# -----------------------------------------------------------------------------
# Sorti-Ka — Agrégateur de sorties & événements (province de Québec)
# Auteur : Simon-Pierre Boucher — contact@spboucher.ai
# connectors/saguenaysorties.py : Sorties Saguenay — portail officiel de la
#   Ville de Saguenay (Phase 3). Remplace le CSV Données Québec mort (2022).
#   Source    : https://sorties.saguenay.ca — portail événementiel municipal
#               (motorisé par les données La Vitrine culturelle, médias
#               lvc-medias.lavitrine.com ; recoupement géré par dedup_key).
#   Extraction: POST /ajax/plus (offset par pas de 36, ~300-400 séances
#               horodatées, vérifié 2026-08-19) — chaque <article> = UNE séance :
#               discipline, date FR, heure « 08 h 30 », titre, « Ville, Salle »,
#               image, data-detail-id. Détails incrémentaux POST /ajax/details
#               (cache data/saguenaysorties_cache.json, SAG_DETAIL_MAX/sync,
#               re-visite 7 j) : description + adresse civique.
#   Accès     : portail public municipal, pas de robots.txt publié (page de
#               maintenance) — volumes modestes, POST throttlés 0,8 s, UA
#               honnête, lien vers la fiche officielle /festivals?id=<id>.
# -----------------------------------------------------------------------------
from __future__ import annotations

import os
import re
import time

from ..normalize import clean_text, parse_date_fr, parse_time
from ..schema import Event
from .base import BaseConnector

BASE = "https://sorties.saguenay.ca"
AJAX_LIST = BASE + "/ajax/plus"
AJAX_DETAIL = BASE + "/ajax/details"
BATCH = 36
BATCH_MAX = int(os.environ.get("SAG_BATCH_MAX", "15"))      # ≤ 540 séances
DETAIL_MAX = int(os.environ.get("SAG_DETAIL_MAX", "40"))
REFRESH_AFTER = 7 * 86400

_DISC_RE = re.compile(r'discipline_name">\s*([^<]+?)\s*</div>', re.S)
_IMG_RE = re.compile(r'<img src="([^"]+)"')
_DATE_RE = re.compile(r'article-date[^>]*>\s*([^<]+?)\s*</p>', re.S)
_TITLE_RE = re.compile(r'data-detail-id="(\d+)"[^>]*>\s*([^<]+?)\s*</a>', re.S)
_PLACE_RE = re.compile(r'</h3>\s*<p>\s*(.*?)\s*</p>', re.S)
_ADDR_RE = re.compile(r'<br\s*/?>\s*([^<]+?)\s*</p>', re.S)


class SaguenaySortiesConnector(BaseConnector):
    source_id = "saguenaysorties"
    request_delay = 0.8

    def _post(self, url: str, data: dict) -> str:
        self._throttle()
        resp = self.session.post(url, data=data, timeout=self.timeout)
        self._last_request = time.time()
        resp.raise_for_status()
        return resp.text

    def _parse_batch(self, html: str) -> list[dict]:
        rows = []
        for art in html.split("<article")[1:]:
            m_title = _TITLE_RE.search(art)
            if not m_title:
                continue
            dates = _DATE_RE.findall(art)
            date_iso = parse_date_fr(dates[0]) if dates else None
            heure = parse_time(dates[1]) if len(dates) > 1 else None
            m_place = _PLACE_RE.search(art)
            city = venue = ""
            if m_place:
                place = clean_text(m_place.group(1))
                if "," in place:
                    city, venue = (p.strip() for p in place.split(",", 1))
                else:
                    city = place
            # événements multi-lieux : la carte liste TOUS les sites séparés
            # par des virgules — un « lieu » de 100+ caractères n'est pas une
            # salle, on ne garde que la ville (jamais un faux GPS de salle)
            if len(venue) > 100:
                venue = ""
            if len(city) > 60:
                city = ""
            disc = _DISC_RE.search(art)
            img = _IMG_RE.search(art)
            rows.append({
                "detail_id": m_title.group(1),
                "title": clean_text(m_title.group(2)),
                "date": date_iso, "time": heure,
                "city": city, "venue": venue,
                "discipline": clean_text(disc.group(1)) if disc else "",
                "image": img.group(1) if img else "",
            })
        return rows

    def _details(self, ids: list[str]) -> dict:
        """description + adresse par detail_id — incrémental, plafonné."""
        cache = self.load_cache()
        now = time.time()
        to_fetch = sorted(
            (i for i in ids
             if i not in cache or now - cache[i].get("ts", 0) > REFRESH_AFTER),
            key=lambda i: cache.get(i, {}).get("ts", 0))[:DETAIL_MAX]
        for did in to_fetch:
            try:
                html = self._post(AJAX_DETAIL, {"id": did})
                m = re.search(r"</h3>\s*(.*?)\s*<h4", html, re.S)
                desc = clean_text(m.group(1)) if m else ""
                m_addr = _ADDR_RE.search(html)
                addr = clean_text(m_addr.group(1)) if m_addr else ""
                cache[did] = {"ts": now, "desc": desc, "addr": addr}
            except Exception:
                cache[did] = {"ts": now, "desc": "", "addr": ""}
        self.save_cache(cache)
        return cache

    def fetch(self) -> list[Event]:
        rows: list[dict] = []
        for i in range(BATCH_MAX):
            html = self._post(AJAX_LIST, {"offset": str(i * BATCH), "page": "1"})
            batch = self._parse_batch(html)
            rows.extend(batch)
            if len(batch) < BATCH // 2:       # dernière page (lots partiels vus)
                break
        details = self._details(sorted({r["detail_id"] for r in rows}))

        events: list[Event] = []
        seen: set[str] = set()
        for r in rows:
            ext = f"{r['detail_id']}-{r['date'] or 'nd'}-{r['time'] or 'nd'}"
            if ext in seen or not r["title"]:
                continue
            seen.add(ext)
            det = details.get(r["detail_id"]) or {}
            events.append(Event(
                source=self.source_id,
                external_id=ext,
                url=f"{BASE}/festivals?id={r['detail_id']}",
                title=r["title"],
                description=det.get("desc", ""),
                raw_categories=[r["discipline"]] if r["discipline"] else [],
                venue=r["venue"],
                address=det.get("addr", ""),
                city=r["city"],
                tourist_region="Saguenay–Lac-Saint-Jean",
                start_date=r["date"],
                start_time=r["time"],
                image=r["image"],
            ))
        return events