fix(llm): 修复模型评估任务中的空指针异常

- 在自动评估过程中，增加了对模型回答为空的处理
-避免了因空回答导致的后续处理异常

yudao-module-llm/yudao-module-llm-biz/src/main/java/cn/iocoder/yudao/module/llm/service/async/AsyncModelAccessManualService.java

@@ -123,7 +123,12 @@ public class AsyncModelAccessManualService {
                 ModelCompletionsRespVO modelCompletionsRespVO = modelService.modelCompletions(modelUrl,modelCompletionsReqVO);
                 log.info("auto response {}",modelCompletionsRespVO);
                 String prompt = modelCompletionsRespVO.getAnswer();
-                String res = trainHttpService.autoEvaluation(modelUrl,prompt, datasetPrompt);
+                String res = "";
+                if(prompt != null && !prompt.equals("")){
+                    res = trainHttpService.autoEvaluation(modelUrl,prompt, datasetPrompt);
+                }else {
+                    res = "bad";
+                }
 
                 ModelAssessTaskAutoInfoDO infoDO = new ModelAssessTaskAutoInfoDO();
                 infoDO.setPrompt(prompt);