# -----------------------------------------------------------------------------
# Lou-Ka — Agrégateur de logements à louer (province de Québec)
# Auteur : Simon-Pierre Boucher — contact@spboucher.ai
# connectors/localys.py : connecteur Gestion Localys (gestionlocalys.com,
#   240+ unités). Site Next.js rendu serveur : /fr/logements liste les
#   logements disponibles (cartes avec adresse, quartier, « 2 ch. · 1 sdb. »
#   et prix « 1 500 $ / mois »), chaque fiche /fr/logements/<slug> fournit la
#   description (avec **DISPONIBLE …**, sections #### Inclusions/Exclusions),
#   les caractéristiques (« 2 ch. 1 sdb. 750 pi² ») et la galerie (photos
#   ws.jumptools.com encodées dans /_next/image?url=…). Une annonce par
#   logement ; external_id = slug ; fiches via le cache détail (clé = hash du
#   texte de la carte).
# -----------------------------------------------------------------------------
from __future__ import annotations

import hashlib
import re
import urllib.parse

from bs4 import BeautifulSoup

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

BASE = "https://www.gestionlocalys.com"
LIST_URL = f"{BASE}/fr/logements"

# « 2558 Rue De Cadillac, Montréal (Mercier/Hochelaga-Maisonneuve), QC, … »
H1_RE = re.compile(
    r"^(?P<street>.+?),\s*(?P<city>[^,(]+?)\s*(?:\((?P<sector>[^)]+)\))?"
    r",\s*QC", re.S)
PRICE_RE = re.compile(r"([\d\s ]+)\$\s*/\s*mois")
CH_RE = re.compile(r"(\d+)\s*ch\.")
SDB_RE = re.compile(r"([\d.]+)\s*sdb\.")
SQFT_RE = re.compile(r"([\d\s ]+)\s*pi²")
NEXT_IMG_RE = re.compile(r'/_next/image\?url=([^&"\']+)')
DISPO_RE = re.compile(r"DISPONIBLE[^*\n]{0,40}", re.I)


def _clean_md(t: str) -> str:
    """Retire les marqueurs markdown (** et ####) du texte rendu."""
    return re.sub(r"\s+", " ", t.replace("**", " ").replace("####", " ")
                  ).strip()


class LocalysConnector(BaseConnector):
    source_id = "localys"
    request_delay = 0.7

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

        # cartes : ancêtre du lien qui contient le prix
        cards: dict[str, str] = {}       # slug -> texte de la carte
        for a in soup.find_all("a", href=True):
            if "/fr/logements/" not in a["href"]:
                continue
            slug = a["href"].rstrip("/").rsplit("/", 1)[-1]
            if slug in cards and cards[slug]:
                continue
            node, text = a, ""
            for _ in range(6):
                node = node.parent
                if node is None:
                    break
                t = re.sub(r"\s+", " ", node.get_text(" ", strip=True))
                if "$" in t:
                    text = t
                    break
            cards[slug] = text

        listings: list[Listing] = []
        for slug, card_text in cards.items():
            try:
                key = hashlib.sha1(card_text.encode("utf-8")).hexdigest()
                d = self.detail(slug, key,
                                lambda s=slug: self._logement(s))
                lst = self._listing(slug, card_text, d)
                if lst:
                    listings.append(lst)
            except Exception:
                continue
        return listings

    def _logement(self, slug: str) -> dict:
        """Scrape la fiche : h1, description, caractéristiques, photos."""
        out: dict = {"h1": "", "desc": "", "carac": "", "images": []}
        html = self.get(f"{LIST_URL}/{slug}").text
        soup = BeautifulSoup(html, "html.parser")
        for s in soup.find_all("script"):
            s.decompose()                # écarter le payload RSC de Next.js
        if soup.h1:
            out["h1"] = soup.h1.get_text(" ", strip=True)
        text = re.sub(r"\s+", " ", soup.get_text(" ", strip=True))
        i = text.find("/ mois")
        j = text.find("Caractéristiques")
        if i >= 0 and j > i:
            out["desc"] = text[i + len("/ mois"):j].strip()[:1500]
        if j >= 0:
            out["carac"] = text[j:j + 120]
        images = []
        for enc in NEXT_IMG_RE.findall(html):
            u = urllib.parse.unquote(enc)
            if u.startswith("http") and "jumptools.com" in u \
                    and u not in images:
                images.append(u)
        out["images"] = images[:15]
        return out

    def _listing(self, slug: str, card_text: str, d: dict) -> Listing | None:
        # fiche morte : la source amont (jumptools) renvoie sa page 404
        if "cannot be found" in d.get("desc", ""):
            return None
        h1 = d.get("h1") or ""
        street = city = sector = ""
        m = H1_RE.match(h1)
        if m:
            street = m.group("street").strip()
            city = m.group("city").strip()
            sector = (m.group("sector") or "").strip()
        address = ", ".join(x for x in (street, city) if x)

        price = None
        pm = PRICE_RE.search(card_text) or PRICE_RE.search(d.get("carac", ""))
        if pm:
            price = parse_price(pm.group(1).strip() + " $")
        bedrooms = bathrooms = None
        blob = f"{card_text} {d.get('carac', '')}"
        bm = CH_RE.search(blob)
        if bm:
            bedrooms = float(bm.group(1))
        sm = SDB_RE.search(blob)
        if sm:
            try:
                bathrooms = float(sm.group(1))
            except ValueError:
                pass
        area = None
        am = SQFT_RE.search(d.get("carac", "")) or SQFT_RE.search(card_text)
        if am:
            try:
                area = float(re.sub(r"[\s ]", "", am.group(1)))
            except ValueError:
                pass

        desc_raw = d.get("desc", "")
        availability = ""
        dm = DISPO_RE.search(desc_raw)
        if dm:
            availability = _clean_md(dm.group(0)).capitalize()

        return Listing(
            source=self.source_id,
            external_id=slug,
            url=f"{LIST_URL}/{slug}",
            title=h1.replace(", QC, Canada", "") or slug,
            address=address,
            sector=sector,
            city=city or "Montréal",
            unit_type="",
            bedrooms=bedrooms,
            bathrooms=bathrooms,
            price=price,
            price_label=f"{pm.group(1).strip()} $ / mois" if pm else "",
            availability=availability,
            area_sqft=area,
            description=_clean_md(desc_raw),
            images=d.get("images", []),
        )