breakpilot-compliance/backend-compliance/compliance/services/doc_staleness_check.py

"""B7 — Doc-Staleness: Datum extrahieren + Aktualität bewerten.

Findings, wenn ein rechtliches Dokument (AGB, Nutzungsbedingungen,
Widerruf, DSE) über N Jahre alt ist. Default-Cap: 3 Jahre für AGB/
Nutzungsbedingungen (TERMS-STALENESS-001), 2 Jahre für DSE.

Heuristik für Datumsextraktion:
  - "Stand: November 2018" / "Stand November 2018" / "Stand: Dezember 2018"
  - "Letzte Aktualisierung: 2018-12-01"
  - "Version vom 1.12.2018"
  - "Last updated: December 2018"
"""

from __future__ import annotations

import logging
import re
from datetime import datetime

logger = logging.getLogger(__name__)

_MONTHS_DE = {
    "januar": 1, "februar": 2, "märz": 3, "maerz": 3, "april": 4,
    "mai": 5, "juni": 6, "juli": 7, "august": 8, "september": 9,
    "oktober": 10, "november": 11, "dezember": 12,
}
_MONTHS_EN = {
    "january": 1, "february": 2, "march": 3, "april": 4, "may": 5,
    "june": 6, "july": 7, "august": 8, "september": 9, "october": 10,
    "november": 11, "december": 12,
}

# Match patterns like "Stand: Dezember 2018" / "Stand November 2018"
_PAT_STAND = re.compile(
    r"(?:stand|version|letzte\s+aktualisierung|last\s+updated|"
    r"last\s+revised)\s*[:\-]?\s*"
    r"(?:vom\s+)?"
    r"(?:(?P<day>\d{1,2})[.\-/])?"
    r"(?P<month>"
    r"januar|februar|m[äa]rz|april|mai|juni|juli|august|september|"
    r"oktober|november|dezember|"
    r"january|february|march|april|may|june|july|august|september|"
    r"october|november|december|"
    r"\d{1,2}"
    r")"
    r"[.\s\-/]+"
    r"(?P<year>\d{4})",
    re.I,
)


_AGE_THRESHOLDS_YEARS = {
    "agb": 3,
    "nutzungsbedingungen": 3,
    "widerruf": 2,
    "dse": 2,
    "impressum": 5,  # less critical
    "cookie": 2,
}


def _extract_date(text: str) -> tuple[int, int, int] | None:
    """Return (year, month, day) of the most recent revision date."""
    if not text:
        return None
    candidates: list[tuple[int, int, int]] = []
    for m in _PAT_STAND.finditer(text):
        try:
            year = int(m.group("year"))
            mon_str = (m.group("month") or "").lower()
            day = int(m.group("day") or 1)
            if mon_str.isdigit():
                month = int(mon_str)
            else:
                month = (_MONTHS_DE.get(mon_str)
                         or _MONTHS_EN.get(mon_str))
            if not month or not (1 <= month <= 12):
                continue
            if year < 2000 or year > 2100:
                continue
            candidates.append((year, month, day))
        except (ValueError, TypeError):
            continue
    if not candidates:
        return None
    # newest date wins
    candidates.sort(reverse=True)
    return candidates[0]


def check_staleness(state: dict) -> list[dict]:
    """Run staleness check across legal docs."""
    findings: list[dict] = []
    doc_texts = state.get("doc_texts") or {}
    today = datetime.utcnow()
    for doc_type, text in doc_texts.items():
        threshold_years = _AGE_THRESHOLDS_YEARS.get(doc_type)
        if not threshold_years:
            continue
        date = _extract_date(text)
        if not date:
            continue
        year, month, day = date
        try:
            doc_date = datetime(year, month, min(day, 28))
        except ValueError:
            continue
        age_years = (today - doc_date).days / 365.25
        if age_years < threshold_years:
            continue
        sev = "HIGH" if age_years > threshold_years * 2 else "MEDIUM"
        findings.append({
            "check_id": f"DOC-STALENESS-{doc_type.upper()}",
            "doc_type": doc_type,
            "severity": sev,
            "severity_reason": "incomplete",
            "title": (
                f"{doc_type.title()} ist {int(age_years)} Jahre alt "
                f"(Stand {year:04d}-{month:02d})"
            ),
            "norm": "Sorgfaltspflicht (laufende Anpassung an Rechtsänderungen)",
            "doc_date": f"{year:04d}-{month:02d}-{day:02d}",
            "age_years": round(age_years, 1),
            "threshold_years": threshold_years,
            "action": (
                f"{doc_type.title()} überprüfen und an aktuelle "
                "Gesetzeslage anpassen (DSGVO-Updates, AI Act, DSA, "
                "neue BGH-Rechtsprechung). Stand-Datum aktualisieren."
            ),
        })
    if findings:
        logger.info("B7 staleness: %d findings", len(findings))
    return findings