update celery schedule and readme

.dockerignore

@@ -14,4 +14,3 @@ coverage.xml
 *.egg-info/
 dist/
 build/
-artifacts/

.env.example

@@ -1,15 +1,11 @@
 IAAI_HEADLESS=true
-IAAI_RAW_OUTPUT_JSON=iaai_raw_network.json
 IAAI_LOG_LEVEL=INFO
 # IAAI_LOG_FILE=iaai_scraper.log
 
-# Conservative first-run mode.
+# Network capture settings.
-IAAI_GENTLE_MODE=true
 IAAI_CAPTURE_SAME_ORIGIN_ONLY=true
 IAAI_MAX_CAPTURED_REQUESTS=40
 IAAI_MAX_CAPTURED_JSON_RESPONSES=30
-IAAI_WARM_SCROLL_ROUNDS=1
-IAAI_POST_OPEN_IDLE_MS=3500
 
 # Sequential listing-first mode.
 IAAI_CARS_LISTING_URL=https://www.iaai.com/Vehiclelisting/Cars
@@ -34,8 +30,8 @@ IAAI_BETWEEN_VEHICLES_MAX_S=6.0
 IAAI_AFTER_PAGE_CHANGE_MIN_S=3.0
 IAAI_AFTER_PAGE_CHANGE_MAX_S=6.0
 
-# Scheduler (default once per hour)
+# Sync settings
-IAAI_SCHEDULER_INTERVAL_MINUTES=60
+IAAI_SYNC_ONLY_NEW=true
 
 # Retry / backoff
 IAAI_RETRY_DELAY_SECONDS=2.5
@@ -49,6 +45,20 @@ IAAI_RETRY_JITTER_SECONDS=0.25
 # IAAI_PROXY_USERNAME=
 # IAAI_PROXY_PASSWORD=
 
-# Database.
+# Database (PostgreSQL).
-IAAI_DATABASE_URL=sqlite:///iaai_scraper.db
+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

.gitignore

@@ -8,9 +8,10 @@ coverage.xml
 *.pyd
 *.log
 *.db
-*.json
 .env
 .vscode/
 *.egg-info/
 dist/
 build/
+artifacts/
+celerybeat-schedule*

Dockerfile

@@ -17,5 +17,6 @@ RUN chmod +x entrypoint.sh
 
 STOPSIGNAL SIGINT
 
+# По умолчанию — API, но worker/beat переопределяют CMD в docker-compose
 ENTRYPOINT ["./entrypoint.sh"]
-CMD ["python", "main.py", "run-daemon"]
+CMD ["uvicorn", "iaai_scraper.api.app:app", "--host", "0.0.0.0", "--port", "8000"]

README.md

@@ -1,173 +1,138 @@
 # IAAI Scraper
 
-Скрапер листинга автомобилей с сайта IAAI.
+Парсер аукционных автомобилей с [iaai.com](https://www.iaai.com). Ходит по листингу, собирает карточки машин, вытаскивает данные из DOM и перехваченных XHR-ответов, складывает всё в PostgreSQL. Работает через Playwright (headless Chromium), крутится в Docker.
-Собирает данные карточек через Playwright, парсит HTML и перехваченные JSON ответы,
-нормализует и сохраняет в PostgreSQL или SQLite через SQLAlchemy.
 
-По умолчанию работает последовательно одна машина за раз, с паузами между запросами.
 
-## Что делает
+## Как устроен сайт и его защита
 
-1. Открывает страницу листинга `Vehiclelisting/Cars`, собирает ссылки на карточки.
+У IAAI стоит **Imperva Incapsula** — внешний WAF и anti-bot.
-2. Переходит на каждую карточку, перехватывает XHR/fetch JSON-ответы.
-3. Парсит DOM-текст, `<title>`, встроенные `<script>` с JSON, сетевые payload'ы.
-4. Маппит всё в единую структуру `CarRecord` (pydantic) с нормализацией полей.
-5. Делает upsert в БД по `origin_id`, сравнивая `content_hash`, чтобы пропускать неизменившиеся записи.
-6. Защищён от дублей: `origin_id` уникален, одинаковые записи пропускаются, а повторные картинки заменяются безопасно.
 
-## Структура проекта
+- **Anti-bot** — headless Chromium без прокси часто режется.
+- **Динамическая подгрузка** — часть данных приходит через XHR (`/Search`, `/VehicleDetail`), часть есть в HTML.
+- **Cookie consent** — при первом заходе показывают баннер.
+
+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`.
+
+По умолчанию `beat` запускает сбор листинга **раз в 1 час** и обрабатывает **до 26 машин за запуск** (`CELERY_BEAT_SYNC_INTERVAL_MINUTES=60`, `CELERY_BEAT_SYNC_LIMIT=26`).
+
+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 — скрапнуть конкретную машину:
+```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 отладки без API/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/
-├── browser/
+  scraper.py              — оркестратор: связывает browser → parser → storage
-│   ├── factory.py     # запуск Chrome/Chromium с desktop-фингерпринтом
+  cli.py                  — CLI-команды (init-db, sync-vehicle, sync-listing, ...)
-│   ├── network.py     # перехват XHR/fetch, фильтрация и категоризация JSON
+  proxy_bridge.py         — HTTP→SOCKS5 мост (Chromium не умеет SOCKS5 с авторизацией)
-│   └── pace.py        # паузы между действиями
+
-├── core/
+  browser/
-│   ├── config.py      # настройки из .env
+    factory.py            — создание браузера, stealth-инъекции, fingerprint
-│   ├── logs.py        # setup logging
+    listing.py            — сбор ссылок на машины с листинга, пагинация
-│   ├── retry.py       # retry-декоратор
+    network.py            — перехват XHR/fetch ответов через Playwright events
-│   └── utils.py       # regex, deep_find_key, save_to_json
+    pace.py               — рандомные паузы и движение мыши
-├── parsing/
+
-│   ├── parser.py      # DOM + JSON парсинг
+  parsing/
-│   └── mapper.py      # нормализация в CarRecord
+    parser.py             — VehicleParser: 3 канала (DOM + XHR JSON + embedded JSON)
-├── storage/
+    mapper.py             — CarMapper: нормализация → CarRecord, content hash
-│   ├── models.py      # ORM: cars, images, sync_runs
+
-│   ├── schemas.py     # pydantic-схемы
+  storage/
-│   ├── db.py          # upsert с content_hash
+    models.py             — SQLAlchemy: Car (30+ полей), Image, SyncRun, ScrapeTask
-│   ├── listing.py     # сбор ссылок из листинга
+    schemas.py            — Pydantic: CarRecord, ImageRecord, CarRead
-│   └── enums.py       # enum-значения для БД
+    enums.py              — допустимые значения (drive, gearbox, body_type, ...)
-├── scraper.py         # главный модуль
+    db.py                 — PersistenceService: upsert (insert/update/skip), sync runs
-└── cli.py             # CLI (argparse)
+
+  api/
+    app.py                — FastAPI factory, lifespan, роутеры
+    deps.py               — dependency injection (Settings, PersistenceService)
+    routes/
+      health.py           — GET /health
+      cars.py             — CRUD по машинам + GET /stats
+      tasks.py            — управление Celery-задачами + sync-runs
+
+  worker/
+    celery_app.py         — конфиг Celery, beat-расписание
+    tasks.py              — sync_vehicle_task, sync_listing_task
+
+  core/
+    config.py             — Settings (dataclass), все env-переменные
+    logs.py               — логирование с trace_id (ContextVar)
+    retry.py              — декоратор @retryable с exponential backoff
+    utils.py              — VIN_RE, deep_find_key, вспомогательные функции
+
+alembic/                  — миграции БД
+tests/                    — 30 тестов (SQLite in-memory)
 ```
 
-## Установка
+## Конфигурация
 
+Всё через env-переменные (полный список 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 накатываются автоматически при старте API-контейнера (`alembic upgrade head` в entrypoint).
+
+Вручную:
 ```bash
-pip install -r requirements.txt
+alembic upgrade head
-python -m playwright install chromium
+alembic revision --autogenerate -m "add_column_x"
-```
-
-## Настройка
-
-Создайте `.env` на основе `.env.example`:
-
-```env
-IAAI_HEADLESS=true
-IAAI_DATABASE_URL=postgresql+psycopg://postgres:postgres@localhost:5432/iaai_scraper
-```
-
-Все настройки (pacing, лимиты, gentle mode, retry/backoff, scheduler) задаются через переменные окружения в `.env.example`.
-
-### БД для рабочего сайта
-
-Для рабочего запуска нужно **PostgreSQL**.
-
-Пример:
-
-```env
-IAAI_DATABASE_URL=postgresql+psycopg://postgres:postgres@localhost:5432/iaai_scraper
-```
-
-SQLite подходит для локальной отладки, но не для production-сценария сайта.
-
-### Scheduler по умолчанию
-
-Режим :
-
-- запуск раз в **1 час**
-- лимит **30 машин за цикл**
-
-Это уже отражено в актуальных env-настройках.
-
-### Прокси и anti-bot
-
-IAAI использует anti-bot / fraud protection.
-
-Что важно:
-
-- предпочтительно использовать **USA residential** или **USA mobile** прокси
-- для Playwright лучше использовать **HTTP/HTTPS proxy**
-- Chromium **не поддерживает SOCKS5 с аутентификацией напрямую**
-- поэтому для production желательно покупать прокси, который отдаёт именно HTTP/HTTPS доступ
-
-Пример:
-
-```env
-IAAI_PROXY_SERVER=http://proxy.example.com:8080
-IAAI_PROXY_USERNAME=username
-IAAI_PROXY_PASSWORD=password
-```
-
-### Captcha / anti-bot detection
-
-В парсере добавлены признаки для определения возможной captcha / anti-bot страницы:
-
-- `possible_captcha`
-- `possible_antibot`
-- `dom_hints.has_captcha_text`
-- `dom_hints.has_antibot_text`
-
-Если сайт начнёт отдавать защитную страницу, это можно увидеть в результате scrape.
-
-## Команды
-
-```bash
-# создать таблицы
-python main.py init-db
-
-# собрать ссылки из листинга
-python main.py collect-listing --make Toyota --model Camry --output listing.json
-
-# scrape одной карточки
-python main.py scrape-vehicle "https://www.iaai.com/VehicleDetail/41180634~US" --output result.json
-
-# scrape + запись в БД
-python main.py sync-vehicle "https://www.iaai.com/VehicleDetail/41180634~US" --lane iaai
-
-# массовая синхронизация листинга
-python main.py sync-listing --make Toyota --model Camry --lane iaai_cars --limit 30
-
-# daemon-режим (цикл каждые N минут)
-python main.py run-daemon --interval 60
 ```
 
 ## Тесты
 
 ```bash
-pytest tests -q
+pytest -q
 ```
 
-## Docker
+30 тестов, SQLite in-memory, без внешних зависимостей.
-
-```bash
-# собрать образ
-docker compose build
-
-# запустить daemon (по умолчанию run-daemon)
-docker compose up -d
-
-# посмотреть логи
-docker compose logs -f
-
-# одноразовая команда
-docker compose run --rm iaai-scraper python main.py sync-listing --limit 5
-
-# остановить (graceful shutdown)
-docker compose down
-```
-
-Контейнер автоматически перезапускается при крашах (`restart: unless-stopped`).
-Для Docker прокси также задаются через `.env`.
-
-## Защита от дублей
-
-Система защищена от дублей на нескольких уровнях:
-
-- `origin_id` уникален в БД
-- при совпадении `content_hash` запись **пропускается** (`skipped`)
-- при обновлении запись не дублируется, а обновляется
-- изображения пересобираются без накопления дублей
-
-
-

alembic.ini

@@ -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

alembic/env.py

@@ -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()

alembic/script.py.mako

@@ -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"}

alembic/versions/001_initial.py

@@ -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")

docker-compose.yml

@@ -1,15 +1,105 @@
 services:
-  iaai-scraper:
+  # ─── 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-scraper
+    container_name: iaai-api
+    restart: unless-stopped
     env_file:
       - path: .env
         required: false
-    volumes:
+    environment:
-      - ./:/app
+      IAAI_DATABASE_URL: postgresql+psycopg2://iaai:iaai@postgres:5432/iaai_scraper
-      - /app/.venv
+      IAAI_REDIS_URL: redis://redis:6379/0
-      - /app/__pycache__
+      CELERY_BROKER_URL: redis://redis:6379/0
-    working_dir: /app
+      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
-    stop_grace_period: 30s
+    env_file:
-    command: python main.py run-daemon
+      - 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:

entrypoint.sh

@@ -20,6 +20,23 @@ if [ -n "$SOCKS5_PROXY_HOST" ]; then
   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
@@ -27,5 +44,14 @@ 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 "$@"

iaai_scraper/api/__init__.py

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

iaai_scraper/api/app.py

@@ -0,0 +1,41 @@
+"""FastAPI app."""
+
+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):
+    """Жизненный цикл."""
+    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
+
+
+# Для запуска через uvicorn.
+app = create_app()

iaai_scraper/api/deps.py

@@ -0,0 +1,9 @@
+"""FastAPI deps."""
+
+from fastapi import Request
+
+from ..storage.db import PersistenceService
+
+
+def get_persistence(request: Request) -> PersistenceService:
+    return request.app.state.persistence

iaai_scraper/api/routes/cars.py

@@ -0,0 +1,144 @@
+"""Cars endpoints."""
+
+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),
+):
+    """Список автомобилей с пагинацией и фильтрами."""
+    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),
+):
+    """Детальная информация об автомобиле с изображениями."""
+    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),
+):
+    """Поиск автомобиля по origin_id."""
+    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)):
+    """Общая статистика по БД."""
+    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:
+    """Сериализация Car в 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

iaai_scraper/api/routes/health.py

@@ -0,0 +1,27 @@
+"""Health endpoint."""
+
+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)):
+    """Проверка API и БД."""
+    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",
+    }

iaai_scraper/api/routes/tasks.py

@@ -0,0 +1,180 @@
+"""Task endpoints."""
+
+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),
+):
+    """Запустить задачу скрапинга одного автомобиля через Celery."""
+    result = sync_vehicle_task.apply_async(
+        kwargs={"vehicle_url": body.vehicle_url, "lane": body.lane},
+        queue="scraping",
+    )
+
+    # Регистрируем задачу 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),
+):
+    """Запустить задачу полного цикла листинга через Celery."""
+    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),
+):
+    """Статус Celery-задачи."""
+    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),
+):
+    """Список задач с пагинацией."""
+    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),
+):
+    """История запусков синхронизации."""
+    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
+            ],
+        }

iaai_scraper/browser/__init__.py

@@ -1,5 +1,6 @@
 from .factory import BrowserFactory
+from .listing import ListingCollector
 from .network import NetworkCapture
 from .pace import HumanPacer
 
-__all__ = ["BrowserFactory", "NetworkCapture", "HumanPacer"]
+__all__ = ["BrowserFactory", "ListingCollector", "NetworkCapture", "HumanPacer"]

iaai_scraper/browser/factory.py

@@ -11,7 +11,6 @@ logger = logging.getLogger("iaai_scraper.browser")
 
 def _build_init_script() -> str:
     """JS-патч признаков автоматизации."""
-  # Подмешиваем базовые browser fingerprint overrides.
     hardware_concurrency = random.choice([4, 8, 12, 16])
     device_memory = random.choice([4, 8, 16])
     languages = ["en-US", "en"]
@@ -62,11 +61,9 @@ def _build_init_script() -> str:
 class BrowserFactory:
 
     def __init__(self, settings: Settings) -> None:
-    # Фабрика браузера и контекстов на основе env-настроек.
         self.settings = settings
 
     def create_browser(self, playwright: Playwright) -> Browser:
-        # пробуем Chrome, если нет — Chromium
         launch_kwargs: dict = {
             "headless": self.settings.headless,
             "args": [
@@ -88,8 +85,6 @@ class BrowserFactory:
             return playwright.chromium.launch(**launch_kwargs)
 
     def create_context(self, browser: Browser, storage_state: str | None = None) -> BrowserContext:
-        # рандом viewport/timezone под каждый контекст
-      # Контекст изолирует cookies, headers и fingerprint-параметры.
         viewport = random.choice(self.settings.fingerprint.viewport_presets)
         timezone_id = random.choice(self.settings.fingerprint.timezone_candidates)
         color_scheme = random.choice(["light", "dark"])

iaai_scraper/browser/listing.py

@@ -1,11 +1,12 @@
 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 ..browser.pace import HumanPacer
+from .pace import HumanPacer
 from ..core.config import Settings
 from ..core.utils import first_non_empty
 
@@ -15,7 +16,7 @@ VEHICLE_HREF_RE = re.compile(r"/VehicleDetail/\d+(?:~[A-Z]{2})?", re.IGNORECASE)
 
 @dataclass(slots=True)
 class ListingVehicleLink:
-	# Ссылка на одну карточку из листинга.
+	# Ссылка на карточку.
 	href: str
 	title: str = ""
 	lot_number: str | None = None
@@ -23,7 +24,7 @@ class ListingVehicleLink:
 
 @dataclass(slots=True)
 class ListingPageResult:
-	# Результат сбора ссылок с одной страницы листинга.
+	# Результат страницы листинга.
 	source_url: str
 	page_number: int
 	vehicle_links: list[ListingVehicleLink] = field(default_factory=list)
@@ -33,12 +34,12 @@ class ListingPageResult:
 
 class ListingCollector:
 	def __init__(self, settings: Settings, pacer: HumanPacer) -> None:
-		# Сборщик ссылок из публичного листинга IAAI.
+		# Сборщик ссылок листинга.
 		self.settings = settings
 		self.pacer = pacer
 
 	def open_cars_listing(self, page: Page) -> None:
-		# Открываем листинг и ждём появления ссылок на карточки.
+		# Открываем листинг и ждем ссылки.
 		logger.info("Opening cars listing page: %s", self.settings.listing.cars_url)
 		page.goto(self.settings.listing.cars_url, wait_until="domcontentloaded")
 		try:
@@ -52,7 +53,7 @@ class ListingCollector:
 		self.pacer.after_listing_open()
 
 	def apply_filters(self, page: Page, make: str | None = None, model: str | None = None) -> dict[str, str | None]:
-		# Пытаемся применить make/model фильтры через UI.
+		# 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
@@ -63,7 +64,7 @@ class ListingCollector:
 		return applied
 
 	def collect_current_page(self, page: Page, page_number: int = 1) -> ListingPageResult:
-		# Собираем ссылки только с текущей страницы.
+		# Собираем ссылки текущей страницы.
 		anchors = page.locator("a[href*='/VehicleDetail/']")
 		total = min(anchors.count(), self.settings.listing.page_link_limit)
 		links: list[ListingVehicleLink] = []
@@ -86,7 +87,7 @@ class ListingCollector:
 		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, если кнопка доступна.
+		# 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
@@ -107,8 +108,8 @@ class ListingCollector:
 			return True
 		return False
 
-	def collect_listing_links(self, page: Page, *, make: str | None = None, model: str | None = None) -> dict[str, object]:
+	def collect_listing_links(self, page: Page, *, make: str | None = None, model: str | None = None) -> dict[str, Any]:
-		# Последовательно собираем ссылки с учётом лимитов и пагинации.
+		# Последовательный сбор с лимитами.
 		self.open_cars_listing(page)
 		applied_filters = self.apply_filters(page, make=make, model=model)
 		pages: list[dict[str, object]] = []
@@ -129,13 +130,15 @@ class ListingCollector:
 
 	@staticmethod
 	def _try_fill_filter_input(page: Page, selectors: list[str], value: str) -> bool:
-		# Пробуем несколько селекторов для одного и того же фильтра.
+		# 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")
+				locator.click()
+				locator.fill(value)
+				page.keyboard.press("Enter")
 				try:
 					page.wait_for_load_state("networkidle", timeout=15000)
 				except Exception:
@@ -147,7 +150,7 @@ class ListingCollector:
 
 	@staticmethod
 	def _has_next_page(page: Page) -> bool:
-		# Грубая проверка наличия кнопки next.
+		# 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

iaai_scraper/browser/network.py

@@ -23,7 +23,7 @@ class NetworkCapture:
     _origin: str | None = None
 
     def attach(self, page: Page, origin_url: str | None = None) -> None:
-        # Подписываемся на request/response события страницы.
+        # Подписка на сетевые события.
         try:
             self._origin = urlparse(origin_url or page.url).netloc.lower() or None
         except Exception:
@@ -32,22 +32,22 @@ class NetworkCapture:
         page.on("response", self._on_response)
 
     def _is_same_origin(self, url: str) -> bool:
-        # При необходимости ограничиваемся same-origin и доменами IAAI.
+        # Фильтр по домену.
-        if not self.settings.gentle.capture_same_origin_only or not self._origin:
+        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:
-        # только xhr/fetch
+        # Берем 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.gentle.max_requests:
+        if len(self.requests) >= self.settings.capture.max_requests:
             return
         key = f"{request.method}:{request.url}:{request.post_data or ''}"
-        # Дедупликация одинаковых запросов.
+        # Убираем дубли.
         if key in self._seen_req:
             return
         self._seen_req.add(key)
@@ -57,13 +57,13 @@ class NetworkCapture:
         })
 
     def _on_response(self, response: Response) -> None:
-        # сохраняем json
+        # 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.gentle.max_json_responses:
+        if len(self.json_responses) >= self.settings.capture.max_json_responses:
             return
         if not self._is_json(response):
             return
@@ -95,7 +95,7 @@ class NetworkCapture:
 
     @staticmethod
     def _categorize(url: str) -> str:
-        # Простая эвристика для разбивки ответов по смыслу.
+        # Простая категория URL.
         low = url.lower()
         mapping = {
             "images": ["image", "media", "photos", "gallery"],
@@ -110,15 +110,13 @@ class NetworkCapture:
         return "other"
 
     def export(self) -> dict[str, Any]:
-        # полезные категории сначала, other в конец
+        responses = sorted(self.json_responses, key=lambda i: (i["category"] == "other", i["url"]))
-        prioritized = sorted(self.json_responses, key=lambda i: (i["category"] == "other", i["url"]))
         return {
             "requests": self.requests,
-            "json_responses": self.json_responses,
+            "json_responses": responses,
-            "prioritized_json_responses": prioritized,
             "capture_limits": {
-                "same_origin_only": self.settings.gentle.capture_same_origin_only,
+                "same_origin_only": self.settings.capture.capture_same_origin_only,
-                "max_requests": self.settings.gentle.max_requests,
+                "max_requests": self.settings.capture.max_requests,
-                "max_json_responses": self.settings.gentle.max_json_responses,
+                "max_json_responses": self.settings.capture.max_json_responses,
             },
         }

iaai_scraper/browser/pace.py

@@ -10,11 +10,9 @@ class HumanPacer:
     """Random паузы между действиями."""
 
     def __init__(self, settings: Settings) -> None:
-	    # Инкапсулирует все задержки между действиями браузера.
         self.settings = settings
 
     def pause(self, min_s: float, max_s: float) -> None:
-	    # Базовая случайная пауза в заданном диапазоне.
         if self.settings.pace.enabled:
             time.sleep(random.uniform(min_s, max_s))
 
@@ -37,13 +35,12 @@ class HumanPacer:
         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:
-	    # Небольшое движение мыши к элементу перед кликом.
         try:
             box = locator.bounding_box()
         except Exception:
             box = None
         if not box:
             return
-        x = box["x"] + min(box["width"] * 0.6, max(5, box["width"] * random.uniform(0.2, 0.8)))
+        x = box["x"] + box["width"] * random.uniform(0.2, 0.8)
-        y = box["y"] + min(box["height"] * 0.6, max(5, box["height"] * 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))

iaai_scraper/cli.py

@@ -7,70 +7,41 @@ from .scraper import IAAIScraper
 
 
 def build_parser() -> argparse.ArgumentParser:
-    # Описание CLI-команд и их аргументов.
+    parser = argparse.ArgumentParser(description="IAAI scraper CLI")
-    parser = argparse.ArgumentParser(description="Public IAAI scraper")
     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)
 
-    init_db_parser = subparsers.add_parser("init-db", help="Create local DB tables")
+    default_output_dir = Path("artifacts/json")
-    init_db_parser.add_argument("--output", default="iaai_db_init.json", help="Path to output JSON")
 
-    listing_parser = subparsers.add_parser("collect-listing", help="Collect vehicle URLs from Vehiclelisting/Cars")
+    subparsers.add_parser("init-db", help="Create DB tables")
-    listing_parser.add_argument("--make", default=None, help="Optional make filter")
-    listing_parser.add_argument("--model", default=None, help="Optional model filter")
-    listing_parser.add_argument("--output", default="iaai_listing_links.json", help="Path to output JSON")
 
-    open_parser = subparsers.add_parser(
+    listing_parser = subparsers.add_parser("collect-listing", help="Collect vehicle URLs from listing page")
-        "open-vehicle",
+    listing_parser.add_argument("--make", default=None)
-        help="Open a vehicle page in gentle mode and save only DOM-based hints",
+    listing_parser.add_argument("--model", default=None)
-    )
+    listing_parser.add_argument("--output", default=str(default_output_dir / "iaai_listing_links.json"))
-    open_parser.add_argument("vehicle_url", help="IAAI vehicle detail URL")
-    open_parser.add_argument("--output", default="iaai_vehicle_opened.json", help="Path to output JSON")
 
-    scrape_parser = subparsers.add_parser(
+    scrape_parser = subparsers.add_parser("scrape-vehicle", help="Scrape a vehicle detail page")
-        "scrape-vehicle",
+    scrape_parser.add_argument("vehicle_url")
-        help="Open a vehicle page and capture a limited set of likely useful JSON responses",
+    scrape_parser.add_argument("--output", default=str(default_output_dir / "iaai_vehicle_detail.json"))
-    )
-    scrape_parser.add_argument("vehicle_url", help="IAAI vehicle detail URL")
-    scrape_parser.add_argument("--output", default="iaai_vehicle_detail.json", help="Path to output JSON")
 
-    export_parser = subparsers.add_parser(
+    sync_vehicle_parser = subparsers.add_parser("sync-vehicle", help="Scrape + upsert one vehicle")
-        "export-db-json",
+    sync_vehicle_parser.add_argument("vehicle_url")
-        help="Scrape a vehicle page and save only the DB-ready car record JSON",
+    sync_vehicle_parser.add_argument("--lane", default="iaai")
-    )
+    sync_vehicle_parser.add_argument("--output", default=str(default_output_dir / "iaai_sync_vehicle.json"))
-    export_parser.add_argument("vehicle_url", help="IAAI vehicle detail URL")
-    export_parser.add_argument("--output", default="iaai_vehicle_db_record.json", help="Path to output JSON")
 
-    sync_vehicle_parser = subparsers.add_parser("sync-vehicle", help="Scrape one vehicle and upsert it into the DB")
+    sync_listing_parser = subparsers.add_parser("sync-listing", help="Collect listing + sync all vehicles")
-    sync_vehicle_parser.add_argument("vehicle_url", help="IAAI vehicle detail URL")
+    sync_listing_parser.add_argument("--make", default=None)
-    sync_vehicle_parser.add_argument("--lane", default="iaai", help="Logical lane name for sync_runs")
+    sync_listing_parser.add_argument("--model", default=None)
-    sync_vehicle_parser.add_argument("--output", default="iaai_sync_vehicle.json", help="Path to output JSON")
+    sync_listing_parser.add_argument("--lane", default="iaai_cars")
-
+    sync_listing_parser.add_argument("--limit", type=int, default=None)
-    sync_listing_parser = subparsers.add_parser(
+    sync_listing_parser.add_argument("--only-new", choices=["true", "false"], default=None)
-        "sync-listing",
+    sync_listing_parser.add_argument("--output", default=str(default_output_dir / "iaai_sync_listing.json"))
-        help="Collect vehicle URLs from the Cars listing and sync them sequentially",
-    )
-    sync_listing_parser.add_argument("--make", default=None, help="Optional make filter")
-    sync_listing_parser.add_argument("--model", default=None, help="Optional model filter")
-    sync_listing_parser.add_argument("--lane", default="iaai_cars", help="Logical lane name for sync_runs")
-    sync_listing_parser.add_argument("--limit", type=int, default=None, help="Limit number of vehicles to sync")
-    sync_listing_parser.add_argument("--output", default="iaai_sync_listing.json", help="Path to output JSON")
-
-    daemon_parser = subparsers.add_parser(
-        "run-daemon",
-        help="Run the scraper in a loop, syncing vehicles every N minutes (default 60)",
-    )
-    daemon_parser.add_argument(
-        "--interval", type=int, default=None,
-        help="Override interval in minutes (default from IAAI_SCHEDULER_INTERVAL_MINUTES or 60)",
-    )
 
     return parser
 
 
 def main() -> None:
-    # Собираем runtime overrides только при необходимости.
     parser = build_parser()
     args = parser.parse_args()
 
@@ -82,36 +53,29 @@ def main() -> None:
         if args.debug:
             runtime_settings.log_level = "DEBUG"
 
-    if args.command == "run-daemon":
-        # daemon: бесконечный цикл
-        runtime_settings = runtime_settings or Settings()
-        if args.interval is not None:
-            runtime_settings.scheduler_interval_minutes = args.interval
     with IAAIScraper(runtime_settings) as scraper:
-            scraper.persistence.create_tables()
-            scraper.run_scheduled()
-        return
-
-    # одноразовый запуск
-    with IAAIScraper(runtime_settings) as scraper:
-        # Каждая команда сводится к одному методу скрапера.
         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 == "open-vehicle":
-            data = scraper.open_vehicle_page(args.vehicle_url)
         elif args.command == "scrape-vehicle":
             data = scraper.scrape_vehicle_detail(args.vehicle_url)
-        elif args.command == "export-db-json":
-            data = scraper.scrape_vehicle_detail(args.vehicle_url).get("db_record", {})
         elif args.command == "sync-vehicle":
             data = scraper.sync_vehicle(args.vehicle_url, lane=args.lane)
         else:
-            data = scraper.sync_listing(make=args.make, model=args.model, lane=args.lane, limit=args.limit)
+            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 result to {Path(args.output).resolve()}")
+        print(f"Saved to {Path(args.output).resolve()}")
 
 
 if __name__ == "__main__":

iaai_scraper/core/__init__.py

@@ -1,4 +1 @@
-from .config import *  # noqa: F401,F403
+__all__: list[str] = []
-from .logs import *  # noqa: F401,F403
-from .retry import *  # noqa: F401,F403
-from .utils import *  # noqa: F401,F403

iaai_scraper/core/config.py

@@ -1,6 +1,5 @@
 import os
 from dataclasses import dataclass, field
-from pathlib import Path
 
 from dotenv import load_dotenv
 
@@ -9,7 +8,7 @@ load_dotenv()
 
 @dataclass(slots=True)
 class FingerprintConfig:
-    # Настройки браузерного отпечатка.
+    # Браузерный отпечаток.
     user_agent: str = (
         "Mozilla/5.0 (Windows NT 10.0; Win64; x64) "
         "AppleWebKit/537.36 (KHTML, like Gecko) "
@@ -32,20 +31,16 @@ class FingerprintConfig:
 
 
 @dataclass(slots=True)
-class GentleModeConfig:
+class CaptureConfig:
-    # Мягкий режим загрузки страницы и capture.
+    # Параметры capture.
-    enabled: bool = os.getenv("IAAI_GENTLE_MODE", "true").strip().lower() in {"1", "true", "yes", "on"}
     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"))
-    warm_scroll_rounds: int = int(os.getenv("IAAI_WARM_SCROLL_ROUNDS", "1"))
-    scroll_pause_ms: int = int(os.getenv("IAAI_SCROLL_PAUSE_MS", "900"))
-    post_open_idle_ms: int = int(os.getenv("IAAI_POST_OPEN_IDLE_MS", "3500"))
 
 
 @dataclass(slots=True)
 class HumanPaceConfig:
-    # Паузы между действиями для более естественного поведения.
+    # Паузы между действиями.
     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"))
@@ -63,7 +58,7 @@ class HumanPaceConfig:
 
 @dataclass(slots=True)
 class ListingConfig:
-    # Ограничения и режим обхода листинга.
+    # Лимиты листинга.
     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"))
@@ -74,32 +69,44 @@ class ListingConfig:
 
 @dataclass(slots=True)
 class DatabaseConfig:
-    # Параметры подключения к БД.
+    # Подключение к БД.
-    url: str = os.getenv("IAAI_DATABASE_URL", "sqlite:///iaai_scraper.db")
+    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:
+    # Подключение к Redis.
+    url: str = os.getenv("IAAI_REDIS_URL", "redis://localhost:6379/0")
+
+
+@dataclass(slots=True)
+class CeleryConfig:
+    # Настройки Celery.
+    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:
-    """Proxy settings for Playwright browser.
+    # Настройки прокси.
-
+    server: str | None = (os.getenv("IAAI_PROXY_SERVER") or "").strip() or None
-    Supports HTTP, HTTPS and SOCKS5 proxies.
+    username: str | None = (os.getenv("IAAI_PROXY_USERNAME") or "").strip() or None
-    Format: ``protocol://[user:password@]host:port``
+    password: str | None = (os.getenv("IAAI_PROXY_PASSWORD") or "").strip() or None
-
-    Examples:
-        - ``http://proxy.example.com:8080``
-        - ``socks5://user:pass@proxy.example.com:1080``
-    """
-    server: str | None = os.getenv("IAAI_PROXY_SERVER") or None
-    username: str | None = os.getenv("IAAI_PROXY_USERNAME") or None
-    password: str | None = os.getenv("IAAI_PROXY_PASSWORD") or None
 
     @property
     def enabled(self) -> bool:
         return bool(self.server)
 
     def to_playwright_dict(self) -> dict[str, str] | None:
-        """Return a dict suitable for Playwright ``proxy=`` kwarg, or *None*."""
+        # Формат для Playwright.
         if not self.server:
             return None
         result: dict[str, str] = {"server": self.server}
@@ -112,7 +119,7 @@ class ProxyConfig:
 
 @dataclass(slots=True)
 class Settings:
-    # Единая точка всех runtime-настроек приложения.
+    # Общие настройки.
     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"))
@@ -121,18 +128,21 @@ class Settings:
     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"}
-    raw_output_json: str | None = os.getenv("IAAI_RAW_OUTPUT_JSON") or None
     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)
-    gentle: GentleModeConfig = field(default_factory=GentleModeConfig)
+    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)
 
 
-# Глобальные настройки по умолчанию.
+# Настройки по умолчанию.
 settings = Settings()

iaai_scraper/core/logs.py

@@ -3,24 +3,20 @@ import sys
 from contextvars import ContextVar
 
 
-# Trace id хранится в контексте текущей операции.
 TRACE_ID: ContextVar[str] = ContextVar("trace_id", default="-")
 
 
 class TraceIdFilter(logging.Filter):
-    # Подмешиваем trace_id в каждую log record.
     def filter(self, record: logging.LogRecord) -> bool:
         record.trace_id = TRACE_ID.get()
         return True
 
 
 def set_trace_id(trace_id: str) -> None:
-    # Обновляем trace_id для текущего потока выполнения.
     TRACE_ID.set(trace_id)
 
 
 def setup_logging(level: str = "INFO", log_file: str | None = None) -> None:
-    # Базовая настройка консольного и файлового логирования.
     handlers: list[logging.Handler] = [logging.StreamHandler(sys.stdout)]
     if log_file:
         handlers.append(logging.FileHandler(log_file, encoding="utf-8"))

iaai_scraper/core/retry.py

@@ -9,7 +9,6 @@ from playwright.sync_api import Error, TimeoutError as PlaywrightTimeoutError
 
 logger = logging.getLogger("iaai_scraper.retry")
 
-# Ошибки, которые считаем временными и пригодными для повтора.
 RETRYABLE_EXCEPTIONS = (
     PlaywrightTimeoutError,
     Error,
@@ -25,7 +24,6 @@ def retryable(
     backoff_multiplier: float = 2.0,
     jitter_seconds: float = 0.0,
 ) -> Callable[[Callable[..., Any]], Callable[..., Any]]:
-    # Универсальный retry с backoff и небольшим случайным jitter.
     def decorator(func: Callable[..., Any]) -> Callable[..., Any]:
         @wraps(func)
         def wrapper(*args: Any, **kwargs: Any) -> Any:

iaai_scraper/core/utils.py

@@ -7,7 +7,9 @@ from typing import Any, Iterable
 
 
 def save_to_json(data: Any, filename: str | Path) -> None:
-    Path(filename).write_text(json.dumps(data, ensure_ascii=False, indent=2), encoding="utf-8")
+    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")
 
 
 def short_sleep(a: float = 0.10, b: float = 0.35) -> None:
@@ -29,7 +31,7 @@ def first_non_empty(values: Iterable[Any]) -> Any | None:
     return None
 
 
-# VIN, lot, price regex
+# Regex для VIN, lot, 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})?)")

iaai_scraper/parsing/__init__.py

@@ -1,2 +1 @@
-from .mapper import *  # noqa: F401,F403
+__all__: list[str] = []
-from .parser import *  # noqa: F401,F403

iaai_scraper/parsing/mapper.py

@@ -48,7 +48,7 @@ class CarMapper:
 	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 из summary и payload insights.
+		# Собираем нормализованную DB-модель.
 		vehicle_summary = vehicle_summary or {}
 		payload_insights = payload_insights or {}
 		notes: list[str] = []
@@ -62,9 +62,21 @@ class CarMapper:
 		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_int(first_non_empty([core.get("year"), vehicle_summary.get("year")]))
+		year = self._to_year(first_non_empty([core.get("year"), vehicle_summary.get("year")]))
-		price = self._to_int(first_non_empty([pricing.get("buy_now"), pricing.get("current_bid"), vehicle_summary.get("buy_now"), vehicle_summary.get("current_bid")]))
+		price = self._first_parsed_int(
-		mileage = self._to_int(first_non_empty([core.get("odometer"), vehicle_summary.get("odometer"), 0])) or 0
+			[
+				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")]))
@@ -83,7 +95,19 @@ class CarMapper:
 		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"), "USD"]))
+		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"
@@ -95,7 +119,7 @@ class CarMapper:
 			notes.append("No image URLs were found in the captured payloads.")
 
 		raw_attributes = {
-			# Здесь сохраняем полезный сырой контекст без жёсткой нормализации.
+			# 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")]),
@@ -126,7 +150,7 @@ class CarMapper:
 		}
 
 		content_hash = hashlib.sha256(json.dumps({
-			# Хеш нужен для пропуска записей без фактических изменений.
+			# 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,
@@ -162,8 +186,78 @@ class CarMapper:
 		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:
-		# Поддерживаем и литры, и уже готовые cc.
+		# Поддержка литров и cc.
 		text = str(value).lower().strip() if value is not None else ""
 		if not text:
 			return None
@@ -174,7 +268,24 @@ class CarMapper:
 		return int(text) if re.match(r"^\d+$", text) else None
 
 	def _normalize_currency(self, value: Any) -> str:
-		text = self._as_str(value).upper() or "USD"
+		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"
@@ -208,7 +319,7 @@ class CarMapper:
 		empty_default: str | None,
 		fallback: str | None,
 	) -> str | None:
-		# Общий helper для enum-нормализации по точному или частичному совпадению.
+		# Общий helper для enum-нормализации.
 		text = self._as_str(value).lower()
 		if not text:
 			return empty_default
@@ -243,7 +354,7 @@ class CarMapper:
 		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 оставляем ссылку с наибольшим размером.
+		# Для imageKeys берем самый большой размер.
 		best_by_key: dict[str, str] = {}
 		key_order: list[str] = []
 		non_keyed: list[str] = []
@@ -284,7 +395,7 @@ class CarMapper:
 		return url
 
 	def _build_origin_id(self, vehicle_url: str, vehicle_summary: dict[str, Any], core: dict[str, Any]) -> str:
-		# Предпочитаем lot_number, затем vin, затем хвост URL.
+		# 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:
@@ -299,6 +410,3 @@ class CarMapper:
 		return re.sub(r"[^a-z0-9]+", "-", value.lower()).strip("-") or "car"
 
 
-class IAAICarMapper(CarMapper):
-	"""Совместимое имя маппера."""
-

iaai_scraper/parsing/parser.py

@@ -74,7 +74,6 @@ class VehicleParser:
 	}
 
 	def _parse_dom_key_value_pairs(self, dom_text: str) -> dict[str, str]:
-		# Вытаскиваем пары label -> value из плоского текста страницы.
 		result: dict[str, str] = {}
 		if not dom_text:
 			return result
@@ -94,7 +93,6 @@ class VehicleParser:
 		return result
 
 	def _parse_title_for_year_make_model(self, page_title: str, dom_text: str) -> dict[str, str | None]:
-		# Title используется как запасной источник year/make/model.
 		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:
@@ -110,7 +108,6 @@ class VehicleParser:
 		return result
 
 	def normalize(self, vehicle_url: str, page_html: str, dom_text: str, network_dump: dict[str, Any]) -> dict[str, Any]:
-		# Собираем итоговую структуру из DOM, title, embedded JSON и network payloads.
 		page_html = page_html or ""
 		dom_text = dom_text or ""
 		network_dump = network_dump or {}
@@ -125,7 +122,6 @@ class VehicleParser:
 
 		summary: dict[str, Any] = {"source_url": vehicle_url}
 		for field, candidate_keys in self.SUMMARY_KEY_MAP.items():
-			# Для каждого поля собираем кандидатов из всех доступных источников.
 			values: list[Any] = []
 			for payload in payloads:
 				values.extend(deep_find_key(payload, candidate_keys))
@@ -160,7 +156,6 @@ class VehicleParser:
 			summary["current_bid"] = prices[1]
 
 		embedded = self._extract_embedded_json(page_html)
-		# Embedded JSON добирает поля, которых не было в DOM и XHR.
 		for item in embedded:
 			p = item.get("payload")
 			if isinstance(p, (dict, list)):
@@ -180,7 +175,6 @@ class VehicleParser:
 		}
 
 	def _build_payload_insights(self, summary: dict[str, Any], responses: list[dict[str, Any]], payloads: list[Any], vehicle_url: str = "") -> dict[str, Any]:
-		# Группируем сырой результат по смысловым блокам для маппера.
 		image_urls = self._extract_image_urls(payloads, "", vehicle_url)
 		return {
 			"vehicle_core": {
@@ -223,7 +217,6 @@ class VehicleParser:
 		}
 
 	def _build_source_endpoints(self, responses: list[dict[str, Any]]) -> dict[str, list[str]]:
-		# Раскладываем observed endpoints по категориям.
 		mapping = {"vehicle": [], "pricing": [], "bids": [], "damage": [], "auction": [], "images": []}
 		for item in responses:
 			url = item.get("url", "")
@@ -242,7 +235,6 @@ class VehicleParser:
 		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]:
-		# Короткие признаки того, что страница была доступна нормально.
 		endpoints = [item.get("url", "") for item in responses]
 		return {
 			"vin_visible": bool(summary.get("vin")),
@@ -289,7 +281,6 @@ class VehicleParser:
 
 	@staticmethod
 	def _extract_embedded_json(html: str) -> list[dict[str, Any]]:
-		# Ищем inline JSON в script-тегах.
 		scripts = re.findall(r"<script[^>]*>(.*?)</script>", html or "", flags=re.DOTALL | re.IGNORECASE)
 		extracted: list[dict[str, Any]] = []
 		for script_text in scripts:
@@ -304,7 +295,6 @@ class VehicleParser:
 
 	@staticmethod
 	def _extract_image_urls(payloads: list[Any], html: str, vehicle_url: str = "") -> list[str]:
-		# Собираем и дедуплицируем ссылки на изображения из JSON и HTML.
 		vehicle_key = ""
 		key_match = re.search(r"VehicleDetail/(\d+)", vehicle_url or "")
 		if key_match:
@@ -354,7 +344,6 @@ class VehicleParser:
 					seen_flat.add(cleaned)
 					flat.append(cleaned)
 		filtered: list[str] = []
-		# Отбрасываем служебные и заведомо нецелевые ссылки.
 		for url in flat:
 			lowered = url.lower()
 			if any(pat in lowered for pat in {"dimensions", "threesixty", "360view", ".js", ".css", ".svg", "/home/", "iframeview"}):
@@ -366,7 +355,6 @@ class VehicleParser:
 
 	@staticmethod
 	def _dom_hints(text: str) -> dict[str, Any]:
-		# Быстрые текстовые признаки полезных данных или anti-bot страницы.
 		lowered = (text or "").lower()
 		return {
 			"has_buy_now_text": "buy now" in lowered,

iaai_scraper/scraper.py

@@ -1,9 +1,8 @@
 import logging
+import re
 import signal
 import time
 import uuid
-from pathlib import Path
-from typing import Any
 
 from playwright.sync_api import Error as PlaywrightError
 from playwright.sync_api import BrowserContext, Page, sync_playwright
@@ -17,14 +16,22 @@ from .core.utils import save_to_json
 from .parsing.mapper import CarMapper
 from .parsing.parser import VehicleParser
 from .storage.db import PersistenceService
-from .storage.listing import ListingCollector
+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
@@ -138,7 +145,6 @@ class IAAIScraper:
                 page.wait_for_load_state("networkidle", timeout=15000)
             except PlaywrightTimeoutError:
                 pass
-            self._warm_page(page)
             self.pacer.after_vehicle_open()
 
             html = page.content()
@@ -150,7 +156,7 @@ class IAAIScraper:
             return {
                 "trace_id": trace_id,
                 "source_url": vehicle_url,
-                "opened_in_gentle_mode": True,
+                "opened_in_light_mode": True,
                 "fetched_at_epoch": int(time.time()),
                 "elapsed_seconds": round(time.perf_counter() - started_at, 3),
                 **parsed,
@@ -260,7 +266,14 @@ class IAAIScraper:
                 error_summary=error_summary,
             )
 
-    def sync_listing(self, make: str | None = None, model: str | None = None, lane: str = "iaai_cars", limit: int | None = None):
+    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")
@@ -271,6 +284,7 @@ class IAAIScraper:
         cars_failed = 0
         images_upserted = 0
         total = 0
+        skipped_existing = 0
         failures: list[dict[str, str]] = []
         listing: dict = {}
 
@@ -280,6 +294,27 @@ class IAAIScraper:
             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)
 
@@ -345,6 +380,7 @@ class IAAIScraper:
             "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,
         }
@@ -422,17 +458,3 @@ class IAAIScraper:
                     slept += chunk
 
         logger.info("Scheduler stopped gracefully after %d cycles.", cycle)
-
-    # helpers
-
-    def _warm_page(self, page: Page) -> None:
-        """Scroll для lazy-load."""
-        # Небольшой прогрев страницы для ленивых блоков и картинок.
-        rounds = max(0, self.settings.gentle.warm_scroll_rounds)
-        pause_seconds = max(0.1, self.settings.gentle.scroll_pause_ms / 1000)
-        for _ in range(rounds):
-            page.mouse.wheel(0, 1600)
-            time.sleep(pause_seconds)
-        if rounds:
-            page.mouse.wheel(0, -3000)
-            time.sleep(min(0.5, pause_seconds))

iaai_scraper/storage/db.py

@@ -4,48 +4,20 @@ from contextlib import contextmanager
 from datetime import datetime, timezone
 from typing import Iterator
 
-from sqlalchemy import create_engine, select
+from sqlalchemy import create_engine, or_, select
 from sqlalchemy.orm import Session, sessionmaker
 
 from ..core.config import Settings
-from .models import Base, Car, Image, SyncRun
+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 = {
-	"parser_id",
+	col.key for col in Car.__table__.columns
-	"brand",
+	if col.key not in ("id",)
-	"model",
-	"year",
-	"price",
-	"currency",
-	"mileage",
-	"country",
-	"is_sold",
-	"color",
-	"drive",
-	"gearbox",
-	"steering_wheel",
-	"body_type",
-	"engine_volume",
-	"selling_type",
-	"one_owner",
-	"new_car",
-	"is_hidden",
-	"origin",
-	"origin_url",
-	"origin_id",
-	"is_damaged",
-	"evaluation",
-	"non_smoking",
-	"rental",
-	"repair_history",
-	"slug",
-	"last_seen_at",
-	"content_hash",
-	"raw_attributes",
 }
 
 
@@ -54,11 +26,23 @@ class PersistenceService:
 	def __init__(self, settings: Settings) -> None:
 		# Инициализация engine и фабрики сессий.
 		self.settings = settings
-		self.engine = create_engine(settings.database.url, echo=settings.database.echo, future=True)
+		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]:
@@ -76,6 +60,16 @@ class PersistenceService:
 	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()
@@ -95,6 +89,22 @@ class PersistenceService:
 			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:
@@ -104,7 +114,7 @@ class PersistenceService:
 	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}
-		# Serialize raw_attributes dict to JSON string for Text column.
+	
 		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
@@ -116,8 +126,11 @@ class PersistenceService:
 		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:
-			# поиск по origin_id
+			# Сначала пытаемся найти по origin_id, а если ранее origin_id был неполный,
-			car = session.execute(select(Car).where(Car.origin_id == record.origin_id)).scalar_one_or_none()
+			# 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)
@@ -152,3 +165,52 @@ class PersistenceService:
 			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,
+			}

iaai_scraper/storage/enums.py

@@ -1,8 +1,8 @@
 CURRENCY_ENUM_VALUES = ("JPY", "USD", "EUR", "RUB", "KRW", "AED", "GBP", "CAD")
-DRIVE_ENUM_VALUES = ("FWD", "RWD", "TWO_WD", "FOUR_WD", "2WD", "4WD", "NA")
+DRIVE_ENUM_VALUES = ("FWD", "RWD", "2WD", "4WD", "NA")
 GEARBOX_ENUM_VALUES = ("AT", "CVT", "MT", "EV", "NA")
-STEERING_WHEEL_ENUM_VALUES = ("LEFT", "RIGHT", "left", "right", "NA")
+STEERING_WHEEL_ENUM_VALUES = ("LEFT", "RIGHT", "NA")
-BODY_TYPE_ENUM_VALUES = ("COUPE", "SUV", "HATCHBACK", "MINIVAN", "SEDAN", "NA", "Station Wagon", "Pickup", "Truck", "Open", "RV", "Other", "STATION_WAGON", "PICKUP", "TRUCK", "OPEN", "OTHER")
+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", "carsensor", "encar", "kuruma_trader", "asnet", "kababa", "ACV", "COPART", "copart", "NA", "ASNET", "KABABA", "IAAI")
+ORIGIN_ENUM_VALUES = ("TAU", "CARSENSOR", "HANAMARU", "ENCAR", "KURUMA_TRADER", "ASNET", "KABABA", "ACV", "COPART", "IAAI", "NA")
-SELLING_TYPE_ENUM_VALUES = ("STOCK", "AUCTION", "TENDER", "stock", "auction", "tender", "NA")
+SELLING_TYPE_ENUM_VALUES = ("STOCK", "AUCTION", "TENDER", "NA")

iaai_scraper/storage/models.py

@@ -16,14 +16,14 @@ from .enums import (
 
 
 class Base(DeclarativeBase):
-	# Базовый класс для всех ORM-моделей.
+	# База ORM.
 	pass
 
 
 class Car(Base):
-	# Основная сущность автомобиля в БД.
+	# Автомобиль.
 	__tablename__ = "cars"
-	id: Mapped[int] = mapped_column(Integer, primary_key=True, autoincrement=True)
+	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)
@@ -59,18 +59,18 @@ class Car(Base):
 
 
 class Image(Base):
-	# Изображения автомобиля, привязанные к записи Car.
+	# 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(Integer, ForeignKey("cars.id", ondelete="CASCADE"), 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):
-	# Служебная таблица для статистики запусков синхронизации.
+	# 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())
@@ -82,3 +82,18 @@ class SyncRun(Base):
 	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)

iaai_scraper/storage/schemas.py

@@ -1,18 +1,16 @@
 from datetime import datetime, timezone
 from typing import Any
 
-from pydantic import BaseModel, Field
+from pydantic import BaseModel, ConfigDict, Field
 
 
 class ImageRecord(BaseModel):
-	# Нормализованная схема одной картинки.
 	fullres_image: str
 	preview_image: str
 	order_index: int = 0
 
 
 class CarRecord(BaseModel):
-	# Основная Pydantic-схема машины перед записью в БД.
 	parser_id: str
 	brand: str
 	model: str
@@ -48,12 +46,33 @@ class CarRecord(BaseModel):
 	mapping_notes: list[str] = Field(default_factory=list)
 
 
-class ScrapeExport(BaseModel):
+class CarRead(BaseModel):
-	# Экспорт результата scrape для JSON-выгрузки.
+	# Ответ API по авто.
-	source_url: str
+	model_config = ConfigDict(from_attributes=True)
-	fetched_at_epoch: int
+
-	vehicle_summary: dict[str, Any] = Field(default_factory=dict)
+	id: int
-	payload_insights: dict[str, Any] = Field(default_factory=dict)
+	parser_id: str
-	db_record: CarRecord | None = None
+	brand: str
-	network: dict[str, Any] = Field(default_factory=dict)
+	model: str
-	access_notes: dict[str, Any] = Field(default_factory=dict)
+	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
+

iaai_scraper/worker/__init__.py

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

iaai_scraper/worker/celery_app.py

@@ -0,0 +1,60 @@
+"""Celery app factory."""
+
+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(
+    # Сериализация.
+    task_serializer="json",
+    accept_content=["json"],
+    result_serializer="json",
+    timezone="UTC",
+    enable_utc=True,
+
+    # Лимиты.
+    task_soft_time_limit=settings.celery.task_soft_time_limit,
+    task_time_limit=settings.celery.task_time_limit,
+    worker_concurrency=settings.celery.worker_concurrency,
+
+    # Playwright sync API: prefork + 1 процесс.
+    worker_pool="prefork",
+    worker_prefetch_multiplier=1,
+
+    # Retry policy брокера.
+    broker_connection_retry_on_startup=True,
+
+    # Beat schedule.
+    beat_schedule={
+        "periodic-sync-listing": {
+            "task": "iaai_scraper.worker.tasks.sync_listing_task",
+            "schedule": settings.celery.beat_sync_interval_minutes * 60.0,
+            "args": (),
+            "kwargs": {"limit": settings.celery.beat_sync_limit},
+            "options": {"queue": "scraping"},
+        },
+    },
+
+    # Роутинг задач.
+    task_routes={
+        "iaai_scraper.worker.tasks.*": {"queue": "scraping"},
+    },
+)
+
+# Автообнаружение задач.
+celery_app.autodiscover_tasks(["iaai_scraper.worker"])

iaai_scraper/worker/tasks.py

@@ -0,0 +1,137 @@
+"""Celery tasks для IAAI."""
+
+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"):
+    """Скрапинг и upsert одного автомобиля."""
+    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,
+):
+    """Полный цикл: листинг + sync всех найденных машин."""
+    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)

requirements.txt

@@ -2,6 +2,11 @@ playwright>=1.53.0
 python-dotenv>=1.0.1
 pydantic>=2.8.2
 SQLAlchemy>=2.0.32
-PySocks>=1.7.1
+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

tests/test_db.py

@@ -8,7 +8,7 @@ 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
+from iaai_scraper.storage.models import Car, Image, SyncRun
 from iaai_scraper.storage.schemas import CarRecord, ImageRecord
 
 
@@ -116,6 +116,36 @@ class TestPersistenceServiceIntegration(unittest.TestCase):
             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()

tests/test_listing.py

@@ -4,7 +4,7 @@ import unittest
 
 from iaai_scraper.browser.pace import HumanPacer
 from iaai_scraper.core.config import Settings
-from iaai_scraper.storage.listing import ListingCollector
+from iaai_scraper.browser.listing import ListingCollector
 
 
 class _FakePage:

tests/test_mappers.py

@@ -22,7 +22,6 @@ class TestCarMapper(unittest.TestCase):
             },
         )
 
-        # Длина hex-представления SHA-256
         self.assertEqual(len(record.content_hash), 64)
 
     def test_content_hash_changes_when_image_set_changes(self) -> None:
@@ -104,6 +103,37 @@ class TestCarMapper(unittest.TestCase):
         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()

tests/test_scraper.py

@@ -23,6 +23,7 @@ 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:
@@ -50,6 +51,8 @@ class TestScraperSync(unittest.TestCase):
         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()
@@ -73,6 +76,8 @@ class TestScraperSync(unittest.TestCase):
         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",
@@ -93,6 +98,8 @@ class TestScraperSync(unittest.TestCase):
         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())
@@ -101,6 +108,35 @@ class TestScraperSync(unittest.TestCase):
 
         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()