knowledge-base/scripts/kb-audit-creds.py

#!/usr/bin/env python3
"""
kb-audit-creds — берёт все URL из credentials.md, пингует каждый,
фиксирует unreachable или нестандартные коды ответа.
Пишет audit/YYYY-MM-DD-creds-drift.md.

Проверка только reachability (HTTP status). Не тестирует реальный логин.
"""

import re
import ssl
import sys
import urllib.error
import urllib.request
from datetime import date
from pathlib import Path

VAULT = Path(__file__).resolve().parent.parent
CREDS = VAULT / "projects/dttb/credentials.md"
OUT_DIR = VAULT / "audit"

_CTX = ssl.create_default_context()
_CTX.check_hostname = False
_CTX.verify_mode = ssl.CERT_NONE


def extract_urls(text: str):
    """URL в обычной записи + в markdown-таблице.
    Пропускаем URL с embedded credentials (http://user:pass@host) — они для git/curl,
    не для reachability-check.
    """
    pattern = re.compile(r"https?://[a-zA-Z0-9._:@-]+(?:/[^\s`|<>\"')]*)?")
    seen = {}
    for m in pattern.finditer(text):
        url = m.group(0).rstrip("/.,;:)")
        # пропускаем URLs с embedded creds
        if re.match(r"https?://[^/]*@", url):
            continue
        if url not in seen:
            start = max(0, m.start() - 30)
            end = min(len(text), m.end() + 30)
            ctx = text[start:end].replace("\n", " ").replace("|", " ").strip()
            seen[url] = ctx[:80]
    return seen


def _request(url: str, method: str, timeout: int):
    req = urllib.request.Request(url, method=method, headers={"User-Agent": "kb-audit/1.0"})
    r = urllib.request.urlopen(req, context=_CTX, timeout=timeout)
    return str(r.status), r.reason or ""


def ping(url: str, timeout: int = 6) -> tuple[str, str]:
    """Возвращает (status, detail). Пробуем HEAD, при 501/405 fallback на GET."""
    try:
        return _request(url, "HEAD", timeout)
    except urllib.error.HTTPError as e:
        if e.code in (501, 405):
            try:
                return _request(url, "GET", timeout)
            except urllib.error.HTTPError as e2:
                return str(e2.code), e2.reason or ""
            except Exception as e2:
                return "ERR", str(e2)[:60]
        return str(e.code), e.reason or ""
    except urllib.error.URLError as e:
        reason = str(e.reason)
        if "ssl" in reason.lower() or "certificate" in reason.lower():
            return "SSL", reason[:60]
        return "FAIL", reason[:60]
    except TimeoutError:
        return "TIMEOUT", ""
    except Exception as e:
        return "ERR", str(e)[:60]


def classify(status: str) -> str:
    """Status → категория для отчёта."""
    if status in ("200", "301", "302", "303", "307", "308"):
        return "✓ reachable"
    if status in ("401", "403"):
        return "✓ auth-required (сервер жив)"
    if status in ("404", "405"):  # 405 на HEAD, 404 ок при ping host root
        return "⚠ 4xx (сервер жив, но путь/метод)"
    if status.startswith("5"):
        return "❌ 5xx server error"
    if status in ("FAIL", "ERR", "TIMEOUT", "SSL"):
        return "❌ недоступен"
    return f"? {status}"


def main():
    today = date.today().isoformat()
    OUT_DIR.mkdir(parents=True, exist_ok=True)
    out = OUT_DIR / f"{today}-creds-drift.md"

    text = CREDS.read_text()
    urls = extract_urls(text)

    if not urls:
        print("no URLs found in credentials.md", file=sys.stderr)
        sys.exit(1)

    results = []
    for url, ctx in urls.items():
        status, detail = ping(url)
        category = classify(status)
        results.append((url, ctx, status, detail, category))

    unreachable = [r for r in results if "❌" in r[4]]
    questionable = [r for r in results if "⚠" in r[4] or "?" in r[4]]
    ok = [r for r in results if "✓" in r[4]]

    lines = [
        "---",
        f"date: {today}",
        "type: audit",
        "source: kb-audit-creds.py",
        "tags: [audit, creds, reachability]",
        "---",
        "",
        f"# Credentials reachability — {today}",
        "",
        f"Ping-проверка URL из [[../projects/dttb/credentials|credentials.md]].",
        f"Проверяется только reachability (HTTP status), не реальный логин.",
        "",
        f"- Всего URL: **{len(results)}**",
        f"- ✓ Reachable: {len(ok)} / ⚠ Questionable: {len(questionable)} / ❌ Unreachable: {len(unreachable)}",
        "",
    ]

    if unreachable:
        lines += ["## ❌ Недоступные (проверить: сервер упал? URL поменялся?)", "",
                  "| URL | Status | Detail | Контекст |", "|---|---|---|---|"]
        for url, ctx, st, det, _ in unreachable:
            lines.append(f"| `{url}` | {st} | {det[:40]} | {ctx[:50]} |")
        lines.append("")

    if questionable:
        lines += ["## ⚠ Нестандартный ответ", "",
                  "| URL | Status | Detail |", "|---|---|---|"]
        for url, ctx, st, det, _ in questionable:
            lines.append(f"| `{url}` | {st} | {det[:40]} |")
        lines.append("")

    lines += ["## ✓ Все ответили нормально", "",
              "| URL | Status | Категория |", "|---|---|---|"]
    for url, ctx, st, det, cat in ok:
        lines.append(f"| `{url}` | {st} | {cat} |")
    lines += ["", "---", "*Автоматически через `scripts/kb-audit-creds.py`.*"]

    out.write_text("\n".join(lines))
    print(f"creds drift: {out}")
    print(f"  ok: {len(ok)} / questionable: {len(questionable)} / unreachable: {len(unreachable)}")


if __name__ == "__main__":
    main()