fix: 5 regex bugs + text extraction scroll + GT update

Root cause: Spiegel DSI text was truncated (lazy-loading) — the
rights/DSB/complaints sections at the bottom were never extracted.

Fixes:
1. Text extraction: scroll to bottom before innerText (dsi_discovery.py)
2. V.i.S.d.P.: add "verantwortlicher i.s.v." + "§18 Abs. N MStV" pattern
3. USt-IdNr: add "umsatzsteuer-id" + "DE 212 442 423" (with spaces)
4. Profiler: remove generic "anwalt"/"praxis" (false positive on Spiegel
   "Redaktionsanwalt"), keep only "rechtsanwalt", "kanzlei" etc.
5. Section splitter: auto_fill_from_dsi() fills empty Cookie/Social-Media
   rows from sections found in the DSI text

Ground Truth 06-spiegel.md fully rewritten with verified data from
live website — 3 L1 False Negatives identified (DSB, Beschwerderecht,
Betroffenenrechte all present on website but not in extracted text).

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

backend-compliance/compliance/api/agent_compliance_check_routes.py

@@ -174,9 +174,14 @@ async def _run_compliance_check(check_id: str, req: ComplianceCheckRequest):
                 "word_count": len(text.split()) if text else 0,
             })
 
-        # Step 1b: If same URL used for multiple doc_types, try section splitting
-        from compliance.services.section_splitter import split_shared_texts
+        # Step 1b: Section splitting — two cases:
+        # 1. Same URL used for multiple doc_types → split by heading
+        # 2. DSI text contains Cookie/Social-Media sections → auto-fill empty rows
+        from compliance.services.section_splitter import (
+            split_shared_texts, auto_fill_from_dsi,
+        )
         split_shared_texts(doc_entries, url_text_cache)
+        auto_fill_from_dsi(doc_entries)
         # Refresh doc_texts after splitting
         for entry in doc_entries:
             if entry.get("text"):

backend-compliance/compliance/services/business_profiler.py

@@ -59,26 +59,34 @@ _NONPROFIT_KEYWORDS = [
 ]
 
 _REGULATED_PROFESSIONS = {
+    # Anwalt — nur spezifische Begriffe, nicht "anwalt" allein
+    # (matcht sonst Redaktionsanwalt, Justiziar etc.)
     "rechtsanwalt": "anwalt",
-    "anwalt": "anwalt",
-    "anwaeltin": "anwalt",
-    "anwältin": "anwalt",
+    "rechtsanwaeltin": "anwalt",
+    "rechtsanwältin": "anwalt",
     "kanzlei": "anwalt",
     "rechtsanwaltskammer": "anwalt",
-    "arzt": "arzt",
-    "ärztin": "arzt",
-    "aerztin": "arzt",
-    "praxis": "arzt",
+    "zugelassener anwalt": "anwalt",
+    # Arzt — "praxis" entfernt (matcht "in der Praxis")
+    "arztpraxis": "arzt",
+    "zahnarzt": "arzt",
+    "facharzt": "arzt",
     "aerztekammer": "arzt",
     "ärztekammer": "arzt",
+    "kassenärztlich": "arzt",
+    "kassenaerztlich": "arzt",
+    # Steuerberater
     "steuerberater": "steuerberater",
     "steuerberaterin": "steuerberater",
     "steuerberaterkammer": "steuerberater",
+    # Architekt
     "architekt": "architekt",
     "architektin": "architekt",
     "architektenkammer": "architekt",
+    # Notar
     "notar": "notar",
     "notariat": "notar",
+    # Apotheker
     "apotheke": "apotheker",
     "apotheker": "apotheker",
 }

backend-compliance/compliance/services/doc_checks/impressum_checks.py

@@ -135,8 +135,9 @@ IMPRESSUM_CHECKLIST = [
         "label": "USt-IdNr.",
         "level": 1, "parent": None,
         "patterns": [
-            r"ust.*id", r"umsatzsteuer.*identifikation",
-            r"vat.*id", r"de\s*\d{9}",
+            r"ust[\s.-]*id", r"umsatzsteuer[\s-]*id",
+            r"umsatzsteuer.*identifikation",
+            r"vat[\s.-]*id", r"de\s*\d{3}\s*\d{3}\s*\d{3}",
         ],
         "severity": "MEDIUM",
         "hint": "§5(1) Nr.6 TMG: Die USt-IdNr. muss angegeben werden, sofern vorhanden. Die Steuernummer ist KEIN Ersatz.",
@@ -146,7 +147,7 @@ IMPRESSUM_CHECKLIST = [
         "label": "USt-IdNr. im Format DE + 9 Ziffern",
         "level": 2, "parent": "vat",
         "patterns": [
-            r"de\s*\d{9}",
+            r"de\s*\d{3}\s*\d{3}\s*\d{3}",
         ],
         "severity": "LOW",
         "hint": "Deutsche USt-IdNr.: 'DE' + exakt 9 Ziffern (z.B. DE123456789). Validierung: https://evatr.bff-online.de/",
@@ -187,7 +188,8 @@ IMPRESSUM_CHECKLIST = [
         "patterns": [
             r"v\.?\s*i\.?\s*s\.?\s*d\.?\s*p",
             r"(?:redaktionell|inhaltlich)\s+verantwortlich",
-            r"§\s*18\s+m(?:edien)?st(?:aat)?v",
+            r"§\s*18\s+(?:abs\.?\s*\d+\s+)?m(?:edien)?st(?:aat)?v",
+            r"verantwortlich\w*\s+i\.?\s*s\.?\s*(?:d\.?\s*)?v\.?",
         ],
         "severity": "INFO",
         "hint": "Nur relevant wenn die Website journalistisch-redaktionelle Inhalte hat (Blog, Ratgeber, News, Fachartikel). Reine Unternehmensseiten ohne redaktionelle Inhalte benoetigen keinen V.i.S.d.P. Pruefen Sie, ob die Website einen Blog oder Ratgeber-Bereich hat.",

backend-compliance/compliance/services/section_splitter.py

@@ -168,3 +168,48 @@ def _find_section_for_type(sections: list[dict], doc_type: str) -> str | None:
             return section["text"]
 
     return None  # No match → keep full text
+
+
+def auto_fill_from_dsi(doc_entries: list[dict]) -> None:
+    """Auto-fill empty document rows from sections found in the DSI text.
+
+    If the user only entered the DSI URL but left Cookie/Social-Media empty,
+    and the DSI text contains those sections, auto-fill them.
+    """
+    # Find the DSI entry
+    dsi_entry = None
+    for entry in doc_entries:
+        if entry["doc_type"] in ("dse", "datenschutz", "privacy") and entry.get("text"):
+            dsi_entry = entry
+            break
+    if not dsi_entry:
+        return
+
+    dsi_text = dsi_entry["text"]
+    if len(dsi_text) < 300:
+        return
+
+    # Split DSI into sections
+    sections = _split_at_headings(dsi_text)
+    if not sections:
+        return
+
+    # Find empty entries that could be filled from DSI sections
+    filled = []
+    for entry in doc_entries:
+        if entry.get("text") or entry.get("url"):
+            continue  # Already has content
+
+        doc_type = entry["doc_type"]
+        section_text = _find_section_for_type(sections, doc_type)
+        if section_text and len(section_text.split()) >= 30:
+            entry["text"] = section_text
+            entry["word_count"] = len(section_text.split())
+            entry["url"] = f"{dsi_entry.get('url', '')} (Abschnitt)"
+            filled.append(doc_type)
+
+    if filled:
+        logger.info(
+            "Auto-filled %d empty rows from DSI sections: %s",
+            len(filled), ", ".join(filled),
+        )