Update app.py

app.py

@@ -11,15 +11,20 @@ load_dotenv()
 genai.configure(api_key=os.getenv("GOOGLE_API_KEY"))
 
 # Function to process the image and get response from Gemini model
-def get_gemini_response(input_prompt, uploaded_file, query):
+def get_gemini_response(input_prompt, uploaded_file_path, query):
     try:
-        # Check if the image is provided
-        if uploaded_file is None:
-            return "Please upload an image."
-        
-        # Prepare the image as bytes
-        image_parts = [{"mime_type": "image/png", "data": uploaded_file}]
-        
+        # Validate the image file path
+        if not uploaded_file_path or not os.path.exists(uploaded_file_path):
+            return "Please upload a valid image."
+
+        # Read the image file as binary data
+        with open(uploaded_file_path, "rb") as f:
+            image_data = f.read()
+
+        # Prepare the image parts for the model
+        mime_type = f"image/{uploaded_file_path.split('.')[-1]}"  # Dynamically detect mime type
+        image_parts = [{"mime_type": mime_type, "data": image_data}]
+
         # Load the Gemini model and get the response
         model = genai.GenerativeModel("gemini-pro-vision")
         response = model.generate_content([input_prompt, image_parts[0], query])
@@ -42,9 +47,9 @@ with gr.Blocks() as invoice_extractor:
         The system uses Google's Gemini model to extract and interpret the invoice details.
         """
     )
-    
+
     input_prompt = gr.Textbox(label="Input Prompt", value=default_prompt, lines=3)
-    image_input = gr.Image(label="Upload Invoice Image", type="binary")
+    image_input = gr.Image(label="Upload Invoice Image", type="filepath")  # Use type="filepath"
     query_input = gr.Textbox(label="Enter your query about the invoice", placeholder="e.g., What is the total amount?")
     output_response = gr.Textbox(label="Response", lines=5)