feat(impressum): Snapshot-Modul-Tab — ImpressumAgent auf gespeichertem Text

Snapshot-Detailseite wird zu Modul-Tabs (Cookies & Tracking | Impressum).
Backend GET /snapshots/{id}/impressum-check laeuft den v3 ImpressumAgent auf
dem gespeicherten Impressum-Text (kein Re-Crawl); Input-Erzeugung in
impressum_input_from_snapshot() ausgelagert (pure + getestet: Text/Scope/
company_name-Fallback/None-Pfad). Frontend laedt lazy beim Tab-Wechsel und
rendert mit dem bestehenden AgentResultTab (keine zweite Engine).

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

backend-compliance/compliance/api/agent_check/_agent_outputs.py

@@ -53,6 +53,31 @@ def _derive_scope(profile_dict: dict) -> list[str]:
     return sorted(scope)
 
 
+def impressum_input_from_snapshot(snap: dict) -> dict | None:
+    """Baut den ImpressumAgent-Input aus einem gespeicherten Snapshot (kein
+    Re-Crawl). Pure + testbar: zieht den Impressum-Text aus doc_entries, leitet
+    den Scope aus scan_context + Profil ab (identisch zur Live-Auswertung) und
+    nimmt site_label als company_name-Fallback. None, wenn kein Impressum-Text.
+    """
+    docs = snap.get("doc_entries") or []
+    text = next((e.get("text") or e.get("content") or ""
+                 for e in docs if e.get("doc_type") == "impressum"), "")
+    if len((text or "").strip()) < _MIN_TEXT:
+        return None
+    profile = snap.get("profile") or {}
+    scope = sorted(
+        set(scan_context_to_scope(snap.get("scan_context")))
+        | set(_derive_scope(profile))
+    )
+    return {
+        "doc_type": "impressum",
+        "text": text,
+        "business_scope": scope,
+        "company_name": (profile.get("company_name") or snap.get("site_label") or ""),
+        "origin_domain": snap.get("site_domain", ""),
+    }
+
+
 async def run_agent_outputs(state: dict) -> None:
     """Für jedes Topic mit registriertem v3-Agent + ausreichend Text:
     Agent laufen lassen, AgentOutput ablegen + als SSE topic-Event

backend-compliance/compliance/api/agent_compliance_check_routes.py

@@ -268,6 +268,33 @@ async def snapshot_cookie_check(snapshot_id: str):
         db.close()
 
 
+@router.get("/snapshots/{snapshot_id}/impressum-check")
+async def snapshot_impressum_check(snapshot_id: str):
+    """Impressum-Analyse aus dem Snapshot (kein Re-Crawl): laeuft den v3
+    ImpressumAgent auf dem gespeicherten Impressum-Text + Profil/Scope und
+    liefert den AgentOutput (Findings/Massnahmen/MC-Coverage) fuer den Tab."""
+    from fastapi import HTTPException
+    from database import SessionLocal
+    from compliance.services.check_snapshot import load_snapshot
+    from compliance.services.specialist_agents import REGISTRY, AgentInput
+    from compliance.api.agent_check._agent_outputs import (
+        impressum_input_from_snapshot,
+    )
+    db = SessionLocal()
+    try:
+        snap = load_snapshot(db, snapshot_id)
+        if not snap:
+            raise HTTPException(status_code=404, detail="snapshot not found")
+        agent_input = impressum_input_from_snapshot(snap)
+        if not agent_input:
+            return {"findings": [], "recommendations": [], "mc_coverage": [],
+                    "notes": "kein Impressum-Text im Snapshot", "confidence": 0.0}
+        out = await REGISTRY.get("impressum").evaluate(AgentInput(**agent_input))
+        return out.model_dump(mode="json")
+    finally:
+        db.close()
+
+
 @router.get("/admin/benchmark")
 async def benchmark(
     industry: str = "",

backend-compliance/compliance/tests/test_impressum_snapshot_input.py

@@ -0,0 +1,47 @@
+"""impressum_input_from_snapshot — Snapshot → ImpressumAgent-Input (pure).
+
+Deckt die Glue des /snapshots/{id}/impressum-check-Endpoints ohne DB/LLM ab:
+Text-Extraktion, Scope-Ableitung (Profil), company_name-Fallback, None-Pfad.
+"""
+
+from __future__ import annotations
+
+from compliance.api.agent_check._agent_outputs import (
+    impressum_input_from_snapshot,
+)
+
+
+def _snap(text: str = "x" * 200, **over) -> dict:
+    s = {
+        "doc_entries": [{"doc_type": "impressum", "text": text}],
+        "profile": {}, "scan_context": None,
+        "site_label": "BMW", "site_domain": "bmw.de",
+    }
+    s.update(over)
+    return s
+
+
+def test_builds_input_from_impressum_text():
+    inp = impressum_input_from_snapshot(_snap())
+    assert inp is not None
+    assert inp["doc_type"] == "impressum"
+    assert inp["text"].startswith("x")
+    assert inp["company_name"] == "BMW"          # site_label-Fallback
+    assert inp["origin_domain"] == "bmw.de"
+    assert isinstance(inp["business_scope"], list)
+
+
+def test_none_when_no_or_short_impressum_text():
+    assert impressum_input_from_snapshot(_snap(doc_entries=[])) is None
+    assert impressum_input_from_snapshot(
+        {"doc_entries": [{"doc_type": "impressum", "text": "zu kurz"}]}) is None
+
+
+def test_scope_includes_profile_derived():
+    inp = impressum_input_from_snapshot(_snap(profile={"has_online_shop": True}))
+    assert "ecommerce" in inp["business_scope"]
+
+
+def test_company_name_prefers_profile_over_site_label():
+    inp = impressum_input_from_snapshot(_snap(profile={"company_name": "ACME AG"}))
+    assert inp["company_name"] == "ACME AG"