# -----------------------------------------------------------------------------
# Auto-Ka — Agrégateur de voitures usagées à vendre (province de Québec)
# Auteur : Simon-Pierre Boucher — contact@spboucher.ai
# connectors/classeauto.py : connecteur Classe Auto (classeauto.ca),
#   marchand d'occasion de l'île de Montréal (boul. Industriel,
#   Montréal-Nord/Anjou) — ~140 véhicules.
#
#   Plateforme : WordPress (thème Astra) + moteur d'inventaire DriveGood
#   (cdn.drivegood.com). Le site expose un export JSON STATIQUE de tout
#   l'inventaire :
#     /wp-content/themes/astra/car_single_page_data/cars_formatted.json
#   — une seule requête pour la liste complète, avec TOUS les champs :
#   prix, km, VIN, stock, transmission, carburant, motricité, carrosserie,
#   couleurs, portes, passagers, moteur, options, photos (cdn.drivegood.com),
#   description. Aucune page détail nécessaire.
#
#   Filtres de sécurité : vehicle_type_id == 1 (automobile), post_status
#   publish/published, condition != NEW, deleted_at vide.
# -----------------------------------------------------------------------------
from __future__ import annotations

import re

from ..schema import Vehicle
from .base import BaseConnector

_JSON_PATH = "/wp-content/themes/astra/car_single_page_data/cars_formatted.json"

# garde-fou non-automobile (le marchand ne vend que des autos)
_EXCLUDE_RE = re.compile(
    r"motoneige|vtt|moto\b|motomarine|bateau|roulotte|remorque|spyder", re.I)


def _s(value) -> str:
    if value is None:
        return ""
    return re.sub(r"\s+", " ", str(value)).strip()


def _num(value) -> float | None:
    try:
        f = float(value)
    except (TypeError, ValueError):
        return None
    return f if f > 0 else None


def _int(value) -> int | None:
    try:
        return int(str(value))
    except (TypeError, ValueError):
        return None


class ClasseAuto(BaseConnector):
    """Classe Auto (Montréal) — inventaire complet en un JSON statique."""

    source_id = "classeauto"
    base_url = "https://www.classeauto.ca"
    dealer_name = "Classe Auto"
    city = "Montréal"
    request_delay = 1.0

    def fetch(self) -> list[Vehicle]:
        # cache-buster comme le fait le site (nouvelle valeur par minute)
        data = self.get(self.base_url + _JSON_PATH).json()
        vehicles: list[Vehicle] = []
        for d in data if isinstance(data, list) else []:
            try:
                veh = self._to_vehicle(d)
            except Exception:
                continue                     # enregistrement malformé isolé
            if veh is not None:
                vehicles.append(veh)
        return vehicles

    def _to_vehicle(self, d: dict) -> Vehicle | None:
        if not isinstance(d, dict):
            return None
        if d.get("vehicle_type_id") not in (1, "1", None):
            return None                      # 1 = automobile
        if _s(d.get("post_status")) not in ("publish", "published", ""):
            return None
        if d.get("deleted_at"):
            return None
        if _s(d.get("condition")).upper() == "NEW":
            return None                      # usagé seulement

        ext_id = _s(d.get("vid")) or _s(d.get("ID"))
        if not ext_id:
            return None

        make = _s(d.get("maker")).title()
        model = _s(d.get("model"))
        trim = _s(d.get("car_trim")) or _s(d.get("car_sub_model"))
        year = _int(d.get("car_year"))
        body = _s(d.get("car_body"))
        if _EXCLUDE_RE.search(f"{make} {model} {body}"):
            return None

        title = " ".join(x for x in (str(year or ""), make, model, trim) if x)

        price = _num(d.get("car_price"))
        km = _num(d.get("car_mileage"))
        unit = _s(d.get("car_mileage_unit")).upper()
        if km and unit.startswith("MI"):
            km = round(km * 1.609344)

        slug = _s(d.get("slug"))
        url = (f"{self.base_url}/cars/{slug}/" if slug
               else _s(d.get("guid")) or self.base_url)

        images = [u.strip() for u in _s(d.get("photos")).split(",")
                  if u.strip().startswith("http")][:20]

        engine = ""
        size = _s(d.get("car_engine_size"))
        cyl = _s(d.get("car_cylinders"))
        if size:
            engine = f"{size}L" + (f" {cyl} cyl." if cyl else "")

        features = [f.strip().replace("_", " ")
                    for f in _s(d.get("car_options")).split(",") if f.strip()]

        details: dict = {}
        if d.get("car_no_accident"):
            details["no_accident"] = True
        if d.get("car_one_owner"):
            details["unique_owner"] = True
        old_price = _num(d.get("car_old_price"))
        if old_price and price and old_price > price:
            details["old_price"] = old_price

        carfax = _s(d.get("carfax_url")) or _s(d.get("carfax_report"))
        if not carfax:                        # parfois rangé dans admin_note
            m = re.search(r"https?://vhr\.carfax\.ca/\S+",
                          _s(d.get("admin_note")))
            if m:
                carfax = m.group(0)

        return Vehicle(
            source=self.source_id,
            external_id=ext_id,
            url=url,
            title=title,
            make=make,
            model=model,
            trim=trim,
            year=year,
            price=price,
            price_label=f"{price:,.0f} $".replace(",", " ") if price else "",
            mileage_km=km,
            mileage_label=f"{km:,.0f} km".replace(",", " ") if km else "",
            transmission=_s(d.get("car_transmission")),
            fuel=_s(d.get("car_fuel_type")),
            drivetrain=_s(d.get("car_drivetrain")),
            body_type=body,
            exterior_color=_s(d.get("car_exterior_color")).title(),
            interior_color=_s(d.get("car_interrior_color")).title(),
            engine=engine,
            doors=_int(d.get("car_doors_count")),
            seats=_int(d.get("number_of_passengers")),
            vin=_s(d.get("car_vin")).upper(),
            stock_number=_s(d.get("stock")),
            dealer_name=self.dealer_name,
            city=self.city,
            description=_s(d.get("post_content"))[:4000],
            features=features[:60],
            details=details,
            images=images,
            carfax_url=carfax,
        )