refactor!: 完全异步化 RAG 系统，移除 LangChain ParentDocumentRetriever 依赖

- 重写 rag_core/vector_store.py：完全异步实现 aadd_documents、asimilarity_search
- 重写 app/rag/retriever.py：异步混合检索，移除同步兼容代码
- 修改 rag_indexer/index_builder.py：全链路异步调用
- 删除 rag_core/retriever_factory.py：不再使用 LangChain ParentDocumentRetriever
- 清理冗余导入和代码：移除 model_services 兼容、不需要的异常导入
- 更新 rag_indexer/README.md：反映新架构

核心改进：
- 完全异步化：索引构建和检索全链路 async/await
- 自定义实现：不再依赖 LangChain 的 ParentDocumentRetriever
- 双向量支持：子文档同时存储 dense + sparse 向量到 Qdrant
- 架构清晰：rag_core 公共组件、rag_indexer 索引、app/rag 检索

tools/test/check_qdrant.py

@@ -7,7 +7,7 @@ import asyncio
 import os
 import sys
 
-from backend.rag_core import QdrantVectorStore
+from backend.rag_core import QdrantHybridStore
 from backend.app.model_services import get_embedding_service
 
 
@@ -18,7 +18,7 @@ def check_qdrant_data():
     print("="*70)
     
     embeddings = get_embedding_service()
-    vs = QdrantVectorStore(collection_name="rag_documents", embeddings=embeddings)
+    vs = QdrantHybridStore(collection_name="rag_documents", embeddings=embeddings)
     client = vs.get_qdrant_client()
     
     # 先获取几个点看看 payload 结构

tools/test/quick_test.py

@@ -8,7 +8,7 @@ import os
 import sys
 
 from qdrant_client import models
-from backend.rag_core import QdrantVectorStore, get_sparse_embedder
+from backend.rag_core import QdrantHybridStore, get_sparse_embedder
 from backend.app.model_services import get_embedding_service
 
 
@@ -19,7 +19,7 @@ def test_dense_retrieval():
     print("="*70)
     
     embeddings = get_embedding_service()
-    vs = QdrantVectorStore(collection_name="rag_documents", embeddings=embeddings)
+    vs = QdrantHybridStore(collection_name="rag_documents", embeddings=embeddings)
     
     query = "黄双银"  # 用文档里真正有的名字查询
     print(f"\n查询: {query}")

tools/test/reset_qdrant.py

@@ -7,7 +7,7 @@ import asyncio
 import os
 import sys
 
-from backend.rag_core import QdrantVectorStore
+from backend.rag_core import QdrantHybridStore
 from backend.app.model_services import get_embedding_service
 
 
@@ -18,7 +18,7 @@ async def delete_and_recreate():
     print("="*70)
     
     embeddings = get_embedding_service()
-    vs = QdrantVectorStore(collection_name="rag_documents", embeddings=embeddings)
+    vs = QdrantHybridStore(collection_name="rag_documents", embeddings=embeddings)
     
     # 删除旧集合
     try:

tools/test/simple_test.py

@@ -8,7 +8,7 @@ import os
 import sys
 
 from qdrant_client import models
-from backend.rag_core import QdrantVectorStore, get_sparse_embedder
+from backend.rag_core import QdrantHybridStore, get_sparse_embedder
 from backend.app.model_services import get_embedding_service
 
 
@@ -19,7 +19,7 @@ def check_qdrant_content():
     print("="*70)
     
     embeddings = get_embedding_service()
-    vs = QdrantVectorStore(collection_name="rag_documents", embeddings=embeddings)
+    vs = QdrantHybridStore(collection_name="rag_documents", embeddings=embeddings)
     client = vs.get_qdrant_client()
     
     # 滚动获取前 5 个点
@@ -51,7 +51,7 @@ def test_dense_retrieval():
     print("="*70)
     
     embeddings = get_embedding_service()
-    vs = QdrantVectorStore(collection_name="rag_documents", embeddings=embeddings)
+    vs = QdrantHybridStore(collection_name="rag_documents", embeddings=embeddings)
     
     query = "蚂蚁"  # 用中文查询
     print(f"\n查询: {query}")
@@ -72,7 +72,7 @@ def test_sparse_retrieval():
     print("="*70)
     
     embeddings = get_embedding_service()
-    vs = QdrantVectorStore(collection_name="rag_documents", embeddings=embeddings)
+    vs = QdrantHybridStore(collection_name="rag_documents", embeddings=embeddings)
     client = vs.get_qdrant_client()
     sparse_embedder = get_sparse_embedder()
     
@@ -109,7 +109,7 @@ def test_hybrid_retrieval():
     print("="*70)
     
     embeddings = get_embedding_service()
-    vs = QdrantVectorStore(collection_name="rag_documents", embeddings=embeddings)
+    vs = QdrantHybridStore(collection_name="rag_documents", embeddings=embeddings)
     client = vs.get_qdrant_client()
     sparse_embedder = get_sparse_embedder()
     

tools/test/test_rag_indexer_result.py

@@ -10,7 +10,7 @@ import os
 from rag_indexer.index_builder import IndexBuilder
 from rag_indexer.splitters import SplitterType
 
-from backend.rag_core import QdrantVectorStore, get_sparse_embedder
+from backend.rag_core import QdrantHybridStore, get_sparse_embedder
 from backend.app.model_services import get_embedding_service
 from qdrant_client import models
 
@@ -61,7 +61,7 @@ def test_dense_retrieval():
     embeddings = get_embedding_service()
     
     # 创建向量存储
-    vs = QdrantVectorStore(collection_name="rag_documents", embeddings=embeddings)
+    vs = QdrantHybridStore(collection_name="rag_documents", embeddings=embeddings)
     
     # 测试查询
     query = "The Ant and the Grasshopper"
@@ -87,7 +87,7 @@ def test_sparse_retrieval_simple():
     
     # 获取嵌入服务和稀疏嵌入器
     embeddings = get_embedding_service()
-    vs = QdrantVectorStore(collection_name="rag_documents", embeddings=embeddings)
+    vs = QdrantHybridStore(collection_name="rag_documents", embeddings=embeddings)
     client = vs.get_qdrant_client()
     sparse_embedder = get_sparse_embedder()
     
@@ -133,7 +133,7 @@ def test_hybrid_retrieval_simple():
     
     # 获取嵌入服务和稀疏嵌入器
     embeddings = get_embedding_service()
-    vs = QdrantVectorStore(collection_name="rag_documents", embeddings=embeddings)
+    vs = QdrantHybridStore(collection_name="rag_documents", embeddings=embeddings)
     client = vs.get_qdrant_client()
     sparse_embedder = get_sparse_embedder()
     
@@ -189,7 +189,7 @@ def test_parent_child_retrieval_simple():
     
     # 获取嵌入服务和稀疏嵌入器
     embeddings = get_embedding_service()
-    vs = QdrantVectorStore(collection_name="rag_documents", embeddings=embeddings)
+    vs = QdrantHybridStore(collection_name="rag_documents", embeddings=embeddings)
     client = vs.get_qdrant_client()
     sparse_embedder = get_sparse_embedder()