# -----------------------------------------------------------------------------
# Lou-Ka — Agrégateur de logements à louer (province de Québec)
# Auteur : Simon-Pierre Boucher — contact@spboucher.ai
# connectors/simplex.py : connecteur Simplex Immobilier (simpleximmobilier.com)
#   — agence/gestionnaire de ~600 logements à Trois-Rivières, Shawinigan et en
#   Mauricie (~25-30 annonces publiées à la fois). Site Astro dont le catalogue
#   vit dans un Supabase public : la RPC `get_public_rental_listings_page`
#   (clé publiable `sb_publishable_…` embarquée dans le bundle JS du site)
#   retourne TOUT en 1-2 requêtes JSON : titre, ville/secteur, type d'unité,
#   chambres/sdb, loyer, statut (available/coming_soon/reserved/rented),
#   date de disponibilité, description, inclusions, animaux, stationnement,
#   photos (bucket public `rental-listing-media`).
#   ⚠️ Aucune adresse civique publiée (même la RPC détail n'en expose pas) :
#   ville/secteur seulement, donc pas de géocodage fin possible.
# -----------------------------------------------------------------------------
from __future__ import annotations

import json
import os

from ..schema import Listing
from .base import BaseConnector

SB_URL = "https://tyrzkvwbiaadajwopdrj.supabase.co"
RPC_URL = f"{SB_URL}/rest/v1/rpc/get_public_rental_listings_page"
STORAGE_URL = f"{SB_URL}/storage/v1/object/public"
SITE = "https://simpleximmobilier.com"
# clé PUBLIABLE (anon) du site — surchargeable si Simplex la fait tourner
API_KEY = os.environ.get(
    "LOUKA_SIMPLEX_KEY", "sb_publishable_HT7wke-8-o_TUxMUeLjWbw_wsVK1FC4")

PAGE = 100
_KEEP = {"available", "coming_soon"}
_STATUS_FR = {"available": "Disponible", "coming_soon": "Bientôt disponible"}
_PETS = {"oui": "oui", "non": "non"}


class SimplexConnector(BaseConnector):
    source_id = "simplex"
    request_delay = 0.8

    def _rpc(self, offset: int) -> list[dict]:
        resp = self.post(RPC_URL, data=json.dumps({
            "p_search": None, "p_status": None, "p_location": None,
            "p_unit_type": None, "p_max_price": None, "p_pets": None,
            "p_parking": None, "p_sort": "recommended",
            "p_offset": offset, "p_limit": PAGE,
        }), headers={"apikey": API_KEY,
                     "Authorization": f"Bearer {API_KEY}",
                     "Content-Type": "application/json"})
        data = resp.json()
        return data if isinstance(data, list) else []

    def fetch(self) -> list[Listing]:
        rows: list[dict] = []
        offset = 0
        while True:
            batch = self._rpc(offset)
            rows.extend(batch)
            total = batch[0].get("total_count", 0) if batch else 0
            offset += PAGE
            if not batch or offset >= total:
                break
        listings: dict[str, Listing] = {}
        for row in rows:
            if row.get("availability_status") not in _KEEP:
                continue
            lst = self._listing(row)
            if lst and lst.uid not in listings:
                listings[lst.uid] = lst
        return list(listings.values())

    def _images(self, row: dict) -> list[str]:
        media = sorted(row.get("media") or [],
                       key=lambda m: (not m.get("isPrimary"),
                                      m.get("sortOrder") or 0))
        out = []
        for m in media:
            if m.get("bucket") and m.get("path"):
                out.append(f"{STORAGE_URL}/{m['bucket']}/{m['path']}")
        for p in [row.get("primary_image_path"),
                  *(row.get("gallery_paths") or [])]:
            if p and p.startswith("http") and p not in out:
                out.append(p)
        return out[:20]

    def _listing(self, row: dict) -> Listing | None:
        price = row.get("rent_amount")
        status = row.get("availability_status") or ""
        avail_txt = _STATUS_FR.get(status, "")
        avail_date = (row.get("available_from") or "")[:10]
        if status == "coming_soon" and avail_date:
            avail_txt = f"{avail_txt} — à compter du {avail_date}"
        pets_raw = (row.get("pets_policy") or "").strip().lower()
        amen = [a for a in (row.get("inclusions") or []) +
                (row.get("features") or []) + (row.get("labels") or []) if a]
        details = {}
        if row.get("parking"):
            details["stationnement"] = str(row["parking"]).strip()
        if row.get("public_reference"):
            details["référence"] = row["public_reference"]
        if row.get("rent_note"):
            details["note loyer"] = row["rent_note"]
        return Listing(
            source=self.source_id,
            external_id=str(row.get("id") or row.get("slug")),
            url=f"{SITE}/logements-a-louer/{row.get('slug', '')}/",
            title=(row.get("title") or "").strip(),
            sector=(row.get("sector") or "").strip(),
            city=(row.get("city") or "").strip(),
            unit_type=(row.get("unit_type") or "").strip(),
            bedrooms=(float(row["bedrooms"])
                      if row.get("bedrooms") is not None else None),
            bathrooms=(float(row["bathrooms"])
                       if row.get("bathrooms") is not None else None),
            price=float(price) if price else None,
            price_label=(f"{price:,.0f} $/mois".replace(",", " ")
                         if price else ""),
            availability=avail_txt,
            availability_date=("now" if status == "available"
                               else (avail_date or None)),
            pets=_PETS.get(pets_raw.rstrip(" .!")),
            description=(row.get("description") or "")[:2000],
            amenities=amen[:30],
            details=details,
            images=self._images(row),
        )