# -----------------------------------------------------------------------------
# Lou-Ka — Location court terme
# connectors/_expediadetail.py : parseur partagé des pages détail de la
# plateforme Expedia (Vrbo + Expedia — même moteur, même SSR).
#
# La page détail (…/location/pXXXXXX[vb] ou …hXXXXXX.Hotel-Information),
# récupérée via Scrapfly ASP SANS rendu JS, embarque :
#   - __APOLLO_STATE__ (JSON.parse("…")) → PropertyInfo :
#       . propertyContentSectionGroups(…).aboutThisProperty → description
#         (blocs PropertyContentItemMarkup, HTML — inclut souvent le n° CITQ)
#       . summary.amenities(…) → commodités localisées (infoItems[].text)
#   - microdonnées schema.org SSR : lat/lng (itemProp latitude/longitude),
#     capacité (occupancy → value), municipalité (addressLocality)
#   - galerie : URLs media.vrbo.com / images.trvl-media.com (~6 photos SSR)
# Vérifié live 2026-08-25 sur p3363083vb (Vrbo) et h130341845 (Expedia).
# -----------------------------------------------------------------------------
from __future__ import annotations

import html as _html
import json
import re

_APOLLO_RE = re.compile(
    r'__APOLLO_STATE__\s*=\s*JSON\.parse\("(.*?)(?<!\\)"\)', re.S)
_LAT_RE = re.compile(r'itemProp="latitude" content="(-?[\d.]+)"')
_LNG_RE = re.compile(r'itemProp="longitude" content="(-?[\d.]+)"')
_OCC_RE = re.compile(
    r'itemProp="occupancy".{0,200}?itemProp="value" content="(\d+)"', re.S)
_CITY_RE = re.compile(r'itemProp="addressLocality" content="([^"]+)"')
_IMG_RE = re.compile(
    r'https://(?:media\.vrbo\.com|images\.trvl-media\.com)/lodging/'
    r'[^"\s\\)&?]+')


def _apollo(html: str) -> dict:
    """Entités du store Apollo SSR ({} si absent/illisible)."""
    m = _APOLLO_RE.search(html or "")
    if not m:
        return {}
    try:
        # la chaîne est un littéral JSON : la re-quoter puis parser deux fois
        return json.loads(json.loads('"' + m.group(1) + '"'))
    except ValueError:
        return {}


def _markup_texts(node, out: list[str]) -> None:
    """Collecte récursive des blocs PropertyContentItemMarkup (description)."""
    if isinstance(node, dict):
        if node.get("__typename") == "PropertyContentItemMarkup":
            txt = ((node.get("content") or {}).get("text") or "").strip()
            if txt:
                out.append(txt)
            return
        for v in node.values():
            _markup_texts(v, out)
    elif isinstance(node, list):
        for v in node:
            _markup_texts(v, out)


def _strip_html(raw: str) -> str:
    t = re.sub(r"<br\s*/?>|</p>", "\n", raw)
    t = re.sub(r"<[^>]+>", " ", t)
    t = _html.unescape(t)
    lines = [re.sub(r"\s+", " ", ln).strip() for ln in t.split("\n")]
    return "\n".join(ln for ln in lines if ln).strip()


def parse_detail(html: str) -> dict:
    """Payload détail {description, amenities, capacity, lat, lng, city,
    images} d'une page hébergement Vrbo/Expedia ({} si page invalide)."""
    store = _apollo(html or "")
    pinfo = next((v for k, v in store.items()
                  if k.startswith("PropertyInfo") and isinstance(v, dict)), {})
    if not pinfo and not _LAT_RE.search(html or ""):
        return {}                       # page vide / redirection hors fiche

    out: dict = {}

    # description : sections « À propos de cet hébergement »
    paras: list[str] = []
    for k, v in pinfo.items():
        if k.startswith("propertyContentSectionGroups") and isinstance(v, dict):
            _markup_texts(v.get("aboutThisProperty"), paras)
    if not paras:                       # repli : éditorial du quartier
        loc = (pinfo.get("summary") or {}).get("location") or {}
        ed = ((loc.get("whatsAround") or {}).get("editorial") or {})
        paras = [t for t in (ed.get("content") or []) if isinstance(t, str)]
    desc = "\n\n".join(_strip_html(p) for p in paras).strip()
    if desc:
        out["description"] = desc[:6000]

    # commodités localisées (summary.amenities → sections → infoItems)
    amenities: list[str] = []
    for k, v in (pinfo.get("summary") or {}).items():
        if not (k.startswith("amenities") and isinstance(v, dict)):
            continue
        for sec in v.get("amenities") or []:
            for cont in (sec or {}).get("contents") or []:
                for it in (cont or {}).get("infoItems") or []:
                    txt = ((it or {}).get("text") or "").strip()
                    if txt and txt not in amenities:
                        amenities.append(txt)
    if amenities:
        out["amenities"] = amenities[:80]

    # microdonnées SSR : géo, capacité, municipalité
    m = _LAT_RE.search(html)
    n = _LNG_RE.search(html)
    if m and n:
        try:
            out["lat"], out["lng"] = float(m.group(1)), float(n.group(1))
        except ValueError:
            pass
    m = _OCC_RE.search(html)
    if m:
        out["capacity"] = float(m.group(1))
    m = _CITY_RE.search(html)
    if m:
        out["city"] = _html.unescape(m.group(1)).strip()

    # galerie SSR : dédupliquée par chemin, servie en 1200 px
    images: list[str] = []
    for u in _IMG_RE.findall(html):
        big = u + "?impolicy=resizecrop&rw=1200&ra=fit"
        if big not in images:
            images.append(big)
        if len(images) >= 15:
            break
    if images:
        out["images"] = images
    return out


def apply_detail(lst, d: dict) -> None:
    """Applique le payload détail sans écraser ce que la carte a fourni
    (sauf galerie : on garde la plus grande)."""
    if not d:
        return
    if d.get("description") and len(d["description"]) > len(lst.description or ""):
        lst.description = d["description"]
    if d.get("amenities"):
        seen = {a.lower() for a in lst.amenities}
        for a in d["amenities"]:
            if a.lower() not in seen:
                lst.amenities.append(a)
                seen.add(a.lower())
    for f in ("capacity", "lat", "lng"):
        if d.get(f) is not None and getattr(lst, f, None) is None:
            setattr(lst, f, d[f])
    if d.get("city") and not lst.city:
        lst.city = d["city"]
    if d.get("images") and len(d["images"]) > len(lst.images):
        lst.images = list(d["images"])