": "MINIVAN", "convertible": "OPEN", "cabriolet": "OPEN", "rv": "RV", "crossover": "SUV",
}
DRIVE_MAP = {
"front wheel drive": "FWD", "fwd": "FWD", "rear wheel drive": "RWD", "rwd": "RWD",
"all wheel drive": "4WD", "awd": "4WD", "4x4": "4WD", "four wheel drive": "4WD",
"4wd": "4WD", "2wd": "2WD", "two wheel drive": "2WD",
}
GEARBOX_MAP = {
"automatic": "AT", "automatic transmission": "AT", "a/t": "AT", "aut": "AT",
"manual": "MT", "manual transmission": "MT", "m/t": "MT", "cvt": "CVT",
"continuously variable transmission": "CVT", "electric": "EV", "ev": "EV",
}
STEERING_MAP = {"left": "LEFT", "left hand drive": "LEFT", "right": "RIGHT", "right hand drive": "RIGHT"}
COUNTRY_MAP = {
"us": "US", "usa": "US", "united states": "US", "ca": "CA", "canada": "CA",
"jp": "JP", "japan": "JP", "kr": "KR", "korea": "KR", "south korea": "KR",
}
NO_DAMAGE_MARKERS = {"normal wear", "normal wear & tear", "normal wear and tear", "n/a", "na", "none", "no damage", "minor dents/scratches"}
def map_to_car_record(self, vehicle_url: str, vehicle_summary: dict[str, Any], payload_insights: dict[str, Any]) -> CarRecord:
# Собираем нормализованную DB-модель.
vehicle_summary = vehicle_summary or {}
payload_insights = payload_insights or {}
notes: list[str] = []
core = payload_insights.get("vehicle_core", {})
pricing = payload_insights.get("pricing", {})
damage = payload_insights.get("damage", {})
auction = payload_insights.get("auction", {})
images = payload_insights.get("images", {})
origin_id = self._build_origin_id(vehicle_url, vehicle_summary, core)
parser_id = f"iaai:{origin_id}"
brand = self._as_str(first_non_empty([core.get("make"), vehicle_summary.get("make")])) or "UNKNOWN"
model = self._as_str(first_non_empty([core.get("model"), vehicle_summary.get("model")])) or "UNKNOWN"
year = self._to_year(first_non_empty([core.get("year"), vehicle_summary.get("year")]))
price = self._first_parsed_int(
[
pricing.get("buy_now"),
pricing.get("current_bid"),
vehicle_summary.get("buy_now"),
vehicle_summary.get("current_bid"),
pricing.get("actual_cash_value"),
],
self._to_money_int,
)
mileage = self._first_parsed_int(
[core.get("odometer"), vehicle_summary.get("odometer"), 0],
self._to_int,
) or 0
color = self._normalize_color(first_non_empty([core.get("color"), vehicle_summary.get("color"), "other"]))
drive = self._normalize_drive(first_non_empty([core.get("drive"), vehicle_summary.get("drive")]))
gearbox = self._normalize_gearbox(first_non_empty([core.get("gearbox"), vehicle_summary.get("gearbox")]))
steering = self._normalize_steering(first_non_empty([core.get("steering_wheel"), vehicle_summary.get("steering_wheel")])) or "LEFT"
body_type = self._normalize_body_type(first_non_empty([core.get("body_type"), vehicle_summary.get("body_type")]))
engine_volume = self._to_engine_cc(first_non_empty([core.get("engine"), core.get("engine_volume"), vehicle_summary.get("engine"), vehicle_summary.get("engine_volume")]))
title_text = self._as_str(first_non_empty([core.get("title"), vehicle_summary.get("title"), ""]))
seller = self._as_str(first_non_empty([core.get("seller"), vehicle_summary.get("seller"), ""]))
location = self._as_str(first_non_empty([core.get("location"), vehicle_summary.get("location"), auction.get("branch"), ""]))
country = self._normalize_country(first_non_empty([core.get("country"), location, "US"]))
is_damaged = self._bool_damage(damage, vehicle_summary)
is_sold = self._bool_sold(auction)
one_owner = self._boolish(first_non_empty([core.get("one_owner"), vehicle_summary.get("one_owner"), False]))
new_car = self._boolish(first_non_empty([core.get("new_car"), vehicle_summary.get("new_car"), False]))
rental = self._boolish(first_non_empty([core.get("rental"), vehicle_summary.get("rental"), False]))
repair_history = self._boolish(first_non_empty([core.get("repair_history"), vehicle_summary.get("repair_history"), False]))
non_smoking = self._boolish(first_non_empty([core.get("non_smoking"), vehicle_summary.get("non_smoking"), True]))
evaluation = self._as_str(first_non_empty([core.get("grade"), core.get("evaluation"), vehicle_summary.get("evaluation")])) or None
currency = self._normalize_currency(
first_non_empty(
[
pricing.get("currency"),
vehicle_summary.get("currency"),
pricing.get("buy_now"),
pricing.get("current_bid"),
vehicle_summary.get("buy_now"),
vehicle_summary.get("current_bid"),
"USD",
]
)
)
slug = self._slugify(" ".join(filter(None, [str(year or ""), brand, model, origin_id])))
images_records = self._build_images(images.get("urls") or vehicle_summary.get("image_urls") or [])
origin = "IAAI" if "IAAI" in ORIGIN_ENUM_VALUES else "NA"
if not price:
notes.append("Price is missing or not parseable from the observed payloads.")
if not vehicle_summary.get("vin"):
notes.append("VIN was not observed in the accessible payloads for this account/session.")
if not images_records:
notes.append("No image URLs were found in the captured payloads.")
raw_attributes = {
# Сохраняем полезный сырой контекст.
"vin": vehicle_summary.get("vin"),
"lot_number": first_non_empty([core.get("lot_number"), vehicle_summary.get("lot_number")]),
"trim": first_non_empty([core.get("trim"), vehicle_summary.get("trim")]),
"fuel_type": first_non_empty([core.get("fuel_type"), vehicle_summary.get("fuel_type")]),
"cylinders": first_non_empty([core.get("cylinders"), vehicle_summary.get("cylinders")]),
"engine": first_non_empty([core.get("engine"), vehicle_summary.get("engine")]),
"manufactured_in": vehicle_summary.get("manufactured_in"),
"vehicle_class": vehicle_summary.get("vehicle_class"),
"run_and_drive": first_non_empty([core.get("run_and_drive"), vehicle_summary.get("run_and_drive")]),
"keys": first_non_empty([core.get("keys"), vehicle_summary.get("keys")]),
"title": title_text,
"title_brand": vehicle_summary.get("title_brand"),
"damage_primary": first_non_empty([damage.get("primary"), vehicle_summary.get("primary_damage")]),
"damage_secondary": damage.get("secondary"),
"damage_description": damage.get("description"),
"buy_now": first_non_empty([pricing.get("buy_now"), vehicle_summary.get("buy_now")]),
"current_bid": first_non_empty([pricing.get("current_bid"), vehicle_summary.get("current_bid")]),
"actual_cash_value": pricing.get("actual_cash_value"),
"estimated_repair_cost": pricing.get("estimated_repair_cost"),
"seller": seller,
"location": location,
"vehicle_location": vehicle_summary.get("vehicle_location"),
"auction_date": auction.get("auction_date"),
"lane": auction.get("lane"),
"branch": auction.get("branch"),
"sale_status": auction.get("sale_status"),
"source_endpoints": payload_insights.get("source_endpoints", {}),
}
content_hash = hashlib.sha256(json.dumps({
# Хеш для пропуска записей без изменений.
"brand": brand, "model": model, "year": year, "price": price, "mileage": mileage,
"color": color, "drive": drive, "gearbox": gearbox, "body_type": body_type,
"engine_volume": engine_volume, "is_damaged": is_damaged, "is_sold": is_sold,
"country": country, "selling_type": "AUCTION", "one_owner": one_owner,
"new_car": new_car, "evaluation": evaluation, "non_smoking": non_smoking,
"rental": rental, "repair_history": repair_history,
"images": [image.fullres_image for image in images_records],
}, sort_keys=True, default=str).encode()).hexdigest()
return CarRecord(
parser_id=parser_id, brand=brand, model=model, year=year, price=price, currency=currency,
mileage=mileage, country=country, is_sold=is_sold, color=color, drive=drive, gearbox=gearbox,
steering_wheel=steering, body_type=body_type, engine_volume=engine_volume,
selling_type=self._normalize_selling_type("AUCTION"), one_owner=one_owner, new_car=new_car,
is_hidden=False, origin=origin, origin_url=vehicle_url, origin_id=origin_id, is_damaged=is_damaged,
evaluation=evaluation, non_smoking=non_smoking, rental=rental, repair_history=repair_history,
slug=slug, last_seen_at=datetime.now(timezone.utc), content_hash=content_hash, images=images_records,
raw_attributes=raw_attributes, mapping_notes=notes,
)
@staticmethod
def _as_str(value: Any) -> str:
return "" if value is None else str(value).strip()
@staticmethod
def _to_int(value: Any) -> int | None:
if value is None:
return None
if isinstance(value, bool):
return int(value)
if isinstance(value, (int, float)):
return int(value)
digits = re.sub(r"[^\d]", "", str(value))
return int(digits) if digits else None
@classmethod
def _to_money_int(cls, value: Any) -> int | None:
# Нормализация стоимости из разных форматов.
if value is None:
return None
if isinstance(value, bool):
return None
if isinstance(value, (int, float)):
return int(value)
text = str(value).strip()
if not text:
return None
lowered = text.lower()
if any(token in lowered for token in ["n/a", "na", "tbd", "unknown", "call", "contact"]):
return None
numbers = re.findall(r"\d[\d\s.,]*", text)
if not numbers:
return None
best: int | None = None
for number in numbers:
clean = number.replace(" ", "")
if "," in clean and "." in clean:
# Поддержка 1,234.56 и 1.234,56.
if clean.rfind(",") > clean.rfind("."):
clean = clean.replace(".", "").replace(",", ".")
else:
clean = clean.replace(",", "")
elif "," in clean:
parts = clean.split(",")
# 123,45 -> 123.45, иначе разделитель тысяч.
if len(parts[-1]) in {1, 2} and len(parts) == 2:
clean = clean.replace(",", ".")
else:
clean = clean.replace(",", "")
elif "." in clean:
parts = clean.split(".")
if not (len(parts[-1]) in {1, 2} and len(parts) == 2):
clean = clean.replace(".", "")
try:
parsed = int(float(clean))
except ValueError:
continue
if parsed > 0 and (best is None or parsed > best):
best = parsed
return best
@staticmethod
def _first_parsed_int(values: list[Any], parser) -> int | None:
for value in values:
parsed = parser(value)
if parsed is not None:
return parsed
return None
@staticmethod
def _to_year(value: Any) -> int | None:
parsed = CarMapper._to_int(value)
if parsed is None:
return None
if 1900 <= parsed <= 2100:
return parsed
return None
def _to_engine_cc(self, value: Any) -> int | None:
# Поддержка литров и cc.
text = str(value).lower().strip() if value is not None else ""
if not text:
return None
if liters_match := re.search(r"(\d+(?:\.\d+)?)\s*l", text):
return int(float(liters_match.group(1)) * 1000)
if cubic_match := re.search(r"(\d{3,5})\s*(?:cc|cm3|cubic)", text):
return int(cubic_match.group(1))
return int(text) if re.match(r"^\d+$", text) else None
def _normalize_currency(self, value: Any) -> str:
text = self._as_str(value)
upper = text.upper() or "USD"
if any(token in text for token in ["€", "EUR"]):
return "EUR"
if any(token in text for token in ["¥", "JPY"]):
return "JPY"
if any(token in text for token in ["₩", "KRW"]):
return "KRW"
if any(token in text for token in ["£", "GBP"]):
return "GBP"
if any(token in text for token in ["₽", "RUB"]):
return "RUB"
if any(token in text for token in ["AED", "د.إ"]):
return "AED"
if any(token in text for token in ["CA$", "CAD"]):
return "CAD"
text = upper
if text in CURRENCY_ENUM_VALUES:
return text
return "USD" if "$" in str(value) else "USD"
def _normalize_drive(self, value: Any) -> str | None:
return self._map_value(value, self.DRIVE_MAP, DRIVE_ENUM_VALUES, empty_default=None, fallback="NA")
def _normalize_gearbox(self, value: Any) -> str | None:
return self._map_value(value, self.GEARBOX_MAP, GEARBOX_ENUM_VALUES, empty_default=None, fallback="NA")
def _normalize_steering(self, value: Any) -> str | None:
return self._map_value(value, self.STEERING_MAP, STEERING_WHEEL_ENUM_VALUES, empty_default=None, fallback=None)
def _normalize_body_type(self, value: Any) -> str:
return self._map_value(value, self.BODY_MAP, BODY_TYPE_ENUM_VALUES, empty_default="OTHER", fallback="OTHER") or "OTHER"
def _normalize_country(self, value: Any) -> str:
fallback = "US" if "US" in COUNTRY_ENUM_VALUES else "NA"
return self._map_value(value, self.COUNTRY_MAP, COUNTRY_ENUM_VALUES, empty_default="US", fallback=fallback) or fallback
def _normalize_selling_type(self, value: Any) -> str:
text = self._as_str(value) or "AUCTION"
return text if text in SELLING_TYPE_ENUM_VALUES else "AUCTION"
def _map_value(
self,
value: Any,
mapping: dict[str, str],
allowed_values: tuple[str, ...],
*,
empty_default: str | None,
fallback: str | None,
) -> str | None:
# Общий helper для enum-нормализации.
text = self._as_str(value).lower()
if not text:
return empty_default
if (mapped := mapping.get(text)) and mapped in allowed_values:
return mapped
for marker, mapped in mapping.items():
if marker in text and mapped in allowed_values:
return mapped
return fallback
@staticmethod
def _normalize_color(value: Any) -> str:
text = str(value).strip() if value is not None else "other"
if not text:
return "other"
if "/" in text:
text = text.split("/")[0].strip()
return text.lower() or "other"
@staticmethod
def _boolish(value: Any) -> bool:
return value if isinstance(value, bool) else str(value).strip().lower() in {"1", "true", "yes", "y", "owner", "one owner", "new"}
def _bool_damage(self, damage: dict[str, Any], vehicle_summary: dict[str, Any]) -> bool:
for value in [damage.get("primary"), damage.get("secondary"), damage.get("description"), vehicle_summary.get("primary_damage")]:
text = self._as_str(value).lower()
if text and text not in self.NO_DAMAGE_MARKERS:
return True
return False
def _bool_sold(self, auction: dict[str, Any]) -> bool:
return any(token in self._as_str(auction.get("sale_status")).lower() for token in ["sold", "closed", "ended"])
def _build_images(self, urls: list[Any]) -> list[ImageRecord]:
# Для imageKeys берем самый большой размер.
best_by_key: dict[str, str] = {}
key_order: list[str] = []
non_keyed: list[str] = []
for item in urls:
if not isinstance(item, str):
continue
url = item.strip()
if not url:
continue
if m := re.search(r'imageKeys=([^&]+)', url):
img_key = m.group(1)
w = int(re.search(r'width=(\d+)', url).group(1)) if re.search(r'width=(\d+)', url) else 0
existing = best_by_key.get(img_key)
if existing is None:
best_by_key[img_key] = url
key_order.append(img_key)
else:
existing_w = int(re.search(r'width=(\d+)', existing).group(1)) if re.search(r'width=(\d+)', existing) else 0
if w > existing_w:
best_by_key[img_key] = url
elif url not in non_keyed:
non_keyed.append(url)
deduped = [best_by_key[k] for k in key_order] + non_keyed
return [ImageRecord(fullres_image=self._make_fullres_url(url), preview_image=self._make_preview_url(url), order_index=index) for index, url in enumerate(deduped)]
@staticmethod
def _make_fullres_url(url: str) -> str:
if "vis.iaai.com" in url:
return re.sub(r'height=\d+', 'height=633', re.sub(r'width=\d+', 'width=845', url))
return url
@staticmethod
def _make_preview_url(url: str) -> str:
if "vis.iaai.com" in url:
preview = re.sub(r'height=\d+', 'height=300', re.sub(r'width=\d+', 'width=400', url))
if preview != url:
return preview
return url
def _build_origin_id(self, vehicle_url: str, vehicle_summary: dict[str, Any], core: dict[str, Any]) -> str:
# Берем lot_number, затем vin, затем хвост URL.
for value in [core.get("lot_number"), vehicle_summary.get("lot_number"), vehicle_summary.get("vin")]:
text = self._as_str(value)
if text:
return text
tail = urlparse(vehicle_url).path.rstrip("/").split("/")[-1]
if "~" in tail:
tail = tail.split("~")[0]
return tail or self._slugify(vehicle_url)
@staticmethod
def _slugify(value: str) -> str:
return re.sub(r"[^a-z0-9]+", "-", value.lower()).strip("-") or "car"

367
iaai_scraper/parsing/parser.py Normal file
View File

@@ -0,0 +1,367 @@
import html as html_module
import json
import logging
import re
from typing import Any
from ..core.utils import LOT_RE, PRICE_RE, VIN_RE, deep_find_key, first_non_empty
logger = logging.getLogger("iaai_scraper.parsers")
class VehicleParser:
"""DOM + JSON парсер страницы авто."""
SUMMARY_KEY_MAP = {
"vin": {"vin", "vehicleidentificationnumber"},
"lot_number": {"lotnumber", "stockno", "itemid", "lotid", "itemnumber"},
"year": {"year"},
"make": {"make", "manufacturer", "brand"},
"model": {"model"},
"trim": {"trim", "series"},
"odometer": {"odometer", "odometermiles", "mileage", "actualcashvalueodometer"},
"primary_damage": {"primarydamage", "damage", "damagetype", "loss"},
"secondary_damage": {"secondarydamage"},
"run_and_drive": {"runanddrive", "canrunanddrive", "rundrive"},
"buy_now": {"buynowprice", "buyitnowprice", "instantpurchaseprice"},
"current_bid": {"currentbid", "highbid", "bidamount", "currenthighbid"},
"actual_cash_value": {"actualcashvalue", "acv"},
"estimated_repair_cost": {"estimatedrepaircost", "repaircost"},
"keys": {"keys", "keystatus"},
"title": {"titletype", "title", "documenttype"},
"seller": {"seller", "sellername"},
"location": {"location", "branchname", "auctionlocation", "branch"},
"auction_date": {"auctiondate", "saledate", "liveauctiondate"},
"body_type": {"bodytype", "bodystyle"},
"drive": {"driveline", "drive"},
"gearbox": {"transmission", "gearbox"},
"engine": {"engine", "enginevolume"},
"fuel_type": {"fueltype", "fuel"},
"cylinders": {"cylinders", "cylindercount"},
"color": {"color", "primarycolor", "exteriorcolor"},
}
DOM_LABEL_MAP: dict[str, str] = {
"stock #": "lot_number",
"vin (status)": "vin",
"vin": "vin",
"primary damage": "primary_damage",
"secondary damage": "secondary_damage",
"odometer": "odometer",
"body style": "body_type",
"engine": "engine",
"transmission": "gearbox",
"drive line type": "drive",
"fuel type": "fuel_type",
"cylinders": "cylinders",
"exterior/interior": "color",
"exterior color": "color",
"model": "model",
"series": "trim",
"selling branch": "location",
"vehicle location": "vehicle_location",
"auction date and time": "auction_date",
"lane/run #": "lane",
"actual cash value": "actual_cash_value",
"estimated repair cost": "estimated_repair_cost",
"seller": "seller",
"title/sale doc": "title",
"title/sale doc brand": "title_brand",
"start code": "run_and_drive",
"key": "keys",
"manufactured in": "manufactured_in",
"vehicle class": "vehicle_class",
}
def _parse_dom_key_value_pairs(self, dom_text: str) -> dict[str, str]:
result: dict[str, str] = {}
if not dom_text:
return result
lines = [line.strip() for line in dom_text.split("\n") if line.strip()]
known_labels = set(self.DOM_LABEL_MAP.keys())
for i, line in enumerate(lines):
clean = line.rstrip(":").lower().strip()
if clean in known_labels and i + 1 < len(lines):
value = lines[i + 1].strip()
if value.rstrip(":").lower().strip() in known_labels:
continue
if value.lower() in ("more actions", "view", "print", "share", "back to results"):
continue
field_name = self.DOM_LABEL_MAP[clean]
if field_name not in result or not result[field_name]:
result[field_name] = value
return result
def _parse_title_for_year_make_model(self, page_title: str, dom_text: str) -> dict[str, str | None]:
result: dict[str, str | None] = {"year": None, "make": None, "model": None}
title_match = re.match(r"(\d{4})\s+(\S+)\s+(.+?)(?:\s+for\s+)", page_title or "")
if title_match:
result["year"] = title_match.group(1)
result["make"] = title_match.group(2)
result["model"] = title_match.group(3)
return result
dom_match = re.search(r"(?:Search|Log In)\s*\n\s*(\d{4})\s+(\S+)\s+(.+?)(?:\n|$)", dom_text or "")
if dom_match:
result["year"] = dom_match.group(1)
result["make"] = dom_match.group(2)
result["model"] = dom_match.group(3).strip()
return result
def normalize(self, vehicle_url: str, page_html: str, dom_text: str, network_dump: dict[str, Any]) -> dict[str, Any]:
page_html = page_html or ""
dom_text = dom_text or ""
network_dump = network_dump or {}
responses = network_dump.get("json_responses", [])
payloads = [item.get("payload") for item in responses if isinstance(item.get("payload"), (dict, list))]
dom_kv = self._parse_dom_key_value_pairs(dom_text)
page_title = ""
title_match = re.search(r"<title[^>]*>(.*?)</title>", page_html or "", re.IGNORECASE | re.DOTALL)
if title_match:
page_title = title_match.group(1).strip()
title_parsed = self._parse_title_for_year_make_model(page_title, dom_text)
summary: dict[str, Any] = {"source_url": vehicle_url}
for field, candidate_keys in self.SUMMARY_KEY_MAP.items():
values: list[Any] = []
for payload in payloads:
values.extend(deep_find_key(payload, candidate_keys))
if field in dom_kv:
values.append(dom_kv[field])
summary[field] = first_non_empty(values)
summary["year"] = summary.get("year") or title_parsed.get("year")
summary["make"] = summary.get("make") or title_parsed.get("make")
summary["model"] = summary.get("model") or title_parsed.get("model")
summary["trim"] = summary.get("trim") or dom_kv.get("trim")
summary["vin"] = summary.get("vin") or self._extract_vin(dom_text) or self._extract_vin(page_html)
if summary.get("vin") and isinstance(summary["vin"], str):
vin_clean = re.sub(r"\s*\(.*?\)\s*$", "", summary["vin"]).strip()
vin_match = VIN_RE.search(vin_clean)
summary["vin"] = vin_match.group(1) if vin_match else vin_clean
summary["lot_number"] = summary.get("lot_number") or self._extract_lot_number(dom_text)
summary["image_urls"] = self._extract_image_urls(payloads, page_html, vehicle_url)
for dom_field, dom_value in dom_kv.items():
if dom_field not in summary or not summary[dom_field]:
summary[dom_field] = dom_value
prices = self._extract_prices_from_text(dom_text)
if not summary.get("actual_cash_value") and prices:
summary["actual_cash_value"] = prices[0]
if not summary.get("buy_now"):
buy_now_match = re.search(r"Buy\s+Now[:\s]*\$\s*([\d,]+(?:\.\d{1,2})?)", dom_text or "", re.IGNORECASE)
if buy_now_match:
summary["buy_now"] = buy_now_match.group(1)
elif prices:
summary["buy_now"] = prices[0]
if not summary.get("current_bid") and len(prices) > 1:
summary["current_bid"] = prices[1]
embedded = self._extract_embedded_json(page_html)
for item in embedded:
p = item.get("payload")
if isinstance(p, (dict, list)):
payloads.append(p)
for field, candidate_keys in self.SUMMARY_KEY_MAP.items():
if not summary.get(field):
val = first_non_empty(deep_find_key(p, candidate_keys))
if val:
summary[field] = val
return {
"vehicle_summary": summary,
"payload_insights": self._build_payload_insights(summary, responses, payloads, vehicle_url),
"embedded_json": embedded,
"dom_hints": self._dom_hints(dom_text),
"access_notes": self._build_access_notes(summary, responses),
}
def _build_payload_insights(self, summary: dict[str, Any], responses: list[dict[str, Any]], payloads: list[Any], vehicle_url: str = "") -> dict[str, Any]:
image_urls = self._extract_image_urls(payloads, "", vehicle_url)
return {
"vehicle_core": {
"vin": summary.get("vin"), "lot_number": summary.get("lot_number"), "year": summary.get("year"),
"make": summary.get("make"), "model": summary.get("model"), "trim": summary.get("trim"),
"odometer": summary.get("odometer"), "run_and_drive": summary.get("run_and_drive"),
"seller": summary.get("seller"), "location": summary.get("location"), "title": summary.get("title"),
"body_type": summary.get("body_type"), "drive": summary.get("drive"), "gearbox": summary.get("gearbox"),
"engine": summary.get("engine"), "fuel_type": summary.get("fuel_type"), "cylinders": summary.get("cylinders"),
"color": summary.get("color"), "keys": summary.get("keys"),
},
"pricing": {
"buy_now": summary.get("buy_now"), "current_bid": summary.get("current_bid"),
"actual_cash_value": summary.get("actual_cash_value"), "estimated_repair_cost": summary.get("estimated_repair_cost"),
"currency": self._guess_currency(summary),
},
"bids": self._build_bid_insights(summary, payloads),
"damage": {
"primary": summary.get("primary_damage"),
"secondary": summary.get("secondary_damage"),
"description": first_non_empty(self._find_in_payloads(payloads, {"damageDescription", "damageDetails"})),
},
"auction": {
"auction_date": summary.get("auction_date"),
"lane": first_non_empty(self._find_in_payloads(payloads, {"lane", "lanename"})),
"branch": first_non_empty([summary.get("location"), *self._find_in_payloads(payloads, {"branch", "branchname"})]),
"sale_status": first_non_empty(self._find_in_payloads(payloads, {"salestatus", "auctionstatus", "status"})),
"item_number": first_non_empty([summary.get("lot_number"), *self._find_in_payloads(payloads, {"itemnumber", "lotnumber", "lotid"})]),
},
"images": {"count": len(image_urls), "urls": image_urls},
"source_endpoints": self._build_source_endpoints(responses),
}
def _build_bid_insights(self, summary: dict[str, Any], payloads: list[Any]) -> dict[str, Any]:
return {
"amount": summary.get("current_bid"),
"currency": self._guess_currency(summary),
"bid_count": first_non_empty(self._find_in_payloads(payloads, {"bidcount", "numberofbids"})),
"status": first_non_empty(self._find_in_payloads(payloads, {"bidstatus", "biddingstatus"})),
}
def _build_source_endpoints(self, responses: list[dict[str, Any]]) -> dict[str, list[str]]:
mapping = {"vehicle": [], "pricing": [], "bids": [], "damage": [], "auction": [], "images": []}
for item in responses:
url = item.get("url", "")
category = item.get("category", "other")
if category == "vehicle":
mapping["vehicle"].append(url)
lowered = url.lower()
if any(token in lowered for token in ["bid", "offer"]):
mapping["bids"].append(url)
if any(token in lowered for token in ["damage", "report"]):
mapping["damage"].append(url)
if any(token in lowered for token in ["auction", "sale", "lane", "branch"]):
mapping["auction"].append(url)
elif category in mapping:
mapping[category].append(url)
return {key: list(dict.fromkeys(urls)) for key, urls in mapping.items()}
def _build_access_notes(self, summary: dict[str, Any], responses: list[dict[str, Any]]) -> dict[str, Any]:
endpoints = [item.get("url", "") for item in responses]
return {
"vin_visible": bool(summary.get("vin")),
"images_visible": bool(summary.get("image_urls")),
"network_json_count": len(responses),
"possible_captcha": False,
"possible_antibot": False,
"observed_endpoints": endpoints[:20],
}
@staticmethod
def _find_in_payloads(payloads: list[Any], keys: set[str]) -> list[Any]:
lowered = {key.lower() for key in keys}
values: list[Any] = []
for payload in payloads:
values.extend(deep_find_key(payload, lowered))
return values
@staticmethod
def _guess_currency(summary: dict[str, Any]) -> str:
for key in ["buy_now", "current_bid", "actual_cash_value", "estimated_repair_cost"]:
value = str(summary.get(key) or "")
if "$" in value:
return "USD"
if "€" in value:
return "EUR"
if "¥" in value:
return "JPY"
return "USD"
@staticmethod
def _extract_vin(text: str) -> str | None:
match = VIN_RE.search(text or "")
return match.group(1) if match else None
@staticmethod
def _extract_lot_number(text: str) -> str | None:
match = LOT_RE.search(text or "")
return match.group(1) if match else None
@staticmethod
def _extract_prices_from_text(text: str) -> list[str]:
return [match.group(1) for match in PRICE_RE.finditer(text or "")]
@staticmethod
def _extract_embedded_json(html: str) -> list[dict[str, Any]]:
scripts = re.findall(r"<script[^>]*>(.*?)</script>", html or "", flags=re.DOTALL | re.IGNORECASE)
extracted: list[dict[str, Any]] = []
for script_text in scripts:
if "{" not in script_text and "[" not in script_text:
continue
try:
parsed = json.loads(script_text.strip())
except Exception:
continue
extracted.append({"type": "inline_json", "payload": parsed})
return extracted
@staticmethod
def _extract_image_urls(payloads: list[Any], html: str, vehicle_url: str = "") -> list[str]:
vehicle_key = ""
key_match = re.search(r"VehicleDetail/(\d+)", vehicle_url or "")
if key_match:
vehicle_key = key_match.group(1)
found: list[str] = []
for payload in payloads:
found.extend(deep_find_key(payload, {"imageurl", "imageurls", "url", "fullsizeurl", "thumbnailurl", "originalurl"}))
flat: list[str] = []
seen_flat: set[str] = set()
for item in found:
if isinstance(item, str) and item.startswith("http"):
cleaned = html_module.unescape(item)
lowered = cleaned.lower()
if vehicle_key and "vis.iaai.com" in lowered and vehicle_key not in cleaned:
continue
if cleaned not in seen_flat:
seen_flat.add(cleaned)
flat.append(cleaned)
elif isinstance(item, list):
for child in item:
if isinstance(child, str) and child.startswith("http"):
cleaned = html_module.unescape(child)
lowered = cleaned.lower()
if vehicle_key and "vis.iaai.com" in lowered and vehicle_key not in cleaned:
continue
if cleaned not in seen_flat:
seen_flat.add(cleaned)
flat.append(cleaned)
for pattern in [r'<img[^>]+(?:src|data-src)\s*=\s*["\']([^"\']+)["\']', r'data-src\s*=\s*["\']([^"\']+)["\']']:
for match in re.finditer(pattern, html or "", re.IGNORECASE):
url = html_module.unescape(match.group(1).strip())
if not url.startswith("http") or url in seen_flat:
continue
lowered = url.lower()
if vehicle_key and vehicle_key in url:
seen_flat.add(url)
flat.append(url)
elif any(token in lowered for token in ["vis.iaai.com", "anvis", "vehicleimage"]):
if vehicle_key and vehicle_key not in url:
continue
seen_flat.add(url)
flat.append(url)
if vehicle_key:
for url in re.findall(r'https?://vis\.iaai\.com[^\s"\'<>]+', html or ""):
cleaned = html_module.unescape(url)
if cleaned not in seen_flat and vehicle_key in cleaned:
seen_flat.add(cleaned)
flat.append(cleaned)
filtered: list[str] = []
for url in flat:
lowered = url.lower()
if any(pat in lowered for pat in {"dimensions", "threesixty", "360view", ".js", ".css", ".svg", "/home/", "iframeview"}):
continue
if "vis.iaai.com" in lowered and "/resizer" not in lowered:
continue
filtered.append(url)
return filtered
@staticmethod
def _dom_hints(text: str) -> dict[str, Any]:
lowered = (text or "").lower()
return {
"has_buy_now_text": "buy now" in lowered,
"has_run_and_drive_text": "run & drive" in lowered or "run and drive" in lowered,
"has_damage_text": "damage" in lowered,
"has_vin_text": "vin" in lowered,
"has_title_text": "title" in lowered,
"has_captcha_text": any(token in lowered for token in ["captcha", "verify you are human", "i am human", "recaptcha", "cloudflare"]),
"has_antibot_text": any(token in lowered for token in ["incapsula", "access denied", "request unsuccessful", "bot detection"]),
}

317
iaai_scraper/proxy_bridge.py Normal file
View File

@@ -0,0 +1,317 @@
from __future__ import annotations
import logging
import os
import select
import socket
import socketserver
import struct
import threading
from urllib.parse import urlsplit
BUFFER_SIZE = 65536
CRLF = b"\r\n"
DEFAULT_LISTEN_HOST = os.getenv("PROXY_BRIDGE_HOST", "0.0.0.0")
DEFAULT_LISTEN_PORT = int(os.getenv("PROXY_BRIDGE_PORT", "8899"))
SOCKS5_HOST = os.getenv("SOCKS5_PROXY_HOST", "")
SOCKS5_PORT = int(os.getenv("SOCKS5_PROXY_PORT", "1002"))
SOCKS5_USER = os.getenv("SOCKS5_PROXY_USER", "")
SOCKS5_PASS = os.getenv("SOCKS5_PROXY_PASS", "")
RELAY_IDLE_TIMEOUT_SECONDS = int(os.getenv("PROXY_BRIDGE_RELAY_IDLE_TIMEOUT_SECONDS", "60"))
MAX_WORKERS = int(os.getenv("PROXY_BRIDGE_MAX_WORKERS", "64"))
logger = logging.getLogger("proxy_bridge")
class ThreadingTCPServer(socketserver.ThreadingMixIn, socketserver.TCPServer):
allow_reuse_address = True
daemon_threads = True
def __init__(self, server_address, request_handler_class):
super().__init__(server_address, request_handler_class)
self._worker_semaphore = threading.BoundedSemaphore(MAX_WORKERS)
def process_request_thread(self, request, client_address):
with self._worker_semaphore:
super().process_request_thread(request, client_address)
def _recv_exact(sock: socket.socket, size: int) -> bytes:
data = b""
while len(data) < size:
chunk = sock.recv(size - len(data))
if not chunk:
raise ConnectionError("Unexpected EOF from SOCKS5 server")
data += chunk
return data
def _socks5_connect(host: str, port: int) -> socket.socket:
if not SOCKS5_HOST:
raise RuntimeError("SOCKS5_PROXY_HOST is not configured")
upstream = socket.create_connection((SOCKS5_HOST, SOCKS5_PORT), timeout=30)
upstream.settimeout(30)
methods = [0x00]
if SOCKS5_USER or SOCKS5_PASS:
methods = [0x02]
upstream.sendall(bytes([0x05, len(methods), *methods]))
version, method = _recv_exact(upstream, 2)
if version != 0x05 or method == 0xFF:
upstream.close()
raise ConnectionError("SOCKS5 authentication negotiation failed")
if method == 0x02:
username = SOCKS5_USER.encode("utf-8")
password = SOCKS5_PASS.encode("utf-8")
if len(username) > 255 or len(password) > 255:
upstream.close()
raise ValueError("SOCKS5 username/password too long")
upstream.sendall(bytes([0x01, len(username)]) + username + bytes([len(password)]) + password)
auth_version, auth_status = _recv_exact(upstream, 2)
if auth_version != 0x01 or auth_status != 0x00:
upstream.close()
raise ConnectionError("SOCKS5 username/password authentication failed")
try:
socket.inet_aton(host)
addr_type = 0x01
addr_payload = socket.inet_aton(host)
except OSError:
host_bytes = host.encode("idna")
if len(host_bytes) > 255:
upstream.close()
raise ValueError("Target host is too long for SOCKS5 domain format")
addr_type = 0x03
addr_payload = bytes([len(host_bytes)]) + host_bytes
request = bytes([0x05, 0x01, 0x00, addr_type]) + addr_payload + struct.pack("!H", port)
upstream.sendall(request)
response_head = _recv_exact(upstream, 4)
version, reply, _reserved, reply_addr_type = response_head
if version != 0x05 or reply != 0x00:
upstream.close()
raise ConnectionError(f"SOCKS5 connect failed with code {reply}")
if reply_addr_type == 0x01:
_recv_exact(upstream, 4)
elif reply_addr_type == 0x03:
domain_len = _recv_exact(upstream, 1)[0]
_recv_exact(upstream, domain_len)
elif reply_addr_type == 0x04:
_recv_exact(upstream, 16)
_recv_exact(upstream, 2)
upstream.settimeout(RELAY_IDLE_TIMEOUT_SECONDS)
return upstream
def _relay_bidirectional(left: socket.socket, right: socket.socket) -> None:
sockets = [left, right]
left.settimeout(RELAY_IDLE_TIMEOUT_SECONDS)
right.settimeout(RELAY_IDLE_TIMEOUT_SECONDS)
try:
while True:
readable, _, exceptional = select.select(sockets, [], sockets, RELAY_IDLE_TIMEOUT_SECONDS)
if exceptional:
break
if not readable:
logger.debug("Relay idle timeout reached; closing sockets")
return
for current in readable:
other = right if current is left else left
data = current.recv(BUFFER_SIZE)
if not data:
return
other.sendall(data)
finally:
for sock in sockets:
try:
sock.shutdown(socket.SHUT_RDWR)
except OSError:
pass
try:
sock.close()
except OSError:
pass
class ProxyHandler(socketserver.StreamRequestHandler):
def handle(self) -> None:
try:
request_line = self.rfile.readline(BUFFER_SIZE).decode("iso-8859-1").strip()
if not request_line:
return
method, target, version = request_line.split()
headers = self._read_headers()
logger.info("%s %s", method, target)
if method.upper() == "CONNECT":
host, port = self._parse_connect_target(target)
logger.info("CONNECT %s:%s", host, port)
upstream = _socks5_connect(host, port)
self.wfile.write(f"{version} 200 Connection Established".encode("ascii") + CRLF + CRLF)
self.wfile.flush()
_relay_bidirectional(self.connection, upstream)
return
host, port, path = self._parse_forward_target(target, headers)
upstream = _socks5_connect(host, port)
self._send_forward_request(upstream, method, path, version, headers)
body = self._read_request_body(headers)
if body:
upstream.sendall(body)
_relay_bidirectional(self.connection, upstream)
except Exception as exc:
logger.exception("Proxy bridge request failed: %s", exc)
try:
self.wfile.write(
b"HTTP/1.1 502 Bad Gateway" + CRLF
+ b"Connection: close" + CRLF
+ b"Content-Type: text/plain; charset=utf-8" + CRLF + CRLF
+ b"Bad Gateway"
)
self.wfile.flush()
except OSError:
pass
def _read_headers(self) -> list[tuple[str, str]]:
headers: list[tuple[str, str]] = []
while True:
line = self.rfile.readline(BUFFER_SIZE)
if line in {CRLF, b"\n", b""}:
break
decoded = line.decode("iso-8859-1")
if ":" not in decoded:
continue
name, value = decoded.split(":", 1)
headers.append((name.strip(), value.strip()))
return headers
@staticmethod
def _parse_connect_target(target: str) -> tuple[str, int]:
if target.startswith("["):
end = target.find("]")
if end == -1 or len(target) <= end + 2 or target[end + 1] != ":":
raise ValueError("Invalid CONNECT target")
host = target[1:end]
port_text = target[end + 2 :]
return host, int(port_text)
host, port_text = target.rsplit(":", 1)
return host, int(port_text)
@staticmethod
def _parse_forward_target(target: str, headers: list[tuple[str, str]]) -> tuple[str, int, str]:
if target.startswith("http://"):
parts = urlsplit(target)
port = parts.port or 80
path = parts.path or "/"
if parts.query:
path += f"?{parts.query}"
return parts.hostname or "", port, path
if target.startswith("https://"):
raise ValueError("HTTPS absolute-form request must use CONNECT")
host_header = next((value for name, value in headers if name.lower() == "host"), "")
if not host_header:
raise ValueError("Missing Host header")
if ":" in host_header:
host, port_text = host_header.rsplit(":", 1)
return host, int(port_text), target
return host_header, 80, target
def _send_forward_request(
self,
upstream: socket.socket,
method: str,
path: str,
version: str,
headers: list[tuple[str, str]],
) -> None:
filtered_headers: list[tuple[str, str]] = []
hop_by_hop = {
"proxy-connection",
"proxy-authorization",
"connection",
"keep-alive",
"te",
"trailer",
"transfer-encoding",
"upgrade",
}
for name, value in headers:
if name.lower() in hop_by_hop:
continue
filtered_headers.append((name, value))
request_head = [f"{method} {path} {version}\r\n"]
request_head.extend(f"{name}: {value}\r\n" for name, value in filtered_headers)
request_head.append("\r\n")
upstream.sendall("".join(request_head).encode("iso-8859-1"))
def _read_request_body(self, headers: list[tuple[str, str]]) -> bytes:
transfer_encoding = next((value for name, value in headers if name.lower() == "transfer-encoding"), "")
if "chunked" in transfer_encoding.lower():
return self._read_chunked_request_body()
content_length = next((value for name, value in headers if name.lower() == "content-length"), None)
if not content_length:
return b""
return self.rfile.read(int(content_length))
def _read_chunked_request_body(self) -> bytes:
chunks: list[bytes] = []
while True:
size_line = self.rfile.readline(BUFFER_SIZE)
if not size_line:
raise ConnectionError("Unexpected EOF in chunked request")
size_text = size_line.strip().split(b";", 1)[0]
chunk_size = int(size_text, 16)
chunks.append(size_line)
if chunk_size == 0:
while True:
trailer_line = self.rfile.readline(BUFFER_SIZE)
if not trailer_line:
raise ConnectionError("Unexpected EOF in chunked trailers")
chunks.append(trailer_line)
if trailer_line in {CRLF, b"\n"}:
return b"".join(chunks)
chunk_data = self.rfile.read(chunk_size)
if len(chunk_data) != chunk_size:
raise ConnectionError("Unexpected EOF in chunk body")
chunks.append(chunk_data)
chunk_end = self.rfile.read(2)
if chunk_end != CRLF:
raise ConnectionError("Invalid chunk terminator")
chunks.append(chunk_end)
def main() -> None:
if not SOCKS5_HOST:
raise SystemExit("SOCKS5_PROXY_HOST is required")
logging.basicConfig(
level=os.getenv("PROXY_BRIDGE_LOG_LEVEL", "INFO").upper(),
format="[%(asctime)s] [proxy_bridge] %(levelname)s: %(message)s",
)
with ThreadingTCPServer((DEFAULT_LISTEN_HOST, DEFAULT_LISTEN_PORT), ProxyHandler) as server:
logger.info(
"Listening on %s:%s -> socks5://%s:%s (max_workers=%s)",
DEFAULT_LISTEN_HOST,
DEFAULT_LISTEN_PORT,
SOCKS5_HOST,
SOCKS5_PORT,
MAX_WORKERS,
)
server.serve_forever()
if __name__ == "__main__":
main()

460
iaai_scraper/scraper.py Normal file
View File

@@ -0,0 +1,460 @@
import logging
import re
import signal
import time
import uuid
from playwright.sync_api import Error as PlaywrightError
from playwright.sync_api import BrowserContext, Page, sync_playwright
from playwright.sync_api import TimeoutError as PlaywrightTimeoutError
from .browser import BrowserFactory, HumanPacer, NetworkCapture
from .core.config import Settings, settings
from .core.logs import set_trace_id, setup_logging
from .core.retry import retryable
from .core.utils import save_to_json
from .parsing.mapper import CarMapper
from .parsing.parser import VehicleParser
from .storage.db import PersistenceService
from .browser.listing import ListingCollector
from .storage.schemas import CarRecord
logger = logging.getLogger("iaai_scraper.scraper")
VEHICLE_ID_RE = re.compile(r"/VehicleDetail/(\d+)(?:~[A-Z]{2})?", re.IGNORECASE)
class IAAIScraper:
@staticmethod
def _extract_origin_id_from_url(vehicle_url: str) -> str | None:
match = VEHICLE_ID_RE.search(vehicle_url)
if not match:
return None
return match.group(1)
def __init__(self, runtime_settings: Settings | None = None) -> None:
# Базовые зависимости и сервисы скрапера
self.settings = runtime_settings or settings
setup_logging(self.settings.log_level, self.settings.log_file)
self.trace_id = uuid.uuid4().hex[:12]
set_trace_id(self.trace_id)
self.playwright = None
self.browser = None
self.context: BrowserContext | None = None
self.browser_factory = BrowserFactory(self.settings)
self.pacer = HumanPacer(self.settings)
self.listing_collector = ListingCollector(self.settings, self.pacer)
self.vehicle_parser = VehicleParser()
self.car_mapper = CarMapper()
self.persistence = PersistenceService(self.settings)
self._shutdown_requested = False
# Жизненный цикл браузера
def _new_trace_id(self, prefix: str) -> str:
# Новый trace id для отдельной операции
trace_id = f"{prefix}-{uuid.uuid4().hex[:8]}"
self.trace_id = trace_id
set_trace_id(trace_id)
return trace_id
def __enter__(self) -> "IAAIScraper":
if self.playwright is None:
self.playwright = sync_playwright().start()
if self.browser is None:
self.browser = self.browser_factory.create_browser(self.playwright)
return self
def __exit__(self, exc_type, exc, tb) -> None:
self.close()
def close(self) -> None:
# Закрываем ресурсы аккуратно даже при ошибках Playwright
if self.context is not None:
try:
self.context.close()
except PlaywrightError:
pass
finally:
self.context = None
if self.browser is not None:
try:
self.browser.close()
except PlaywrightError:
pass
finally:
self.browser = None
if self.playwright is not None:
try:
self.playwright.stop()
except PlaywrightError:
pass
finally:
self.playwright = None
def _new_context(self, storage_state: str | None = None) -> BrowserContext:
# Контекст пересоздаётся, чтобы не тянуть старое состояние страницы
if self.browser is None:
self.__enter__()
if self.context:
self.context.close()
self.context = self.browser_factory.create_context(self.browser, storage_state=storage_state)
return self.context
def init_db(self):
self.persistence.create_tables()
return {"status": "ok", "database_url": self.settings.database.url}
def _get_unauthenticated_page(self) -> Page:
context = self._new_context()
return context.new_page()
# Сбор листинга
def collect_listing(self, make: str | None = None, model: str | None = None):
# Собираем ссылки карточек с публичного листинга
page = self._get_unauthenticated_page()
try:
listing = self.listing_collector.collect_listing_links(page, make=make, model=model)
finally:
page.close()
return {
"status": "ok",
**listing,
}
def _get_page(self) -> Page:
if not self.context:
self._new_context()
return self.context.new_page()
# Открытие и скрейп карточки
@retryable(max_attempts=3, jitter_seconds=0.25)
def open_vehicle_page(self, vehicle_url: str):
# Лёгкое открытие страницы без сетевого дампа и сохранения в БД
trace_id = self._new_trace_id("open")
started_at = time.perf_counter()
page = self._get_page()
try:
self.pacer.before_vehicle_open()
page.goto(vehicle_url, wait_until="domcontentloaded", timeout=60_000)
try:
page.wait_for_load_state("networkidle", timeout=15000)
except PlaywrightTimeoutError:
pass
self.pacer.after_vehicle_open()
html = page.content()
try:
dom_text = page.locator("body").inner_text(timeout=10_000)
except Exception:
dom_text = ""
parsed = self.vehicle_parser.normalize(vehicle_url, html, dom_text, {"json_responses": []})
return {
"trace_id": trace_id,
"source_url": vehicle_url,
"opened_in_light_mode": True,
"fetched_at_epoch": int(time.time()),
"elapsed_seconds": round(time.perf_counter() - started_at, 3),
**parsed,
}
finally:
page.close()
@retryable(max_attempts=3, jitter_seconds=0.25)
def scrape_vehicle_detail(self, vehicle_url: str):
"""Открыть страницу, перехватить JSON, вернуть данные."""
page = self._get_page()
try:
return self._scrape_on_page(page, vehicle_url)
finally:
page.close()
def _scrape_on_page(self, page: Page, vehicle_url: str):
# Полный проход по карточке с network capture и маппингом в DB-модель
trace_id = self._new_trace_id("scrape")
started_at = time.perf_counter()
capture = NetworkCapture(self.settings)
capture.attach(page, origin_url=vehicle_url)
page.goto(vehicle_url, wait_until="domcontentloaded", timeout=60_000)
try:
page.wait_for_load_state("networkidle", timeout=10000)
except PlaywrightTimeoutError:
pass
time.sleep(self.settings.network_settle_ms / 1000)
html = page.content()
try:
dom_text = page.locator("body").inner_text(timeout=10_000)
except Exception:
dom_text = ""
network_dump = capture.export()
parsed = self.vehicle_parser.normalize(vehicle_url, html, dom_text, network_dump)
db_record = self.car_mapper.map_to_car_record(
vehicle_url=vehicle_url,
vehicle_summary=parsed.get("vehicle_summary", {}),
payload_insights=parsed.get("payload_insights", {}),
)
result = {
"trace_id": trace_id,
"source_url": vehicle_url,
"fetched_at_epoch": int(time.time()),
"elapsed_seconds": round(time.perf_counter() - started_at, 3),
"network": network_dump,
**parsed,
"db_record": db_record.model_dump(mode="json"),
}
if self.settings.raw_output_json:
save_to_json(network_dump, self.settings.raw_output_json)
return result
# Синхронизация с БД
def sync_vehicle(self, vehicle_url: str, lane: str = "iaai"):
"""Scrape + upsert одного авто."""
# Один URL -> один scrape -> один upsert
trace_id = self._new_trace_id("sync-vehicle")
started_at = time.perf_counter()
self.persistence.create_tables()
run_id = self.persistence.start_sync_run(lane=lane)
ids_fetched = 1
cars_upserted = 0
cars_failed = 0
images_upserted = 0
status = "failed"
error_summary = None
try:
scrape_result = self.scrape_vehicle_detail(vehicle_url)
db_record = scrape_result.get("db_record")
if not db_record:
raise RuntimeError("Scrape result does not contain db_record")
record = CarRecord.model_validate(db_record)
upsert = self.persistence.upsert_car(record)
cars_upserted = 1
images_upserted = int(upsert.get("images_upserted", 0))
status = "success"
return {
"trace_id": trace_id,
"status": status,
"run_id": run_id,
"vehicle_url": vehicle_url,
"db_action": upsert.get("action"),
"images_upserted": images_upserted,
"elapsed_seconds": round(time.perf_counter() - started_at, 3),
"db_record": db_record,
}
except Exception as exc:
cars_failed = 1
status = "failed"
error_summary = str(exc)
raise
finally:
self.persistence.finish_sync_run(
run_id,
status=status,
ids_fetched=ids_fetched,
cars_upserted=cars_upserted,
cars_failed=cars_failed,
images_upserted=images_upserted,
error_summary=error_summary,
)
def sync_listing(
self,
make: str | None = None,
model: str | None = None,
lane: str = "iaai_cars",
limit: int | None = None,
only_new: bool | None = None,
):
"""Листинг + sync всех найденных машин."""
# Массовая синхронизация с общим run_id и сбором ошибок
trace_id = self._new_trace_id("sync-listing")
started_at = time.perf_counter()
self.persistence.create_tables()
run_id = self.persistence.start_sync_run(lane=lane)
cars_upserted = 0
cars_failed = 0
images_upserted = 0
total = 0
skipped_existing = 0
failures: list[dict[str, str]] = []
listing: dict = {}
try:
listing = self.collect_listing(make=make, model=model)
vehicle_urls = list(listing.get("vehicle_urls", []))
if limit is not None:
vehicle_urls = vehicle_urls[:max(0, limit)]
effective_only_new = self.settings.sync_only_new if only_new is None else only_new
if effective_only_new:
existing_urls = self.persistence.get_existing_origin_urls(vehicle_urls)
url_to_origin_id = {
url: self._extract_origin_id_from_url(url)
for url in vehicle_urls
}
candidate_origin_ids = [origin_id for origin_id in url_to_origin_id.values() if origin_id]
existing_ids = self.persistence.get_existing_origin_ids(candidate_origin_ids)
known_urls = {
url
for url in vehicle_urls
if (url in existing_urls) or (url_to_origin_id.get(url) in existing_ids)
}
skipped_existing = len(known_urls)
if skipped_existing:
logger.info("Filtering already known vehicles: skipped %d", skipped_existing)
vehicle_urls = [url for url in vehicle_urls if url not in known_urls]
total = len(vehicle_urls)
logger.info("Starting sync: %d vehicles to process", total)
for index, vehicle_url in enumerate(vehicle_urls, start=1):
page = self._get_page()
try:
logger.info("[%d/%d] Scraping %s", index, total, vehicle_url)
scrape_result = self._scrape_on_page(page, vehicle_url)
db_record = scrape_result.get("db_record")
if not db_record:
raise RuntimeError("Scrape result does not contain db_record")
record = CarRecord.model_validate(db_record)
upsert = self.persistence.upsert_car(record)
if upsert.get("action") != "skipped":
cars_upserted += 1
img_count = int(upsert.get("images_upserted", 0))
images_upserted += img_count
logger.info(
"[%d/%d] %s %s: %s %s %s — %s, %d images",
index, total, upsert.get("action", "?"),
record.origin_id, record.brand, record.model,
record.year or "?", record.price or "N/A", img_count,
)
except Exception as exc:
cars_failed += 1
failures.append({"vehicle_url": vehicle_url, "error": str(exc)})
logger.error("[%d/%d] Failed %s: %s", index, total, vehicle_url, exc)
finally:
try:
page.close()
except Exception:
pass
if index < total:
self.pacer.between_vehicles()
except Exception as exc:
# Если падает collect_listing, считаем это общей ошибкой запуска
if not failures:
failures.append({"vehicle_url": "collect_listing", "error": str(exc)})
logger.error("sync_listing failed: %s", exc)
finally:
status = "success" if not failures else ("partial_success" if cars_upserted else "failed")
error_summary = "; ".join(item["error"] for item in failures[:10]) if failures else None
self.persistence.finish_sync_run(
run_id,
status=status,
ids_fetched=total,
cars_upserted=cars_upserted,
cars_failed=cars_failed,
images_upserted=images_upserted,
error_summary=error_summary,
)
logger.info(
"Sync run #%d finished: %d/%d upserted, %d failed, %d images",
run_id, cars_upserted, total, cars_failed, images_upserted,
)
return {
"trace_id": trace_id,
"status": status,
"run_id": run_id,
"listing": listing,
"cars_upserted": cars_upserted,
"cars_failed": cars_failed,
"images_upserted": images_upserted,
"skipped_existing": skipped_existing,
"elapsed_seconds": round(time.perf_counter() - started_at, 3),
"failures": failures,
}
# Встроенный планировщик
def run_scheduled(self) -> None:
# Простой бесконечный цикл без внешнего планировщика
# Graceful shutdown по SIGINT/SIGTERM
def _handle_shutdown(signum, frame):
logger.info("Received signal %s, shutting down gracefully...", signum)
self._shutdown_requested = True
signal.signal(signal.SIGINT, _handle_shutdown)
signal.signal(signal.SIGTERM, _handle_shutdown)
interval = self.settings.scheduler_interval_minutes * 60
logger.info(
"Scheduler started: syncing every %d minutes",
self.settings.scheduler_interval_minutes,
)
cycle = 0
while not self._shutdown_requested:
cycle += 1
logger.info("=== Scheduler cycle #%d starting ===", cycle)
start = time.time()
try:
# Сбрасываем контекст перед каждым циклом
if self.context:
try:
self.context.close()
except PlaywrightError:
pass
self.context = None
# Проверяем, что browser/playwright живы, и пересоздаём при необходимости
if self.browser is None or self.playwright is None:
logger.info("Browser/Playwright not available, re-initializing...")
self.close()
self.__enter__()
result = self.sync_listing()
elapsed = time.time() - start
logger.info(
"Cycle #%d done in %.1fs: %d upserted, %d failed",
cycle, elapsed,
result.get("cars_upserted", 0),
result.get("cars_failed", 0),
)
except Exception as exc:
elapsed = time.time() - start
logger.error("Cycle #%d failed after %.1fs: %s", cycle, elapsed, exc)
# Полный сброс при любой ошибке цикла, следующий цикл пересоздаст всё
try:
self.close()
except Exception:
pass
# Пересоздаём browser для следующего цикла
try:
self.__enter__()
except Exception as reinit_exc:
logger.error("Failed to re-initialize browser: %s", reinit_exc)
if self._shutdown_requested:
break
sleep_time = max(0, interval - (time.time() - start))
if sleep_time > 0:
logger.info("Sleeping %.0f seconds until next cycle...", sleep_time)
# Прерываемый sleep, проверяем shutdown каждые 5 секунд
slept = 0.0
while slept < sleep_time and not self._shutdown_requested:
chunk = min(5.0, sleep_time - slept)
time.sleep(chunk)
slept += chunk
logger.info("Scheduler stopped gracefully after %d cycles.", cycle)

1
iaai_scraper/storage/__init__.py Normal file
View File

@@ -0,0 +1 @@
__all__: list[str] = []

216
iaai_scraper/storage/db.py Normal file
View File

@@ -0,0 +1,216 @@
import json
import logging
from contextlib import contextmanager
from datetime import datetime, timezone
from typing import Iterator
from sqlalchemy import create_engine, or_, select
from sqlalchemy.orm import Session, sessionmaker
from ..core.config import Settings
from .models import Base, Car, Image, ScrapeTask, SyncRun
from .schemas import CarRecord
logger = logging.getLogger("iaai_scraper.db")
# Поля Car, которые приходят из CarRecord (без id, images, relationship).
CAR_DB_FIELDS = {
col.key for col in Car.__table__.columns
if col.key not in ("id",)
}
class PersistenceService:
def __init__(self, settings: Settings) -> None:
# Инициализация engine и фабрики сессий.
self.settings = settings
engine_kwargs = {
"echo": settings.database.echo,
"future": True,
}
# Пул только для PostgreSQL.
if "postgresql" in settings.database.url:
engine_kwargs["pool_size"] = settings.database.pool_size
engine_kwargs["max_overflow"] = settings.database.max_overflow
self.engine = create_engine(settings.database.url, **engine_kwargs)
self.session_factory = sessionmaker(bind=self.engine, expire_on_commit=False, future=True)
def create_tables(self) -> None:
try:
Base.metadata.create_all(self.engine)
except Exception:
# Alembic уже создал таблицы/ENUM — пропускаем.
logger.debug("create_tables skipped (schema already exists)")
@contextmanager
def session_scope(self) -> Iterator[Session]:
# Единая точка commit/rollback для операций записи.
session = self.session_factory()
try:
yield session
session.commit()
except Exception:
session.rollback()
raise
finally:
session.close()
def start_sync_run(self, lane: str) -> int:
# Создаём запись о запуске синхронизации.
with self.session_scope() as session:
# Если предыдущий процесс умер, оставив run в `running`,
# помечаем его как failed перед новым запуском.
now = datetime.now(timezone.utc)
stale_runs = session.execute(select(SyncRun).where(SyncRun.status == "running")).scalars().all()
for stale in stale_runs:
stale.status = "failed"
stale.finished_at = now
if not stale.error_summary:
stale.error_summary = "Recovered stale running sync run before starting a new run"
run = SyncRun(status="running", lane=lane, ids_fetched=0, cars_upserted=0, cars_failed=0, images_upserted=0)
session.add(run)
session.flush()
return int(run.id)
def finish_sync_run(self, run_id: int, *, status: str, ids_fetched: int, cars_upserted: int, cars_failed: int, images_upserted: int, error_summary: str | None = None) -> None:
# Завершаем sync_run и фиксируем итоговую статистику.
with self.session_scope() as session:
run = session.get(SyncRun, run_id)
if run is None:
return
run.finished_at = datetime.now(timezone.utc)
run.status = status
run.ids_fetched = ids_fetched
run.cars_upserted = cars_upserted
run.cars_failed = cars_failed
run.images_upserted = images_upserted
run.error_summary = error_summary
def get_existing_origin_urls(self, origin_urls: list[str]) -> set[str]:
# Возвращает уже существующие в БД origin_url для фильтрации only-new запусков.
if not origin_urls:
return set()
with self.session_scope() as session:
rows = session.execute(select(Car.origin_url).where(Car.origin_url.in_(origin_urls))).all()
return {str(row[0]) for row in rows if row and row[0]}
def get_existing_origin_ids(self, origin_ids: list[str]) -> set[str]:
# Возвращает уже существующие в БД origin_id для фильтрации только новых авто.
if not origin_ids:
return set()
with self.session_scope() as session:
rows = session.execute(select(Car.origin_id).where(Car.origin_id.in_(origin_ids))).all()
return {str(row[0]) for row in rows if row and row[0]}
@staticmethod
def _add_images(session: Session, car_id: int, images: list[dict[str, object]]) -> None:
for image_payload in images:
session.add(Image(fullres_image=str(image_payload["fullres_image"]), preview_image=str(image_payload["preview_image"]), order_index=int(image_payload.get("order_index", 0)), car_id=car_id))
@staticmethod
def _car_payload(record: CarRecord) -> dict[str, object]:
payload = record.model_dump(mode="python")
result = {key: value for key, value in payload.items() if key in CAR_DB_FIELDS}
if "raw_attributes" in result and isinstance(result["raw_attributes"], dict):
result["raw_attributes"] = json.dumps(result["raw_attributes"], ensure_ascii=False, default=str)
return result
def upsert_car(self, record: CarRecord):
"""Insert/update/skip по content_hash."""
# В БД отправляем только поля, реально существующие в финальной схеме cars.
payload = self._car_payload(record)
images = [image.model_dump(mode="python") for image in record.images]
content_hash = str(payload.get("content_hash") or "")
with self.session_scope() as session:
# Сначала пытаемся найти по origin_id, а если ранее origin_id был неполный,
# подхватываем существующую запись по origin_url, чтобы не плодить дубли.
car = session.execute(
select(Car).where(or_(Car.origin_id == record.origin_id, Car.origin_url == record.origin_url))
).scalar_one_or_none()
action = "inserted"
if car is None:
car = Car(**payload)
session.add(car)
session.flush()
else:
# Если контент не менялся, просто обновляем last_seen_at.
if content_hash and car.content_hash == content_hash:
car.last_seen_at = record.last_seen_at
session.flush()
return {"car_id": int(car.id), "images_upserted": 0, "action": "skipped"}
action = "updated"
# Обновляем поля машины и затем безопасно пересобираем картинки.
for key, value in payload.items():
setattr(car, key, value)
car.last_seen_at = record.last_seen_at
session.flush()
# замена картинок в savepoint
nested = session.begin_nested()
try:
for image in list(car.images):
session.delete(image)
session.flush()
self._add_images(session, int(car.id), images)
session.flush()
nested.commit()
except Exception:
nested.rollback()
logger.warning("Image replacement failed for car %s, keeping old images", record.origin_id)
images = []
return {"car_id": int(car.id), "images_upserted": len(images), "action": action}
self._add_images(session, int(car.id), images)
session.flush()
return {"car_id": int(car.id), "images_upserted": len(images), "action": action}
# --- ScrapeTask.
def create_scrape_task(self, celery_task_id: str, task_type: str, vehicle_url: str | None = None) -> int:
"""Создаёт запись задачи Celery."""
with self.session_scope() as session:
task = ScrapeTask(
celery_task_id=celery_task_id,
task_type=task_type,
vehicle_url=vehicle_url,
status="pending",
)
session.add(task)
session.flush()
return int(task.id)
def update_scrape_task(self, celery_task_id: str, **kwargs) -> None:
"""Обновляет поля задачи по celery_task_id."""
with self.session_scope() as session:
task = session.execute(
select(ScrapeTask).where(ScrapeTask.celery_task_id == celery_task_id)
).scalars().first()
if task is None:
return
for key, value in kwargs.items():
if hasattr(task, key):
setattr(task, key, value)
session.flush()
def get_scrape_task(self, celery_task_id: str) -> dict | None:
"""Возвращает информацию о задаче."""
with self.session_scope() as session:
task = session.execute(
select(ScrapeTask).where(ScrapeTask.celery_task_id == celery_task_id)
).scalars().first()
if task is None:
return None
return {
"id": task.id,
"celery_task_id": task.celery_task_id,
"task_type": task.task_type,
"status": task.status,
"vehicle_url": task.vehicle_url,
"created_at": task.created_at.isoformat() if task.created_at else None,
"started_at": task.started_at.isoformat() if task.started_at else None,
"finished_at": task.finished_at.isoformat() if task.finished_at else None,
"result_summary": task.result_summary,
"error_message": task.error_message,
}

8
iaai_scraper/storage/enums.py Normal file
View File

@@ -0,0 +1,8 @@
CURRENCY_ENUM_VALUES = ("JPY", "USD", "EUR", "RUB", "KRW", "AED", "GBP", "CAD")
DRIVE_ENUM_VALUES = ("FWD", "RWD", "2WD", "4WD", "NA")
GEARBOX_ENUM_VALUES = ("AT", "CVT", "MT", "EV", "NA")
STEERING_WHEEL_ENUM_VALUES = ("LEFT", "RIGHT", "NA")
BODY_TYPE_ENUM_VALUES = ("COUPE", "SUV", "HATCHBACK", "MINIVAN", "SEDAN", "STATION_WAGON", "PICKUP", "TRUCK", "OPEN", "RV", "OTHER", "NA")
COUNTRY_ENUM_VALUES = ("JP", "KR", "US", "CA", "NA")
ORIGIN_ENUM_VALUES = ("TAU", "CARSENSOR", "HANAMARU", "ENCAR", "KURUMA_TRADER", "ASNET", "KABABA", "ACV", "COPART", "IAAI", "NA")
SELLING_TYPE_ENUM_VALUES = ("STOCK", "AUCTION", "TENDER", "NA")

99
iaai_scraper/storage/models.py Normal file
View File

@@ -0,0 +1,99 @@
from datetime import datetime
from sqlalchemy import BigInteger, Boolean, DateTime, Enum, ForeignKey, Integer, String, Text, func
from sqlalchemy.orm import DeclarativeBase, Mapped, mapped_column, relationship
from .enums import (
BODY_TYPE_ENUM_VALUES,
COUNTRY_ENUM_VALUES,
CURRENCY_ENUM_VALUES,
DRIVE_ENUM_VALUES,
GEARBOX_ENUM_VALUES,
ORIGIN_ENUM_VALUES,
SELLING_TYPE_ENUM_VALUES,
STEERING_WHEEL_ENUM_VALUES,
)
class Base(DeclarativeBase):
# База ORM.
pass
class Car(Base):
# Автомобиль.
__tablename__ = "cars"
id: Mapped[int] = mapped_column(BigInteger().with_variant(Integer, "sqlite"), primary_key=True, autoincrement=True)
parser_id: Mapped[str] = mapped_column(String(50), nullable=False, unique=True)
brand: Mapped[str] = mapped_column(String(50), nullable=False)
model: Mapped[str] = mapped_column(String(50), nullable=False)
year: Mapped[int | None] = mapped_column(Integer, nullable=True)
price: Mapped[int | None] = mapped_column(BigInteger, nullable=True)
currency: Mapped[str] = mapped_column(Enum(*CURRENCY_ENUM_VALUES, name="currencyenum", native_enum=True, create_constraint=False), nullable=False, default="USD")
mileage: Mapped[int] = mapped_column(Integer, nullable=False, default=0)
country: Mapped[str] = mapped_column(Enum(*COUNTRY_ENUM_VALUES, name="countryenum", native_enum=True, create_constraint=False), nullable=False, default="NA")
is_sold: Mapped[bool] = mapped_column(Boolean, nullable=False, default=False)
color: Mapped[str] = mapped_column(String(), nullable=False, default="other")
drive: Mapped[str | None] = mapped_column(Enum(*DRIVE_ENUM_VALUES, name="driveenum", native_enum=True, create_constraint=False), nullable=True)
gearbox: Mapped[str | None] = mapped_column(Enum(*GEARBOX_ENUM_VALUES, name="gearboxenum", native_enum=True, create_constraint=False), nullable=True)
steering_wheel: Mapped[str | None] = mapped_column(Enum(*STEERING_WHEEL_ENUM_VALUES, name="steeringwheelenum", native_enum=True, create_constraint=False), nullable=True)
body_type: Mapped[str] = mapped_column(Enum(*BODY_TYPE_ENUM_VALUES, name="bodytypeenum", native_enum=True, create_constraint=False), nullable=False, default="OTHER")
engine_volume: Mapped[int | None] = mapped_column(Integer, nullable=True)
selling_type: Mapped[str] = mapped_column(Enum(*SELLING_TYPE_ENUM_VALUES, name="sellingtypeenum", native_enum=True, create_constraint=False), nullable=False, default="NA")
one_owner: Mapped[bool] = mapped_column(Boolean, nullable=False, default=False)
new_car: Mapped[bool] = mapped_column(Boolean, nullable=False, default=False)
is_hidden: Mapped[bool] = mapped_column(Boolean, nullable=False, default=False)
origin: Mapped[str] = mapped_column(Enum(*ORIGIN_ENUM_VALUES, name="originenum", native_enum=True, create_constraint=False), nullable=False, default="NA")
origin_url: Mapped[str] = mapped_column(String(), nullable=False, index=True)
origin_id: Mapped[str] = mapped_column(String(), nullable=False, unique=True, index=True)
is_damaged: Mapped[bool] = mapped_column(Boolean, nullable=False, default=False)
evaluation: Mapped[str | None] = mapped_column(String(), nullable=True)
non_smoking: Mapped[bool] = mapped_column(Boolean, nullable=False, default=True)
rental: Mapped[bool] = mapped_column(Boolean, nullable=False, default=False)
repair_history: Mapped[bool] = mapped_column(Boolean, nullable=False, default=False)
slug: Mapped[str] = mapped_column(String(), nullable=False)
last_seen_at: Mapped[datetime] = mapped_column(DateTime(timezone=True), nullable=False, default=func.now())
content_hash: Mapped[str] = mapped_column(String(64), nullable=False, default="", index=True)
raw_attributes: Mapped[str | None] = mapped_column(Text, nullable=True)
images: Mapped[list["Image"]] = relationship("Image", back_populates="car", cascade="all, delete-orphan")
class Image(Base):
# Картинки автомобиля.
__tablename__ = "images"
id: Mapped[int] = mapped_column(BigInteger().with_variant(Integer, "sqlite"), primary_key=True, autoincrement=True)
fullres_image: Mapped[str] = mapped_column(String(), nullable=False)
preview_image: Mapped[str] = mapped_column(String(), nullable=False)
order_index: Mapped[int] = mapped_column(Integer, nullable=False)
car_id: Mapped[int] = mapped_column(BigInteger, ForeignKey("cars.id", ondelete="CASCADE"), nullable=False)
car: Mapped[Car] = relationship("Car", back_populates="images")
class SyncRun(Base):
# Статистика синхронизаций.
__tablename__ = "sync_runs"
id: Mapped[int] = mapped_column(BigInteger().with_variant(Integer, "sqlite"), primary_key=True, autoincrement=True)
started_at: Mapped[datetime] = mapped_column(DateTime(timezone=True), nullable=False, default=func.now())
finished_at: Mapped[datetime | None] = mapped_column(DateTime(timezone=True), nullable=True)
status: Mapped[str] = mapped_column(Text, nullable=False)
lane: Mapped[str] = mapped_column(Text, nullable=False)
ids_fetched: Mapped[int] = mapped_column(Integer, nullable=False, default=0)
cars_upserted: Mapped[int] = mapped_column(Integer, nullable=False, default=0)
cars_failed: Mapped[int] = mapped_column(Integer, nullable=False, default=0)
images_upserted: Mapped[int] = mapped_column(Integer, nullable=False, default=0)
error_summary: Mapped[str | None] = mapped_column(Text, nullable=True)
class ScrapeTask(Base):
# Задачи Celery.
__tablename__ = "scrape_tasks"
id: Mapped[int] = mapped_column(BigInteger().with_variant(Integer, "sqlite"), primary_key=True, autoincrement=True)
celery_task_id: Mapped[str] = mapped_column(String(255), nullable=False, unique=True, index=True)
task_type: Mapped[str] = mapped_column(String(50), nullable=False) # sync_vehicle/sync_listing
status: Mapped[str] = mapped_column(String(20), nullable=False, default="pending") # pending/running/success/failed
vehicle_url: Mapped[str | None] = mapped_column(Text, nullable=True)
created_at: Mapped[datetime] = mapped_column(DateTime(timezone=True), nullable=False, default=func.now())
started_at: Mapped[datetime | None] = mapped_column(DateTime(timezone=True), nullable=True)
finished_at: Mapped[datetime | None] = mapped_column(DateTime(timezone=True), nullable=True)
result_summary: Mapped[str | None] = mapped_column(Text, nullable=True)
error_message: Mapped[str | None] = mapped_column(Text, nullable=True)

78
iaai_scraper/storage/schemas.py Normal file
View File

@@ -0,0 +1,78 @@
from datetime import datetime, timezone
from typing import Any
from pydantic import BaseModel, ConfigDict, Field
class ImageRecord(BaseModel):
fullres_image: str
preview_image: str
order_index: int = 0
class CarRecord(BaseModel):
parser_id: str
brand: str
model: str
year: int | None = None
price: int | None = None
currency: str = "USD"
mileage: int = 0
country: str = "US"
is_sold: bool = False
color: str = "other"
drive: str | None = None
gearbox: str | None = None
steering_wheel: str | None = None
body_type: str = "OTHER"
engine_volume: int | None = None
selling_type: str = "AUCTION"
one_owner: bool = False
new_car: bool = False
is_hidden: bool = False
origin: str = "NA"
origin_url: str
origin_id: str
is_damaged: bool = False
evaluation: str | None = None
non_smoking: bool = True
rental: bool = False
repair_history: bool = False
slug: str
last_seen_at: datetime = Field(default_factory=lambda: datetime.now(timezone.utc))
content_hash: str = ""
images: list[ImageRecord] = Field(default_factory=list)
raw_attributes: dict[str, Any] = Field(default_factory=dict)
mapping_notes: list[str] = Field(default_factory=list)
class CarRead(BaseModel):
# Ответ API по авто.
model_config = ConfigDict(from_attributes=True)
id: int
parser_id: str
brand: str
model: str
year: int | None = None
price: int | None = None
currency: str = "USD"
mileage: int = 0
country: str = "US"
is_sold: bool = False
color: str = "other"
drive: str | None = None
gearbox: str | None = None
steering_wheel: str | None = None
body_type: str = "OTHER"
engine_volume: int | None = None
selling_type: str = "AUCTION"
origin: str = "NA"
origin_url: str = ""
origin_id: str = ""
is_damaged: bool = False
slug: str = ""
last_seen_at: datetime | None = None
created_at: datetime | None = None
updated_at: datetime | None = None

3
iaai_scraper/worker/__init__.py Normal file
View File

@@ -0,0 +1,3 @@
from .celery_app import celery_app
__all__ = ["celery_app"]

51
iaai_scraper/worker/celery_app.py Normal file
View File

@@ -0,0 +1,51 @@
from celery import Celery
from ..core.config import settings
def _broker_url() -> str:
return settings.celery.broker_url or settings.redis.url
def _result_backend() -> str:
return settings.celery.result_backend or settings.redis.url
celery_app = Celery(
"iaai_scraper",
broker=_broker_url(),
backend=_result_backend(),
)
celery_app.conf.update(
task_serializer="json",
accept_content=["json"],
result_serializer="json",
timezone="UTC",
enable_utc=True,
task_soft_time_limit=settings.celery.task_soft_time_limit,
task_time_limit=settings.celery.task_time_limit,
worker_concurrency=settings.celery.worker_concurrency,
# prefork + 1 процесс, тк playwright sync api
worker_pool="prefork",
worker_prefetch_multiplier=1,
broker_connection_retry_on_startup=True,
beat_schedule={
"periodic-sync-listing": {
"task": "iaai_scraper.worker.tasks.sync_listing_task",
"schedule": settings.celery.beat_sync_interval_minutes * 60.0,
"kwargs": {"limit": settings.celery.beat_sync_limit},
"options": {"queue": "scraping"},
},
},
task_routes={
"iaai_scraper.worker.tasks.*": {"queue": "scraping"},
},
)
celery_app.autodiscover_tasks(["iaai_scraper.worker"])

133
iaai_scraper/worker/tasks.py Normal file
View File

@@ -0,0 +1,133 @@
import json
import logging
from datetime import datetime, timezone
from celery import shared_task
from ..core.config import Settings
from ..scraper import IAAIScraper
from ..storage.db import PersistenceService
logger = logging.getLogger("iaai_scraper.worker.tasks")
def _get_persistence() -> PersistenceService:
return PersistenceService(Settings())
@shared_task(
name="iaai_scraper.worker.tasks.sync_vehicle_task",
bind=True,
max_retries=2,
default_retry_delay=30,
acks_late=True,
)
def sync_vehicle_task(self, vehicle_url: str, lane: str = "iaai"):
task_id = self.request.id
persistence = _get_persistence()
persistence.create_tables()
# регистрируем запуск
persistence.update_scrape_task(
task_id,
status="running",
started_at=datetime.now(timezone.utc),
)
try:
with IAAIScraper() as scraper:
result = scraper.sync_vehicle(vehicle_url, lane=lane)
persistence.update_scrape_task(
task_id,
status="success",
finished_at=datetime.now(timezone.utc),
result_summary=json.dumps({
"car_upserted": True,
"db_action": result.get("db_action"),
"images_upserted": result.get("images_upserted", 0),
"elapsed_seconds": result.get("elapsed_seconds"),
}, default=str),
)
logger.info("sync_vehicle_task completed: %s", vehicle_url)
return {
"status": "success",
"vehicle_url": vehicle_url,
"db_action": result.get("db_action"),
"images_upserted": result.get("images_upserted", 0),
}
except Exception as exc:
persistence.update_scrape_task(
task_id,
status="failed",
finished_at=datetime.now(timezone.utc),
error_message=str(exc)[:2000],
)
logger.error("sync_vehicle_task failed: %s — %s", vehicle_url, exc)
raise self.retry(exc=exc)
@shared_task(
name="iaai_scraper.worker.tasks.sync_listing_task",
bind=True,
max_retries=1,
default_retry_delay=120,
acks_late=True,
)
def sync_listing_task(
self,
make: str | None = None,
model: str | None = None,
lane: str = "iaai_cars",
limit: int | None = None,
only_new: bool | None = None,
):
task_id = self.request.id
persistence = _get_persistence()
persistence.create_tables()
persistence.update_scrape_task(
task_id,
status="running",
started_at=datetime.now(timezone.utc),
)
try:
with IAAIScraper() as scraper:
result = scraper.sync_listing(
make=make,
model=model,
lane=lane,
limit=limit,
only_new=only_new,
)
summary = {
"cars_upserted": result.get("cars_upserted", 0),
"cars_failed": result.get("cars_failed", 0),
"images_upserted": result.get("images_upserted", 0),
"skipped_existing": result.get("skipped_existing", 0),
"elapsed_seconds": result.get("elapsed_seconds"),
}
persistence.update_scrape_task(
task_id,
status="success",
finished_at=datetime.now(timezone.utc),
result_summary=json.dumps(summary, default=str),
)
logger.info(
"sync_listing_task completed: %d upserted, %d failed",
summary["cars_upserted"], summary["cars_failed"],
)
return {"status": "success", **summary}
except Exception as exc:
persistence.update_scrape_task(
task_id,
status="failed",
finished_at=datetime.now(timezone.utc),
error_message=str(exc)[:2000],
)
logger.error("sync_listing_task failed: %s", exc)
raise self.retry(exc=exc)

5
main.py Normal file
View File

@@ -0,0 +1,5 @@
from iaai_scraper.cli import main
if __name__ == "__main__":
main()

4
pytest.ini Normal file
View File

@@ -0,0 +1,4 @@
[pytest]
addopts = -q --disable-warnings
python_files = tests/test_*.py
log_cli = false

12
requirements.txt Normal file
View File

@@ -0,0 +1,12 @@
playwright>=1.53.0
python-dotenv>=1.0.1
pydantic>=2.8.2
SQLAlchemy>=2.0.32
fastapi>=0.115.0
uvicorn>=0.34.0
celery>=5.4.0
redis>=5.2.0
psycopg2-binary>=2.9.9
alembic>=1.14.0
pytest>=8.3.0
pytest-cov>=5.0.0

7
tests/conftest.py Normal file
View File

@@ -0,0 +1,7 @@
from __future__ import annotations
import logging
def pytest_configure() -> None:
logging.disable(logging.CRITICAL)

151
tests/test_db.py Normal file
View File

@@ -0,0 +1,151 @@
from __future__ import annotations
import tempfile
import unittest
from pathlib import Path
from sqlalchemy import select
from iaai_scraper.core.config import Settings
from iaai_scraper.storage.db import PersistenceService
from iaai_scraper.storage.models import Car, Image, SyncRun
from iaai_scraper.storage.schemas import CarRecord, ImageRecord
class TestPersistenceServiceIntegration(unittest.TestCase):
def setUp(self) -> None:
self.tmp_dir = tempfile.TemporaryDirectory()
db_path = Path(self.tmp_dir.name) / "test.sqlite"
self.settings = Settings()
self.settings.database.url = f"sqlite:///{db_path.as_posix()}"
self.settings.database.echo = False
self.persistence = PersistenceService(self.settings)
self.persistence.create_tables()
def tearDown(self) -> None:
self.persistence.engine.dispose()
self.tmp_dir.cleanup()
@staticmethod
def _record(origin_id: str, *, price: int = 1000, content_hash: str = "hash1") -> CarRecord:
return CarRecord(
parser_id=f"iaai:{origin_id}",
brand="Toyota",
model="Camry",
year=2014,
price=price,
origin_url=f"https://www.iaai.com/VehicleDetail/{origin_id}~US",
origin_id=origin_id,
slug=f"toyota-camry-{origin_id}",
content_hash=content_hash,
images=[
ImageRecord(
fullres_image="https://vis.iaai.com/resizer?imageKeys=1&width=845&height=633",
preview_image="https://vis.iaai.com/resizer?imageKeys=1&width=400&height=300",
order_index=0,
)
],
raw_attributes={"foo": "bar"},
)
def test_insert_update_and_skip_flow(self) -> None:
first = self._record("777", price=1000, content_hash="same")
inserted = self.persistence.upsert_car(first)
self.assertEqual(inserted["action"], "inserted")
self.assertEqual(inserted["images_upserted"], 1)
same = self._record("777", price=1000, content_hash="same")
updated_same = self.persistence.upsert_car(same)
self.assertEqual(updated_same["action"], "skipped")
self.assertEqual(updated_same["images_upserted"], 0)
changed = self._record("777", price=1500, content_hash="changed")
updated = self.persistence.upsert_car(changed)
self.assertEqual(updated["action"], "updated")
self.assertEqual(updated["images_upserted"], 1)
with self.persistence.session_scope() as session:
cars = session.execute(select(Car)).scalars().all()
images = session.execute(select(Image)).scalars().all()
self.assertEqual(len(cars), 1)
self.assertEqual(cars[0].price, 1500)
self.assertEqual(len(images), 1)
def test_update_replaces_old_images(self) -> None:
first = self._record("888", content_hash="v1")
self.persistence.upsert_car(first)
second = self._record("888", content_hash="v2")
second.images = [
ImageRecord(
fullres_image="https://vis.iaai.com/resizer?imageKeys=2&width=845&height=633",
preview_image="https://vis.iaai.com/resizer?imageKeys=2&width=400&height=300",
order_index=0,
)
]
self.persistence.upsert_car(second)
with self.persistence.session_scope() as session:
images = session.execute(select(Image)).scalars().all()
self.assertEqual(len(images), 1)
self.assertIn("imageKeys=2", images[0].fullres_image)
def test_same_content_hash_is_skipped(self) -> None:
first = self._record("999", content_hash="same-hash")
self.persistence.upsert_car(first)
second = self._record("999", content_hash="same-hash")
result = self.persistence.upsert_car(second)
self.assertEqual(result["action"], "skipped")
self.assertEqual(result["images_upserted"], 0)
def test_persistence_ignores_non_db_fields(self) -> None:
record = self._record("1000", content_hash="schema-test")
record.raw_attributes = {"vin": "123"}
record.mapping_notes = ["note"]
result = self.persistence.upsert_car(record)
self.assertEqual(result["action"], "inserted")
with self.persistence.session_scope() as session:
car = session.execute(select(Car).where(Car.origin_id == "1000")).scalar_one()
self.assertEqual(car.origin_id, "1000")
def test_start_sync_run_marks_stale_running_runs_as_failed(self) -> None:
first_run_id = self.persistence.start_sync_run("lane-a")
second_run_id = self.persistence.start_sync_run("lane-b")
self.assertNotEqual(first_run_id, second_run_id)
with self.persistence.session_scope() as session:
first = session.get(SyncRun, first_run_id)
second = session.get(SyncRun, second_run_id)
self.assertEqual(first.status, "failed")
self.assertIsNotNone(first.finished_at)
self.assertEqual(second.status, "running")
def test_upsert_falls_back_to_origin_url_to_prevent_duplicates(self) -> None:
first = self._record("OLD-ID", content_hash="v1")
first.origin_url = "https://www.iaai.com/VehicleDetail/45089484~US"
self.persistence.upsert_car(first)
second = self._record("NEW-ID", content_hash="v2")
second.origin_url = "https://www.iaai.com/VehicleDetail/45089484~US"
result = self.persistence.upsert_car(second)
self.assertEqual(result["action"], "updated")
with self.persistence.session_scope() as session:
cars = session.execute(select(Car)).scalars().all()
self.assertEqual(len(cars), 1)
self.assertEqual(cars[0].origin_id, "NEW-ID")
if __name__ == "__main__":
unittest.main()

42
tests/test_listing.py Normal file
View File

@@ -0,0 +1,42 @@
from __future__ import annotations
import unittest
from iaai_scraper.browser.pace import HumanPacer
from iaai_scraper.core.config import Settings
from iaai_scraper.browser.listing import ListingCollector
class _FakePage:
def __init__(self, counts: dict[str, int]) -> None:
self._counts = counts
class _Locator:
def __init__(self, count_value: int) -> None:
self._count_value = count_value
def count(self) -> int:
return self._count_value
def locator(self, selector: str) -> "_FakePage._Locator":
return _FakePage._Locator(self._counts.get(selector, 0))
class TestListingUnit(unittest.TestCase):
def test_has_next_page_true_for_known_selector(self) -> None:
page = _FakePage({"a[aria-label*='Next']": 1})
self.assertTrue(ListingCollector._has_next_page(page))
def test_has_next_page_false_when_no_selectors(self) -> None:
page = _FakePage({})
self.assertFalse(ListingCollector._has_next_page(page))
def test_constructor_with_settings_and_pacer(self) -> None:
settings = Settings()
pacer = HumanPacer(settings)
collector = ListingCollector(settings, pacer)
self.assertIsNotNone(collector)
if __name__ == "__main__":
unittest.main()

139
tests/test_mappers.py Normal file
View File

@@ -0,0 +1,139 @@
from __future__ import annotations
import unittest
from iaai_scraper.parsing.mapper import CarMapper
class TestCarMapper(unittest.TestCase):
def setUp(self) -> None:
self.mapper = CarMapper()
def test_content_hash_is_sha256(self) -> None:
record = self.mapper.map_to_car_record(
vehicle_url="https://www.iaai.com/VehicleDetail/45089484~US",
vehicle_summary={"make": "Toyota", "model": "Camry", "year": "2014"},
payload_insights={
"vehicle_core": {"odometer": "120,000", "body_type": "sedan"},
"pricing": {"buy_now": "$4,500"},
"damage": {"primary": "normal wear"},
"auction": {},
"images": {"urls": []},
},
)
self.assertEqual(len(record.content_hash), 64)
def test_content_hash_changes_when_image_set_changes(self) -> None:
record_one = self.mapper.map_to_car_record(
vehicle_url="https://www.iaai.com/VehicleDetail/45089484~US",
vehicle_summary={
"make": "Toyota",
"model": "Camry",
"year": "2014",
"image_urls": ["https://vis.iaai.com/resizer?imageKeys=1&width=845&height=633"],
},
payload_insights={"vehicle_core": {}, "pricing": {}, "damage": {}, "auction": {}, "images": {}},
)
record_two = self.mapper.map_to_car_record(
vehicle_url="https://www.iaai.com/VehicleDetail/45089484~US",
vehicle_summary={
"make": "Toyota",
"model": "Camry",
"year": "2014",
"image_urls": ["https://vis.iaai.com/resizer?imageKeys=2&width=845&height=633"],
},
payload_insights={"vehicle_core": {}, "pricing": {}, "damage": {}, "auction": {}, "images": {}},
)
self.assertNotEqual(record_one.content_hash, record_two.content_hash)
def test_deduplicates_images_by_image_key(self) -> None:
urls = [
"https://vis.iaai.com/resizer?imageKeys=1&width=200&height=150",
"https://vis.iaai.com/resizer?imageKeys=1&width=845&height=633",
"https://vis.iaai.com/resizer?imageKeys=2&width=400&height=300",
]
record = self.mapper.map_to_car_record(
vehicle_url="https://www.iaai.com/VehicleDetail/123~US",
vehicle_summary={"make": "Honda", "model": "Civic", "image_urls": urls},
payload_insights={"vehicle_core": {}, "pricing": {}, "damage": {}, "auction": {}, "images": {}},
)
self.assertEqual(len(record.images), 2)
self.assertIn("width=845", record.images[0].fullres_image)
self.assertIn("height=633", record.images[0].fullres_image)
def test_no_damage_marker_is_not_damaged(self) -> None:
record = self.mapper.map_to_car_record(
vehicle_url="https://www.iaai.com/VehicleDetail/123~US",
vehicle_summary={"make": "Ford", "model": "Focus"},
payload_insights={
"vehicle_core": {},
"pricing": {},
"damage": {"primary": "normal wear"},
"auction": {},
"images": {},
},
)
self.assertFalse(record.is_damaged)
def test_unknown_and_empty_values_fallbacks(self) -> None:
record = self.mapper.map_to_car_record(
vehicle_url="https://www.iaai.com/VehicleDetail/999~US",
vehicle_summary={"make": " ", "model": None, "drive": "???", "gearbox": "unknown"},
payload_insights={"vehicle_core": {}, "pricing": {}, "damage": {}, "auction": {}, "images": {}},
)
self.assertEqual(record.brand, "UNKNOWN")
self.assertEqual(record.model, "UNKNOWN")
self.assertIsNone(record.steering_wheel if record.steering_wheel not in {"LEFT", None} else None)
self.assertEqual(record.drive, "NA")
self.assertEqual(record.gearbox, "NA")
def test_mapper_handles_case_and_spaces(self) -> None:
record = self.mapper.map_to_car_record(
vehicle_url="https://www.iaai.com/VehicleDetail/888~US",
vehicle_summary={"make": "Honda", "model": "Civic", "drive": " Front Wheel Drive ", "gearbox": " AUTOMATIC "},
payload_insights={"vehicle_core": {}, "pricing": {}, "damage": {}, "auction": {}, "images": {}},
)
self.assertEqual(record.drive, "FWD")
self.assertEqual(record.gearbox, "AT")
def test_price_parsing_dirty_formats(self) -> None:
record = self.mapper.map_to_car_record(
vehicle_url="https://www.iaai.com/VehicleDetail/777~US",
vehicle_summary={"make": "Toyota", "model": "Corolla"},
payload_insights={
"vehicle_core": {},
"pricing": {"buy_now": "USD 4,500 - 5,200"},
"damage": {},
"auction": {},
"images": {},
},
)
self.assertEqual(record.price, 5200)
def test_currency_detection_from_symbol(self) -> None:
record = self.mapper.map_to_car_record(
vehicle_url="https://www.iaai.com/VehicleDetail/778~US",
vehicle_summary={"make": "Toyota", "model": "Corolla"},
payload_insights={
"vehicle_core": {},
"pricing": {"buy_now": "€4.500,00"},
"damage": {},
"auction": {},
"images": {},
},
)
self.assertEqual(record.currency, "EUR")
self.assertEqual(record.price, 4500)
if __name__ == "__main__":
unittest.main()

62
tests/test_parser.py Normal file
View File

@@ -0,0 +1,62 @@
from __future__ import annotations
import unittest
from iaai_scraper.parsing.parser import VehicleParser
class TestVehicleParserUnit(unittest.TestCase):
def setUp(self) -> None:
self.parser = VehicleParser()
def test_parse_dom_key_value_pairs_extracts_known_fields(self) -> None:
dom_text = """
Stock #:
45089484
VIN (Status):
1HGCM82633A123456 (OK)
Primary Damage:
Front End
"""
result = self.parser._parse_dom_key_value_pairs(dom_text)
self.assertEqual(result.get("lot_number"), "45089484")
self.assertEqual(result.get("vin"), "1HGCM82633A123456 (OK)")
self.assertEqual(result.get("primary_damage"), "Front End")
def test_parse_title_for_year_make_model(self) -> None:
parsed = self.parser._parse_title_for_year_make_model("2014 TOYOTA CAMRY for sale", "")
self.assertEqual(parsed["year"], "2014")
self.assertEqual(parsed["make"], "TOYOTA")
self.assertEqual(parsed["model"], "CAMRY")
def test_extract_image_urls_filters_other_vehicle(self) -> None:
vehicle_url = "https://www.iaai.com/VehicleDetail/45089484~US"
payloads = [
{
"imageUrls": [
"https://vis.iaai.com/resizer?imageKeys=45089484~SID1&width=845&height=633",
"https://vis.iaai.com/resizer?imageKeys=99999999~SID2&width=845&height=633",
]
}
]
urls = self.parser._extract_image_urls(payloads, "", vehicle_url)
self.assertEqual(len(urls), 1)
self.assertIn("45089484", urls[0])
def test_extract_image_urls_deduplicates_same_url(self) -> None:
vehicle_url = "https://www.iaai.com/VehicleDetail/45089484~US"
payloads = [{"imageUrls": [
"https://vis.iaai.com/resizer?imageKeys=45089484~SID1&width=845&height=633",
"https://vis.iaai.com/resizer?imageKeys=45089484~SID1&width=845&height=633",
]}]
urls = self.parser._extract_image_urls(payloads, "", vehicle_url)
self.assertEqual(len(urls), 1)
def test_dom_hints_detect_captcha_and_antibot(self) -> None:
hints = self.parser._dom_hints("Please verify you are human. CAPTCHA. Incapsula access denied.")
self.assertTrue(hints["has_captcha_text"])
self.assertTrue(hints["has_antibot_text"])
if __name__ == "__main__":
unittest.main()

154
tests/test_scraper.py Normal file
View File

@@ -0,0 +1,154 @@
from __future__ import annotations
import unittest
from unittest.mock import MagicMock
from iaai_scraper.core.config import Settings
from iaai_scraper.scraper import IAAIScraper
from iaai_scraper.storage.schemas import CarRecord
def make_db_record(origin_id: str) -> dict[str, object]:
return CarRecord(
parser_id=f"iaai:{origin_id}",
brand="Toyota",
model="Camry",
origin_url=f"https://www.iaai.com/VehicleDetail/{origin_id}~US",
origin_id=origin_id,
slug=f"toyota-camry-{origin_id}",
).model_dump(mode="json")
class TestScraperSync(unittest.TestCase):
def _make_scraper(self) -> IAAIScraper:
s = Settings()
s.log_level = "CRITICAL"
s.database.url = "sqlite://"
return IAAIScraper(s)
def test_sync_vehicle_uses_db_record_without_remapping(self) -> None:
scraper = self._make_scraper()
scraper.persistence.create_tables = MagicMock()
scraper.persistence.start_sync_run = MagicMock(return_value=1)
scraper.persistence.finish_sync_run = MagicMock()
scraper.persistence.upsert_car = MagicMock(return_value={"action": "inserted", "images_upserted": 0})
scraper.scrape_vehicle_detail = MagicMock(return_value={"db_record": make_db_record("111")})
scraper.car_mapper.map_to_car_record = MagicMock(side_effect=AssertionError("should not be called"))
result = scraper.sync_vehicle("https://www.iaai.com/VehicleDetail/111~US")
self.assertEqual(result["status"], "success")
self.assertIn("trace_id", result)
self.assertIn("elapsed_seconds", result)
self.assertEqual(scraper.persistence.upsert_car.call_count, 1)
def test_sync_listing_uses_db_record_without_remapping(self) -> None:
scraper = self._make_scraper()
scraper.persistence.create_tables = MagicMock()
scraper.persistence.start_sync_run = MagicMock(return_value=2)
scraper.persistence.finish_sync_run = MagicMock()
scraper.persistence.upsert_car = MagicMock(return_value={"action": "inserted", "images_upserted": 1})
scraper.persistence.get_existing_origin_urls = MagicMock(return_value=set())
scraper.persistence.get_existing_origin_ids = MagicMock(return_value=set())
scraper.collect_listing = MagicMock(return_value={"vehicle_urls": ["https://www.iaai.com/VehicleDetail/222~US"]})
page = MagicMock()
scraper._scrape_on_page = MagicMock(return_value={"db_record": make_db_record("222")})
scraper._get_page = MagicMock(return_value=page)
scraper.car_mapper.map_to_car_record = MagicMock(side_effect=AssertionError("should not be called"))
result = scraper.sync_listing()
self.assertEqual(result["cars_upserted"], 1)
self.assertEqual(result["cars_failed"], 0)
self.assertIn("trace_id", result)
self.assertIn("elapsed_seconds", result)
self.assertEqual(scraper.persistence.upsert_car.call_count, 1)
page.close.assert_called_once()
def test_sync_listing_respects_limit(self) -> None:
scraper = self._make_scraper()
scraper.persistence.create_tables = MagicMock()
scraper.persistence.start_sync_run = MagicMock(return_value=3)
scraper.persistence.finish_sync_run = MagicMock()
scraper.persistence.upsert_car = MagicMock(return_value={"action": "inserted", "images_upserted": 1})
scraper.persistence.get_existing_origin_urls = MagicMock(return_value=set())
scraper.persistence.get_existing_origin_ids = MagicMock(return_value=set())
scraper.collect_listing = MagicMock(return_value={
"vehicle_urls": [
"https://www.iaai.com/VehicleDetail/222~US",
"https://www.iaai.com/VehicleDetail/333~US",
]
})
scraper._scrape_on_page = MagicMock(return_value={"db_record": make_db_record("222")})
scraper._get_page = MagicMock(return_value=MagicMock())
scraper.sync_listing(limit=1)
self.assertEqual(scraper._scrape_on_page.call_count, 1)
def test_sync_listing_does_not_count_skipped_as_upserted(self) -> None:
scraper = self._make_scraper()
scraper.persistence.create_tables = MagicMock()
scraper.persistence.start_sync_run = MagicMock(return_value=4)
scraper.persistence.finish_sync_run = MagicMock()
scraper.persistence.upsert_car = MagicMock(return_value={"action": "skipped", "images_upserted": 0})
scraper.persistence.get_existing_origin_urls = MagicMock(return_value=set())
scraper.persistence.get_existing_origin_ids = MagicMock(return_value=set())
scraper.collect_listing = MagicMock(return_value={"vehicle_urls": ["https://www.iaai.com/VehicleDetail/444~US"]})
scraper._scrape_on_page = MagicMock(return_value={"db_record": make_db_record("444")})
scraper._get_page = MagicMock(return_value=MagicMock())
result = scraper.sync_listing()
self.assertEqual(result["cars_upserted"], 0)
def test_sync_listing_only_new_filters_existing_by_url_and_origin_id(self) -> None:
scraper = self._make_scraper()
scraper.persistence.create_tables = MagicMock()
scraper.persistence.start_sync_run = MagicMock(return_value=5)
scraper.persistence.finish_sync_run = MagicMock()
scraper.persistence.upsert_car = MagicMock(return_value={"action": "inserted", "images_upserted": 0})
scraper.persistence.get_existing_origin_urls = MagicMock(return_value={"https://www.iaai.com/VehicleDetail/111~US"})
scraper.persistence.get_existing_origin_ids = MagicMock(return_value={"222"})
scraper.collect_listing = MagicMock(return_value={
"vehicle_urls": [
"https://www.iaai.com/VehicleDetail/111~US", # exists by URL
"https://www.iaai.com/VehicleDetail/222~US", # exists by ID
"https://www.iaai.com/VehicleDetail/333~US", # new
]
})
page = MagicMock()
scraper._get_page = MagicMock(return_value=page)
scraper._scrape_on_page = MagicMock(return_value={"db_record": make_db_record("333")})
result = scraper.sync_listing(only_new=True)
self.assertEqual(result["skipped_existing"], 2)
self.assertEqual(result["cars_upserted"], 1)
self.assertEqual(scraper._scrape_on_page.call_count, 1)
scraper.persistence.get_existing_origin_urls.assert_called_once()
scraper.persistence.get_existing_origin_ids.assert_called_once()
def test_close_resets_browser_state(self) -> None:
scraper = self._make_scraper()
scraper.context = MagicMock()
scraper.browser = MagicMock()
scraper.playwright = MagicMock()
scraper.close()
self.assertIsNone(scraper.context)
self.assertIsNone(scraper.browser)
self.assertIsNone(scraper.playwright)
if __name__ == "__main__":
unittest.main()

31
tests/test_utils.py Normal file
View File

@@ -0,0 +1,31 @@
from __future__ import annotations
import unittest
from iaai_scraper.core.utils import deep_find_key
class TestDeepFindKey(unittest.TestCase):
def test_finds_key_in_nested_structure(self) -> None:
payload = {
"root": {
"target": "a",
"nested": [{"target": "b"}, {"x": 1}],
}
}
found = deep_find_key(payload, {"target"})
self.assertEqual(found, ["a", "b"])
def test_respects_max_depth(self) -> None:
payload = {"l1": {"l2": {"l3": {"target": "value"}}}}
found_too_shallow = deep_find_key(payload, {"target"}, max_depth=2)
found_enough_depth = deep_find_key(payload, {"target"}, max_depth=8)
self.assertEqual(found_too_shallow, [])
self.assertEqual(found_enough_depth, ["value"])
if __name__ == "__main__":
unittest.main()