Update app.py

app.py

@@ -35,13 +35,29 @@ Never invent information. Cite sources for all facts. Use neutral, academic tone
 """
 
 def process_searches(response):
-    formatted_response = response.replace("<thinking>", "\n💭 THINKING PROCESS:\n").replace("</thinking>", "\n")
-    searches = re.findall(r'<search>(.*?)</search>', formatted_response, re.DOTALL)
+    searches = re.findall(r'<search>(.*?)</search>', response, re.DOTALL)
     if searches:
         queries = [q.strip() for q in searches[0].split('\n') if q.strip()]
         return queries
     return None
 
+def process_thinking(response, duration):
+    thinking_blocks = re.findall(r'<thinking>(.*?)</thinking>', response, re.DOTALL)
+    if not thinking_blocks:
+        return response, False
+    
+    formatted_response = response
+    for content in thinking_blocks:
+        formatted_think = (
+            f"\n💭 THINKING PROCESS:\n{content.strip()}\n"
+            f"Thought for {duration:.1f} seconds.\n"
+        )
+        formatted_response = formatted_response.replace(
+            f'<thinking>{content}</thinking>', 
+            formatted_think
+        )
+    return formatted_response, True
+
 def search_with_retry(query, max_retries=3, delay=2):
     for attempt in range(max_retries):
         try:
@@ -53,6 +69,14 @@ def search_with_retry(query, max_retries=3, delay=2):
             raise
     return None
 
+def animate_thinking():
+    dots = ["", ".", "..", "..."]
+    i = 0
+    while True:
+        yield f"Thinking{dots[i]}"
+        i = (i + 1) % 4
+        time.sleep(0.5)
+
 def respond(
     message,
     history: list[tuple[str, str]],
@@ -79,12 +103,16 @@ def respond(
     
     full_response = ""
     search_cycle = True
+    thinking_animation = animate_thinking()
     
     try:
         while search_cycle:
             search_cycle = False
+            show_thinking = False
+            thinking_start = None
             
             try:
+                start_time = time.time()
                 completion = client.chat.completions.create(
                     model="qwen/qwq-32b:free",
                     messages=messages,
@@ -102,10 +130,42 @@ def respond(
                 return
 
             response = ""
+            thinking_buffer = ""
+            in_thinking_tag = False
+            
             for chunk in completion:
                 token = chunk.choices[0].delta.content or ""
                 response += token
                 full_response += token
+
+                # Detect thinking tags
+                if not in_thinking_tag and '<thinking>' in token:
+                    in_thinking_tag = True
+                    thinking_start = time.time()
+                    show_thinking = True
+                
+                if in_thinking_tag:
+                    thinking_buffer += token
+                    if '</thinking>' in token:
+                        in_thinking_tag = False
+                        thinking_duration = time.time() - thinking_start
+                
+                # Show animated thinking if needed
+                if show_thinking and not in_thinking_tag:
+                    formatted, has_thinking = process_thinking(full_response, time.time() - start_time)
+                    if has_thinking:
+                        full_response = formatted
+                        show_thinking = False
+                        yield full_response
+                    else:
+                        yield next(thinking_animation)
+                else:
+                    yield full_response
+
+            # Process final thinking state
+            final_response, has_thinking = process_thinking(full_response, time.time() - start_time)
+            if has_thinking:
+                full_response = final_response
                 yield full_response
 
             queries = process_searches(response)