#!/usr/bin/env python3
"""Build unique HTTP(S) probe list from resources.conf + connect_check catalog (C builtins).

Output TSV: tag\\turl  (tag = section|name|host)
Prefer resources.conf paths; catalog hosts get https://host/ (captive specials as in CC).
"""
from __future__ import annotations

import argparse
import json
import sys
from pathlib import Path
from urllib.parse import urlparse

ROOT = Path(__file__).resolve().parents[1]


def _captive_url(host: str, name: str = "") -> str:
    h = host.lower()
    n = (name or "").lower()
    if "gstatic" in h or "android.com" in h or "clients" in h and "google" in h:
        return f"http://{h}/generate_204"
    if "captive.apple" in h:
        return f"http://{h}/hotspot-detect.html"
    if "msftconnecttest" in h or "msftncsi" in h:
        return f"http://{h}/connecttest.txt"
    if "detectportal" in h:
        return f"http://{h}/success.txt"
    if "network-test.debian" in h:
        return f"http://{h}/"
    if "generate_204" in n or "captive" in n:
        return f"http://{h}/"
    return f"https://{h}/"


def build(conf_path: Path, catalog_path: Path) -> list[tuple[str, str]]:
    by_host: dict[str, tuple[str, str]] = {}

    section = None
    for ln in conf_path.read_text(errors="replace").splitlines():
        ln = ln.strip()
        if not ln or ln.startswith("#"):
            continue
        if ln.startswith("[") and ln.endswith("]"):
            section = ln[1:-1]
            continue
        parts = [p.strip() for p in ln.split("|")]
        if len(parts) < 2 or not parts[1].startswith("http"):
            continue
        name, url = parts[0], parts[1]
        host = (urlparse(url).hostname or "").lower()
        if not host:
            continue
        tag = f"{section or 'conf'}|{name}|{host}"
        by_host[host] = (tag, url)

    cat = json.loads(catalog_path.read_text())
    for item in cat.get("items") or []:
        host = (item.get("host") or "").lower().strip()
        if not host or host in by_host:
            continue
        name = item.get("name") or host
        section = item.get("section") or "catalog"
        url = _captive_url(host, name) if section == "captive" else f"https://{host}/"
        # DoH / known http endpoints
        if section == "doh" and "cloudflare-dns" in host:
            url = f"https://{host}/dns-query"
        if section == "doh" and "dns.google" in host:
            url = f"https://{host}/dns-query"
        if "gstatic.com" in host or host.endswith("android.com"):
            url = f"http://{host}/generate_204"
        tag = f"{section}|{name}|{host}"
        by_host[host] = (tag, url)

    rows = sorted(by_host.values(), key=lambda x: (x[0].split("|")[0], x[1]))
    return rows


def main() -> int:
    ap = argparse.ArgumentParser()
    ap.add_argument("--conf", type=Path, default=ROOT / "var" / "resources.conf")
    ap.add_argument(
        "--catalog", type=Path, default=ROOT / "var" / "connect_check_catalog.json"
    )
    ap.add_argument("-o", "--output", type=Path, required=True)
    args = ap.parse_args()
    rows = build(args.conf, args.catalog)
    args.output.parent.mkdir(parents=True, exist_ok=True)
    args.output.write_text("".join(f"{t}\t{u}\n" for t, u in rows))
    print(f"wrote {len(rows)} urls → {args.output}", file=sys.stderr)
    return 0


if __name__ == "__main__":
    sys.exit(main())