# -----------------------------------------------------------------------------
# Rent-Ka — Rental listings aggregator (Canada, outside Québec)
# Author: Simon-Pierre Boucher — contact@spboucher.ai
# scripts/probe_rentfaster_cities.py : discover RentFaster numeric city_ids.
#   /api/search.json only filters by numeric city_id (1=Calgary, 2=Edmonton,
#   6=Vancouver…) but /api/map.json's `cities` array exposes names without ids.
#   This script probes city_id 1..N once, records {id, city, province, total}
#   and writes data/rentfaster_cities.json (consumed by connectors/rentfaster.py).
#   Province comes from the map.json cities array (name match) with a bbox
#   fallback on the first listing's coordinates. Resumable via the output file.
# Usage: python scripts/probe_rentfaster_cities.py [max_id=700]
# -----------------------------------------------------------------------------
from __future__ import annotations

import json
import sys
import time
from pathlib import Path

import requests

ROOT = Path(__file__).resolve().parents[1]
OUT = ROOT / "data" / "rentfaster_cities.json"
UA = {"User-Agent": "RentKaBot/1.0 (+https://www.rent-ka.com/bot; contact@spboucher.ai)",
      "Accept": "application/json"}

sys.path.insert(0, str(ROOT))
from rentka.schema import PROVINCE_BBOX  # noqa: E402


def bbox_province(lat: float, lng: float) -> str:
    for prov, (lo_lat, hi_lat, lo_lng, hi_lng) in PROVINCE_BBOX.items():
        if lo_lat <= lat <= hi_lat and lo_lng <= lng <= hi_lng:
            return prov
    return ""


def city_provinces() -> dict[str, str]:
    """name(lower) -> province from map.json's cities array (incl. markers)."""
    try:
        d = requests.post("https://www.rentfaster.ca/api/map.json",
                          data={"x": "1"}, headers=UA, timeout=30).json()
    except Exception:
        return {}
    out: dict[str, str] = {}

    def walk(items):
        for c in items or []:
            name = (c.get("city") or "").strip().lower()
            if name and c.get("province"):
                out.setdefault(name, c["province"])
            walk(c.get("markers"))
    walk(d.get("cities"))
    return out


def main() -> None:
    max_id = int(sys.argv[1]) if len(sys.argv) > 1 else 700
    known: dict[str, dict] = {}
    if OUT.exists():
        try:
            known = {str(c["id"]): c
                     for c in json.loads(OUT.read_text())["cities"]}
        except Exception:
            known = {}
    provs = city_provinces()
    print(f"[rf-probe] {len(provs)} city names with province from map.json")
    for cid in range(1, max_id + 1):
        if str(cid) in known:
            continue
        try:
            r = requests.get("https://www.rentfaster.ca/api/search.json",
                             params={"city_id": str(cid), "cur_page": "0"},
                             headers=UA, timeout=25)
            d = r.json()
        except Exception:
            time.sleep(2)
            continue
        ls = d.get("listings") or []
        total = int(d.get("total") or 0)
        if not ls:
            known[str(cid)] = {"id": cid, "city": "", "province": "", "total": 0}
        else:
            city = (ls[0].get("city") or "").strip()
            prov = provs.get(city.lower(), "")
            if not prov:
                try:
                    prov = bbox_province(float(ls[0]["latitude"]),
                                         float(ls[0]["longitude"]))
                except (KeyError, TypeError, ValueError):
                    prov = ""
            known[str(cid)] = {"id": cid, "city": city,
                               "province": prov, "total": total}
        if cid % 50 == 0:
            print(f"[rf-probe] {cid}/{max_id}")
            OUT.write_text(json.dumps(
                {"_comment": "RentFaster city_id map — probed live; consumed "
                             "by connectors/rentfaster.py (QC skipped there).",
                 "updated": time.strftime("%Y-%m-%d"),
                 "cities": sorted(known.values(), key=lambda c: c["id"])},
                ensure_ascii=False, indent=1))
        time.sleep(0.4)
    OUT.write_text(json.dumps(
        {"_comment": "RentFaster city_id map — probed live; consumed by "
                     "connectors/rentfaster.py (QC skipped there).",
         "updated": time.strftime("%Y-%m-%d"),
         "cities": sorted(known.values(), key=lambda c: c["id"])},
        ensure_ascii=False, indent=1))
    active = [c for c in known.values() if c["total"] > 0]
    print(f"[rf-probe] done: {len(active)} active city_ids → {OUT}")


if __name__ == "__main__":
    main()