Update app.py

app.py

@@ -25,8 +25,10 @@ def preprocess_image(image: Image.Image) -> Image.Image:
     image_height = int(image.height * IMAGE_WIDTH / image.width)
     return image.resize((IMAGE_WIDTH, image_height))
 
-def user(text_prompt: str, chatbot: List[Tuple[str, str]]):
-    return "", chatbot + [[text_prompt, None]]
+def user(text_prompt: str, chatbot):
+    if chatbot is None:
+        chatbot = []
+    return "", chatbot + [{"role": "user", "content": text_prompt}, {"role": "assistant", "content": ""}]
 
 def bot(
     google_key: str,
@@ -36,24 +38,22 @@ def bot(
     stop_sequences: str,
     top_k: int,
     top_p: float,
-    chatbot: List[Tuple[str, str]]
+    chatbot: List[dict]
 ):
     google_key = google_key or GOOGLE_API_KEY
     if not google_key:
         raise ValueError("GOOGLE_API_KEY is not set. Please set it up.")
 
-    text_prompt = chatbot[-1][0].strip() if chatbot[-1][0] else None
+    text_prompt = chatbot[-2]["content"].strip() if len(chatbot) >= 2 and chatbot[-2]["content"] else None
 
     # Handle cases for text and/or image input
     if not text_prompt and not image_prompt:
-        chatbot[-1][1] = "Prompt cannot be empty. Please provide input text or an image."
+        chatbot[-1]["content"] = "Prompt cannot be empty. Please provide input text or an image."
         yield chatbot
         return
     elif image_prompt and not text_prompt:
-        # If only an image is provided
         text_prompt = "Describe the image"
     elif image_prompt and text_prompt:
-        # If both text and image are provided, combine them
         text_prompt = f"{text_prompt}. Also, analyze the provided image."
 
     # Configure the model
@@ -66,23 +66,20 @@ def bot(
         top_p=top_p,
     )
 
-    # Prepare inputs
     inputs = [text_prompt] if image_prompt is None else [text_prompt, preprocess_image(image_prompt)]
 
-    # Generate response
     try:
         response = model.generate_content(inputs, stream=True, generation_config=generation_config)
         response.resolve()
     except Exception as e:
-        chatbot[-1][1] = f"Error occurred: {str(e)}"
+        chatbot[-1]["content"] = f"Error occurred: {str(e)}"
         yield chatbot
         return
 
-    # Stream the response back to the chatbot
-    chatbot[-1][1] = ""
+    chatbot[-1]["content"] = ""
     for chunk in response:
         for i in range(0, len(chunk.text), 10):
-            chatbot[-1][1] += chunk.text[i:i + 10]
+            chatbot[-1]["content"] += chunk.text[i:i + 10]
             time.sleep(0.01)
             yield chatbot
 # Components
@@ -94,7 +91,7 @@ google_key_component = gr.Textbox(
 )
 
 image_prompt_component = gr.Image(type="pil", label="Input Image (Optional: Figure/Graph)")
-chatbot_component = gr.Chatbot(label="Chatbot",type="tuples")
+chatbot_component = gr.Chatbot(label="Chatbot")
 text_prompt_component = gr.Textbox(
     placeholder="Type your question here...",
     label="Ask",