Update app.py

app.py

@@ -2,7 +2,7 @@ import os
 import base64
 import requests
 import gradio as gr
-from huggingface_hub import InferenceClient
+from huggingface_hub import InferenceClient, ChatMessage
 
 class XylariaChat:
     def __init__(self):
@@ -11,9 +11,9 @@ class XylariaChat:
         if not self.hf_token:
             raise ValueError("HuggingFace token not found in environment variables")
         
-        # Initialize the inference client
+        # Initialize the inference client with the Qwen model
         self.client = InferenceClient(
-            model=os.getenv("MODEL_NAME"), 
+            model="Qwen/QwQ-32B-Preview", 
             api_key=self.hf_token
         )
         
@@ -26,16 +26,16 @@ class XylariaChat:
         self.persistent_memory = {}
         
         # System prompt with more detailed instructions
-        self.system_prompt = """You are a helpful and harmless assistant. You are Xylaria developed by Sk Md Saad Amin(india, 12 year old). You should think step-by-step. 
-"""
+        self.system_prompt = """You are a helpful and harmless assistant. You are Xylaria developed by Sk Md Saad Amin(india, 12 year old). You should think step-by-step."""
     
     def store_information(self, key, value):
         """Store important information in persistent memory"""
         self.persistent_memory[key] = value
+        return f"Stored: {key} = {value}"
 
     def retrieve_information(self, key):
         """Retrieve information from persistent memory"""
-        return self.persistent_memory.get(key)
+        return self.persistent_memory.get(key, "No information found for this key.")
 
     def reset_conversation(self):
         """
@@ -46,9 +46,8 @@ class XylariaChat:
         self.conversation_history = []
         self.persistent_memory.clear()
         
-        # Clear API-side memory by resetting the conversation
+        # Reinitialize the client
         try:
-            # Attempt to clear any API-side session or context
             self.client = InferenceClient(
                 model="Qwen/QwQ-32B-Preview", 
                 api_key=self.hf_token
@@ -101,32 +100,57 @@ class XylariaChat:
             return f"Error processing image: {str(e)}"
 
     def get_response(self, user_input, image=None):
-        # Prepare messages with conversation context and persistent memory
-        messages = [
-            {"role": "system", "content": self.system_prompt},
-        ]
-        
-        # Add persistent memory context if available
-        if self.persistent_memory:
-            memory_context = "Remembered Information:\n" + "\n".join(
-                [f"{k}: {v}" for k, v in self.persistent_memory.items()]
-            )
-            messages.append({"role": "system", "content": memory_context})
-        
-        # Add conversation history
-        messages.extend(self.conversation_history)
-        
-        # Process image if uploaded
-        if image:
-            image_caption = self.caption_image(image)
-            user_input = f"Image description: {image_caption}\n\nUser's message: {user_input}"
+        """
+        Generate a response using chat completions with improved error handling
         
-        # Add user input
-        messages.append({"role": "user", "content": user_input})
+        Args:
+            user_input (str): User's message
+            image (optional): Uploaded image
         
-        # Generate response with streaming
+        Returns:
+            Stream of chat completions or error message
+        """
         try:
+            # Prepare messages with conversation context and persistent memory
+            messages = []
+            
+            # Add system prompt as first message
+            messages.append(ChatMessage(
+                role="system", 
+                content=self.system_prompt
+            ))
+            
+            # Add persistent memory context if available
+            if self.persistent_memory:
+                memory_context = "Remembered Information:\n" + "\n".join(
+                    [f"{k}: {v}" for k, v in self.persistent_memory.items()]
+                )
+                messages.append(ChatMessage(
+                    role="system", 
+                    content=memory_context
+                ))
+            
+            # Convert existing conversation history to ChatMessage objects
+            for msg in self.conversation_history:
+                messages.append(ChatMessage(
+                    role=msg['role'], 
+                    content=msg['content']
+                ))
+            
+            # Process image if uploaded
+            if image:
+                image_caption = self.caption_image(image)
+                user_input = f"Image description: {image_caption}\n\nUser's message: {user_input}"
+            
+            # Add user input
+            messages.append(ChatMessage(
+                role="user", 
+                content=user_input
+            ))
+            
+            # Generate response with streaming
             stream = self.client.chat.completions.create(
+                model="Qwen/QwQ-32B-Preview",
                 messages=messages,
                 temperature=0.5,
                 max_tokens=10240,
@@ -137,6 +161,7 @@ class XylariaChat:
             return stream
         
         except Exception as e:
+            print(f"Detailed error in get_response: {e}")
             return f"Error generating response: {str(e)}"
 
     def create_interface(self):
@@ -153,14 +178,18 @@ class XylariaChat:
             updated_history = chat_history + [[message, ""]]
             
             # Streaming output
-            for chunk in response_stream:
-                if chunk.choices[0].delta.content:
-                    chunk_content = chunk.choices[0].delta.content
-                    full_response += chunk_content
-                    
-                    # Update the last message in chat history with partial response
-                    updated_history[-1][1] = full_response
-                    yield "", updated_history, None
+            try:
+                for chunk in response_stream:
+                    if chunk.choices and chunk.choices[0].delta and chunk.choices[0].delta.content:
+                        chunk_content = chunk.choices[0].delta.content
+                        full_response += chunk_content
+                        
+                        # Update the last message in chat history with partial response
+                        updated_history[-1][1] = full_response
+                        yield "", updated_history, None
+            except Exception as e:
+                print(f"Streaming error: {e}")
+                yield "", updated_history + [["", f"Error during response: {e}"]], None
             
             # Update conversation history
             self.conversation_history.append(
@@ -197,10 +226,10 @@ class XylariaChat:
             # Chat interface with improved styling
             with gr.Column():
                 chatbot = gr.Chatbot(
-                    label="Xylaria 1.4 Senoa",
+                    label="Xylaria 1.4 Senoa (Qwen Model)",
                     height=500,
                     show_copy_button=True,
-                    type="messages"  # Updated to use messages type
+                    type="messages"
                 )
                 
                 # Input row with improved layout and image upload