Merge pull request #105 from Arcueld/perf/rag-optimize-embeddings

perf(rag): 优化嵌入服务性能并改进批处理大小配置

backend/app/services/rag/embeddings.py

@@ -767,7 +767,8 @@ class EmbeddingService:
                         logger.warning(f"Progress callback error: {e}")
 
                 # 添加小延迟避免限流
-                await asyncio.sleep(0.1)
+                if self.provider not in ["ollama"]:    
+                    await asyncio.sleep(0.1)  # 本地不延时
 
         # 确保没有 None
         return [e if e is not None else [0.0] * self.dimension for e in embeddings]

backend/app/services/rag/indexer.py

@@ -1290,7 +1290,7 @@ class CodeIndexer:
         # 批量嵌入（带进度回调和取消检查）
         embeddings = await self.embedding_service.embed_batch(
             texts,
-            batch_size=50,
+            batch_size=getattr(self.embedding_service, 'batch_size', 50), 
             progress_callback=embedding_progress_callback,
             cancel_check=cancel_check,
         )