# -----------------------------------------------------------------------------
# Lou-Ka — Agrégateur de logements à louer (province de Québec)
# Auteur : Simon-Pierre Boucher — contact@spboucher.ai
# connectors/pro_urbain.py : connecteur Gestion Pro-Urbain
#   (gestionpro-urbain.com). Gestionnaire locatif de Lanaudière / Rive-Nord /
#   Grand Montréal (Charlemagne, Repentigny, Berthierville, Saint-Paul,
#   Saint-Jacques, Saint-Charles-Borromée…). Site October CMS (plugin maison
#   « nerd/realestate ») rendu serveur : /fr/proprietes liste des cartes par
#   UNITÉ (adresse, ville, code postal, prix/mois, type n½, sdb, chambres,
#   stationnements, disponibilité), page /fr/propriete/<slug> pour la
#   description, les inclusions (puces « • ») et les photos (via cache
#   detail). Granularité UNITÉ, external_id = slug de la fiche.
# -----------------------------------------------------------------------------
from __future__ import annotations

import re

from bs4 import BeautifulSoup

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

BASE = "https://gestionpro-urbain.com"
LIST_URL = f"{BASE}/fr/proprietes"

TYPE_RE = re.compile(r"^\d\s*(?:½|1/2)$")
COUNT_RE = re.compile(r"^\d+(?:[.,]\d)?$")
AVAIL_RE = re.compile(r"Disponibilit[ée]\s*:\s*([^|<]{2,40})", re.I)
IMG_RE = re.compile(
    r"https://gestionpro-urbain\.com/storage/app/uploads/[^\"'\s\\)]+?"
    r"\.(?:jpg|jpeg|png|webp)", re.I)
BULLET_RE = re.compile(r"•\s*([^|<>•]{3,90})")


class ProUrbainConnector(BaseConnector):
    source_id = "pro_urbain"
    request_delay = 0.7

    def _detail_payload(self, url: str) -> dict:
        """Description, inclusions et photos de la fiche /fr/propriete/…"""
        try:
            html = self.get(url).text
        except Exception:
            return {}
        soup = BeautifulSoup(html, "html.parser")
        payload: dict = {}
        images = [u for u in dict.fromkeys(IMG_RE.findall(html))][:20]
        if images:
            payload["images"] = images
        # bloc Description : premier paragraphe substantiel après le titre
        desc = ""
        node = soup.find(string=re.compile(r"^\s*Description\s*$"))
        if node is not None:
            for p in node.find_all_next(["p", "div"]):
                t = p.get_text(" ", strip=True)
                if len(t) > 60:
                    desc = t
                    break
                if "Caractéristiques" in t:
                    break
        if desc:
            payload["description"] = desc[:1500]
        text = soup.get_text("|", strip=True)
        amenities = []
        for b in BULLET_RE.findall(text):
            b = b.strip(" .")
            if b and b not in amenities:
                amenities.append(b)
        if amenities:
            payload["amenities"] = amenities[:25]
        return payload

    def fetch(self) -> list[Listing]:
        listings: list[Listing] = []
        # Échec du GET de la liste = erreur transitoire à propager (journalisée
        # en échec par ingest), pas un « succès » à 0 annonce.
        html = self.get(LIST_URL).text
        soup = BeautifulSoup(html, "html.parser")

        seen: set[str] = set()
        for card in soup.select("div.mygrid"):
            link = card.select_one("a[href*='/fr/propriete/']")
            if link is None:
                continue
            url = link["href"]
            slug = url.rstrip("/").rsplit("/", 1)[-1]
            if not slug or slug in seen:
                continue
            seen.add(slug)

            head = card.select_one(".gsMain h2")
            head_txt = head.get_text(" ", strip=True) if head else ""
            if head_txt and not head_txt.lower().startswith("appartement"):
                continue                     # commercial/autre : hors sujet
            am = AVAIL_RE.search(head_txt)
            availability = f"Disponibilité : {am.group(1).strip()}" if am else ""

            addr_p = card.select_one(".gsMain p")
            address = city = postal = ""
            if addr_p is not None:
                parts = [x.strip() for x in
                         addr_p.get_text("|", strip=True).split("|") if x.strip()]
                if parts:
                    address = parts[0]
                if len(parts) > 1:
                    city = parts[1]
                if len(parts) > 2:
                    postal = parts[2]

            price = None
            pr = card.select_one(".currentPrice")
            if pr is not None:
                digits = re.sub(r"[^\d]", "", pr.get_text())
                if digits:
                    price = float(digits)

            # pictogrammes du pied de carte : type n½, sdb, chambres, autos
            unit_type, bathrooms, bedrooms, parking = "", None, None, None
            for ic in card.select(".gsFooter i"):
                img = ic.find("img")
                span = ic.find("span")
                if img is None or span is None:
                    continue
                src = img.get("src", "")
                val = span.get_text(" ", strip=True)
                if "room-svgrepo" in src and TYPE_RE.match(val):
                    unit_type = normalize_unit_type(val)
                elif "bathtub" in src and COUNT_RE.match(val):
                    bathrooms = float(val.replace(",", "."))
                elif "bed-svgrepo" in src and COUNT_RE.match(val):
                    bedrooms = float(val)
                elif "car-svgrepo" in src and COUNT_RE.match(val):
                    parking = int(val)

            img = card.select_one(".gsHead img")
            cover = [img["src"]] if img is not None and img.get("src") else []

            key = f"{price or 0:.0f}-{availability}-{unit_type}"
            payload = self.detail(slug, key,
                                  lambda u=url: self._detail_payload(u))

            details: dict = {}
            if parking is not None:
                details["parking"] = parking
            if postal:
                details["postal_code"] = postal
            listings.append(Listing(
                source=self.source_id,
                external_id=slug,
                url=url,
                title=(f"{unit_type} — {address}, {city}"
                       if unit_type else f"{address}, {city}"),
                address=address,
                city=city,
                unit_type=unit_type,
                bedrooms=bedrooms,
                bathrooms=bathrooms,
                price=price,
                price_label=f"{price:.0f} $ /mois" if price else "",
                availability=availability,
                description=payload.get("description", ""),
                amenities=payload.get("amenities", []),
                details=details,
                images=payload.get("images") or cover,
            ))
        return listings