"""Волна по китайским/азиатским источникам: форумы, реверс, security-новости.

Проверяет кандидатов живыми запросами (лента / список тем), пишет живые в
omni/sources_cn.py. Запуск:  python tools/wave_cn.py [--apply]
"""
import sys
import time
from pathlib import Path

import requests

ROOT = Path(__file__).resolve().parents[1]
OUT = ROOT / "omni" / "sources_cn.py"

UA = ("Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 "
      "(KHTML, like Gecko) Chrome/131.0 Safari/537.36")

# name, url, adapter, category, tags, limit
CAND = [
    ("cn_52pojie", "https://www.52pojie.cn/forum.php?mod=guide&view=new", "html_listing", "forum", ("cn", "reverse", "abuse"), 20),
    ("cn_nodeseek", "https://www.nodeseek.com/", "html_listing", "forum", ("cn", "hosting"), 20),
    ("cn_freebuf", "https://www.freebuf.com/", "html_listing", "security", ("cn", "security"), 20),
    ("cn_seebug", "https://paper.seebug.org/", "html_listing", "security", ("cn", "papers", "security"), 15),
    ("cn_anquanke", "https://www.anquanke.com/", "html_listing", "security", ("cn", "security"), 20),
    ("cn_4hou", "https://www.4hou.com/", "html_listing", "security", ("cn", "security"), 15),
    ("cn_kanxue", "https://bbs.kanxue.com/", "html_listing", "forum", ("cn", "reverse"), 20),
    ("cn_ithome", "https://www.ithome.com/rss/", "rss", "news", ("cn", "tech"), 20),
    ("cn_cnbeta", "https://www.cnbeta.com.tw/backend.php", "rss", "news", ("cn", "tech"), 20),
    ("cn_sspai", "https://sspai.com/feed", "rss", "news", ("cn", "tech"), 15),
    ("cn_chiphell", "https://www.chiphell.com/forum.php?mod=guide&view=new", "html_listing", "forum", ("cn", "hardware"), 15),
    ("cn_52pojie_rss", "https://www.52pojie.cn/forum.php?mod=rss", "rss", "forum", ("cn", "reverse"), 15),
    ("cn_v2ex_rss", "https://www.v2ex.com/index.xml", "rss", "forum", ("cn", "community"), 20),
    ("cn_linuxdo_rss", "https://linux.do/latest.rss", "rss", "forum", ("cn", "community", "ai"), 20),
    ("cn_hostloc_rss", "https://hostloc.com/forum.php?mod=rss", "rss", "forum", ("cn", "hosting"), 15),
    ("cn_baidu_sec", "https://xz.aliyun.com/", "html_listing", "security", ("cn", "security"), 15),
    ("cn_secfree", "https://www.secfree.com/", "html_listing", "security", ("cn", "security"), 10),
    ("cn_tttang", "https://tttang.com/", "html_listing", "security", ("cn", "security"), 10),
]


def probe(url, adapter="html_listing"):
    try:
        r = requests.get(url, headers={"User-Agent": UA}, timeout=18, allow_redirects=True)
    except Exception as e:
        return 0, type(e).__name__, 0
    if adapter == "rss":
        n_links = r.text.count("<item") + r.text.count("<entry")
    else:
        n_links = r.text.count("<a ")
    return r.status_code, "", n_links


def main():
    apply = "--apply" in sys.argv
    live, dead = [], []
    for name, url, adapter, cat, tags, lim in CAND:
        code, err, links = probe(url, adapter)
        ok = code == 200 and links >= 30
        print("%-18s %-3s links %-5d %s" % (name, code or err, links, "LIVE" if ok else "—"))
        if ok:
            live.append((name, url, adapter, cat, tags, lim))
        else:
            dead.append(name)
        time.sleep(0.8)

    print("\nживых %d, мёртвых %d" % (len(live), len(dead)))
    if not apply:
        print("dry-run: --apply чтобы записать omni/sources_cn.py")
        return 0

    body = ['"""Китайские и азиатские источники (проверены живыми запросами tools/wave_cn.py).',
            '', 'Форумы реверса/хостинга, security-новости, tech-ленты. Все — публичные списки,',
            'без ключей и кук. Подключаются в реестр omni как категории forum/security/news.',
            '"""', '', 'CN = [']
    for name, url, adapter, cat, tags, lim in live:
        t = ", ".join('"%s"' % x for x in tags)
        body.append('    {"name": "%s", "category": "%s", "kind": "site", "url": "%s",'
                    ' "adapter": "%s", "tags": (%s), "limit": %d},' % (name, cat, url, adapter, t, lim))
    body.append(']')
    OUT.write_text("\n".join(body) + "\n", encoding="utf-8", newline="\n")
    print("записано", OUT, "| источников", len(live))
    return 0


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