Update app.py

app.py

@@ -29,44 +29,39 @@ def bot_streaming(message, history, together_api_key, max_new_tokens=250, temper
         try:
             initialize_client(together_api_key)
         except Exception as e:
-            yield f"Error initializing client: {str(e)}"
+            yield [("Error initializing client", str(e))]
             return
 
     prompt = "You are a helpful AI assistant. Analyze the image provided (if any) and respond to the user's query or comment."
-    
+
     messages = [{"role": "system", "content": prompt}]
-    
-    # Add history to messages
+
+    # Build the conversation history
     for user_msg, assistant_msg in history:
-        if isinstance(user_msg, tuple):  # Handle image messages
-            text = user_msg[1] if len(user_msg) > 1 else ""
-            messages.append({"role": "user", "content": [
-                {"type": "text", "text": text},
-                {"type": "image_url", "image_url": {"url": f"data:image/png;base64,{encode_image(user_msg[0])}"}}
-            ]})
-        else:
-            messages.append({"role": "user", "content": [{"type": "text", "text": user_msg}]})
+        messages.append({"role": "user", "content": [{"type": "text", "text": user_msg}]})
         messages.append({"role": "assistant", "content": [{"type": "text", "text": assistant_msg}]})
-    
+
     # Prepare the current message
-    current_message = {"role": "user", "content": []}
-    
-    # Add text content
-    if isinstance(message, dict) and message.get("text"):
-        current_message["content"].append({"type": "text", "text": message["text"]})
-    elif isinstance(message, str):
-        current_message["content"].append({"type": "text", "text": message})
-    
-    # Add image content if present
-    if isinstance(message, dict) and message.get("files") and len(message["files"]) > 0:
-        image_path = message["files"][0]["path"] if isinstance(message["files"][0], dict) else message["files"][0]
-        image_base64 = encode_image(image_path)
-        current_message["content"].append({
-            "type": "image_url",
-            "image_url": {"url": f"data:image/png;base64,{image_base64}"}
-        })
-    
-    messages.append(current_message)
+    content = []
+    user_text = ""
+    if isinstance(message, dict):
+        if 'text' in message:
+            user_text = message['text']
+            content.append({"type": "text", "text": user_text})
+        if 'files' in message and len(message['files']) > 0:
+            image_path = message['files'][0]['name'] if isinstance(message['files'][0], dict) else message['files'][0]
+            image_base64 = encode_image(image_path)
+            content.append({"type": "image_url", "image_url": {"url": f"data:image/png;base64,{image_base64}"}})
+            user_text += "\n[User uploaded an image]"
+    else:
+        user_text = message
+        content.append({"type": "text", "text": user_text})
+
+    messages.append({"role": "user", "content": content})
+
+    # Update the history with the new user message (with empty assistant response)
+    history = history + [[user_text, ""]]
+    yield history
 
     try:
         stream = client.chat.completions.create(
@@ -80,16 +75,20 @@ def bot_streaming(message, history, together_api_key, max_new_tokens=250, temper
         response = ""
         for chunk in stream:
             response += chunk.choices[0].delta.content or ""
-            yield response
+            history[-1][1] = response
+            yield history
 
         if not response:
-            yield "No response generated. Please try again."
+            history[-1][1] = "No response generated. Please try again."
+            yield history
 
     except Exception as e:
         if "Request Entity Too Large" in str(e):
-            yield "The image is too large. Please try with a smaller image or compress the existing one."
+            history[-1][1] = "The image is too large. Please try with a smaller image or compress the existing one."
+            yield history
         else:
-            yield f"An error occurred: {str(e)}"
+            history[-1][1] = f"An error occurred: {str(e)}"
+            yield history
 
 with gr.Blocks() as demo:
     gr.Markdown("# Meta Llama-3.2-11B-Vision-Instruct (FREE)")
@@ -122,8 +121,12 @@ with gr.Blocks() as demo:
     msg = gr.MultimodalTextbox(label="Enter text or upload an image")
     clear = gr.Button("Clear")
 
-    msg.submit(bot_streaming, [msg, chatbot, together_api_key, max_new_tokens, temperature], chatbot)
-    clear.click(lambda: None, None, chatbot, queue=False)
+    msg.submit(
+        bot_streaming, 
+        inputs=[msg, chatbot, together_api_key, max_new_tokens, temperature], 
+        outputs=chatbot
+    )
+    clear.click(lambda: [], None, chatbot, queue=False)
 
 if __name__ == "__main__":
-    demo.launch(debug=True)
+    demo.launch(debug=True)