# -----------------------------------------------------------------------------
# Lou-Ka — Agrégateur de logements à louer (province de Québec)
# Auteur : Simon-Pierre Boucher — contact@spboucher.ai
# connectors/lambert.py : connecteur Société Immobilière Lambert
#   (lambertimmobilier.com — Louiseville, Yamachiche, Mauricie). WordPress
#   avec thème custom « cognitif-starter » : la page /logements-disponibles/
#   liste des <article class="apartment"> (type h2, prix, <address>, photo en
#   background-image). On ne garde que la section « Nos logements à louer »
#   (les <article> suivant le titre « Projets à venir » = terrains, exclus).
#   Aucune page détail ; external_id = slug de l'adresse. 1 requête par sync.
# -----------------------------------------------------------------------------
from __future__ import annotations

import re
import unicodedata

from bs4 import BeautifulSoup

from ..schema import Listing, normalize_unit_type, parse_price
from .base import BaseConnector

BASE = "https://lambertimmobilier.com"
LIST_URL = f"{BASE}/logements-disponibles/"

_BG_URL_RE = re.compile(r"background-image\s*:\s*url\(['\"]?([^'\")]+)")
_CITIES = [
    ("louiseville", "Louiseville"),
    ("yamachiche", "Yamachiche"),
    ("trois-rivieres", "Trois-Rivières"),
]


def _strip_accents(s: str) -> str:
    return "".join(c for c in unicodedata.normalize("NFD", s)
                   if unicodedata.category(c) != "Mn")


def _slug(s: str) -> str:
    s = _strip_accents(s.lower())
    return re.sub(r"[^a-z0-9]+", "-", s).strip("-")


class LambertConnector(BaseConnector):
    source_id = "lambert"
    request_delay = 0.7

    def fetch(self) -> list[Listing]:
        html = self.get(LIST_URL).text
        soup = BeautifulSoup(html, "html.parser")
        listings: dict[str, Listing] = {}
        in_projects = False
        for el in soup.find_all(["h1", "h2", "article"]):
            if el.name in ("h1", "h2"):
                txt = el.get_text(" ", strip=True)
                if re.search(r"projets? à venir", txt, re.I):
                    in_projects = True        # terrains/projets : hors annonces
                continue
            if in_projects or "apartment" not in (el.get("class") or []):
                continue
            try:
                self._parse_card(el, listings)
            except Exception:
                continue
        return list(listings.values())

    def _parse_card(self, card, listings: dict[str, Listing]) -> None:
        head = card.select_one("h2")
        head_txt = re.sub(r"\s+", " ", head.get_text(" ", strip=True)) if head else ""
        if re.search(r"terrain|commercial|local\b", head_txt, re.I):
            return
        addr_el = card.find("address")
        address_full = re.sub(r"\s+", " ",
                              addr_el.get_text(" ", strip=True)) if addr_el else ""
        if not address_full and not head_txt:
            return

        # ville en fin d'adresse (« 131 St-Ubald Louiseville »)
        city, address = "", address_full
        low = _strip_accents(address_full.lower())
        for key, name in _CITIES:
            if key in low:
                city = name
                address = re.sub(rf",?\s*{key}\s*$", "", address_full,
                                 flags=re.I).strip(" ,")
                break

        price_el = card.select_one(".apartment--price")
        price_label = re.sub(r"\s+", " ",
                             price_el.get_text(" ", strip=True)) if price_el else ""

        images: list[str] = []
        img_div = card.select_one("div.img[style]")
        if img_div:
            m = _BG_URL_RE.search(img_div["style"])
            if m and m.group(1).startswith("http"):
                images.append(m.group(1))

        ext = _slug(address_full or head_txt)
        if not ext or ext in listings:
            return
        title = f"{head_txt} — {address_full}" if head_txt else address_full
        listings[ext] = Listing(
            source=self.source_id,
            external_id=ext,
            url=LIST_URL,
            title=title,
            address=address,
            city=city,
            unit_type=normalize_unit_type(head_txt),
            price=parse_price(price_label),
            price_label=price_label,
            images=images,
        )