Update Gradio app with multiple files

app.py

@@ -114,13 +114,13 @@ def process_chat_message(
     
     return response
 
-def chat_fn(message: Dict[str, Any], history: List[List[Any]]) -> Tuple[str, List[List[Any]]]:
+def chat_fn(message: Dict[str, Any], history: List[Dict[str, Any]]) -> Tuple[str, List[Dict[str, Any]]]:
     """
     Main chat function that processes user input and returns response.
     
     Args:
         message: Dictionary containing text and optional files
-        history: Chat history as list of [user_msg, assistant_msg] pairs
+        history: Chat history in messages format
     
     Returns:
         Empty string and updated history
@@ -144,14 +144,11 @@ def chat_fn(message: Dict[str, Any], history: List[List[Any]]) -> Tuple[str, Lis
     
     # Convert history to format expected by model
     model_history = []
-    for user_msg, assistant_msg in history:
-        if isinstance(user_msg, dict):
-            model_history.append({"role": "user", "content": user_msg.get("text", "")})
-        elif isinstance(user_msg, str):
-            model_history.append({"role": "user", "content": user_msg})
-        
-        if assistant_msg:
-            model_history.append({"role": "assistant", "content": assistant_msg})
+    for msg in history:
+        if msg.get("role") == "user":
+            model_history.append({"role": "user", "content": msg.get("content", "")})
+        elif msg.get("role") == "assistant":
+            model_history.append({"role": "assistant", "content": msg.get("content", "")})
     
     # Get response from model
     try:
@@ -159,14 +156,15 @@ def chat_fn(message: Dict[str, Any], history: List[List[Any]]) -> Tuple[str, Lis
     except Exception as e:
         response = f"Sorry, I encountered an error: {str(e)}"
     
-    # Update history
+    # Update history with proper message format
     if image is not None:
-        # Store message with image indicator
-        user_message = {"text": text, "image": "[Image uploaded]"}
+        # Include image indicator in the content
+        user_content = f"{text}\n[Image uploaded]" if text else "[Image uploaded]"
     else:
-        user_message = text
+        user_content = text
     
-    history.append([user_message, response])
+    history.append({"role": "user", "content": user_content})
+    history.append({"role": "assistant", "content": response})
     
     return "", history
 
@@ -175,27 +173,39 @@ def retry_fn(history: List[Dict[str, Any]]) -> Tuple[str, List[Dict[str, Any]]]:
     if not history or len(history) < 2:
         return "", history
     
-    # Remove last assistant response and regenerate
-    last_user_msg = history[-2]
+    # Remove last assistant response
+    history = history[:-1]
+    
+    # Get the last user message
+    last_user_msg = history[-1] if history else None
+    if not last_user_msg:
+        return "", history
+    
+    # Remove the last user message too (we'll re-add it with new response)
     history = history[:-1]
     
     # Recreate the message dict
-    if isinstance(last_user_msg["content"], dict):
-        message = {"text": last_user_msg["content"].get("text", "")}
+    user_content = last_user_msg.get("content", "")
+    # Extract text without image indicator
+    if "[Image uploaded]" in user_content:
+        text = user_content.replace("\n[Image uploaded]", "").replace("[Image uploaded]", "")
     else:
-        message = {"text": last_user_msg["content"]}
+        text = user_content
+    
+    message = {"text": text}
     
     return chat_fn(message, history)
 
 def undo_fn(history: List[Dict[str, Any]]) -> List[Dict[str, Any]]:
-    """Undo the last message."""
-    if history:
-        return history[:-2] if len(history) >= 2 else []
-    return history
+    """Undo the last message pair (user + assistant)."""
+    if history and len(history) >= 2:
+        # Remove last user and assistant messages
+        return history[:-2]
+    return []
 
-def clear_fn() -> Tuple[None, List]:
+def clear_fn() -> Tuple[str, List]:
     """Clear the chat."""
-    return None, []
+    return "", []
 
 # Create the Gradio interface
 with gr.Blocks(theme=gr.themes.Soft(), fill_height=True) as demo: