feat；优化了记忆检索的速度和token消耗（将question提出交给planenr)

src/bw_learner/expression_selector.py

@@ -407,8 +407,8 @@ class ExpressionSelector:
             # 4. 调用LLM
             content, (reasoning_content, model_name, _) = await self.llm_model.generate_response_async(prompt=prompt)
 
-            print(prompt)
-            print(content)
+            # print(prompt)
+            # print(content)
 
             if not content:
                 logger.warning("LLM返回空结果")