Update app.py

app.py

@@ -162,7 +162,7 @@ def get_context(message):
     print(f"\nFinal context length: {total_chars} characters")
     return context
 
-def chat_response(message, history):
+def chat_response(message, history, model_name):
     """Chat response function for Gradio interface"""
     try:
         # Get context
@@ -215,20 +215,51 @@ Context: {context}"""
         # Get response
         response = ""
         completion = client.chat.completions.create(
-            model="llama-3.3-70b-versatile",
+            model=model_name,
             messages=messages,
             temperature=0.7,
-            max_tokens=2000,
+            max_tokens=2500,
             top_p=0.95,
             stream=True
         )
         
         print("\n=== LLM Response Start ===")
+        thinking_process = ""
+        final_response = ""
+        is_thinking = False
+        
         for chunk in completion:
             if chunk.choices[0].delta.content:
-                response += chunk.choices[0].delta.content
-                print(chunk.choices[0].delta.content, end='', flush=True)
-                yield response
+                content = chunk.choices[0].delta.content
+                print(content, end='', flush=True)
+                
+                # Check for thinking tags
+                if "<think>" in content:
+                    is_thinking = True
+                    continue
+                elif "</think>" in content:
+                    is_thinking = False
+                    # Create collapsible thinking section
+                    if thinking_process:
+                        final_response = f"""<details>
+<summary>🤔 Click here to see the 'thinking' process</summary>
+<hr>
+<div style="font-size: 0.9em;">
+<i>💭{thinking_process}</i>
+</div>
+<hr>
+</details>
+
+{final_response}"""
+                    continue
+                
+                # Append content to appropriate section
+                if is_thinking:
+                    thinking_process += content
+                else:
+                    final_response += content
+                    yield final_response
+                    
         print("\n=== LLM Response End ===\n")
 
     except Exception as e:
@@ -251,24 +282,47 @@ if __name__ == "__main__":
         demo = gr.Blocks()
         
         with demo:
-            gr.Markdown("# Website Chat Assistant")
-            gr.Markdown("Ask questions about the website.")
+            gr.Markdown("# British Antarctic Survey Website Chat Assistant 🌍")
+            gr.Markdown("Ask questions about the BAS website. This system accesses text data from 11,982 unique BAS URLs (6GB vector database) 📚")
+            # Add model selector
+            model_selector = gr.Dropdown(
+                choices=[
+                    "llama-3.3-70b-versatile",
+                    "llama-3.1-8b-instant",
+                    "mixtral-8x7b-32768",
+                    "deepseek-r1-distill-llama-70b"
+                ],
+                value="llama-3.3-70b-versatile",
+                label="Select AI Model 🤖",
+                info="Choose which AI model to use for responses - select 'deepseek-r1-distill-llama-70b' for the 'thinking' AI search"
+            )
             
             chatbot = gr.Chatbot(height=600)
-            msg = gr.Textbox(placeholder="Ask a question...", label="Your question")
-            clear = gr.Button("Clear")
+            msg = gr.Textbox(placeholder="Ask a question or select an example question... 🤔", label="Your question")
+            clear = gr.Button("Clear 🧹")
+            
+            # Add example questions
+            gr.Examples(
+                examples=[
+                    "What research stations does BAS operate in Antarctica? 🏔️",
+                    "Tell me about the RRS Sir David Attenborough 🚢",
+                    "What kind of science and research does BAS do? 🔬",
+                    "What is BAS doing about climate change? 🌡️",
+                ],
+                inputs=msg,
+            )
             
             def user(user_message, history):
                 return "", history + [[user_message, None]]
             
-            def bot(history):
+            def bot(history, model_name):
                 if history and history[-1][1] is None:
-                    for response in chat_response(history[-1][0], history[:-1]):
+                    for response in chat_response(history[-1][0], history[:-1], model_name):
                         history[-1][1] = response
                         yield history
             
             msg.submit(user, [msg, chatbot], [msg, chatbot], queue=False).then(
-                bot, chatbot, chatbot
+                bot, [chatbot, model_selector], chatbot
             )
             
             clear.click(lambda: None, None, chatbot, queue=False)