feat: RAG 임베딩 자동 동기화 + AI 서비스 개선

- 부적합 라이프사이클 전 과정에서 Qdrant 임베딩 자동 동기화
  - 관리함 5개 저장 함수 + 수신함 상태 변경 시 fire-and-forget sync
  - 30분 주기 전체 재동기화 안전망 (FastAPI lifespan 백그라운드 태스크)
  - build_document_text에 카테고리(final_category/category) 포함
- RAG 질의에 DB 통계 집계 지원 (카테고리별/부서별 건수)
- Qdrant client.search → query_points API 마이그레이션
- AI 어시스턴트 페이지 권한 추가 (tkuser)

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

ai-service/db/vector_store.py

@@ -1,8 +1,11 @@
+import logging
 import uuid
 from qdrant_client import QdrantClient
 from qdrant_client.models import Distance, VectorParams, PointStruct, Filter, FieldCondition, MatchValue
 from config import settings
 
+logger = logging.getLogger(__name__)
+
 
 class VectorStore:
     def __init__(self):
@@ -51,17 +54,19 @@ class VectorStore:
     ) -> list[dict]:
         query_filter = self._build_filter(where) if where else None
         try:
-            results = self.client.search(
+            response = self.client.query_points(
                 collection_name=self.collection,
-                query_vector=embedding,
+                query=embedding,
                 limit=n_results,
                 query_filter=query_filter,
+                with_payload=True,
             )
-        except Exception:
+        except Exception as e:
+            logger.error(f"Qdrant search failed: {e}", exc_info=True)
             return []
 
         items = []
-        for hit in results:
+        for hit in response.points:
             payload = hit.payload or {}
             item = {
                 "id": payload.get("original_id", str(hit.id)),