Update app.py

app.py

@@ -41,10 +41,6 @@ def wikipedia_search(query: str) -> str:
         wikipedia.set_lang("en")
         summary = wikipedia.summary(query, sentences=3, auto_suggest=True)
         return f"Wikipedia: {summary}"
-    except wikipedia.exceptions.DisambiguationError as e:
-        return f"Wikipedia: Multiple results found. Options: {', '.join(e.options[:5])}"
-    except wikipedia.exceptions.PageError:
-        return f"Wikipedia: No page found for '{query}'."
     except Exception as e:
         return f"Wikipedia error: {str(e)}"
 
@@ -123,7 +119,7 @@ def download_and_load_model(progress=gr.Progress()):
         progress(0.95, desc="Finalizing...")
         model_loaded = True
         progress(1.0, desc="Model loaded!")
-        return f"Model loaded successfully!"
+        return "Model loaded successfully!"
     except Exception as e:
         return f"Error: {str(e)}"
 
@@ -151,96 +147,35 @@ def call_llm(messages: List[Dict], max_tokens: int = 500) -> str:
     except Exception as e:
         return f"Error: {str(e)}"
 
-def chat_think_only(message, history, mode):
-    """Think-Only mode: Pure reasoning without tools"""
-    
-    system_prompt = """You are a helpful AI assistant that thinks step-by-step. For each question:
-1. Think through the problem carefully
-2. Show your reasoning process
-3. Provide a clear answer
-
-Always respond conversationally and naturally."""
-    
-    messages = [{"role": "system", "content": system_prompt}]
-    
-    for user_msg, assistant_msg in history:
-        messages.append({"role": "user", "content": user_msg})
-        messages.append({"role": "assistant", "content": assistant_msg})
-    
-    messages.append({"role": "user", "content": message})
-    
-    response = call_llm(messages, max_tokens=800)
-    return response
-
-def chat_act_only(message, history, mode):
-    """Act-Only mode: Use tools to answer"""
+def chat_function(message, history, mode):
+    """Main chat function"""
     
     if not model_loaded:
-        return "Error: Model not loaded."
+        return history + [[message, "Error: Model not loaded. Please click 'Load Model' button first."]]
     
-    system_prompt = f"""You are a helpful AI assistant with access to tools. Use tools to find accurate information.
+    if mode == "Think-Only":
+        system_prompt = "You are a helpful AI assistant. Think step-by-step and provide clear answers."
+        
+    elif mode == "Act-Only":
+        system_prompt = f"""You are a helpful AI assistant with tools.
 
 Available tools:
 {get_tool_descriptions()}
 
-When you need to use a tool, respond with:
+Use tools when needed. Format:
 Action: tool_name
-Action Input: your input
-
-After getting results, provide a natural answer to the user."""
-    
-    messages = [{"role": "system", "content": system_prompt}]
-    
-    for user_msg, assistant_msg in history:
-        messages.append({"role": "user", "content": user_msg})
-        messages.append({"role": "assistant", "content": assistant_msg})
-    
-    messages.append({"role": "user", "content": message})
-    
-    max_iterations = 3
-    response_parts = []
-    
-    for iteration in range(max_iterations):
-        response = call_llm(messages, max_tokens=400)
+Action Input: input"""
         
-        action_name, action_input = parse_action(response)
-        
-        if action_name and action_input:
-            response_parts.append(f"🔧 Using tool: **{action_name}**")
-            observation = call_tool(action_name, action_input)
-            response_parts.append(f"📊 Result: {observation}\n")
-            
-            messages.append({"role": "assistant", "content": response})
-            messages.append({"role": "user", "content": f"Tool result: {observation}\n\nNow provide a natural answer to the user based on this information."})
-        else:
-            response_parts.append(response)
-            break
-    
-    return "\n\n".join(response_parts)
-
-def chat_react(message, history, mode):
-    """ReAct mode: Interleaved thinking and tool use"""
-    
-    if not model_loaded:
-        return "Error: Model not loaded."
-    
-    system_prompt = f"""You are a helpful AI assistant that thinks and uses tools. 
+    else:
+        system_prompt = f"""You are a helpful AI assistant with tools.
 
 Available tools:
 {get_tool_descriptions()}
 
-For each question:
-1. Think about what you need to do
-2. Use tools when you need information
-3. Think about the results
-4. Provide a clear answer
-
-Format for tool use:
-Thought: [your reasoning]
+Think step-by-step and use tools when needed. Format:
+Thought: [reasoning]
 Action: tool_name
-Action Input: input
-
-After tool results, think again and either use another tool or provide the final answer."""
+Action Input: input"""
     
     messages = [{"role": "system", "content": system_prompt}]
     
@@ -250,102 +185,81 @@ After tool results, think again and either use another tool or provide the final
     
     messages.append({"role": "user", "content": message})
     
-    max_iterations = 3
     response_parts = []
+    max_iterations = 3
     
     for iteration in range(max_iterations):
         response = call_llm(messages, max_tokens=400)
         
-        thoughts = re.findall(r'Thought:\s*(.+?)(?=\n|Action:|$)', response, re.IGNORECASE | re.DOTALL)
-        for thought in thoughts:
-            response_parts.append(f"💭 **Thought:** {thought.strip()}")
+        if mode == "ReAct":
+            thoughts = re.findall(r'Thought:\s*(.+?)(?=\nAction:|$)', response, re.IGNORECASE | re.DOTALL)
+            for thought in thoughts:
+                response_parts.append(f"💭 {thought.strip()}")
         
         action_name, action_input = parse_action(response)
         
-        if action_name and action_input:
-            response_parts.append(f"🔧 **Action:** {action_name} with input: {action_input}")
+        if action_name and action_input and mode in ["Act-Only", "ReAct"]:
+            response_parts.append(f"🔧 Using: {action_name}")
             observation = call_tool(action_name, action_input)
-            response_parts.append(f"📊 **Result:** {observation}\n")
+            response_parts.append(f"📊 {observation}")
             
             messages.append({"role": "assistant", "content": response})
-            messages.append({"role": "user", "content": f"Observation: {observation}\n\nContinue thinking or provide final answer."})
+            messages.append({"role": "user", "content": f"Result: {observation}\n\nProvide final answer."})
         else:
-            if not thoughts:
-                response_parts.append(response)
+            response_parts.append(response)
             break
     
-    return "\n\n".join(response_parts)
-
-def chat_response(message, history, mode):
-    """Route to appropriate mode"""
-    if mode == "Think-Only":
-        return chat_think_only(message, history, mode)
-    elif mode == "Act-Only":
-        return chat_act_only(message, history, mode)
-    elif mode == "ReAct":
-        return chat_react(message, history, mode)
-    else:
-        return "Please select a reasoning mode."
+    final_response = "\n\n".join(response_parts)
+    return history + [[message, final_response]]
 
-with gr.Blocks(title="LLM Reasoning Chat", theme=gr.themes.Soft()) as demo:
+with gr.Blocks(title="LLM Reasoning Chat") as demo:
     gr.Markdown("""
-    # 🤖 LLM Reasoning Chat
+    # LLM Reasoning Chat
     
-    **Model:** openai/gpt-oss-20b | **Tools:** DuckDuckGo, Wikipedia, Weather, Calculator, Python
+    **Model:** openai/gpt-oss-20b
     
-    Select a reasoning mode and start chatting!
+    **Modes:**
+    - **Think-Only**: Pure reasoning
+    - **Act-Only**: Uses tools
+    - **ReAct**: Thinks and uses tools
     """)
     
     with gr.Row():
-        load_btn = gr.Button("🚀 Load Model", variant="primary", size="lg")
-        status = gr.Textbox(label="Status", value="Click 'Load Model' to start", interactive=False, scale=2)
+        load_btn = gr.Button("Load Model", variant="primary", size="lg")
+        status = gr.Textbox(label="Status", value="Click 'Load Model'", interactive=False)
     
-    with gr.Row():
-        mode_selector = gr.Radio(
-            choices=["Think-Only", "Act-Only", "ReAct"],
-            value="ReAct",
-            label="Reasoning Mode",
-            info="Think-Only: Pure reasoning | Act-Only: Tools only | ReAct: Think + Tools"
-        )
-    
-    chatbot = gr.Chatbot(
-        label="Chat",
-        height=500,
-        show_label=True,
-        avatar_images=(None, "https://huggingface.co/datasets/huggingface/brand-assets/resolve/main/hf-logo.png")
+    mode_selector = gr.Radio(
+        choices=["Think-Only", "Act-Only", "ReAct"],
+        value="ReAct",
+        label="Reasoning Mode"
     )
     
+    chatbot = gr.Chatbot(label="Chat", height=400)
+    
     with gr.Row():
-        msg = gr.Textbox(
-            label="Message",
-            placeholder="Ask me anything...",
-            scale=4
-        )
+        msg = gr.Textbox(label="Message", placeholder="Ask anything...", scale=4)
         submit = gr.Button("Send", variant="primary", scale=1)
     
+    clear = gr.Button("Clear Chat")
+    
     gr.Examples(
         examples=[
-            "What is the capital of France and what's the weather there?",
-            "Who wrote 'To Kill a Mockingbird' and when was it published?",
-            "Calculate the compound interest on $1000 at 5% for 3 years",
-            "What is the population of Tokyo?",
-            "Explain quantum mechanics in simple terms",
+            "What is the capital of France and its weather?",
+            "Calculate: 1000 * (1.05 ** 3)",
+            "Who wrote To Kill a Mockingbird?",
         ],
-        inputs=msg,
-        label="Example Questions"
+        inputs=msg
     )
     
-    clear = gr.Button("🗑️ Clear Chat")
-    
     load_btn.click(fn=download_and_load_model, outputs=status)
     
-    msg.submit(fn=chat_response, inputs=[msg, chatbot, mode_selector], outputs=chatbot)
-    submit.click(fn=chat_response, inputs=[msg, chatbot, mode_selector], outputs=chatbot)
+    submit.click(fn=chat_function, inputs=[msg, chatbot, mode_selector], outputs=chatbot)
+    msg.submit(fn=chat_function, inputs=[msg, chatbot, mode_selector], outputs=chatbot)
     
-    msg.submit(lambda: "", None, msg)
-    submit.click(lambda: "", None, msg)
+    submit.click(lambda: "", outputs=msg)
+    msg.submit(lambda: "", outputs=msg)
     
-    clear.click(lambda: None, None, chatbot)
+    clear.click(lambda: [], outputs=chatbot)
 
 if __name__ == "__main__":
     demo.launch(share=True)