feat(b17): Stufe 4 banner-tour + Stufe 5 annotierte Screenshots + V2-default

Stufe 4 — Cookie-Banner-Tour vor dem Accept-Klick:
  - audit_walk_banner_tour.tour_cookie_banner(): öffnet Settings
    (16 Phrase-Varianten), scrollt vertikal, aktiviert jedes
    [role=tab], expandet jedes [aria-expanded=false] / details /
    summary + 14 CMP-spezifische Selektoren. Max 35 Klicks,
    Best-Effort.
  - audit_walk_recorder ruft tour_cookie_banner() VOR
    _try_accept_banner auf — Reviewer sieht den vollen Consent-
    Katalog im Video (Vendor-Liste, Kategorien, Zwecke).
  - Recorder unter 500 LOC (412+155 split).

Stufe 5 — Annotierte Screenshots pro Finding:
  - finding_annotator.annotate_url(): WebKit headless, JS-Inject
    eines rot-banner-Labels oben + roter Outline um das Element
    (Selector oder Text-Match).
  - finding_annotator.annotate_findings(): dispatched 3 Cases —
    B1 Tap-Target (Anchor markiert mit "Tap-Target X×Y px"),
    B16 URL-Slug-Drift (404-Seite mit "/<slug> 404"),
    B13 Widerruf (Footer markiert "Widerruf-Link fehlt").
  - routes_audit_walk.POST /annotate-findings (consent-tester).
  - _b17_wiring ruft annotate-findings nach record_audit_walk und
    speichert annotations in walk.annotations.
  - audit_walk_zip_builder packt PNGs nach findings/<name>.png ins
    ZIP — Reviewer hat Beweis-Bilder im Postfach.

Plausibility Circuit-Breaker:
  - Nach 6 consecutive empty batches (PLAUSIBILITY_EMPTY_BUDGET=6)
    bricht die ganze Phase ab statt 200 Calls zu warten. Fix für
    qwen3-down + große DSE-Sites (BMW: ohne Breaker 21min, mit
    Breaker ~3min).

audit_walk_zip_builder fängt walk.annotations ab und legt sie unter
  findings/<fname>.png im ZIP-Anhang ab.

V2-Default:
  - docker-compose.yml backend-compliance.environment.MAIL_RENDER_V2:
    default 'true'. Ohne diesen Override liefert die Engine
    weiterhin das alte Legacy-Mail-Layout, in dem die B-Wiring-
    Blöcke nicht sichtbar sind.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
This commit is contained in:
Benjamin Admin
2026-06-07 20:44:42 +02:00
parent e8ff75cbfe
commit b16130369a
8 changed files with 540 additions and 0 deletions
@@ -0,0 +1,183 @@
"""Finding-Annotator — pro Befund ein PNG mit Marker.
Erzeugt für die Audit-Mail einen visuellen Beweis pro Finding:
roter Rand um das verletzende Element + Label-Banner darüber.
Aktuell unterstützte Annotation-Typen:
- "tap_target_too_small" B1 Mobile-Reachability — markiert den
Footer-Anchor mit Größenanforderung
- "slug_404" B16 URL-Slug-Drift — Screenshot des
404-Slugs mit "Standard-Slug 404"
- "missing_link_in_footer" B13 Widerruf — Footer-Screenshot mit
"→ Widerruf-Link erwartet, fehlt"
Output: bytes (PNG). Best-Effort: Fehler beim Rendern werfen nicht,
sondern returnen leere bytes (caller dropt das Asset).
"""
from __future__ import annotations
import base64
import logging
from typing import Any
logger = logging.getLogger(__name__)
# Banner-Style für Label oben im Bild (rot mit weißer Schrift).
_LABEL_JS = """
(label) => {
const div = document.createElement('div');
div.style.cssText = `
position:fixed;top:0;left:0;right:0;z-index:2147483647;
background:#dc2626;color:#fff;padding:14px 24px;
font:bold 18px/1.2 -apple-system,sans-serif;
box-shadow:0 4px 12px rgba(0,0,0,0.4);
border-bottom:3px solid #7f1d1d;text-align:center;
`;
div.textContent = label;
document.body.appendChild(div);
}
"""
# JS zum Markieren eines Elements: roter Rand + 0.4-Schatten innen.
_MARK_JS = """
(args) => {
const sel = args.selector;
const text = args.text;
let el = null;
if (sel) {
try { el = document.querySelector(sel); } catch(e) {}
}
if (!el && text) {
const all = document.querySelectorAll('a, button, [role=button]');
for (const c of all) {
if ((c.innerText || '').trim().toLowerCase()
.includes(text.toLowerCase())) {
el = c; break;
}
}
}
if (!el) return false;
el.scrollIntoView({behavior:'instant', block:'center'});
el.style.outline = '4px solid #dc2626';
el.style.outlineOffset = '4px';
el.style.boxShadow = 'inset 0 0 0 9999px rgba(220,38,38,0.15)';
return true;
}
"""
async def annotate_url(
url: str,
label: str,
selector: str | None = None,
text_match: str | None = None,
viewport_w: int = 1280,
viewport_h: int = 800,
timeout_s: float = 20.0,
) -> bytes:
"""Open `url`, mark element by selector or text-match, save PNG."""
try:
from playwright.async_api import async_playwright
except Exception as e:
logger.warning("playwright unavailable: %s", e)
return b""
async with async_playwright() as p:
try:
browser = await p.webkit.launch(headless=True)
context = await browser.new_context(
viewport={"width": viewport_w, "height": viewport_h},
locale="de-DE",
)
page = await context.new_page()
try:
await page.goto(url, wait_until="domcontentloaded",
timeout=timeout_s * 1000)
except Exception as e:
logger.info("annotate_url goto failed for %s: %s",
url, e)
await page.wait_for_timeout(1500)
await page.evaluate(_LABEL_JS, label)
if selector or text_match:
try:
await page.evaluate(
_MARK_JS,
{"selector": selector or "",
"text": text_match or ""},
)
except Exception as e:
logger.info("annotate mark failed: %s", e)
await page.wait_for_timeout(500)
png = await page.screenshot(full_page=False, type="png")
await context.close()
await browser.close()
return png
except Exception as e:
logger.warning("annotate_url failed: %s", e)
return b""
async def annotate_findings(
findings: list[dict],
home_url: str,
) -> list[dict[str, Any]]:
"""Per finding produce a PNG. Returns list of {filename, png_b64}.
Supported finding shapes:
- {check_id: 'COOKIE-CONSENT-UX-001', mobile_playwright:
{anchor_text: '...'}}
- {check_id: 'URL-SLUG-DRIFT-001', alt_slugs_404: [...],
doc_type: '...', source_url: '...'} (Stufe-1: nimmt source_url
aus state, hier optional)
- {check_id: 'WIDERRUF-REACH-001'} → Footer-Screenshot der
Homepage mit "→ Widerruf-Link fehlt"
Skip silently if a finding doesn't match a supported shape.
"""
out: list[dict[str, Any]] = []
for f in findings:
cid = (f.get("check_id") or "").upper()
png = b""
fname = ""
if cid == "COOKIE-CONSENT-UX-001":
mob = f.get("mobile_playwright") or {}
anchor_text = mob.get("anchor_text") or ""
if anchor_text:
tp = mob.get("tap_target_px") or {}
label = (
f"B1 — Tap-Target {tp.get('w','?')}×{tp.get('h','?')} "
"px (Anforderung ≥ 44×44)"
)
png = await annotate_url(
home_url, label, text_match=anchor_text,
)
fname = f"finding-tap-target.png"
elif cid == "URL-SLUG-DRIFT-001":
alts = f.get("alt_slugs_404") or []
if alts:
# Pick first 404-slug + screenshot the 404 page
doc_type = f.get("doc_type") or "x"
from urllib.parse import urlparse
p = urlparse(home_url)
origin = f"{p.scheme}://{p.netloc}"
slug_url = f"{origin}/{alts[0]}"
label = (
f"B16 — Standard-Slug 404 ({doc_type}): /{alts[0]}"
)
png = await annotate_url(slug_url, label)
fname = f"finding-slug-404-{doc_type}.png"
elif cid == "WIDERRUF-REACH-001":
label = "B13 — Widerrufsbelehrung im Footer erwartet, fehlt"
png = await annotate_url(
home_url, label, selector="footer",
)
fname = "finding-widerruf-missing.png"
if png and fname:
out.append({
"filename": fname,
"png_b64": base64.b64encode(png).decode("ascii"),
})
return out