#!/usr/bin/env python
"""Probe a wiki host and report how (or whether) it can be mirrored.

Phase 0 of the build, as a command:

    python scripts/detect_wiki.py https://some.wiki.example.com

Reports the source kind, the endpoint, the article URL pattern, and — for
MediaWiki — the engine version and page count. Exits non-zero if the host cannot
be mirrored at all, so it can gate a scaffold.
"""

from __future__ import annotations

import sys

import httpx

from wikikb import config
from wikikb.sources import MEDIAWIKI, SitemapHtmlSource, detect_source_kind


def main(argv: list[str] | None = None) -> int:
    args = sys.argv[1:] if argv is None else argv
    if not args:
        print("usage: detect_wiki.py <https://host>", file=sys.stderr)
        return 2
    host = args[0].rstrip("/")

    with httpx.Client(
        headers={"User-Agent": config.USER_AGENT}, timeout=30, follow_redirects=True
    ) as client:
        try:
            kind, endpoint = detect_source_kind(client.get, host)
        except RuntimeError as exc:
            print(f"[FAIL] {exc}")
            return 1

        print(f"[ OK ] source kind : {kind}")
        print(f"       endpoint    : {endpoint}")

        if kind == MEDIAWIKI:
            info = client.get(
                f"{endpoint}?action=query&meta=siteinfo&siprop=general|statistics&format=json"
            ).json()["query"]
            general, stats = info["general"], info["statistics"]
            base = general.get("base", "")
            pattern = "{host}/wiki/{title}" if "/wiki/" in base else "{host}/{title}"
            print(f"       engine      : {general.get('generator')}")
            print(f"       content     : {stats.get('articles')} articles")
            print(f"       article URL : {pattern}")
            print(f"       licence     : {info.get('rightsinfo', {}).get('text', 'unknown')}")
            print("\nScaffold with:")
            print(f"  WIKI_KB_HOST={host} WIKI_KB_API_URL={endpoint}")
            if pattern.endswith("/wiki/{title}"):
                print(f"  WIKI_KB_PAGE_URL='{host}/wiki/{{title}}'")
        else:
            urls = list(SitemapHtmlSource(client.get, sitemap_url=endpoint).iter_titles())
            print(f"       content     : {len(urls)} sitemap URLs")
            print(f"       sample      : {urls[:5]}")
            print("\n[WARN] No API. Revision ids and maxlag are unavailable, so fidelity")
            print("       checks fall back to content diffing and rate limiting is purely")
            print("       self-imposed. Check the site's robots.txt and terms before use.")
            print("\nScaffold with:")
            print(f"  WIKI_KB_HOST={host} WIKI_KB_SOURCE=html WIKI_KB_SITEMAP={endpoint}")

        robots = client.get(f"{host}/robots.txt")
        if robots.status_code == 200 and "gptbot" in robots.text.lower():
            print("\n[NOTE] robots.txt names an AI crawler (e.g. GPTBot). It does not name")
            print("       this tool, but it is the operators' clearest statement about AI")
            print("       ingestion of their content. Your call — read it before scraping.")
    return 0


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