Update app.py

app.py

@@ -829,7 +829,9 @@ with gr.Blocks(
         cognitive_state = update_cognitive_state_from_message(message, cognitive_state)
 
         # RAG 始终基于 (预加载 Module10 + 可选上传)
-        rag_context = retrieve_relevant_chunks(message, rag_chunks or [])
+        rag_context_text, rag_used_chunks = retrieve_relevant_chunks(
+            message, rag_chunks or []
+        )
 
         start_ts = time.time()
         answer, new_history = chat_with_clare(
@@ -842,13 +844,11 @@ with gr.Blocks(
             course_outline=course_outline,
             weaknesses=weaknesses,
             cognitive_state=cognitive_state,
-            rag_context=rag_context,
+            rag_context=rag_context_text,   # 这里传文本
         )
-        end_ts = time.time()
-        latency_ms = (end_ts - start_ts) * 1000.0
-
+        
         # === 在这里附上 References ===
-        ref_text = format_references(rag_context)
+        ref_text = format_references(rag_used_chunks)
         if ref_text and new_history:
             last_user, last_assistant = new_history[-1]
             if "References (RAG context used):" not in (last_assistant or ""):
@@ -955,6 +955,13 @@ with gr.Blocks(
         resolved_lang = lang_pref
 
         start_ts = time.time()
+
+
+        quiz_ctx_text, _quiz_ctx_chunks = retrieve_relevant_chunks(
+            "Module 10 quiz", rag_chunks or []
+        )
+
+        
         answer, new_history = chat_with_clare(
             message=quiz_instruction,
             history=chat_history,
@@ -965,9 +972,7 @@ with gr.Blocks(
             course_outline=course_outline,
             weaknesses=weaknesses,
             cognitive_state=cognitive_state,
-            rag_context=retrieve_relevant_chunks(
-                "Module 10 quiz", rag_chunks or []
-            ),
+            rag_context=quiz_ctx_text,
         )
         end_ts = time.time()
         latency_ms = (end_ts - start_ts) * 1000.0