# -----------------------------------------------------------------------------
# Lou-Ka — Agrégateur de logements à louer (province de Québec)
# Auteur : Simon-Pierre Boucher — contact@spboucher.ai
# connectors/bons_locataires.py : Bons Locataires (bonslocataires.com) — agence
#   de location Montréal/Laval/Longueuil. WordPress, aucun anti-bot (seul
#   mod_security bloque les sitemaps XML — on ne les utilise pas).
#   Source de vérité : la page carte MapListPro (~8 Mo) dont le tableau JSON
#   inline contient les 1 600+ emplacements avec lat/lng, taxonomies
#   (région/quartier/taille) et description complète. Le site garde les
#   logements loués sur la carte (« DÉJÀ LOUÉ », pin PouceVert) : on ne
#   conserve que les annonces encore offertes, et on valide que leur page
#   (locationUrl) répond toujours 200 (certains vieux liens sont morts).
#   Prix et disponibilité en texte libre -> regex. ~20-40 annonces actives.
# -----------------------------------------------------------------------------
from __future__ import annotations

import html as H
import json
import re

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

MAP_PAGE = ("https://www.bonslocataires.com/logement-appartement-a-louer-"
            "enquete-proprio-credit-locataire-visite-virtuelle-360/")

_PRICE_RE = re.compile(r"(\d{3,4})\s*\$\s*(?:/|par)?\s*mois", re.I)
_PRICE_ALT_RE = re.compile(r"(\d{3,4})\s*\$")
_IMG_RE = re.compile(r'src="(https?://www\.bonslocataires\.com/wp-content/'
                     r'uploads/[^"]+)"')
_LOC_ID_RE = re.compile(r"loc-(\d+)")
_TAG_RE = re.compile(r"<[^>]+>")


def _texte(fragment: str) -> str:
    return re.sub(r"\s+\n", "\n", H.unescape(_TAG_RE.sub("", fragment or ""))).strip()


class BonsLocatairesConnector(BaseConnector):
    source_id = "bons_locataires"
    request_delay = 1.0
    timeout = 120                      # la page carte fait ~8 Mo

    def _map_objects(self) -> list[dict]:
        import time
        html = ""
        for attempt in range(3):        # mod_security coupe parfois la page 8 Mo
            try:
                html = self.get(MAP_PAGE).text
                break
            except Exception:
                if attempt == 2:
                    raise
                time.sleep(20 * (attempt + 1))
        i = html.find('[{"title":')
        if i < 0:
            return []
        depth, end = 0, None
        for j in range(i, len(html)):
            if html[j] == "[":
                depth += 1
            elif html[j] == "]":
                depth -= 1
                if depth == 0:
                    end = j + 1
                    break
        try:
            return json.loads(html[i:end])
        except ValueError:
            return []

    @staticmethod
    def _available(obj: dict) -> bool:
        blob = (obj.get("title", "") + (obj.get("description") or "")
                + (obj.get("icon") or "") + (obj.get("pinImageUrl") or ""))
        return "LOU" not in blob.upper() or "PouceVert" not in blob

    def fetch(self) -> list[Listing]:
        out: list[Listing] = []
        for obj in self._map_objects():
            desc_html = obj.get("description") or ""
            blob = (obj.get("title", "") + desc_html).upper()
            pin = (obj.get("pinImageUrl") or "") + (obj.get("icon") or "")
            if "DÉJÀ LOU" in blob or "DEJA LOU" in blob or "LOUÉ!" in blob \
                    or "PouceVert" in pin:
                continue
            url = (obj.get("locationUrl") or "").replace("http://", "https://")
            if not url:
                continue
            try:                        # vieux liens morts -> annonce écartée
                page_html = self.get(url).text
            except Exception:
                continue
            m = _LOC_ID_RE.search(obj.get("cssClass") or "")
            eid = m.group(1) if m else url.rstrip("/").rsplit("/", 1)[-1]

            cats = obj.get("customCategories") or {}
            def _first(tax: str) -> str:
                terms = cats.get(tax) or []
                return terms[0].get("name", "") if terms else ""
            city = _first("map_location_categories_region") or "Montréal"
            sector = _first("map_location_categories_city")
            taille = _first("map_location_categories_taille")

            title = _texte(obj.get("title", "")).strip("* ")
            unit_type = normalize_unit_type(taille or title)
            if not re.match(r"^\d½$|^6½\+$|^Studio$|^Loft$|^Chambre", unit_type or ""):
                unit_type = ""

            # la fiche WordPress (déjà téléchargée pour la validation) est plus
            # riche que le popup carte : prix, galerie, disponibilité
            page_main = ""
            m = re.search(r'<div class="entry-content[^"]*">(.*?)</div>\s*'
                          r'(?:<footer|<div class="entry-footer)', page_html, re.S)
            if not m:
                m = re.search(r"<article\b.*?</article>", page_html, re.S)
            if m:
                page_main = m.group(0)
            page_text = _texte(page_main)

            description = _texte(desc_html)
            if 100 < len(page_text) and len(page_text) > len(description):
                description = page_text
            price = None                # le popup carte prime : texte ciblé
            for texte in (_texte(desc_html), page_text):
                m = _PRICE_RE.search(texte) or _PRICE_ALT_RE.search(texte)
                if m and 400 <= float(m.group(1)) <= 5000:
                    price = float(m.group(1))
                    break
            availability = ""
            m = re.search(r"[Dd]isponible[^.\n<]{0,70}", description)
            if m:
                availability = m.group(0).strip()
            images = []
            for u in _IMG_RE.findall(page_main) + _IMG_RE.findall(desc_html):
                u = u.replace("http://", "https://")
                if u not in images:
                    images.append(u)
            try:
                lat, lng = float(obj.get("latitude")), float(obj.get("longitude"))
            except (TypeError, ValueError):
                lat = lng = None
            out.append(Listing(
                source=self.source_id,
                external_id=str(eid),
                url=url,
                title=title,
                sector=sector,
                city=city,
                unit_type=unit_type,
                price=price,
                price_label=(f"{price:,.0f} $/mois".replace(",", " ")
                             if price else ""),
                availability=availability,
                description=description[:6000],
                images=images,
                lat=lat,
                lng=lng,
            ))
        return out