Update app.py

app.py

@@ -78,16 +78,16 @@ def update_vectors(files, parser):
     
     return f"Vector store updated successfully. Processed {total_chunks} chunks from {len(files)} files using {parser}."
 
-def generate_chunked_response(prompt, model, max_tokens=1000, num_calls=3, temperature=0.2):
-    print(f"Starting generate_chunked_response with {num_calls} calls")  # Debug log
+def generate_chunked_response(prompt, model, max_tokens=1000, num_calls=3, temperature=0.2, stop_clicked=False):
+    print(f"Starting generate_chunked_response with {num_calls} calls")
     client = InferenceClient(model, token=huggingface_token)
     full_responses = []
     messages = [{"role": "user", "content": prompt}]
     
     for i in range(num_calls):
-        print(f"Starting API call {i+1}")  # Debug log
-        if stop_clicked:
-            print("Stop clicked, breaking loop")  # Debug log
+        print(f"Starting API call {i+1}")
+        if stop_clicked():  # Assume stop_clicked is now a function that returns the current state
+            print("Stop clicked, breaking loop")
             break
         try:
             response = ""
@@ -97,25 +97,25 @@ def generate_chunked_response(prompt, model, max_tokens=1000, num_calls=3, tempe
                 temperature=temperature,
                 stream=True,
             ):
-                if stop_clicked:
-                    print("Stop clicked during streaming, breaking")  # Debug log
+                if stop_clicked():
+                    print("Stop clicked during streaming, breaking")
                     break
                 if message.choices and message.choices[0].delta and message.choices[0].delta.content:
                     chunk = message.choices[0].delta.content
                     response += chunk
-            print(f"API call {i+1} response: {response[:100]}...")  # Debug log
+            print(f"API call {i+1} response: {response[:100]}...")
             full_responses.append(response)
         except Exception as e:
             print(f"Error in generating response: {str(e)}")
     
     combined_response = " ".join(full_responses)
-    print(f"Combined response: {combined_response[:100]}...")  # Debug log
+    print(f"Combined response: {combined_response[:100]}...")
     
     clean_response = re.sub(r'<s>\[INST\].*?\[/INST\]\s*', '', combined_response, flags=re.DOTALL)
     clean_response = clean_response.replace("Using the following context:", "").strip()
     clean_response = clean_response.replace("Using the following context from the PDF documents:", "").strip()
     
-    print(f"Final clean response: {clean_response[:100]}...")  # Debug log
+    print(f"Final clean response: {clean_response[:100]}...")
     return clean_response
 
 def duckduckgo_search(query):
@@ -251,37 +251,37 @@ with gr.Blocks() as demo:
         clear_btn = gr.Button("Clear")
 
     def protected_generate_response(message, history, use_web_search, model, temperature, num_calls, is_generating, stop_clicked):
-        print("Starting protected_generate_response")  # Debug log
+        print("Starting protected_generate_response")
         if is_generating:
-            print("Already generating, returning")  # Debug log
+            print("Already generating, returning")
             return message, history, is_generating, stop_clicked
         is_generating = True
-        stop_clicked = False
+        stop_clicked = False  # Reset stop_clicked at the start of generation
         
         try:
-            print(f"Generating response for: {message}")  # Debug log
+            print(f"Generating response for: {message}")
             if use_web_search:
-                print("Using web search")  # Debug log
-                main_content, sources = get_response_with_search(message, model, num_calls=num_calls, temperature=temperature)
+                print("Using web search")
+                main_content, sources = get_response_with_search(message, model, num_calls=num_calls, temperature=temperature, stop_clicked=stop_clicked)
                 formatted_response = f"{main_content}\n\nSources:\n{sources}"
             else:
-                print("Using PDF search")  # Debug log
-                response = get_response_from_pdf(message, model, num_calls=num_calls, temperature=temperature)
+                print("Using PDF search")
+                response = get_response_from_pdf(message, model, num_calls=num_calls, temperature=temperature, stop_clicked=stop_clicked)
                 formatted_response = response
     
-            print(f"Generated response: {formatted_response[:100]}...")  # Debug log
+            print(f"Generated response: {formatted_response[:100]}...")
     
             if not stop_clicked:
-                print("Appending to history")  # Debug log
+                print("Appending to history")
                 history.append((message, formatted_response))
             else:
-                print("Stop clicked, not appending to history")  # Debug log
+                print("Stop clicked, not appending to history")
         except Exception as e:
             print(f"Error generating response: {str(e)}")
             history.append((message, "I'm sorry, but I encountered an error while generating the response. Please try again."))
         
         is_generating = False
-        print(f"Returning history with {len(history)} items")  # Debug log
+        print(f"Returning history with {len(history)} items")
         return "", history, is_generating, stop_clicked
 
     submit_btn.click(