"""Scrape the CIS AWS Foundations Benchmark page.

The page is a mapping table:
  | Control ID and title (FSBP-style) | CIS v5.0.0 | CIS v3.0.0 | CIS v1.4.0 | CIS v1.2.0 |

We emit one CISRow per unified control. The `requirement` field summarises
which CIS versions reference it.
"""
from __future__ import annotations

import re
from dataclasses import dataclass, field

from bs4 import BeautifulSoup

from scripts.control_resource_map import control_id_to_types


CIS_URL = (
    "https://docs.aws.amazon.com/securityhub/latest/userguide/"
    "cis-aws-foundations-benchmark.html"
)


@dataclass
class CISRow:
    control_id: str
    title: str
    severity: str
    requirement: str
    source_url: str
    resource_types: list[str] = field(default_factory=list)


_TITLE_RE = re.compile(r"^\s*\[(?P<id>[A-Za-z][A-Za-z0-9]*\.\d+)\]\s*(?P<title>.+)$")
_VERSION_RE = re.compile(r"CIS\s+v(?P<ver>[\d.]+)", re.IGNORECASE)


def _extract_versions(headers: list[str]) -> list[str]:
    """From header cells, pull out version strings like '5.0.0' for each
    column. Non-version columns yield empty string."""
    versions: list[str] = []
    for h in headers:
        m = _VERSION_RE.search(h)
        versions.append(m.group("ver") if m else "")
    return versions


def parse_cis_page(html: str, source_url: str) -> list[CISRow]:
    soup = BeautifulSoup(html, "html.parser")
    rows: list[CISRow] = []
    for table in soup.find_all("table"):
        header_cells = table.find("tr").find_all(["th", "td"]) if table.find("tr") else []
        header_texts = [c.get_text(strip=True) for c in header_cells]
        versions = _extract_versions(header_texts)
        if not any(versions):
            continue
        for tr in table.find_all("tr")[1:]:
            cells = tr.find_all("td")
            if len(cells) < 2:
                continue
            title_text = cells[0].get_text(strip=True)
            m = _TITLE_RE.match(title_text)
            if not m:
                continue
            unified_id = m.group("id")
            title = m.group("title").strip()
            version_refs: list[str] = []
            for ver, cell in zip(versions[1:], cells[1:]):
                if not ver:
                    continue
                num = cell.get_text(strip=True)
                if num:
                    version_refs.append(f"v{ver} §{num}")
            requirement = "CIS " + ", ".join(version_refs) if version_refs else ""
            rows.append(CISRow(
                control_id=f"CIS {unified_id}",
                title=title,
                severity="medium",
                requirement=requirement,
                source_url=source_url,
                resource_types=control_id_to_types(f"CIS {unified_id}"),
            ))
    return rows