# -----------------------------------------------------------------------------
# Sorti-Ka — Agrégateur de sorties & événements (province de Québec)
# Auteur : Simon-Pierre Boucher — contact@spboucher.ai
# connectors/evenko.py : connecteur evenko — grands concerts & festivals
#   Source    : evenko.ca (promoteur : Centre Bell, Place Bell, MTELUS, Osheaga…)
#   Extraction: endpoint /api/search (proxy Algolia interne) — EXPLICITEMENT
#               permis par robots.txt (« Allow: /api/search* »). Format percé
#               dans leur bundle JS : GET ?body=base64(JSON{params:[query,
#               options], lang:"fr-CA"}). ~14 requêtes/sync (100 hits/page).
#   Accès     : robots.txt Allow /api/search* ; GET throttlés, identification
#               honnête ; lien billets = URL Ticketmaster officielle du hit.
#   Prix      : NON PUBLIÉS par la source. Champ `free` → is_free, MAIS sondé
#               null sur 100 % des hits (462/462 vérifié 2026-08 ; re-vérifié
#               660/660 le 2026-08-25) et l'index
#               ne porte aucun montant (additional_information.regular_ticket
#               = libellés + URLs Ticketmaster seulement, filtre free:true →
#               0 hit ; event_discount = {percentage, amount, minimum} d'un
#               rabais, pas un tarif). Re-sondé live 2026-08-25 sur l'index
#               COMPLET (660 hits, tous attributs) ET sur les pages événement
#               /fr/evenements/<salle>/<slug> de 6 salles : le JSON-LD Event
#               y porte offers SANS price/priceCurrency (URL Ticketmaster +
#               availability seulement) — les montants ne vivent que chez
#               Ticketmaster (couvert par le connecteur ticketmaster, API
#               officielle priceRanges). Complément honnête : étiquettes
#               explicites « gratuit » dans event_tag / promotional_tag →
#               is_free=True. On n'invente rien.
# -----------------------------------------------------------------------------
from __future__ import annotations

import base64
import json
from datetime import datetime, timezone

from ..schema import Event, normalize_status
from .base import BaseConnector

API = "https://evenko.ca/api/search"
PAGE_SIZE = 100

ATTRS = ["objectID", "title", "subtitle", "headliners", "supports", "venue",
         "show_date", "show_time", "hide_show_time", "door_time", "tour_name",
         "category", "genre", "description",
         "free", "age", "thumbnail", "status", "additional_information",
         "presented_by", "event_tag", "promotional_tag"]

# les salles evenko hors Québec sont exclues du périmètre (Sorti-Ka = QC)
_EXCLUDED_REGIONS = ("ottawa", "ontario", "nouveau-brunswick", "nouvelle-écosse",
                     "nouvelle-ecosse", "terre-neuve", "états-unis", "etats-unis",
                     "île-du-prince-édouard", "ile-du-prince-edouard", "manitoba")


def _unix_to_iso(ts) -> str | None:
    try:
        return datetime.fromtimestamp(int(ts), tz=timezone.utc).strftime("%Y-%m-%d")
    except (TypeError, ValueError, OSError):
        return None


def _unix_to_local_time(ts) -> str | None:
    """show_time (epoch) → heure locale du spectacle ("19:30"), Québec."""
    from zoneinfo import ZoneInfo
    try:
        t = datetime.fromtimestamp(int(ts), tz=ZoneInfo("America/Montreal"))
        return None if t.strftime("%H:%M") == "00:00" else t.strftime("%H:%M")
    except (TypeError, ValueError, OSError):
        return None


def _ticket_url(hit: dict) -> str:
    """URL billets officielle (fr_CA) depuis additional_information."""
    reg = (hit.get("additional_information") or {}).get("regular_ticket") or {}
    for entry in reg.get("regular_ticket_url") or []:
        if entry.get("locale") == "fr_CA" and entry.get("data"):
            return entry["data"]
    for entry in reg.get("regular_ticket_url") or []:
        if entry.get("data"):
            return entry["data"]
    return ""


class EvenkoConnector(BaseConnector):
    source_id = "evenko"

    def _search_page(self, page: int) -> dict:
        payload = {"params": ["", {"hitsPerPage": PAGE_SIZE, "page": page,
                                   "filters": "entity_type:evenko_show",
                                   "attributesToRetrieve": ATTRS}],
                   "lang": "fr-CA"}
        body = base64.b64encode(json.dumps(payload).encode()).decode()
        return self.get_json(API, params={"body": body})

    def fetch(self) -> list[Event]:
        events: list[Event] = []
        page, nb_pages = 0, 1
        while page < nb_pages:
            data = self._search_page(page)
            nb_pages = min(int(data.get("nbPages", 0)), 50)  # garde-fou
            for hit in data.get("hits", []):
                ext_id = (hit.get("objectID") or "").removesuffix("_fr")
                title = hit.get("title") or ""
                if not ext_id or not title:
                    continue
                venue = hit.get("venue") or {}
                region_lbl = (venue.get("region") or "").lower()
                if any(x in region_lbl for x in _EXCLUDED_REGIONS):
                    continue
                if hit.get("subtitle"):
                    title = f"{title} — {hit['subtitle']}"
                cats = [g.get("name", "") for g in hit.get("genre") or []]
                cat = hit.get("category")
                if isinstance(cat, list):
                    cats += [str(c) for c in cat]
                elif cat:
                    cats.append(str(cat))
                thumb = hit.get("thumbnail") or ""
                if thumb.startswith("//"):
                    thumb = "https:" + thumb
                day = _unix_to_iso(hit.get("show_date")) or _unix_to_iso(hit.get("show_time"))
                # heure précise : show_time (epoch) sauf si la source demande
                # explicitement de la cacher (hide_show_time)
                start_time = (None if hit.get("hide_show_time")
                              else _unix_to_local_time(hit.get("show_time")))
                # artistes : têtes d'affiche (`headliners`) + premières
                # parties (`supports`) — champs explicites de la source
                artists = [str(a) for a in hit.get("headliners") or [] if a]
                artists += [str(a) for a in hit.get("supports") or [] if a]
                # statut billetterie : additional_information.representation_status
                # (['buy_now'], ['sold_out'], ['cancelled']…) — jamais inventé
                rep = (hit.get("additional_information") or {}) \
                    .get("representation_status") or []
                status = ""
                for token in (rep if isinstance(rep, list) else [rep]):
                    status = normalize_status(str(token))
                    if status:
                        break
                # description enrichie : tournée + heure des portes (factuels,
                # publiés par la source ; aucun champ dédié au schéma)
                description = hit.get("description") or ""
                extras = []
                if hit.get("tour_name"):
                    extras.append(f"Tournée : {str(hit['tour_name']).strip()}")
                door = _unix_to_local_time(hit.get("door_time"))
                if door:
                    extras.append(f"Portes : {door.replace(':', ' h ')}")
                if extras:
                    description = " — ".join(
                        p for p in (description, " · ".join(extras)) if p)
                # gratuité : champ `free` (aujourd'hui toujours null côté
                # source, gardé au cas où il serait re-rempli) + étiquettes
                # explicites « gratuit » (event_tag / promotional_tag).
                is_free = bool(hit["free"]) if hit.get("free") is not None else None
                tags = f"{hit.get('event_tag') or ''},{hit.get('promotional_tag') or ''}"
                if is_free is None and "gratuit" in tags.lower():
                    is_free = True
                events.append(Event(
                    source=self.source_id,
                    external_id=ext_id,
                    url=_ticket_url(hit) or (venue.get("website") or "https://evenko.ca/fr/calendrier"),
                    title=title,
                    description=description,
                    raw_categories=[c for c in cats if c] or ["Concert"],
                    audience={"everyone": "Pour tous", "all_ages": "Pour tous",
                              "18+": "18 ans et plus", "16+": "16 ans et plus",
                              }.get(hit.get("age") or "", hit.get("age") or ""),
                    venue=venue.get("name") or "",
                    city=venue.get("city") or "",
                    start_date=day,
                    start_time=start_time,
                    end_date=day,
                    status=status,
                    artists=artists,
                    is_free=is_free,
                    organizer="evenko",
                    image=thumb,
                ))
            page += 1
        return events