文件变更

rag_indexer/index_builder.py

@@ -16,6 +16,7 @@ from langchain_core.embeddings import Embeddings
 from langchain_core.stores import BaseStore
 from langchain_text_splitters import RecursiveCharacterTextSplitter, TextSplitter
 from langchain_classic.retrievers import ParentDocumentRetriever
+from qdrant_client.http.exceptions import ResponseHandlingException
 
 from .loaders import DocumentLoader
 from .splitters import SplitterType, get_splitter, SemanticChunkerAdapter
@@ -223,18 +224,26 @@ class IndexBuilder:
 
     async def _add_batch_with_retry(self, batch: List[Document], batch_no: int) -> None:
         """添加批次，失败时自动重试（处理网络波动）。"""
-        max_retries = 3
+        max_retries = 5
+        base_delay = 2
         for attempt in range(max_retries):
             try:
                 await self.retriever.aadd_documents(batch)   # type: ignore[union-attr]
+                logger.info("批次 %d 成功添加 %d 个文档", batch_no, len(batch))
                 return
-            except (RemoteProtocolError, ConnectionError, OSError) as e:
+            except (RemoteProtocolError, ConnectionError, OSError, ResponseHandlingException) as e:
                 if attempt == max_retries - 1:
+                    logger.error("批次 %d 重试 %d 次后仍然失败: %s", batch_no, max_retries, e)
                     raise
-                logger.warning("批次 %d 连接断开，重试 (%d/%d): %s",
-                               batch_no, attempt + 1, max_retries, e)
+                wait_time = base_delay * (2 ** attempt)
+                error_type = type(e).__name__
+                logger.warning(
+                    "批次 %d 遇到网络异常 [%s]，%d秒后重试 (%d/%d): %s",
+                    batch_no, error_type, wait_time, attempt + 1, max_retries, e
+                )
                 self.vector_store.refresh_client()
-                await asyncio.sleep(1)
+                logger.debug("批次 %d 已刷新 Qdrant 客户端连接", batch_no)
+                await asyncio.sleep(wait_time)
 
     # ---------- 信息获取方法 ----------
     def get_collection_info(self) -> Any: