IAAI scraper

iaai_scraper/__init__.py

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

iaai_scraper/browser/__init__.py

@@ -0,0 +1,5 @@
+from .factory import BrowserFactory
+from .network import NetworkCapture
+from .pace import HumanPacer
+
+__all__ = ["BrowserFactory", "NetworkCapture", "HumanPacer"]

iaai_scraper/browser/factory.py

@@ -0,0 +1,115 @@
+import json
+import logging
+import random
+
+from playwright.sync_api import Browser, BrowserContext, Playwright
+
+from ..core.config import Settings
+
+logger = logging.getLogger("iaai_scraper.browser")
+
+
+def _build_init_script() -> str:
+    """JS-патч признаков автоматизации."""
+    hardware_concurrency = random.choice([4, 8, 12, 16])
+    device_memory = random.choice([4, 8, 16])
+    languages = ["en-US", "en"]
+
+    return f"""
+(() => {{
+  const define = (obj, prop, value) => {{
+    try {{
+      Object.defineProperty(obj, prop, {{ get: () => value, configurable: true }});
+    }} catch (e) {{}}
+  }};
+
+  define(navigator, 'webdriver', undefined);
+  define(navigator, 'platform', 'Win32');
+  define(navigator, 'vendor', 'Google Inc.');
+  define(navigator, 'language', '{languages[0]}');
+  define(navigator, 'languages', {json.dumps(languages)});
+  define(navigator, 'hardwareConcurrency', {hardware_concurrency});
+  define(navigator, 'deviceMemory', {device_memory});
+  define(navigator, 'maxTouchPoints', 0);
+
+  if (!window.chrome) {{
+    Object.defineProperty(window, 'chrome', {{
+      value: {{ runtime: {{}}, app: {{}}, csi: () => ({{}}), loadTimes: () => ({{}}) }},
+      configurable: true
+    }});
+  }}
+
+  const originalQuery = navigator.permissions && navigator.permissions.query;
+  if (originalQuery) {{
+    navigator.permissions.query = (params) => (
+      params && params.name === 'notifications'
+        ? Promise.resolve({{ state: Notification.permission }})
+        : originalQuery(params)
+    );
+  }}
+
+  const originalGetParameter = WebGLRenderingContext.prototype.getParameter;
+  WebGLRenderingContext.prototype.getParameter = function(parameter) {{
+    if (parameter === 37445) return 'Intel Inc.';
+    if (parameter === 37446) return 'Intel Iris OpenGL Engine';
+    return originalGetParameter.call(this, parameter);
+  }};
+}})();
+"""
+
+
+class BrowserFactory:
+
+    def __init__(self, settings: Settings) -> None:
+        self.settings = settings
+
+    def create_browser(self, playwright: Playwright) -> Browser:
+        # пробуем Chrome, если нет — Chromium
+        launch_kwargs = {
+            "headless": self.settings.headless,
+            "args": [
+                "--disable-blink-features=AutomationControlled",
+                "--no-default-browser-check",
+                "--disable-dev-shm-usage",
+                "--disable-features=IsolateOrigins,site-per-process",
+            ],
+        }
+        try:
+            logger.info("Trying to launch real Chrome channel")
+            return playwright.chromium.launch(channel="chrome", **launch_kwargs)
+        except Exception:
+            logger.warning("Chrome channel launch failed, falling back to Chromium")
+            return playwright.chromium.launch(**launch_kwargs)
+
+    def create_context(self, browser: Browser, storage_state: str | None = None) -> BrowserContext:
+        # рандом viewport/timezone под каждый контекст
+        viewport = random.choice(self.settings.fingerprint.viewport_presets)
+        timezone_id = random.choice(self.settings.fingerprint.timezone_candidates)
+        color_scheme = random.choice(["light", "dark"])
+
+        context = browser.new_context(
+            storage_state=storage_state or None,
+            user_agent=self.settings.fingerprint.user_agent,
+            viewport=viewport,
+            screen=viewport,
+            device_scale_factor=random.choice([1, 1.25]),
+            is_mobile=False,
+            has_touch=False,
+            locale=self.settings.fingerprint.locale,
+            timezone_id=timezone_id,
+            color_scheme=color_scheme,
+            java_script_enabled=True,
+            ignore_https_errors=False,
+            extra_http_headers={
+                "Accept-Language": "en-US,en;q=0.9",
+                "DNT": "1",
+                "Upgrade-Insecure-Requests": "1",
+                "Sec-CH-UA": self.settings.fingerprint.sec_ch_ua,
+                "Sec-CH-UA-Mobile": "?0",
+                "Sec-CH-UA-Platform": '"Windows"',
+            },
+        )
+        context.set_default_timeout(self.settings.default_timeout_ms)
+        context.set_default_navigation_timeout(self.settings.default_timeout_ms)
+        context.add_init_script(_build_init_script())
+        return context

iaai_scraper/browser/network.py

@@ -0,0 +1,120 @@
+import json
+import logging
+from dataclasses import dataclass, field
+from typing import Any
+from urllib.parse import urlparse
+
+from playwright.sync_api import Page, Request, Response
+
+from ..core.config import Settings
+
+logger = logging.getLogger("iaai_scraper.network")
+
+
+@dataclass
+class NetworkCapture:
+    """XHR/fetch перехватчик."""
+
+    settings: Settings
+    requests: list[dict[str, Any]] = field(default_factory=list)
+    json_responses: list[dict[str, Any]] = field(default_factory=list)
+    _seen_req: set[str] = field(default_factory=set)
+    _seen_resp: set[str] = field(default_factory=set)
+    _origin: str | None = None
+
+    def attach(self, page: Page) -> None:
+        try:
+            self._origin = urlparse(page.url).netloc.lower() or None
+        except Exception:
+            self._origin = None
+        page.on("request", self._on_request)
+        page.on("response", self._on_response)
+
+    def _is_same_origin(self, url: str) -> bool:
+        if not self.settings.gentle.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
+        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:
+            return
+        key = f"{request.method}:{request.url}:{request.post_data or ''}"
+        if key in self._seen_req:
+            return
+        self._seen_req.add(key)
+        self.requests.append({
+            "url": request.url, "method": request.method,
+            "resource_type": request.resource_type, "post_data": request.post_data,
+        })
+
+    def _on_response(self, response: Response) -> None:
+        # сохраняем json
+        request = response.request
+        if request.resource_type not in {"xhr", "fetch"}:
+            return
+        if not self._is_same_origin(response.url):
+            return
+        if len(self.json_responses) >= self.settings.gentle.max_json_responses:
+            return
+        if not self._is_json(response):
+            return
+        key = f"{request.method}:{response.url}:{response.status}"
+        if key in self._seen_resp:
+            return
+        self._seen_resp.add(key)
+        try:
+            payload = response.json()
+        except Exception:
+            try:
+                payload = json.loads(response.text())
+            except Exception:
+                return
+        self.json_responses.append({
+            "url": response.url, "status": response.status,
+            "request_method": request.method, "post_data": request.post_data,
+            "resource_type": request.resource_type, "payload": payload,
+            "category": self._categorize(response.url),
+        })
+
+    @staticmethod
+    def _is_json(response: Response) -> bool:
+        ct = (response.headers.get("content-type") or "").lower()
+        if "application/json" in ct or "+json" in ct:
+            return True
+        url = response.url.lower()
+        return any(m in url for m in ["/api/", "/graphql", "vehicledetail", "vehicle", "auction", "bid", "images", "media"])
+
+    @staticmethod
+    def _categorize(url: str) -> str:
+        low = url.lower()
+        mapping = {
+            "images": ["image", "media", "photos", "gallery"],
+            "bids": ["bid", "offer", "buy-now", "buynow"],
+            "auction": ["auction", "sale", "lane", "branch"],
+            "vehicle": ["vehicle", "detail", "vin", "damage", "runanddrive"],
+            "documents": ["title", "document", "report"],
+        }
+        for cat, markers in mapping.items():
+            if any(m in low for m in markers):
+                return cat
+        return "other"
+
+    def export(self) -> dict[str, Any]:
+        # полезные категории сначала, other в конец
+        prioritized = sorted(self.json_responses, key=lambda i: (i["category"] == "other", i["url"]))
+        return {
+            "requests": self.requests,
+            "json_responses": self.json_responses,
+            "prioritized_json_responses": prioritized,
+            "capture_limits": {
+                "same_origin_only": self.settings.gentle.capture_same_origin_only,
+                "max_requests": self.settings.gentle.max_requests,
+                "max_json_responses": self.settings.gentle.max_json_responses,
+            },
+        }

iaai_scraper/browser/pace.py

@@ -0,0 +1,46 @@
+import random
+import time
+
+from playwright.sync_api import Locator, Page
+
+from ..core.config import Settings
+
+
+class HumanPacer:
+    """Random паузы между действиями."""
+
+    def __init__(self, settings: Settings) -> None:
+        self.settings = settings
+
+    def pause(self, min_s: float, max_s: float) -> None:
+        if self.settings.pace.enabled:
+            time.sleep(random.uniform(min_s, max_s))
+
+    def after_listing_open(self) -> None:
+        self.pause(self.settings.pace.after_listing_open_min_s, self.settings.pace.after_listing_open_max_s)
+
+    def after_filter_action(self) -> None:
+        self.pause(self.settings.pace.after_filter_action_min_s, self.settings.pace.after_filter_action_max_s)
+
+    def before_vehicle_open(self) -> None:
+        self.pause(self.settings.pace.before_vehicle_open_min_s, self.settings.pace.before_vehicle_open_max_s)
+
+    def after_vehicle_open(self) -> None:
+        self.pause(self.settings.pace.after_vehicle_open_min_s, self.settings.pace.after_vehicle_open_max_s)
+
+    def between_vehicles(self) -> None:
+        self.pause(self.settings.pace.between_vehicles_min_s, self.settings.pace.between_vehicles_max_s)
+
+    def after_page_change(self) -> None:
+        self.pause(self.settings.pace.after_page_change_min_s, self.settings.pace.after_page_change_max_s)
+
+    def move_mouse_to(self, page: Page, locator: Locator) -> None:
+        try:
+            box = locator.bounding_box()
+        except Exception:
+            box = None
+        if not box:
+            return
+        x = box["x"] + min(box["width"] * 0.6, max(5, 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)))
+        page.mouse.move(x, y, steps=random.randint(8, 18))

iaai_scraper/cli.py

@@ -0,0 +1,104 @@
+import argparse
+from pathlib import Path
+
+from .core.utils import save_to_json
+from .scraper import IAAIScraper
+
+
+def build_parser() -> argparse.ArgumentParser:
+    parser = argparse.ArgumentParser(description="Public IAAI scraper")
+    subparsers = parser.add_subparsers(dest="command", required=True)
+
+    init_db_parser = subparsers.add_parser("init-db", help="Create local DB tables")
+    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")
+    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(
+        "open-vehicle",
+        help="Open a vehicle page in gentle mode and save only DOM-based hints",
+    )
+    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-vehicle",
+        help="Open a vehicle page and capture a limited set of likely useful JSON responses",
+    )
+    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(
+        "export-db-json",
+        help="Scrape a vehicle page and save only the DB-ready car record 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_vehicle_parser.add_argument("vehicle_url", help="IAAI vehicle detail URL")
+    sync_vehicle_parser.add_argument("--lane", default="iaai", help="Logical lane name for sync_runs")
+    sync_vehicle_parser.add_argument("--output", default="iaai_sync_vehicle.json", help="Path to output JSON")
+
+    sync_listing_parser = subparsers.add_parser(
+        "sync-listing",
+        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("--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:
+    parser = build_parser()
+    args = parser.parse_args()
+
+    if args.command == "run-daemon":
+        # daemon: бесконечный цикл
+        from .core.config import Settings
+        runtime_settings = 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() as scraper:
+        if args.command == "init-db":
+            data = scraper.init_db()
+        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)
+
+        save_to_json(data, Path(args.output))
+        print(f"Saved result to {Path(args.output).resolve()}")
+
+
+if __name__ == "__main__":
+    main()

iaai_scraper/core/__init__.py

@@ -0,0 +1,4 @@
+from .config import *  # noqa: F401,F403
+from .logs import *  # noqa: F401,F403
+from .retry import *  # noqa: F401,F403
+from .utils import *  # noqa: F401,F403

iaai_scraper/core/config.py

@@ -0,0 +1,95 @@
+import os
+from dataclasses import dataclass, field
+from pathlib import Path
+
+from dotenv import load_dotenv
+
+load_dotenv()
+
+
+@dataclass(slots=True)
+class FingerprintConfig:
+    user_agent: str = (
+        "Mozilla/5.0 (Windows NT 10.0; Win64; x64) "
+        "AppleWebKit/537.36 (KHTML, like Gecko) "
+        "Chrome/135.0.0.0 Safari/537.36"
+    )
+    viewport_presets: tuple[dict[str, int], ...] = (
+        {"width": 1920, "height": 1080},
+        {"width": 1600, "height": 900},
+        {"width": 1536, "height": 864},
+        {"width": 1440, "height": 900},
+        {"width": 1366, "height": 768},
+    )
+    timezone_candidates: tuple[str, ...] = (
+        "America/New_York",
+        "America/Chicago",
+        "America/Los_Angeles",
+    )
+    locale: str = "en-US"
+    sec_ch_ua: str = '"Google Chrome";v="135", "Chromium";v="135", "Not.A/Brand";v="24"'
+
+
+@dataclass(slots=True)
+class GentleModeConfig:
+    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:
+    enabled: bool = os.getenv("IAAI_HUMAN_PACE_ENABLED", "true").strip().lower() in {"1", "true", "yes", "on"}
+    after_listing_open_min_s: float = float(os.getenv("IAAI_AFTER_LISTING_OPEN_MIN_S", "2.5"))
+    after_listing_open_max_s: float = float(os.getenv("IAAI_AFTER_LISTING_OPEN_MAX_S", "4.5"))
+    after_filter_action_min_s: float = float(os.getenv("IAAI_AFTER_FILTER_ACTION_MIN_S", "2.0"))
+    after_filter_action_max_s: float = float(os.getenv("IAAI_AFTER_FILTER_ACTION_MAX_S", "4.0"))
+    before_vehicle_open_min_s: float = float(os.getenv("IAAI_BEFORE_VEHICLE_OPEN_MIN_S", "0.5"))
+    before_vehicle_open_max_s: float = float(os.getenv("IAAI_BEFORE_VEHICLE_OPEN_MAX_S", "1.5"))
+    after_vehicle_open_min_s: float = float(os.getenv("IAAI_AFTER_VEHICLE_OPEN_MIN_S", "0.3"))
+    after_vehicle_open_max_s: float = float(os.getenv("IAAI_AFTER_VEHICLE_OPEN_MAX_S", "1.0"))
+    between_vehicles_min_s: float = float(os.getenv("IAAI_BETWEEN_VEHICLES_MIN_S", "0.5"))
+    between_vehicles_max_s: float = float(os.getenv("IAAI_BETWEEN_VEHICLES_MAX_S", "1.5"))
+    after_page_change_min_s: float = float(os.getenv("IAAI_AFTER_PAGE_CHANGE_MIN_S", "3.0"))
+    after_page_change_max_s: float = float(os.getenv("IAAI_AFTER_PAGE_CHANGE_MAX_S", "6.0"))
+
+
+@dataclass(slots=True)
+class ListingConfig:
+    cars_url: str = os.getenv("IAAI_CARS_LISTING_URL", "https://www.iaai.com/Vehiclelisting/Cars")
+    max_pages_per_run: int = int(os.getenv("IAAI_MAX_PAGES_PER_RUN", "5"))
+    max_vehicles_per_run: int = int(os.getenv("IAAI_MAX_VEHICLES_PER_RUN", "100"))
+    page_link_limit: int = int(os.getenv("IAAI_PAGE_LINK_LIMIT", "200"))
+    include_pagination: bool = os.getenv("IAAI_INCLUDE_PAGINATION", "true").strip().lower() in {"1", "true", "yes", "on"}
+    collect_current_page_only: bool = os.getenv("IAAI_COLLECT_CURRENT_PAGE_ONLY", "false").strip().lower() in {"1", "true", "yes", "on"}
+
+
+@dataclass(slots=True)
+class DatabaseConfig:
+    url: str = os.getenv("IAAI_DATABASE_URL", "sqlite:///iaai_scraper.db")
+    echo: bool = os.getenv("IAAI_DATABASE_ECHO", "false").strip().lower() in {"1", "true", "yes", "on"}
+
+
+@dataclass(slots=True)
+class Settings:
+    home_url: str = "https://www.iaai.com/"
+    default_timeout_ms: int = int(os.getenv("IAAI_TIMEOUT_MS", "45000"))
+    network_settle_ms: int = int(os.getenv("IAAI_NETWORK_SETTLE_MS", "800"))
+    max_retries: int = int(os.getenv("IAAI_MAX_RETRIES", "3"))
+    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
+    scheduler_interval_minutes: int = int(os.getenv("IAAI_SCHEDULER_INTERVAL_MINUTES", "60"))
+    fingerprint: FingerprintConfig = field(default_factory=FingerprintConfig)
+    gentle: GentleModeConfig = field(default_factory=GentleModeConfig)
+    pace: HumanPaceConfig = field(default_factory=HumanPaceConfig)
+    listing: ListingConfig = field(default_factory=ListingConfig)
+    database: DatabaseConfig = field(default_factory=DatabaseConfig)
+
+
+settings = Settings()

iaai_scraper/core/logs.py

@@ -0,0 +1,14 @@
+import logging
+import sys
+
+
+def setup_logging(level: str = "INFO", log_file: str | None = None) -> None:
+    handlers: list[logging.Handler] = [logging.StreamHandler(sys.stdout)]
+    if log_file:
+        handlers.append(logging.FileHandler(log_file, encoding="utf-8"))
+    logging.basicConfig(
+        level=getattr(logging, level.upper(), logging.INFO),
+        format="%(asctime)s | %(levelname)s | %(name)s | %(message)s",
+        handlers=handlers,
+        force=True,
+    )

iaai_scraper/core/retry.py

@@ -0,0 +1,31 @@
+import logging
+import time
+from collections.abc import Callable
+from functools import wraps
+from typing import Any
+
+from playwright.sync_api import Error, TimeoutError as PlaywrightTimeoutError
+
+logger = logging.getLogger("iaai_scraper.retry")
+
+
+def retryable(max_attempts: int, delay_seconds: float = 2.5) -> Callable[[Callable[..., Any]], Callable[..., Any]]:
+    def decorator(func: Callable[..., Any]) -> Callable[..., Any]:
+        @wraps(func)
+        def wrapper(*args: Any, **kwargs: Any) -> Any:
+            last_error: Exception | None = None
+            for attempt in range(1, max_attempts + 1):
+                try:
+                    return func(*args, **kwargs)
+                except (PlaywrightTimeoutError, Error, ConnectionError, OSError, RuntimeError) as exc:
+                    last_error = exc
+                    logger.warning("%s failed on attempt %s/%s: %s", func.__name__, attempt, max_attempts, exc)
+                    if attempt < max_attempts:
+                        time.sleep(delay_seconds * (2 ** (attempt - 1)))
+            if last_error is not None:
+                raise last_error
+            raise RuntimeError("Retry wrapper failed without a captured exception")
+
+        return wrapper
+
+    return decorator

iaai_scraper/core/utils.py

@@ -0,0 +1,50 @@
+import json
+import random
+import re
+import time
+from pathlib import Path
+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")
+
+
+def short_sleep(a: float = 0.10, b: float = 0.35) -> None:
+    time.sleep(random.uniform(a, b))
+
+
+def mask_email(email: str) -> str:
+    if "@" not in email:
+        return "***"
+    local, domain = email.split("@", 1)
+    safe_local = local[:2] + "***" if len(local) > 2 else local[:1] + "*"
+    return f"{safe_local}@{domain}"
+
+
+def first_non_empty(values: Iterable[Any]) -> Any | None:
+    for value in values:
+        if value not in (None, "", [], {}, ()):
+            return value
+    return None
+
+
+# VIN, lot, price regex
+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})?)")
+
+
+def deep_find_key(obj, target_keys: set[str], max_depth: int = 64, _depth: int = 0) -> list:
+    found = []
+    if _depth >= max_depth:
+        return found
+    if isinstance(obj, dict):
+        for key, value in obj.items():
+            if key.lower() in target_keys:
+                found.append(value)
+            found.extend(deep_find_key(value, target_keys, max_depth=max_depth, _depth=_depth + 1))
+    elif isinstance(obj, list):
+        for item in obj:
+            found.extend(deep_find_key(item, target_keys, max_depth=max_depth, _depth=_depth + 1))
+    return found

iaai_scraper/parsing/__init__.py

@@ -0,0 +1,2 @@
+from .mapper import *  # noqa: F401,F403
+from .parser import *  # noqa: F401,F403

iaai_scraper/parsing/mapper.py

@@ -0,0 +1,309 @@
+import hashlib
+import json
+import re
+from datetime import datetime, timezone
+from typing import Any
+from urllib.parse import urlparse
+
+from ..core.utils import first_non_empty
+from ..storage.enums import (
+	BODY_TYPE_ENUM_VALUES,
+	COUNTRY_ENUM_VALUES,
+	CURRENCY_ENUM_VALUES,
+	DRIVE_ENUM_VALUES,
+	GEARBOX_ENUM_VALUES,
+	ORIGIN_ENUM_VALUES,
+	SELLING_TYPE_ENUM_VALUES,
+	STEERING_WHEEL_ENUM_VALUES,
+)
+from ..storage.schemas import CarRecord, ImageRecord
+
+
+class CarMapper:
+	"""IAAI data → CarRecord."""
+
+	BODY_MAP = {
+		"sedan": "SEDAN", "coupe": "COUPE", "hatchback": "HATCHBACK", "sport utility": "SUV",
+		"suv": "SUV", "wagon": "STATION_WAGON", "station wagon": "STATION_WAGON", "pickup": "PICKUP",
+		"pickup truck": "PICKUP", "crew cab": "PICKUP", "extended cab": "PICKUP", "regular cab": "PICKUP",
+		"quad cab": "PICKUP", "double cab": "PICKUP", "king cab": "PICKUP", "mega cab": "PICKUP",
+		"supercab": "PICKUP", "supercrew": "PICKUP", "truck": "TRUCK", "van": "MINIVAN",
+		"minivan": "MINIVAN", "convertible": "OPEN", "cabriolet": "OPEN", "rv": "RV", "crossover": "SUV",
+	}
+	DRIVE_MAP = {
+		"front wheel drive": "FWD", "fwd": "FWD", "rear wheel drive": "RWD", "rwd": "RWD",
+		"all wheel drive": "4WD", "awd": "4WD", "4x4": "4WD", "four wheel drive": "4WD",
+		"4wd": "4WD", "2wd": "2WD", "two wheel drive": "2WD",
+	}
+	GEARBOX_MAP = {
+		"automatic": "AT", "automatic transmission": "AT", "a/t": "AT", "aut": "AT",
+		"manual": "MT", "manual transmission": "MT", "m/t": "MT", "cvt": "CVT",
+		"continuously variable transmission": "CVT", "electric": "EV", "ev": "EV",
+	}
+	STEERING_MAP = {"left": "LEFT", "left hand drive": "LEFT", "right": "RIGHT", "right hand drive": "RIGHT"}
+	COUNTRY_MAP = {
+		"us": "US", "usa": "US", "united states": "US", "ca": "CA", "canada": "CA",
+		"jp": "JP", "japan": "JP", "kr": "KR", "korea": "KR", "south korea": "KR",
+	}
+	NO_DAMAGE_MARKERS = {"normal wear", "normal wear & tear", "normal wear and tear", "n/a", "na", "none", "no damage", "minor dents/scratches"}
+
+	def map_to_car_record(self, vehicle_url: str, vehicle_summary: dict[str, Any], payload_insights: dict[str, Any]) -> CarRecord:
+		vehicle_summary = vehicle_summary or {}
+		payload_insights = payload_insights or {}
+		notes: list[str] = []
+		core = payload_insights.get("vehicle_core", {})
+		pricing = payload_insights.get("pricing", {})
+		damage = payload_insights.get("damage", {})
+		auction = payload_insights.get("auction", {})
+		images = payload_insights.get("images", {})
+
+		origin_id = self._build_origin_id(vehicle_url, vehicle_summary, core)
+		parser_id = f"iaai:{origin_id}"
+		brand = self._as_str(first_non_empty([core.get("make"), vehicle_summary.get("make"), "UNKNOWN"]))
+		model = self._as_str(first_non_empty([core.get("model"), vehicle_summary.get("model"), "UNKNOWN"]))
+		year = self._to_int(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")]))
+		mileage = self._to_int(first_non_empty([core.get("odometer"), vehicle_summary.get("odometer"), 0])) or 0
+		color = self._normalize_color(first_non_empty([core.get("color"), vehicle_summary.get("color"), "other"]))
+		drive = self._normalize_drive(first_non_empty([core.get("drive"), vehicle_summary.get("drive")]))
+		gearbox = self._normalize_gearbox(first_non_empty([core.get("gearbox"), vehicle_summary.get("gearbox")]))
+		steering = self._normalize_steering(first_non_empty([core.get("steering_wheel"), vehicle_summary.get("steering_wheel")])) or "LEFT"
+		body_type = self._normalize_body_type(first_non_empty([core.get("body_type"), vehicle_summary.get("body_type")]))
+		engine_volume = self._to_engine_cc(first_non_empty([core.get("engine"), core.get("engine_volume"), vehicle_summary.get("engine"), vehicle_summary.get("engine_volume")]))
+		title_text = self._as_str(first_non_empty([core.get("title"), vehicle_summary.get("title"), ""]))
+		seller = self._as_str(first_non_empty([core.get("seller"), vehicle_summary.get("seller"), ""]))
+		location = self._as_str(first_non_empty([core.get("location"), vehicle_summary.get("location"), auction.get("branch"), ""]))
+		country = self._normalize_country(first_non_empty([core.get("country"), location, "US"]))
+		is_damaged = self._bool_damage(damage, vehicle_summary)
+		is_sold = self._bool_sold(auction)
+		one_owner = self._boolish(first_non_empty([core.get("one_owner"), vehicle_summary.get("one_owner"), False]))
+		new_car = self._boolish(first_non_empty([core.get("new_car"), vehicle_summary.get("new_car"), False]))
+		rental = self._boolish(first_non_empty([core.get("rental"), vehicle_summary.get("rental"), False]))
+		repair_history = self._boolish(first_non_empty([core.get("repair_history"), vehicle_summary.get("repair_history"), False]))
+		non_smoking = self._boolish(first_non_empty([core.get("non_smoking"), vehicle_summary.get("non_smoking"), True]))
+		evaluation = self._as_str(first_non_empty([core.get("grade"), core.get("evaluation"), vehicle_summary.get("evaluation")])) or None
+		currency = self._normalize_currency(first_non_empty([pricing.get("currency"), vehicle_summary.get("currency"), "USD"]))
+		slug = self._slugify(" ".join(filter(None, [str(year or ""), brand, model, origin_id])))
+		images_records = self._build_images(images.get("urls") or vehicle_summary.get("image_urls") or [])
+		origin = "IAAI" if "IAAI" in ORIGIN_ENUM_VALUES else "NA"
+		if not price:
+			notes.append("Price is missing or not parseable from the observed payloads.")
+		if not vehicle_summary.get("vin"):
+			notes.append("VIN was not observed in the accessible payloads for this account/session.")
+		if not images_records:
+			notes.append("No image URLs were found in the captured payloads.")
+
+		raw_attributes = {
+			"vin": vehicle_summary.get("vin"),
+			"lot_number": first_non_empty([core.get("lot_number"), vehicle_summary.get("lot_number")]),
+			"trim": first_non_empty([core.get("trim"), vehicle_summary.get("trim")]),
+			"fuel_type": first_non_empty([core.get("fuel_type"), vehicle_summary.get("fuel_type")]),
+			"cylinders": first_non_empty([core.get("cylinders"), vehicle_summary.get("cylinders")]),
+			"engine": first_non_empty([core.get("engine"), vehicle_summary.get("engine")]),
+			"manufactured_in": vehicle_summary.get("manufactured_in"),
+			"vehicle_class": vehicle_summary.get("vehicle_class"),
+			"run_and_drive": first_non_empty([core.get("run_and_drive"), vehicle_summary.get("run_and_drive")]),
+			"keys": first_non_empty([core.get("keys"), vehicle_summary.get("keys")]),
+			"title": title_text,
+			"title_brand": vehicle_summary.get("title_brand"),
+			"damage_primary": first_non_empty([damage.get("primary"), vehicle_summary.get("primary_damage")]),
+			"damage_secondary": damage.get("secondary"),
+			"damage_description": damage.get("description"),
+			"buy_now": first_non_empty([pricing.get("buy_now"), vehicle_summary.get("buy_now")]),
+			"current_bid": first_non_empty([pricing.get("current_bid"), vehicle_summary.get("current_bid")]),
+			"actual_cash_value": pricing.get("actual_cash_value"),
+			"estimated_repair_cost": pricing.get("estimated_repair_cost"),
+			"seller": seller,
+			"location": location,
+			"vehicle_location": vehicle_summary.get("vehicle_location"),
+			"auction_date": auction.get("auction_date"),
+			"lane": auction.get("lane"),
+			"branch": auction.get("branch"),
+			"sale_status": auction.get("sale_status"),
+			"source_endpoints": payload_insights.get("source_endpoints", {}),
+		}
+
+		content_hash = hashlib.sha256(json.dumps({
+			"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,
+			"image_count": len(images_records),
+		}, sort_keys=True, default=str).encode()).hexdigest()
+
+		return CarRecord(
+			parser_id=parser_id, brand=brand, model=model, year=year, price=price, currency=currency,
+			mileage=mileage, country=country, is_sold=is_sold, color=color, drive=drive, gearbox=gearbox,
+			steering_wheel=steering, body_type=body_type, engine_volume=engine_volume,
+			selling_type=self._normalize_selling_type("AUCTION"), one_owner=one_owner, new_car=new_car,
+			is_hidden=False, origin=origin, origin_url=vehicle_url, origin_id=origin_id, is_damaged=is_damaged,
+			evaluation=evaluation, non_smoking=non_smoking, rental=rental, repair_history=repair_history,
+			slug=slug, last_seen_at=datetime.now(timezone.utc), content_hash=content_hash, images=images_records,
+			raw_attributes=raw_attributes, mapping_notes=notes,
+		)
+
+	@staticmethod
+	def _as_str(value: Any) -> str:
+		return "" if value is None else str(value).strip()
+
+	@staticmethod
+	def _to_int(value: Any) -> int | None:
+		if value is None:
+			return None
+		if isinstance(value, bool):
+			return int(value)
+		if isinstance(value, (int, float)):
+			return int(value)
+		digits = re.sub(r"[^\d]", "", str(value))
+		return int(digits) if digits else None
+
+	def _to_engine_cc(self, value: Any) -> int | None:
+		text = str(value).lower().strip() if value is not None else ""
+		if not text:
+			return None
+		if liters_match := re.search(r"(\d+(?:\.\d+)?)\s*l", text):
+			return int(float(liters_match.group(1)) * 1000)
+		if cubic_match := re.search(r"(\d{3,5})\s*(?:cc|cm3|cubic)", text):
+			return int(cubic_match.group(1))
+		return int(text) if re.match(r"^\d+$", text) else None
+
+	def _normalize_currency(self, value: Any) -> str:
+		text = self._as_str(value).upper() or "USD"
+		if text in CURRENCY_ENUM_VALUES:
+			return text
+		return "USD" if "$" in str(value) else "USD"
+
+	def _normalize_drive(self, value: Any) -> str | None:
+		text = self._as_str(value).lower()
+		if not text:
+			return None
+		if (mapped := self.DRIVE_MAP.get(text)) and mapped in DRIVE_ENUM_VALUES:
+			return mapped
+		for marker, mapped in self.DRIVE_MAP.items():
+			if marker in text and mapped in DRIVE_ENUM_VALUES:
+				return mapped
+		return "NA"
+
+	def _normalize_gearbox(self, value: Any) -> str | None:
+		text = self._as_str(value).lower()
+		if not text:
+			return None
+		if (mapped := self.GEARBOX_MAP.get(text)) and mapped in GEARBOX_ENUM_VALUES:
+			return mapped
+		for marker, mapped in self.GEARBOX_MAP.items():
+			if marker in text and mapped in GEARBOX_ENUM_VALUES:
+				return mapped
+		return "NA"
+
+	def _normalize_steering(self, value: Any) -> str | None:
+		text = self._as_str(value).lower()
+		if not text:
+			return None
+		if (mapped := self.STEERING_MAP.get(text)) and mapped in STEERING_WHEEL_ENUM_VALUES:
+			return mapped
+		for marker, mapped in self.STEERING_MAP.items():
+			if marker in text and mapped in STEERING_WHEEL_ENUM_VALUES:
+				return mapped
+		return None
+
+	def _normalize_body_type(self, value: Any) -> str:
+		text = self._as_str(value).lower()
+		if not text:
+			return "OTHER"
+		for marker, mapped in self.BODY_MAP.items():
+			if marker in text and mapped in BODY_TYPE_ENUM_VALUES:
+				return mapped
+		return "OTHER"
+
+	def _normalize_country(self, value: Any) -> str:
+		text = self._as_str(value).lower()
+		if not text:
+			return "US"
+		for marker, mapped in self.COUNTRY_MAP.items():
+			if marker in text and mapped in COUNTRY_ENUM_VALUES:
+				return mapped
+		return "US" if "US" in COUNTRY_ENUM_VALUES else "NA"
+
+	def _normalize_selling_type(self, value: Any) -> str:
+		text = self._as_str(value) or "AUCTION"
+		return text if text in SELLING_TYPE_ENUM_VALUES else "AUCTION"
+
+	@staticmethod
+	def _normalize_color(value: Any) -> str:
+		text = str(value).strip() if value is not None else "other"
+		if not text:
+			return "other"
+		if "/" in text:
+			text = text.split("/")[0].strip()
+		return text.lower() or "other"
+
+	@staticmethod
+	def _boolish(value: Any) -> bool:
+		return value if isinstance(value, bool) else str(value).strip().lower() in {"1", "true", "yes", "y", "owner", "one owner", "new"}
+
+	def _bool_damage(self, damage: dict[str, Any], vehicle_summary: dict[str, Any]) -> bool:
+		for value in [damage.get("primary"), damage.get("secondary"), damage.get("description"), vehicle_summary.get("primary_damage")]:
+			text = self._as_str(value).lower()
+			if text and text not in self.NO_DAMAGE_MARKERS:
+				return True
+		return False
+
+	def _bool_sold(self, auction: dict[str, Any]) -> bool:
+		return any(token in self._as_str(auction.get("sale_status")).lower() for token in ["sold", "closed", "ended"])
+
+	def _build_images(self, urls: list[Any]) -> list[ImageRecord]:
+		best_by_key: dict[str, str] = {}
+		key_order: list[str] = []
+		non_keyed: list[str] = []
+		for item in urls:
+			if not isinstance(item, str):
+				continue
+			url = item.strip()
+			if not url:
+				continue
+			if m := re.search(r'imageKeys=([^&]+)', url):
+				img_key = m.group(1)
+				w = int(re.search(r'width=(\d+)', url).group(1)) if re.search(r'width=(\d+)', url) else 0
+				existing = best_by_key.get(img_key)
+				if existing is None:
+					best_by_key[img_key] = url
+					key_order.append(img_key)
+				else:
+					existing_w = int(re.search(r'width=(\d+)', existing).group(1)) if re.search(r'width=(\d+)', existing) else 0
+					if w > existing_w:
+						best_by_key[img_key] = url
+			elif url not in non_keyed:
+				non_keyed.append(url)
+		deduped = [best_by_key[k] for k in key_order] + non_keyed
+		return [ImageRecord(fullres_image=self._make_fullres_url(url), preview_image=self._make_preview_url(url), order_index=index) for index, url in enumerate(deduped)]
+
+	@staticmethod
+	def _make_fullres_url(url: str) -> str:
+		if "vis.iaai.com" in url:
+			return re.sub(r'height=\d+', 'height=633', re.sub(r'width=\d+', 'width=845', url))
+		return url
+
+	@staticmethod
+	def _make_preview_url(url: str) -> str:
+		if "vis.iaai.com" in url:
+			preview = re.sub(r'height=\d+', 'height=300', re.sub(r'width=\d+', 'width=400', url))
+			if preview != url:
+				return preview
+		return url
+
+	def _build_origin_id(self, vehicle_url: str, vehicle_summary: dict[str, Any], core: dict[str, Any]) -> str:
+		for value in [core.get("lot_number"), vehicle_summary.get("lot_number"), vehicle_summary.get("vin")]:
+			text = self._as_str(value)
+			if text:
+				return text
+		tail = urlparse(vehicle_url).path.rstrip("/").split("/")[-1]
+		if "~" in tail:
+			tail = tail.split("~")[0]
+		return tail or self._slugify(vehicle_url)
+
+	@staticmethod
+	def _slugify(value: str) -> str:
+		return re.sub(r"[^a-z0-9]+", "-", value.lower()).strip("-") or "car"
+
+
+class IAAICarMapper(CarMapper):
+	"""Совместимое имя маппера."""

iaai_scraper/parsing/parser.py

@@ -0,0 +1,358 @@
+import html as html_module
+import json
+import logging
+import re
+from typing import Any
+
+from ..core.utils import LOT_RE, PRICE_RE, VIN_RE, deep_find_key, first_non_empty
+
+logger = logging.getLogger("iaai_scraper.parsers")
+
+
+class VehicleParser:
+	"""DOM + JSON парсер страницы авто."""
+
+	SUMMARY_KEY_MAP = {
+		"vin": {"vin", "vehicleidentificationnumber"},
+		"lot_number": {"lotnumber", "stockno", "itemid", "lotid", "itemnumber"},
+		"year": {"year"},
+		"make": {"make", "manufacturer", "brand"},
+		"model": {"model"},
+		"trim": {"trim", "series"},
+		"odometer": {"odometer", "odometermiles", "mileage", "actualcashvalueodometer"},
+		"primary_damage": {"primarydamage", "damage", "damagetype", "loss"},
+		"secondary_damage": {"secondarydamage"},
+		"run_and_drive": {"runanddrive", "canrunanddrive", "rundrive"},
+		"buy_now": {"buynowprice", "buyitnowprice", "instantpurchaseprice"},
+		"current_bid": {"currentbid", "highbid", "bidamount", "currenthighbid"},
+		"actual_cash_value": {"actualcashvalue", "acv"},
+		"estimated_repair_cost": {"estimatedrepaircost", "repaircost"},
+		"keys": {"keys", "keystatus"},
+		"title": {"titletype", "title", "documenttype"},
+		"seller": {"seller", "sellername"},
+		"location": {"location", "branchname", "auctionlocation", "branch"},
+		"auction_date": {"auctiondate", "saledate", "liveauctiondate"},
+		"body_type": {"bodytype", "bodystyle"},
+		"drive": {"driveline", "drive"},
+		"gearbox": {"transmission", "gearbox"},
+		"engine": {"engine", "enginevolume"},
+		"fuel_type": {"fueltype", "fuel"},
+		"cylinders": {"cylinders", "cylindercount"},
+		"color": {"color", "primarycolor", "exteriorcolor"},
+	}
+
+	DOM_LABEL_MAP: dict[str, str] = {
+		"stock #": "lot_number",
+		"vin (status)": "vin",
+		"vin": "vin",
+		"primary damage": "primary_damage",
+		"secondary damage": "secondary_damage",
+		"odometer": "odometer",
+		"body style": "body_type",
+		"engine": "engine",
+		"transmission": "gearbox",
+		"drive line type": "drive",
+		"fuel type": "fuel_type",
+		"cylinders": "cylinders",
+		"exterior/interior": "color",
+		"exterior color": "color",
+		"model": "model",
+		"series": "trim",
+		"selling branch": "location",
+		"vehicle location": "vehicle_location",
+		"auction date and time": "auction_date",
+		"lane/run #": "lane",
+		"actual cash value": "actual_cash_value",
+		"estimated repair cost": "estimated_repair_cost",
+		"seller": "seller",
+		"title/sale doc": "title",
+		"title/sale doc brand": "title_brand",
+		"start code": "run_and_drive",
+		"key": "keys",
+		"manufactured in": "manufactured_in",
+		"vehicle class": "vehicle_class",
+	}
+
+	def _parse_dom_key_value_pairs(self, dom_text: str) -> dict[str, str]:
+		result: dict[str, str] = {}
+		if not dom_text:
+			return result
+		lines = [line.strip() for line in dom_text.split("\n") if line.strip()]
+		known_labels = set(self.DOM_LABEL_MAP.keys())
+		for i, line in enumerate(lines):
+			clean = line.rstrip(":").lower().strip()
+			if clean in known_labels and i + 1 < len(lines):
+				value = lines[i + 1].strip()
+				if value.rstrip(":").lower().strip() in known_labels:
+					continue
+				if value.lower() in ("more actions", "view", "print", "share", "back to results"):
+					continue
+				field_name = self.DOM_LABEL_MAP[clean]
+				if field_name not in result or not result[field_name]:
+					result[field_name] = value
+		return result
+
+	def _parse_title_for_year_make_model(self, page_title: str, dom_text: str) -> dict[str, str | None]:
+		result: dict[str, str | None] = {"year": None, "make": None, "model": None}
+		title_match = re.match(r"(\d{4})\s+(\S+)\s+(.+?)(?:\s+for\s+)", page_title or "")
+		if title_match:
+			result["year"] = title_match.group(1)
+			result["make"] = title_match.group(2)
+			result["model"] = title_match.group(3)
+			return result
+		dom_match = re.search(r"(?:Search|Log In)\s*\n\s*(\d{4})\s+(\S+)\s+(.+?)(?:\n|$)", dom_text or "")
+		if dom_match:
+			result["year"] = dom_match.group(1)
+			result["make"] = dom_match.group(2)
+			result["model"] = dom_match.group(3).strip()
+		return result
+
+	def normalize(self, vehicle_url: str, page_html: str, dom_text: str, network_dump: dict[str, Any]) -> dict[str, Any]:
+		page_html = page_html or ""
+		dom_text = dom_text or ""
+		network_dump = network_dump or {}
+		responses = network_dump.get("json_responses", [])
+		payloads = [item.get("payload") for item in responses if isinstance(item.get("payload"), (dict, list))]
+		dom_kv = self._parse_dom_key_value_pairs(dom_text)
+		page_title = ""
+		title_match = re.search(r"<title[^>]*>(.*?)</title>", page_html or "", re.IGNORECASE | re.DOTALL)
+		if title_match:
+			page_title = title_match.group(1).strip()
+		title_parsed = self._parse_title_for_year_make_model(page_title, dom_text)
+
+		summary: dict[str, Any] = {"source_url": vehicle_url}
+		for field, candidate_keys in self.SUMMARY_KEY_MAP.items():
+			values: list[Any] = []
+			for payload in payloads:
+				values.extend(deep_find_key(payload, candidate_keys))
+			if field in dom_kv:
+				values.append(dom_kv[field])
+			summary[field] = first_non_empty(values)
+
+		summary["year"] = summary.get("year") or title_parsed.get("year")
+		summary["make"] = summary.get("make") or title_parsed.get("make")
+		summary["model"] = summary.get("model") or title_parsed.get("model")
+		summary["trim"] = summary.get("trim") or dom_kv.get("trim")
+		summary["vin"] = summary.get("vin") or self._extract_vin(dom_text) or self._extract_vin(page_html)
+		if summary.get("vin") and isinstance(summary["vin"], str):
+			vin_clean = re.sub(r"\s*\(.*?\)\s*$", "", summary["vin"]).strip()
+			vin_match = VIN_RE.search(vin_clean)
+			summary["vin"] = vin_match.group(1) if vin_match else vin_clean
+		summary["lot_number"] = summary.get("lot_number") or self._extract_lot_number(dom_text)
+		summary["image_urls"] = self._extract_image_urls(payloads, page_html, vehicle_url)
+		for dom_field, dom_value in dom_kv.items():
+			if dom_field not in summary or not summary[dom_field]:
+				summary[dom_field] = dom_value
+		prices = self._extract_prices_from_text(dom_text)
+		if not summary.get("actual_cash_value") and prices:
+			summary["actual_cash_value"] = prices[0]
+		if not summary.get("buy_now"):
+			buy_now_match = re.search(r"Buy\s+Now[:\s]*\$\s*([\d,]+(?:\.\d{1,2})?)", dom_text or "", re.IGNORECASE)
+			if buy_now_match:
+				summary["buy_now"] = buy_now_match.group(1)
+			elif prices:
+				summary["buy_now"] = prices[0]
+		if not summary.get("current_bid") and len(prices) > 1:
+			summary["current_bid"] = prices[1]
+
+		embedded = self._extract_embedded_json(page_html)
+		for item in embedded:
+			p = item.get("payload")
+			if isinstance(p, (dict, list)):
+				payloads.append(p)
+				for field, candidate_keys in self.SUMMARY_KEY_MAP.items():
+					if not summary.get(field):
+						val = first_non_empty(deep_find_key(p, candidate_keys))
+						if val:
+							summary[field] = val
+
+		return {
+			"vehicle_summary": summary,
+			"payload_insights": self._build_payload_insights(summary, responses, payloads, vehicle_url),
+			"embedded_json": embedded,
+			"dom_hints": self._dom_hints(dom_text),
+			"access_notes": self._build_access_notes(summary, responses),
+		}
+
+	def _build_payload_insights(self, summary: dict[str, Any], responses: list[dict[str, Any]], payloads: list[Any], vehicle_url: str = "") -> dict[str, Any]:
+		image_urls = self._extract_image_urls(payloads, "", vehicle_url)
+		return {
+			"vehicle_core": {
+				"vin": summary.get("vin"), "lot_number": summary.get("lot_number"), "year": summary.get("year"),
+				"make": summary.get("make"), "model": summary.get("model"), "trim": summary.get("trim"),
+				"odometer": summary.get("odometer"), "run_and_drive": summary.get("run_and_drive"),
+				"seller": summary.get("seller"), "location": summary.get("location"), "title": summary.get("title"),
+				"body_type": summary.get("body_type"), "drive": summary.get("drive"), "gearbox": summary.get("gearbox"),
+				"engine": summary.get("engine"), "fuel_type": summary.get("fuel_type"), "cylinders": summary.get("cylinders"),
+				"color": summary.get("color"), "keys": summary.get("keys"),
+			},
+			"pricing": {
+				"buy_now": summary.get("buy_now"), "current_bid": summary.get("current_bid"),
+				"actual_cash_value": summary.get("actual_cash_value"), "estimated_repair_cost": summary.get("estimated_repair_cost"),
+				"currency": self._guess_currency(summary),
+			},
+			"bids": self._build_bid_insights(summary, payloads),
+			"damage": {
+				"primary": summary.get("primary_damage"),
+				"secondary": summary.get("secondary_damage"),
+				"description": first_non_empty(self._find_in_payloads(payloads, {"damageDescription", "damageDetails"})),
+			},
+			"auction": {
+				"auction_date": summary.get("auction_date"),
+				"lane": first_non_empty(self._find_in_payloads(payloads, {"lane", "lanename"})),
+				"branch": first_non_empty([summary.get("location"), *self._find_in_payloads(payloads, {"branch", "branchname"})]),
+				"sale_status": first_non_empty(self._find_in_payloads(payloads, {"salestatus", "auctionstatus", "status"})),
+				"item_number": first_non_empty([summary.get("lot_number"), *self._find_in_payloads(payloads, {"itemnumber", "lotnumber", "lotid"})]),
+			},
+			"images": {"count": len(image_urls), "urls": image_urls},
+			"source_endpoints": self._build_source_endpoints(responses),
+		}
+
+	def _build_bid_insights(self, summary: dict[str, Any], payloads: list[Any]) -> dict[str, Any]:
+		return {
+			"amount": summary.get("current_bid"),
+			"currency": self._guess_currency(summary),
+			"bid_count": first_non_empty(self._find_in_payloads(payloads, {"bidcount", "numberofbids"})),
+			"status": first_non_empty(self._find_in_payloads(payloads, {"bidstatus", "biddingstatus"})),
+		}
+
+	def _build_source_endpoints(self, responses: list[dict[str, Any]]) -> dict[str, list[str]]:
+		mapping = {"vehicle": [], "pricing": [], "bids": [], "damage": [], "auction": [], "images": []}
+		for item in responses:
+			url = item.get("url", "")
+			category = item.get("category", "other")
+			if category == "vehicle":
+				mapping["vehicle"].append(url)
+				lowered = url.lower()
+				if any(token in lowered for token in ["bid", "offer"]):
+					mapping["bids"].append(url)
+				if any(token in lowered for token in ["damage", "report"]):
+					mapping["damage"].append(url)
+				if any(token in lowered for token in ["auction", "sale", "lane", "branch"]):
+					mapping["auction"].append(url)
+			elif category in mapping:
+				mapping[category].append(url)
+		return {key: list(dict.fromkeys(urls)) for key, urls in mapping.items()}
+
+	def _build_access_notes(self, summary: dict[str, Any], responses: list[dict[str, Any]]) -> dict[str, Any]:
+		endpoints = [item.get("url", "") for item in responses]
+		return {
+			"vin_visible": bool(summary.get("vin")),
+			"images_visible": bool(summary.get("image_urls")),
+			"network_json_count": len(responses),
+
+			"observed_endpoints": endpoints[:20],
+		}
+
+	@staticmethod
+	def _find_in_payloads(payloads: list[Any], keys: set[str]) -> list[Any]:
+		lowered = {key.lower() for key in keys}
+		values: list[Any] = []
+		for payload in payloads:
+			values.extend(deep_find_key(payload, lowered))
+		return values
+
+	@staticmethod
+	def _guess_currency(summary: dict[str, Any]) -> str:
+		for key in ["buy_now", "current_bid", "actual_cash_value", "estimated_repair_cost"]:
+			value = str(summary.get(key) or "")
+			if "$" in value:
+				return "USD"
+			if "€" in value:
+				return "EUR"
+			if "¥" in value:
+				return "JPY"
+		return "USD"
+
+	@staticmethod
+	def _extract_vin(text: str) -> str | None:
+		match = VIN_RE.search(text or "")
+		return match.group(1) if match else None
+
+	@staticmethod
+	def _extract_lot_number(text: str) -> str | None:
+		match = LOT_RE.search(text or "")
+		return match.group(1) if match else None
+
+	@staticmethod
+	def _extract_prices_from_text(text: str) -> list[str]:
+		return [match.group(1) for match in PRICE_RE.finditer(text or "")]
+
+	@staticmethod
+	def _extract_embedded_json(html: str) -> list[dict[str, Any]]:
+		scripts = re.findall(r"<script[^>]*>(.*?)</script>", html or "", flags=re.DOTALL | re.IGNORECASE)
+		extracted: list[dict[str, Any]] = []
+		for script_text in scripts:
+			if "{" not in script_text and "[" not in script_text:
+				continue
+			try:
+				parsed = json.loads(script_text.strip())
+			except Exception:
+				continue
+			extracted.append({"type": "inline_json", "payload": parsed})
+		return extracted
+
+	@staticmethod
+	def _extract_image_urls(payloads: list[Any], html: str, vehicle_url: str = "") -> list[str]:
+		vehicle_key = ""
+		key_match = re.search(r"VehicleDetail/(\d+)", vehicle_url or "")
+		if key_match:
+			vehicle_key = key_match.group(1)
+		found: list[str] = []
+		for payload in payloads:
+			found.extend(deep_find_key(payload, {"imageurl", "imageurls", "url", "fullsizeurl", "thumbnailurl", "originalurl"}))
+		flat: list[str] = []
+		for item in found:
+			if isinstance(item, str) and item.startswith("http"):
+				cleaned = html_module.unescape(item)
+				lowered = cleaned.lower()
+				if vehicle_key and "vis.iaai.com" in lowered and vehicle_key not in cleaned:
+					continue
+				if cleaned not in flat:
+					flat.append(cleaned)
+			elif isinstance(item, list):
+				for child in item:
+					if isinstance(child, str) and child.startswith("http"):
+						cleaned = html_module.unescape(child)
+						lowered = cleaned.lower()
+						if vehicle_key and "vis.iaai.com" in lowered and vehicle_key not in cleaned:
+							continue
+						if cleaned not in flat:
+							flat.append(cleaned)
+		for pattern in [r'<img[^>]+(?:src|data-src)\s*=\s*["\']([^"\']+)["\']', r'data-src\s*=\s*["\']([^"\']+)["\']']:
+			for match in re.finditer(pattern, html or "", re.IGNORECASE):
+				url = html_module.unescape(match.group(1).strip())
+				if not url.startswith("http") or url in flat:
+					continue
+				lowered = url.lower()
+				if vehicle_key and vehicle_key in url:
+					flat.append(url)
+				elif any(token in lowered for token in ["vis.iaai.com", "anvis", "vehicleimage"]):
+					if vehicle_key and vehicle_key not in url:
+						continue
+					flat.append(url)
+		if vehicle_key:
+			for url in re.findall(r'https?://vis\.iaai\.com[^\s"\'<>]+', html or ""):
+				cleaned = html_module.unescape(url)
+				if cleaned not in flat and vehicle_key in cleaned:
+					flat.append(cleaned)
+		filtered: list[str] = []
+		for url in flat:
+			lowered = url.lower()
+			if any(pat in lowered for pat in {"dimensions", "threesixty", "360view", ".js", ".css", ".svg", "/home/", "iframeview"}):
+				continue
+			if "vis.iaai.com" in lowered and "/resizer" not in lowered:
+				continue
+			filtered.append(url)
+		return filtered
+
+	@staticmethod
+	def _dom_hints(text: str) -> dict[str, Any]:
+		lowered = (text or "").lower()
+		return {
+			"has_buy_now_text": "buy now" in lowered,
+			"has_run_and_drive_text": "run & drive" in lowered or "run and drive" in lowered,
+			"has_damage_text": "damage" in lowered,
+			"has_vin_text": "vin" in lowered,
+			"has_title_text": "title" in lowered,
+		}

iaai_scraper/scraper.py

@@ -0,0 +1,342 @@
+import logging
+import time
+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
+from playwright.sync_api import TimeoutError as PlaywrightTimeoutError
+
+from .browser import BrowserFactory, HumanPacer, NetworkCapture
+from .core.config import Settings, settings
+from .core.logs import setup_logging
+from .core.retry import retryable
+from .core.utils import save_to_json
+from .parsing.mapper import CarMapper
+from .parsing.parser import VehicleParser
+from .storage.db import PersistenceService
+from .storage.listing import ListingCollector
+from .storage.schemas import CarRecord
+
+logger = logging.getLogger("iaai_scraper.scraper")
+
+
+class IAAIScraper:
+
+    def __init__(self, runtime_settings: Settings | None = None) -> None:
+        self.settings = runtime_settings or settings
+        setup_logging(self.settings.log_level, self.settings.log_file)
+        self.playwright = None
+        self.browser = None
+        self.context: BrowserContext | None = None
+        self.browser_factory = BrowserFactory(self.settings)
+        self.pacer = HumanPacer(self.settings)
+        self.listing_collector = ListingCollector(self.settings, self.pacer)
+        self.vehicle_parser = VehicleParser()
+        self.car_mapper = CarMapper()
+        self.persistence = PersistenceService(self.settings)
+
+    # browser lifecycle
+
+    def __enter__(self) -> "IAAIScraper":
+        self.playwright = sync_playwright().start()
+        self.browser = self.browser_factory.create_browser(self.playwright)
+        return self
+
+    def __exit__(self, exc_type, exc, tb) -> None:
+        if self.context:
+            self.context.close()
+        if self.browser:
+            self.browser.close()
+        if self.playwright:
+            self.playwright.stop()
+
+    def _new_context(self, storage_state: str | None = None) -> BrowserContext:
+        if self.context:
+            self.context.close()
+        self.context = self.browser_factory.create_context(self.browser, storage_state=storage_state)
+        return self.context
+
+    def init_db(self):
+        self.persistence.create_tables()
+        return {"status": "ok", "database_url": self.settings.database.url}
+
+    def _get_unauthenticated_page(self) -> Page:
+        context = self._new_context()
+        return context.new_page()
+
+    # listing
+
+    def collect_listing(self, make: str | None = None, model: str | None = None):
+        page = self._get_unauthenticated_page()
+
+        try:
+            listing = self.listing_collector.collect_listing_links(page, make=make, model=model)
+        finally:
+            page.close()
+
+        return {
+            "status": "ok",
+            **listing,
+        }
+
+    def _get_page(self) -> Page:
+        if not self.context:
+            self._new_context()
+        return self.context.new_page()
+
+    # scrape
+
+    @retryable(max_attempts=3)
+    def open_vehicle_page(self, vehicle_url: str):
+        page = self._get_page()
+        try:
+            self.pacer.before_vehicle_open()
+            page.goto(vehicle_url, wait_until="domcontentloaded", timeout=60_000)
+            try:
+                page.wait_for_load_state("networkidle", timeout=15000)
+            except PlaywrightTimeoutError:
+                pass
+            self._warm_page(page)
+            self.pacer.after_vehicle_open()
+
+            html = page.content()
+            try:
+                dom_text = page.locator("body").inner_text(timeout=10_000)
+            except Exception:
+                dom_text = ""
+            parsed = self.vehicle_parser.normalize(vehicle_url, html, dom_text, {"json_responses": []})
+            return {
+                "source_url": vehicle_url,
+                "opened_in_gentle_mode": True,
+                "fetched_at_epoch": int(time.time()),
+                **parsed,
+            }
+        finally:
+            page.close()
+
+    @retryable(max_attempts=3)
+    def scrape_vehicle_detail(self, vehicle_url: str):
+        """Открыть страницу, перехватить JSON, вернуть данные."""
+        page = self._get_page()
+        try:
+            return self._scrape_on_page(page, vehicle_url)
+        finally:
+            page.close()
+
+    def _scrape_on_page(self, page: Page, vehicle_url: str):
+        capture = NetworkCapture(self.settings)
+        capture.attach(page)
+
+        page.goto(vehicle_url, wait_until="domcontentloaded", timeout=60_000)
+        try:
+            page.wait_for_load_state("networkidle", timeout=10000)
+        except PlaywrightTimeoutError:
+            pass
+        time.sleep(self.settings.network_settle_ms / 1000)
+
+        html = page.content()
+        try:
+            dom_text = page.locator("body").inner_text(timeout=10_000)
+        except Exception:
+            dom_text = ""
+        network_dump = capture.export()
+        parsed = self.vehicle_parser.normalize(vehicle_url, html, dom_text, network_dump)
+
+        db_record = self.car_mapper.map_to_car_record(
+            vehicle_url=vehicle_url,
+            vehicle_summary=parsed.get("vehicle_summary", {}),
+            payload_insights=parsed.get("payload_insights", {}),
+        )
+
+        result = {
+            "source_url": vehicle_url,
+            "fetched_at_epoch": int(time.time()),
+            "network": network_dump,
+            **parsed,
+            "db_record": db_record.model_dump(mode="json"),
+        }
+
+        if self.settings.raw_output_json:
+            save_to_json(network_dump, self.settings.raw_output_json)
+        return result
+
+    # db sync
+
+    def sync_vehicle(self, vehicle_url: str, lane: str = "iaai"):
+        """Scrape + upsert одного авто."""
+        self.persistence.create_tables()
+        run_id = self.persistence.start_sync_run(lane=lane)
+        ids_fetched = 1
+        cars_upserted = 0
+        cars_failed = 0
+        images_upserted = 0
+        status = "failed"
+        error_summary = None
+        try:
+            scrape_result = self.scrape_vehicle_detail(vehicle_url)
+            db_record = scrape_result.get("db_record")
+            if not db_record:
+                raise RuntimeError("Scrape result does not contain db_record")
+            record = CarRecord.model_validate(db_record)
+            upsert = self.persistence.upsert_car(record)
+            cars_upserted = 1
+            images_upserted = int(upsert.get("images_upserted", 0))
+            status = "success"
+            return {
+                "status": status,
+                "run_id": run_id,
+                "vehicle_url": vehicle_url,
+                "db_action": upsert.get("action"),
+                "images_upserted": images_upserted,
+                "db_record": db_record,
+            }
+        except Exception as exc:
+            cars_failed = 1
+            status = "failed"
+            error_summary = str(exc)
+            raise
+        finally:
+            self.persistence.finish_sync_run(
+                run_id,
+                status=status,
+                ids_fetched=ids_fetched,
+                cars_upserted=cars_upserted,
+                cars_failed=cars_failed,
+                images_upserted=images_upserted,
+                error_summary=error_summary,
+            )
+
+    def sync_listing(self, make: str | None = None, model: str | None = None, lane: str = "iaai_cars"):
+        """Листинг + sync всех найденных машин."""
+        self.persistence.create_tables()
+        listing = self.collect_listing(make=make, model=model)
+        vehicle_urls = list(listing.get("vehicle_urls", []))
+        run_id = self.persistence.start_sync_run(lane=lane)
+        cars_upserted = 0
+        cars_failed = 0
+        images_upserted = 0
+        failures: list[dict[str, str]] = []
+
+        total = len(vehicle_urls)
+        logger.info("Starting sync: %d vehicles to process", total)
+
+        page = self._get_page()
+        try:
+            for index, vehicle_url in enumerate(vehicle_urls, start=1):
+                try:
+                    logger.info("[%d/%d] Scraping %s", index, total, vehicle_url)
+                    scrape_result = self._scrape_on_page(page, vehicle_url)
+                    db_record = scrape_result.get("db_record")
+                    if not db_record:
+                        raise RuntimeError("Scrape result does not contain db_record")
+                    record = CarRecord.model_validate(db_record)
+                    upsert = self.persistence.upsert_car(record)
+                    cars_upserted += 1
+                    img_count = int(upsert.get("images_upserted", 0))
+                    images_upserted += img_count
+                    logger.info(
+                        "[%d/%d] %s %s: %s %s %s — %s, %d images",
+                        index, total, upsert.get("action", "?"),
+                        record.origin_id, record.brand, record.model,
+                        record.year or "?", record.price or "N/A", img_count,
+                    )
+                except Exception as exc:
+                    cars_failed += 1
+                    failures.append({"vehicle_url": vehicle_url, "error": str(exc)})
+                    logger.error("[%d/%d] Failed %s: %s", index, total, vehicle_url, exc)
+                    try:
+                        page.close()
+                    except Exception:
+                        pass
+                    page = self._get_page()
+
+                if index < total:
+                    time.sleep(1.0)
+        finally:
+            try:
+                page.close()
+            except Exception:
+                pass
+
+        status = "success" if not failures else ("partial_success" if cars_upserted else "failed")
+        error_summary = "; ".join(item["error"] for item in failures[:10]) if failures else None
+        self.persistence.finish_sync_run(
+            run_id,
+            status=status,
+            ids_fetched=total,
+            cars_upserted=cars_upserted,
+            cars_failed=cars_failed,
+            images_upserted=images_upserted,
+            error_summary=error_summary,
+        )
+        logger.info(
+            "Sync run #%d finished: %d/%d upserted, %d failed, %d images",
+            run_id, cars_upserted, total, cars_failed, images_upserted,
+        )
+        return {
+            "status": status,
+            "run_id": run_id,
+            "listing": listing,
+            "cars_upserted": cars_upserted,
+            "cars_failed": cars_failed,
+            "images_upserted": images_upserted,
+            "failures": failures,
+        }
+
+    # scheduler
+
+    def run_scheduled(self) -> None:
+        interval = self.settings.scheduler_interval_minutes * 60
+        logger.info(
+            "Scheduler started: syncing every %d minutes",
+            self.settings.scheduler_interval_minutes,
+        )
+        cycle = 0
+        while True:
+            cycle += 1
+            logger.info("=== Scheduler cycle #%d starting ===", cycle)
+            start = time.time()
+            try:
+                if self.context:
+                    try:
+                        self.context.close()
+                    except PlaywrightError:
+                        pass
+                    self.context = None
+
+                result = self.sync_listing()
+                elapsed = time.time() - start
+                logger.info(
+                    "Cycle #%d done in %.1fs: %d upserted, %d failed",
+                    cycle, elapsed,
+                    result.get("cars_upserted", 0),
+                    result.get("cars_failed", 0),
+                )
+            except Exception as exc:
+                elapsed = time.time() - start
+                logger.error("Cycle #%d failed after %.1fs: %s", cycle, elapsed, exc)
+                if self.context:
+                    try:
+                        self.context.close()
+                    except PlaywrightError:
+                        pass
+                    self.context = None
+
+            sleep_time = max(0, interval - (time.time() - start))
+            if sleep_time > 0:
+                logger.info("Sleeping %.0f seconds until next cycle...", sleep_time)
+                time.sleep(sleep_time)
+
+    # helpers
+
+    def _warm_page(self, page: Page) -> None:
+        """Scroll для lazy-load."""
+        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/__init__.py

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

iaai_scraper/storage/db.py

@@ -0,0 +1,100 @@
+import logging
+from contextlib import contextmanager
+from datetime import datetime, timezone
+from typing import Iterator
+
+from sqlalchemy import create_engine, select
+from sqlalchemy.orm import Session, sessionmaker
+
+from ..core.config import Settings
+from .models import Base, Car, Image, SyncRun
+from .schemas import CarRecord
+
+logger = logging.getLogger("iaai_scraper.db")
+
+
+class PersistenceService:
+
+	def __init__(self, settings: Settings) -> None:
+		self.settings = settings
+		self.engine = create_engine(settings.database.url, echo=settings.database.echo, future=True)
+		self.session_factory = sessionmaker(bind=self.engine, expire_on_commit=False, future=True)
+
+	def create_tables(self) -> None:
+		Base.metadata.create_all(self.engine)
+
+	@contextmanager
+	def session_scope(self) -> Iterator[Session]:
+		session = self.session_factory()
+		try:
+			yield session
+			session.commit()
+		except Exception:
+			session.rollback()
+			raise
+		finally:
+			session.close()
+
+	def start_sync_run(self, lane: str) -> int:
+		with self.session_scope() as session:
+			run = SyncRun(status="running", lane=lane, ids_fetched=0, cars_upserted=0, cars_failed=0, images_upserted=0)
+			session.add(run)
+			session.flush()
+			return int(run.id)
+
+	def finish_sync_run(self, run_id: int, *, status: str, ids_fetched: int, cars_upserted: int, cars_failed: int, images_upserted: int, error_summary: str | None = None) -> None:
+		with self.session_scope() as session:
+			run = session.get(SyncRun, run_id)
+			if run is None:
+				return
+			run.finished_at = datetime.now(timezone.utc)
+			run.status = status
+			run.ids_fetched = ids_fetched
+			run.cars_upserted = cars_upserted
+			run.cars_failed = cars_failed
+			run.images_upserted = images_upserted
+			run.error_summary = error_summary
+
+	@staticmethod
+	def _add_images(session: Session, car_id: int, images: list[dict[str, object]]) -> None:
+		for image_payload in images:
+			session.add(Image(fullres_image=str(image_payload["fullres_image"]), preview_image=str(image_payload["preview_image"]), order_index=int(image_payload.get("order_index", 0)), car_id=car_id))
+
+	def upsert_car(self, record: CarRecord):
+		"""Insert/update/skip по content_hash."""
+		payload = record.model_dump(mode="python")
+		images = payload.pop("images", [])
+		payload.pop("raw_attributes", None)
+		payload.pop("mapping_notes", None)
+		content_hash = payload.pop("content_hash", "")
+		with self.session_scope() as session:
+			# поиск по origin_id
+			car = session.execute(select(Car).where(Car.origin_id == record.origin_id)).scalar_one_or_none()
+			action = "inserted"
+			if car is None:
+				car = Car(**payload)
+				session.add(car)
+				session.flush()
+			else:
+				action = "updated"
+				for key, value in payload.items():
+					setattr(car, key, value)
+				car.last_seen_at = record.last_seen_at
+				session.flush()
+				# замена картинок в savepoint
+				nested = session.begin_nested()
+				try:
+					for image in list(car.images):
+						session.delete(image)
+					session.flush()
+					self._add_images(session, int(car.id), images)
+					session.flush()
+					nested.commit()
+				except Exception:
+					nested.rollback()
+					logger.warning("Image replacement failed for car %s, keeping old images", record.origin_id)
+					images = []
+				return {"car_id": int(car.id), "images_upserted": len(images), "action": action}
+			self._add_images(session, int(car.id), images)
+			session.flush()
+			return {"car_id": int(car.id), "images_upserted": len(images), "action": action}

iaai_scraper/storage/enums.py

@@ -0,0 +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")
+GEARBOX_ENUM_VALUES = ("AT", "CVT", "MT", "EV", "NA")
+STEERING_WHEEL_ENUM_VALUES = ("LEFT", "RIGHT", "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")
+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")
+SELLING_TYPE_ENUM_VALUES = ("STOCK", "AUCTION", "TENDER", "stock", "auction", "tender", "NA")

iaai_scraper/storage/listing.py

@@ -0,0 +1,144 @@
+import logging
+import re
+from dataclasses import asdict, dataclass, field
+from urllib.parse import urljoin
+
+from playwright.sync_api import Page
+
+from ..browser.pace import HumanPacer
+from ..core.config import Settings
+from ..core.utils import first_non_empty
+
+logger = logging.getLogger("iaai_scraper.listing")
+VEHICLE_HREF_RE = re.compile(r"/VehicleDetail/\d+(?:~[A-Z]{2})?", re.IGNORECASE)
+
+
+@dataclass(slots=True)
+class ListingVehicleLink:
+	href: str
+	title: str = ""
+	lot_number: str | None = None
+
+
+@dataclass(slots=True)
+class ListingPageResult:
+	source_url: str
+	page_number: int
+	vehicle_links: list[ListingVehicleLink] = field(default_factory=list)
+	pagination_available: bool = False
+	next_page_detected: bool = False
+
+
+class ListingCollector:
+	def __init__(self, settings: Settings, pacer: HumanPacer) -> None:
+		self.settings = settings
+		self.pacer = pacer
+
+	def open_cars_listing(self, page: Page) -> None:
+		logger.info("Opening cars listing page: %s", self.settings.listing.cars_url)
+		page.goto(self.settings.listing.cars_url, wait_until="domcontentloaded")
+		try:
+			page.wait_for_load_state("networkidle", timeout=15000)
+		except Exception:
+			logger.debug("networkidle timeout on listing page, continuing with current state")
+		try:
+			page.wait_for_selector("a[href*='/VehicleDetail/']", timeout=20000)
+		except Exception:
+			logger.warning("Vehicle links did not appear within timeout; page may not have rendered fully")
+		self.pacer.after_listing_open()
+
+	def apply_filters(self, page: Page, make: str | None = None, model: str | None = None) -> dict[str, str | None]:
+		applied = {"make": None, "model": None}
+		if make and self._try_fill_filter_input(page, ["input[placeholder*='Make']", "input[aria-label*='Make']"], make):
+			applied["make"] = make
+			self.pacer.after_filter_action()
+		if model and self._try_fill_filter_input(page, ["input[placeholder*='Model']", "input[aria-label*='Model']"], model):
+			applied["model"] = model
+			self.pacer.after_filter_action()
+		return applied
+
+	def collect_current_page(self, page: Page, page_number: int = 1) -> ListingPageResult:
+		anchors = page.locator("a[href*='/VehicleDetail/']")
+		total = min(anchors.count(), self.settings.listing.page_link_limit)
+		links: list[ListingVehicleLink] = []
+		seen: set[str] = set()
+		for idx in range(total):
+			anchor = anchors.nth(idx)
+			href = anchor.get_attribute("href") or ""
+			match = VEHICLE_HREF_RE.search(href)
+			if not match:
+				continue
+			absolute = urljoin(self.settings.home_url, match.group(0))
+			if absolute in seen:
+				continue
+			seen.add(absolute)
+			title = first_non_empty([anchor.get_attribute("title"), anchor.text_content(), ""]) or ""
+			links.append(ListingVehicleLink(href=absolute, title=str(title).strip()))
+			if len(links) >= self.settings.listing.max_vehicles_per_run:
+				break
+		next_page_detected = self._has_next_page(page)
+		return ListingPageResult(source_url=page.url, page_number=page_number, vehicle_links=links, pagination_available=next_page_detected, next_page_detected=next_page_detected)
+
+	def go_to_next_page(self, page: Page) -> bool:
+		selectors = ["a[aria-label*='Next']", "button[aria-label*='Next']", "a.pagination-next", "button.pagination-next", "a:has-text('Next')", "button:has-text('Next')"]
+		for selector in selectors:
+			locator = page.locator(selector).first
+			if locator.count() == 0:
+				continue
+			disabled = (locator.get_attribute("disabled") or "").lower()
+			aria_disabled = (locator.get_attribute("aria-disabled") or "").lower()
+			classes = (locator.get_attribute("class") or "").lower()
+			if disabled or aria_disabled == "true" or "disabled" in classes:
+				continue
+			self.pacer.move_mouse_to(page, locator)
+			locator.click()
+			try:
+				page.wait_for_load_state("networkidle", timeout=15000)
+			except Exception:
+				pass
+			self.pacer.after_page_change()
+			return True
+		return False
+
+	def collect_listing_links(self, page: Page, *, make: str | None = None, model: str | None = None) -> dict[str, object]:
+		self.open_cars_listing(page)
+		applied_filters = self.apply_filters(page, make=make, model=model)
+		pages: list[dict[str, object]] = []
+		all_links: list[str] = []
+		for page_number in range(1, max(1, self.settings.listing.max_pages_per_run) + 1):
+			page_result = self.collect_current_page(page, page_number=page_number)
+			pages.append({"page_number": page_result.page_number, "source_url": page_result.source_url, "links_found": len(page_result.vehicle_links), "vehicle_links": [asdict(item) for item in page_result.vehicle_links], "next_page_detected": page_result.next_page_detected})
+			for item in page_result.vehicle_links:
+				if item.href not in all_links:
+					all_links.append(item.href)
+				if len(all_links) >= self.settings.listing.max_vehicles_per_run:
+					break
+			if len(all_links) >= self.settings.listing.max_vehicles_per_run or self.settings.listing.collect_current_page_only or not self.settings.listing.include_pagination or not page_result.next_page_detected:
+				break
+			if not self.go_to_next_page(page):
+				break
+		return {"listing_url": self.settings.listing.cars_url, "applied_filters": applied_filters, "pages_collected": len(pages), "vehicles_collected": len(all_links), "vehicle_urls": all_links, "pages": pages, "strategy": {"sequential": True, "collect_current_page_only": self.settings.listing.collect_current_page_only, "include_pagination": self.settings.listing.include_pagination, "max_pages_per_run": self.settings.listing.max_pages_per_run, "max_vehicles_per_run": self.settings.listing.max_vehicles_per_run}}
+
+	@staticmethod
+	def _try_fill_filter_input(page: Page, selectors: list[str], value: str) -> bool:
+		for selector in selectors:
+			locator = page.locator(selector).first
+			if locator.count() == 0:
+				continue
+			try:
+				locator.click(); locator.fill(value); page.keyboard.press("Enter")
+				try:
+					page.wait_for_load_state("networkidle", timeout=15000)
+				except Exception:
+					pass
+				return True
+			except Exception:
+				continue
+		return False
+
+	@staticmethod
+	def _has_next_page(page: Page) -> bool:
+		for selector in ["a[aria-label*='Next']", "button[aria-label*='Next']", "a.pagination-next", "button.pagination-next", "a:has-text('Next')", "button:has-text('Next')"]:
+			if page.locator(selector).count() > 0:
+				return True
+		return False

iaai_scraper/storage/models.py

@@ -0,0 +1,78 @@
+from datetime import datetime
+
+from sqlalchemy import BigInteger, Boolean, DateTime, Enum, ForeignKey, Integer, String, Text, func
+from sqlalchemy.orm import DeclarativeBase, Mapped, mapped_column, relationship
+
+from .enums import (
+	BODY_TYPE_ENUM_VALUES,
+	COUNTRY_ENUM_VALUES,
+	CURRENCY_ENUM_VALUES,
+	DRIVE_ENUM_VALUES,
+	GEARBOX_ENUM_VALUES,
+	ORIGIN_ENUM_VALUES,
+	SELLING_TYPE_ENUM_VALUES,
+	STEERING_WHEEL_ENUM_VALUES,
+)
+
+
+class Base(DeclarativeBase):
+	pass
+
+
+class Car(Base):
+	__tablename__ = "cars"
+	id: Mapped[int] = mapped_column(Integer, primary_key=True, autoincrement=True)
+	parser_id: Mapped[str] = mapped_column(String(50), nullable=False, unique=True)
+	brand: Mapped[str] = mapped_column(String(50), nullable=False)
+	model: Mapped[str] = mapped_column(String(50), nullable=False)
+	year: Mapped[int | None] = mapped_column(Integer, nullable=True)
+	price: Mapped[int | None] = mapped_column(BigInteger, nullable=True)
+	currency: Mapped[str] = mapped_column(Enum(*CURRENCY_ENUM_VALUES, name="currencyenum", native_enum=True, create_constraint=False), nullable=False, default="USD")
+	mileage: Mapped[int] = mapped_column(Integer, nullable=False, default=0)
+	country: Mapped[str] = mapped_column(Enum(*COUNTRY_ENUM_VALUES, name="countryenum", native_enum=True, create_constraint=False), nullable=False, default="NA")
+	is_sold: Mapped[bool] = mapped_column(Boolean, nullable=False, default=False)
+	color: Mapped[str] = mapped_column(String(), nullable=False, default="other")
+	drive: Mapped[str | None] = mapped_column(Enum(*DRIVE_ENUM_VALUES, name="driveenum", native_enum=True, create_constraint=False), nullable=True)
+	gearbox: Mapped[str | None] = mapped_column(Enum(*GEARBOX_ENUM_VALUES, name="gearboxenum", native_enum=True, create_constraint=False), nullable=True)
+	steering_wheel: Mapped[str | None] = mapped_column(Enum(*STEERING_WHEEL_ENUM_VALUES, name="steeringwheelenum", native_enum=True, create_constraint=False), nullable=True)
+	body_type: Mapped[str] = mapped_column(Enum(*BODY_TYPE_ENUM_VALUES, name="bodytypeenum", native_enum=True, create_constraint=False), nullable=False, default="OTHER")
+	engine_volume: Mapped[int | None] = mapped_column(Integer, nullable=True)
+	selling_type: Mapped[str] = mapped_column(Enum(*SELLING_TYPE_ENUM_VALUES, name="sellingtypeenum", native_enum=True, create_constraint=False), nullable=False, default="NA")
+	one_owner: Mapped[bool] = mapped_column(Boolean, nullable=False, default=False)
+	new_car: Mapped[bool] = mapped_column(Boolean, nullable=False, default=False)
+	is_hidden: Mapped[bool] = mapped_column(Boolean, nullable=False, default=False)
+	origin: Mapped[str] = mapped_column(Enum(*ORIGIN_ENUM_VALUES, name="originenum", native_enum=True, create_constraint=False), nullable=False, default="NA")
+	origin_url: Mapped[str] = mapped_column(String(), nullable=False)
+	origin_id: Mapped[str] = mapped_column(String(), nullable=False, unique=True)
+	is_damaged: Mapped[bool] = mapped_column(Boolean, nullable=False, default=False)
+	evaluation: Mapped[str | None] = mapped_column(String(), nullable=True)
+	non_smoking: Mapped[bool] = mapped_column(Boolean, nullable=False, default=True)
+	rental: Mapped[bool] = mapped_column(Boolean, nullable=False, default=False)
+	repair_history: Mapped[bool] = mapped_column(Boolean, nullable=False, default=False)
+	slug: Mapped[str] = mapped_column(String(), nullable=False)
+	last_seen_at: Mapped[datetime] = mapped_column(DateTime(timezone=True), nullable=False, default=func.now())
+	images: Mapped[list["Image"]] = relationship("Image", back_populates="car", cascade="all, delete-orphan")
+
+
+class Image(Base):
+	__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: Mapped[Car] = relationship("Car", back_populates="images")
+
+
+class SyncRun(Base):
+	__tablename__ = "sync_runs"
+	id: Mapped[int] = mapped_column(BigInteger().with_variant(Integer, "sqlite"), primary_key=True, autoincrement=True)
+	started_at: Mapped[datetime] = mapped_column(DateTime(timezone=True), nullable=False, default=func.now())
+	finished_at: Mapped[datetime | None] = mapped_column(DateTime(timezone=True), nullable=True)
+	status: Mapped[str] = mapped_column(Text, nullable=False)
+	lane: Mapped[str] = mapped_column(Text, nullable=False)
+	ids_fetched: Mapped[int] = mapped_column(Integer, nullable=False, default=0)
+	cars_upserted: Mapped[int] = mapped_column(Integer, nullable=False, default=0)
+	cars_failed: Mapped[int] = mapped_column(Integer, nullable=False, default=0)
+	images_upserted: Mapped[int] = mapped_column(Integer, nullable=False, default=0)
+	error_summary: Mapped[str | None] = mapped_column(Text, nullable=True)

iaai_scraper/storage/schemas.py

@@ -0,0 +1,56 @@
+from datetime import datetime, timezone
+from typing import Any
+
+from pydantic import BaseModel, Field
+
+
+class ImageRecord(BaseModel):
+	fullres_image: str
+	preview_image: str
+	order_index: int = 0
+
+
+class CarRecord(BaseModel):
+	parser_id: str
+	brand: str
+	model: str
+	year: int | None = None
+	price: int | None = None
+	currency: str = "USD"
+	mileage: int = 0
+	country: str = "US"
+	is_sold: bool = False
+	color: str = "other"
+	drive: str | None = None
+	gearbox: str | None = None
+	steering_wheel: str | None = None
+	body_type: str = "OTHER"
+	engine_volume: int | None = None
+	selling_type: str = "AUCTION"
+	one_owner: bool = False
+	new_car: bool = False
+	is_hidden: bool = False
+	origin: str = "NA"
+	origin_url: str
+	origin_id: str
+	is_damaged: bool = False
+	evaluation: str | None = None
+	non_smoking: bool = True
+	rental: bool = False
+	repair_history: bool = False
+	slug: str
+	last_seen_at: datetime = Field(default_factory=lambda: datetime.now(timezone.utc))
+	content_hash: str = ""
+	images: list[ImageRecord] = Field(default_factory=list)
+	raw_attributes: dict[str, Any] = Field(default_factory=dict)
+	mapping_notes: list[str] = Field(default_factory=list)
+
+
+class ScrapeExport(BaseModel):
+	source_url: str
+	fetched_at_epoch: int
+	vehicle_summary: dict[str, Any] = Field(default_factory=dict)
+	payload_insights: dict[str, Any] = Field(default_factory=dict)
+	db_record: CarRecord | None = None
+	network: dict[str, Any] = Field(default_factory=dict)
+	access_notes: dict[str, Any] = Field(default_factory=dict)