Add Ground Truth regression test system for OCR pipeline

Extract _build_grid_core() from build_grid() endpoint for reuse.
New ocr_pipeline_regression.py with endpoints to mark sessions as
ground truth, list them, and run regression comparisons after code
changes. Frontend button in StepGroundTruth.tsx to mark/update GT.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

klausur-service/backend/ocr_pipeline_session_store.py

@@ -270,6 +270,26 @@ async def get_sub_sessions(parent_session_id: str) -> List[Dict[str, Any]]:
         return [_row_to_dict(row) for row in rows]
 
 
+async def list_ground_truth_sessions_db() -> List[Dict[str, Any]]:
+    """List sessions that have a build_grid_reference in ground_truth."""
+    pool = await get_pool()
+    async with pool.acquire() as conn:
+        rows = await conn.fetch("""
+            SELECT id, name, filename, status, current_step,
+                   document_category, doc_type,
+                   ground_truth,
+                   parent_session_id, box_index,
+                   created_at, updated_at
+            FROM ocr_pipeline_sessions
+            WHERE ground_truth IS NOT NULL
+              AND ground_truth::text LIKE '%build_grid_reference%'
+              AND parent_session_id IS NULL
+            ORDER BY created_at DESC
+        """)
+
+        return [_row_to_dict(row) for row in rows]
+
+
 async def delete_session_db(session_id: str) -> bool:
     """Delete a session."""
     pool = await get_pool()