# -----------------------------------------------------------------------------
# Lou-Ka — Agrégateur de logements à louer (province de Québec)
# Auteur : Simon-Pierre Boucher — contact@spboucher.ai
# connectors/trams.py : connecteur T.R.A.M.S Property Management (tramsmgmt.com)
#   Gestionnaire montréalais : 3 immeubles locatifs résidentiels au Québec —
#   Chateau Ste-Marie (Kirkland), L'Ambiance Ouest (Dollard-des-Ormeaux),
#   Palacio D'Anjou (Anjou). WordPress rendu serveur : la page
#   /residential-properties-for-rent/ liste les cartes (les propriétés de
#   Floride sont EXCLUES) ; chaque page /property/<slug>/ donne l'adresse,
#   la description, les caractéristiques (ul) et le contact. AUCUN prix ni
#   inventaire par unité publié (le lien securecafe/Yardi n'est que le
#   portail des résidents) → une annonce par immeuble, comme brochu.py.
# -----------------------------------------------------------------------------
from __future__ import annotations

import hashlib
import re

from bs4 import BeautifulSoup

from ..schema import Listing
from .base import BaseConnector

BASE = "https://tramsmgmt.com"
LIST_URL = f"{BASE}/residential-properties-for-rent/"

QC_RE = re.compile(r",\s*(?:QC|Qu[ée]bec)\b", re.I)
CITY_RE = re.compile(r"([A-Za-zÀ-ÿ'’.\- ]{3,40}),\s*(?:QC|Qu[ée]bec)\b")
IMG_RE = re.compile(r"https://tramsmgmt\.com/wp-content/uploads/"
                    r'[^"\s\\)]+\.(?:jpe?g|png|webp)', re.I)


class TramsConnector(BaseConnector):
    source_id = "trams"
    request_delay = 0.7

    def fetch(self) -> list[Listing]:
        listings: list[Listing] = []
        try:
            html = self.get(LIST_URL).text
        except Exception:
            return listings
        soup = BeautifulSoup(html, "html.parser")

        seen: set[str] = set()
        for a in soup.select('a[href*="/property/"]'):
            url = (a.get("href") or "").split("?")[0]
            slug = url.rstrip("/").split("/")[-1]
            if not slug or slug in seen:
                continue
            # carte englobante : ne garder que les immeubles québécois
            card = a
            for _ in range(5):
                if card.parent is None:
                    break
                card = card.parent
                if len(card.select('a[href*="/property/"]')) > 3:
                    card = None
                    break
                text = card.get_text(" ", strip=True)
                if QC_RE.search(text):
                    break
            else:
                continue
            if card is None or not QC_RE.search(
                    card.get_text(" ", strip=True)):
                continue
            seen.add(slug)
            card_key = hashlib.sha1(
                card.get_text(" ", strip=True).encode("utf-8")).hexdigest()
            try:
                d = self.detail(slug, card_key,
                                lambda url=url: self._fetch_detail(url))
            except Exception:
                continue
            if not d.get("title"):
                continue
            city = d.get("city", "")
            listings.append(Listing(
                source=self.source_id,
                external_id=slug,
                url=url,
                title=f"{d['title']} — appartements à louer"
                      f"{' (' + city + ')' if city else ''}",
                address=d.get("address", ""),
                city=city,
                availability="Prix et visites : communiquer avec l'immeuble",
                description=d.get("description", ""),
                amenities=d.get("amenities") or [],
                details={"contact": d["contact"]} if d.get("contact") else {},
                images=d.get("images") or [],
            ))
        return listings

    def _fetch_detail(self, url: str) -> dict:
        out: dict = {}
        html = self.get(url).text
        soup = BeautifulSoup(html, "html.parser")
        article = soup.find("article") or soup

        h1 = article.find("h1")
        if h1:
            out["title"] = h1.get_text(" ", strip=True)

        # bloc adresse sous le titre (rue / ville, QC / code postal)
        addr_div = article.select_one("div.text-center.pb-1")
        if addr_div:
            addr = re.sub(r"\s+", " ",
                          addr_div.get_text(", ", strip=True))
            out["address"] = re.sub(r"\s*,\s*(,\s*)+", ", ", addr).strip(" ,")
            m = CITY_RE.search(out["address"])
            if m:
                out["city"] = m.group(1).strip()

        # description : paragraphes substantiels ; caractéristiques : 1re liste
        paras = [p.get_text(" ", strip=True) for p in article.find_all("p")
                 if len(p.get_text(strip=True)) > 60
                 and "call" not in p.get_text().lower()[:40]]
        out["description"] = re.sub(r"\s+", " ",
                                    " ".join(paras[:3])).strip()[:1200]
        amenities: list[str] = []
        for li in article.select("ul li"):
            t = re.sub(r"\s+", " ", li.get_text(" ", strip=True)).rstrip(".")
            if 0 < len(t) <= 80 and t not in amenities:
                amenities.append(t)
        out["amenities"] = amenities[:15]

        contact: dict = {}
        m = re.search(r'href="tel:([\d\- ()]{7,})"', html)
        if m:
            contact["phone"] = m.group(1).strip()
        m = re.search(r'href="mailto:([^"?]+)"', html)
        if m:
            contact["email"] = m.group(1).strip().lower()
        out["contact"] = contact

        out["images"] = [
            u for u in dict.fromkeys(IMG_RE.findall(html))
            if not re.search(r"logo|icon|favicon|cropped-|-\d{2,3}x\d{2,3}\.",
                             u, re.I)][:20]
        return out