# -----------------------------------------------------------------------------
# Lou-Ka — Agrégateur de logements à louer (province de Québec)
# Auteur : Simon-Pierre Boucher — contact@spboucher.ai
# connectors/gerik.py : connecteur Gerik (gerik.ca) — Outaouais
#   Promoteur/gestionnaire de Gatineau (Hull-Aylmer, Gatineau, Cantley,
#   Masson-Angers). WordPress/WPBakery rendu serveur : la page
#   /appartements-a-louer-gatineau/ liste les projets locatifs en cartes
#   (h2 nom du projet + méta « Hull-Aylmer · Studios, 1 et 2 chambres ·
#   Disponible dès maintenant » + description + lien /projets/<slug>/).
#   AUCUN prix publié nulle part (ni carte, ni fiche projet) : on remplit
#   availability et on laisse price/price_label vides. Granularité = projet ×
#   typologie (studio / n chambres) — le site n'expose pas d'inventaire par
#   unité. Seuls les projets marqués « Disponible » sont retenus (les projets
#   « À VENIR » / « En construction » / « en développement » sont exclus).
#   Photos : classes CSS vc_custom_* (background-image) mappées aux cartes.
#   external_id = <slug du projet>-<typologie> — stable.
# -----------------------------------------------------------------------------
from __future__ import annotations

import re

from bs4 import BeautifulSoup

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

BASE = "https://gerik.ca"
LIST_URL = f"{BASE}/appartements-a-louer-gatineau/"

BG_RE = re.compile(r"\.(vc_custom_\w+)\{background-image:\s*url\(([^)?\s]+)")
TYPES_RE = re.compile(r"Studios?|(\d+)(?:,\s*\d+)*(?:\s*et\s*\d+)?\s*chambres?",
                      re.I)
SEP = "·"

# secteur affiché sur la carte -> (secteur Lou-Ka, ville)
PLACES = {
    "hull-aylmer": ("Aylmer", "Gatineau"),
    "aylmer": ("Aylmer", "Gatineau"),
    "hull": ("Hull", "Gatineau"),
    "gatineau": ("", "Gatineau"),
    "masson-angers": ("Masson-Angers", "Gatineau"),
    "cantley": ("", "Cantley"),
    "chelsea": ("", "Chelsea"),
}


def _slug(text: str) -> str:
    s = strip_accents(text.lower())
    return re.sub(r"[^a-z0-9]+", "-", s).strip("-")


class GerikConnector(BaseConnector):
    source_id = "gerik"
    request_delay = 0.8

    def fetch(self) -> list[Listing]:
        html = self.get(LIST_URL).text
        soup = BeautifulSoup(html, "html.parser")
        bg_map = dict(BG_RE.findall(html))     # classe vc_custom_* -> photo

        listings: list[Listing] = []
        seen: set[str] = set()
        for h2 in soup.find_all("h2"):
            name = re.sub(r"\s+", " ", h2.get_text(" ", strip=True))
            if not name or "Aucun projet" in name or "Découvrez" in name \
                    or "partenaire" in name:
                continue
            row = h2
            for _ in range(8):                 # remonter à la rangée du projet
                row = row.parent
                if row is None:
                    break
                if "vc_row" in (row.get("class") or []):
                    break
            if row is None or "vc_row" not in (row.get("class") or []):
                continue
            try:
                listings.extend(self._parse_project(name, row, bg_map, seen))
            except Exception:
                continue
        return listings

    def _parse_project(self, name: str, row, bg_map: dict,
                       seen: set[str]) -> list[Listing]:
        proj_slug = _slug(name)
        if proj_slug in seen:
            return []
        seen.add(proj_slug)

        text = re.sub(r"\s+", " ", row.get_text(" | ", strip=True))
        # méta « Hull-Aylmer · Studios, 1 et 2 chambres · Disponible … »
        meta_el = row.find(string=re.compile(SEP))
        meta = re.sub(r"\s+", " ", str(meta_el)).strip() if meta_el else ""
        parts = [p.strip() for p in meta.split(SEP) if p.strip()]

        availability = next((p for p in parts
                             if re.search(r"disponible", p, re.I)), "")
        if not availability:
            return []                          # à venir / en construction

        sector, city = "", "Gatineau"
        if parts:
            key = strip_accents(parts[0].lower()).strip()
            sector, city = PLACES.get(key, (parts[0], "Gatineau"))

        # typologies : « Studios, 1 et 2 chambres » -> Studio, 1 ch, 2 ch
        types: list[tuple[str, float | None]] = []
        types_part = next((p for p in parts
                           if re.search(r"studio|chambre", p, re.I)), "")
        if re.search(r"studio", types_part, re.I):
            types.append(("Studio", None))
        nums = re.findall(r"\d+", re.sub(r"\d+\s*unités?", "", types_part))
        for n in nums:
            types.append((f"{n} chambres" if int(n) > 1 else "1 chambre",
                          float(n)))
        if not types:
            types = [("", None)]

        # description : paragraphe le plus long de la carte
        description = ""
        for p in row.find_all("p"):
            t = re.sub(r"\s+", " ", p.get_text(" ", strip=True))
            if len(t) > len(description) and SEP not in t:
                description = t
        description = description[:1200]

        url = LIST_URL
        a = row.find("a", href=re.compile(r"/projets/"))
        if a:
            url = a["href"]
            if not url.startswith("http"):
                url = BASE + url

        images: list[str] = []
        for el in row.select("[class*='vc_custom_']"):
            for cls in (el.get("class") or []):
                u = bg_map.get(cls)
                if u and u not in images:
                    images.append(u)

        details: dict = {}
        units_part = next((p for p in parts
                           if re.search(r"\d+\s*unités?", p, re.I)), "")
        m = re.search(r"(\d+)\s*unités?", units_part, re.I)
        if m:
            details["building_units"] = int(m.group(1))

        listings = []
        for label, beds in types:
            suffix = "studio" if label == "Studio" else \
                f"{int(beds)}ch" if beds else "logement"
            listings.append(Listing(
                source=self.source_id,
                external_id=f"{proj_slug}-{suffix}",
                url=url,
                title=f"{name} — {label}" if label else name,
                address="",
                sector=sector,
                city=city,
                unit_type="Studio" if label == "Studio" else "",
                bedrooms=beds,
                availability=availability,
                description=description,
                details=dict(details),
                images=list(images),
            ))
        return listings