merge: rag chat stabilization from wren-dev

fastapi_modules/fastapi_leaudit/services/impl/ragDatasetServiceImpl.py

@@ -36,7 +36,7 @@ from fastapi_modules.fastapi_leaudit.domian.vo.ragDatasetVo import (
 )
 from fastapi_modules.fastapi_leaudit.domian.vo.ragChatVo import RagOperationResultVO
 from fastapi_modules.fastapi_leaudit.rag_engine.chroma_client import get_chroma
-from fastapi_modules.fastapi_leaudit.rag_engine.config import RAG_CONFIG
+from fastapi_modules.fastapi_leaudit.rag_engine.config import RAG_CONFIG, build_openai_embeddings_url
 from fastapi_modules.fastapi_leaudit.services.ragDatasetService import IRagDatasetService
 
 
@@ -1503,7 +1503,7 @@ class RagDatasetServiceImpl(IRagDatasetService):
         return chunks
 
     async def _embed_texts(self, texts: list[str], model_name: str) -> list[list[float]]:
-        embed_url = (RAG_CONFIG.get("EMBED_URL") or "").strip() or f"{RAG_CONFIG['LLM_BASE_URL'].rstrip('/')}/embeddings"
+        embed_url = (RAG_CONFIG.get("EMBED_URL") or "").strip() or build_openai_embeddings_url(RAG_CONFIG["LLM_BASE_URL"])
         embed_key = (RAG_CONFIG.get("EMBED_KEY") or "").strip() or RAG_CONFIG["LLM_API_KEY"]
         embed_model = model_name or (RAG_CONFIG.get("EMBED_MODEL") or "").strip() or "text-embedding-v4"
         batch_size = max(1, int(RAG_CONFIG.get("EMBED_BATCH_SIZE") or 10))

fastapi_modules/fastapi_leaudit/services/ragChatService.py

@@ -6,6 +6,7 @@ from typing import AsyncGenerator
 from fastapi_modules.fastapi_leaudit.domian.Dto.ragChatDto import (
     RagConversationRenameDTO,
     RagMessageFeedbackDTO,
+    RagStopMessageDTO,
 )
 from fastapi_modules.fastapi_leaudit.domian.vo.ragChatVo import (
     RagAppParametersVO,
@@ -52,6 +53,9 @@ class IRagChatService(ABC):
     @abstractmethod
     async def UpdateFeedback(self, CurrentUserId: int, MessageId: str, Body: RagMessageFeedbackDTO) -> RagOperationResultVO: ...
 
+    @abstractmethod
+    async def StopMessage(self, CurrentUserId: int, MessageId: str, Body: RagStopMessageDTO | None = None) -> RagOperationResultVO: ...
+
     @abstractmethod
     async def GetAppParameters(
         self,