Update app.py

app.py

@@ -13,14 +13,14 @@ def encode_image(img):
     encoded_string = base64.b64encode(buffered.getvalue()).decode("utf-8")
     return encoded_string
 
-def get_api_response(api_key, user_message, user_image):
+def get_api_response(api_key, user_inputs):
     """
     Sends the user message and image to the Hyperbolic API and retrieves the response.
     """
     if not api_key:
         return {"error": "API key is required."}
     
-    if not user_message and not user_image:
+    if not user_inputs.get("text") and not user_inputs.get("image"):
         return {"error": "Please provide a text message, an image, or both."}
     
     try:
@@ -31,15 +31,15 @@ def get_api_response(api_key, user_message, user_image):
         
         messages = []
         
-        if user_message:
+        if user_inputs.get("text"):
             messages.append({
                 "type": "text",
-                "text": user_message
+                "text": user_inputs["text"]
             })
         
-        if user_image:
+        if user_inputs.get("image"):
             # Open the uploaded image (already a PIL Image)
-            base64_img = encode_image(user_image)
+            base64_img = encode_image(user_inputs["image"])
             messages.append({
                 "type": "image_url",
                 "image_url": {"url": f"data:image/png;base64,{base64_img}"}
@@ -73,19 +73,19 @@ def get_api_response(api_key, user_message, user_image):
     except Exception as e:
         return {"error": str(e)}
 
-def chatbot_response(api_key, user_message, user_image, history):
+def chatbot_response(api_key, user_inputs, history):
     """
     Handles the chatbot interaction by updating the conversation history.
     """
+    user_text = user_inputs.get("text")
+    user_image = user_inputs.get("image")
+    
     # Append the user's message to the history
-    if user_message or user_image:
-        if user_image:
-            history.append(("User", user_message, user_image))
-        else:
-            history.append(("User", user_message, None))
+    if user_text or user_image:
+        history.append(("User", user_text, user_image))
     
     # Get the API response
-    api_result = get_api_response(api_key, user_message, user_image)
+    api_result = get_api_response(api_key, user_inputs)
     
     if "error" in api_result:
         ai_message = f"Error: {api_result['error']}"
@@ -118,19 +118,12 @@ with gr.Blocks() as demo:
     chatbot = gr.Chatbot(label="💬 Chatbot")  # Removed `.style()` method
     
     with gr.Row():
-        with gr.Column(scale=4):
-            user_text = gr.Textbox(
-                label="Your Message",
-                placeholder="Type your message here...",
-                lines=1
-            )
-        with gr.Column(scale=1):
-            user_image = gr.Image(
-                label="Upload Image",
-                type="pil",  # Changed from 'file' to 'pil'
-                # Removed the 'tool' parameter
-                interactive=True
-            )
+        chat_input = gr.MultimodalTextbox(
+            label="Your Input",
+            placeholder="Type your message and/or upload an image...",
+            file_count="multiple",  # Allows multiple files if needed
+            interactive=True
+        )
     
     send_button = gr.Button("📤 Send")
     
@@ -139,7 +132,7 @@ with gr.Blocks() as demo:
     
     send_button.click(
         fn=chatbot_response,
-        inputs=[api_key_input, user_text, user_image, state],
+        inputs=[api_key_input, chat_input, state],
         outputs=[chatbot, state]
     )