feat: BreakPilot PWA - Full codebase (clean push without large binaries)

All services: admin-v2, studio-v2, website, ai-compliance-sdk,
consent-service, klausur-service, voice-service, and infrastructure.
Large PDFs and compiled binaries excluded via .gitignore.

backend/debug_visualization.py

@@ -0,0 +1,64 @@
+"""
+Debug visualization to see what the AI detected
+"""
+
+import cv2
+import json
+from pathlib import Path
+
+def visualize_detection(image_path: Path, analysis_json_path: Path, output_path: Path):
+    """
+    Zeichne Bounding Boxes auf das Bild um zu sehen was die AI erkannt hat
+    """
+    # Load image
+    img = cv2.imread(str(image_path))
+    if img is None:
+        print(f"❌ Cannot load image: {image_path}")
+        return
+
+    # Load analysis
+    with open(analysis_json_path, 'r', encoding='utf-8') as f:
+        analysis = json.load(f)
+
+    # Draw handwriting regions in RED
+    hw_regions = analysis.get('handwriting_regions', [])
+    for i, region in enumerate(hw_regions, 1):
+        bbox = region.get('bounding_box', {})
+        x, y, w, h = bbox.get('x', 0), bbox.get('y', 0), bbox.get('width', 0), bbox.get('height', 0)
+
+        if w > 0 and h > 0:
+            # Draw rectangle
+            cv2.rectangle(img, (x, y), (x + w, y + h), (0, 0, 255), 3)
+
+            # Draw label
+            label = f"HW{i}: {region.get('type')} ({region.get('color_hint')})"
+            cv2.putText(img, label, (x, y - 10), cv2.FONT_HERSHEY_SIMPLEX, 0.6, (0, 0, 255), 2)
+
+    # Draw diagram regions in GREEN
+    diagrams = analysis.get('layout', {}).get('diagram_elements', [])
+    for i, diagram in enumerate(diagrams, 1):
+        bbox = diagram.get('bounding_box', {})
+        x, y, w, h = bbox.get('x', 0), bbox.get('y', 0), bbox.get('width', 0), bbox.get('height', 0)
+
+        if w > 0 and h > 0:
+            cv2.rectangle(img, (x, y), (x + w, y + h), (0, 255, 0), 2)
+            label = f"DIAGRAM{i}"
+            cv2.putText(img, label, (x, y - 10), cv2.FONT_HERSHEY_SIMPLEX, 0.5, (0, 255, 0), 2)
+
+    # Save
+    cv2.imwrite(str(output_path), img)
+    print(f"✅ Debug visualization saved: {output_path}")
+    print(f"   📍 {len(hw_regions)} handwriting regions (RED)")
+    print(f"   📍 {len(diagrams)} diagram regions (GREEN)")
+
+if __name__ == "__main__":
+    eingang = Path.home() / "Arbeitsblaetter" / "Eingang"
+    bereinigt = Path.home() / "Arbeitsblaetter" / "Bereinigt"
+
+    image_file = "2025-12-11_Auge1.JPG"
+
+    visualize_detection(
+        eingang / image_file,
+        bereinigt / "2025-12-11_Auge1_analyse.json",
+        bereinigt / "2025-12-11_Auge1_DEBUG.jpg"
+    )