#!/usr/bin/env python3
# -----------------------------------------------------------------------------
# Immo-Ka — régénère data/via_capitale_agencies.json en sondant les domaines
# des sous-agences Via Capitale (plateforme source.immo). Étend facilement la
# couverture : ajoute des bannières/domaines à CANDIDATES et relance.
#   Usage : .venv/bin/python scripts/refresh_via_capitale.py
# -----------------------------------------------------------------------------
from __future__ import annotations

import json
import re
import urllib.request
from pathlib import Path

UA = {"User-Agent": "Mozilla/5.0 Chrome/126"}
CFG = "/wp-content/uploads/_sourceimmo/_configs.json"
OUT = Path(__file__).resolve().parent.parent / "data" / "via_capitale_agencies.json"

# Domaines connus + bannières à tester (viacapitale<x>.{immo,com,ca}).
KNOWN_DOMAINS = [
    "viacapitaleelite.immo", "viacapitaleselect.com",
    "viacapitaledumontroyal.com", "viacapitaleacces.com",
    "viacapitalepartenaires.com",
]
BANNERS = [
    "elite", "select", "dumontroyal", "acces", "horizon", "diamant", "expertise",
    "cite", "vendu", "action", "distinction", "premium", "platine", "signature",
    "prestige", "excellence", "sommet", "boreal", "alliance", "synergie", "expert",
    "optima", "concept", "capitale", "3000", "2000", "or", "plus", "inter",
]
TLDS = [".immo", ".com", ".ca"]


def fetch_cfg(domain: str) -> dict | None:
    for host in (domain, f"www.{domain}"):
        try:
            b = urllib.request.urlopen(
                urllib.request.Request(f"https://{host}{CFG}", headers=UA), timeout=12
            ).read().decode("utf-8", "ignore")
            d = json.loads(b)
            if d.get("account_id") and d.get("api_key") and d.get("default_view"):
                return {"domain": host, "cfg": d}
        except Exception:
            continue
    return None


def main() -> None:
    domains = set(KNOWN_DOMAINS)
    for b in BANNERS:
        for t in TLDS:
            domains.add(f"viacapitale{b}{t}")
    reg, seen = [], set()
    for dom in sorted(domains):
        r = fetch_cfg(dom)
        if not r or r["domain"] in seen:
            continue
        seen.add(r["domain"])
        d = r["cfg"]
        banner = re.sub(r"^www\.viacapitale|\.(immo|com|ca)$", "", r["domain"])
        reg.append({
            "banner": banner, "domain": r["domain"],
            "account": d["account_id"], "api": d["api_key"],
            "app": d["app_id"], "view": d["default_view"],
        })
        print(f"OK {banner:16} {r['domain']}")
    OUT.write_text(json.dumps(reg, ensure_ascii=False, indent=1), encoding="utf-8")
    print(f"→ {len(reg)} sous-agences écrites dans {OUT}")


if __name__ == "__main__":
    main()