Update app.py

app.py

@@ -1,19 +1,21 @@
+
 import os
 import time
 import gradio as gr
 import google.generativeai as genai
+from google.generativeai.types import Part
 
-# Configure the Gemini API key
+# Configure Gemini API key
 genai.configure(api_key=os.environ["GEMINI_API_KEY"])
 
-def upload_to_gemini(file_path, mime_type=None):
-    """Uploads a file to the Gemini API and returns the file object."""
-    file = genai.upload_file(file_path, mime_type=mime_type)
+def upload_to_gemini(filepath, mime_type=None):
+    """Uploads a file to Gemini and returns a Part object."""
+    file = genai.upload_file(filepath, mime_type=mime_type)
     print(f"Uploaded file '{file.display_name}' with URI: {file.uri}")
-    return file
+    return Part.from_uri(file.uri, mime_type=mime_type)
 
 def wait_for_files_active(files):
-    """Waits until all uploaded files are processed and active."""
+    """Waits until all files are active."""
     print("Waiting for files to become active...")
     for file in files:
         while True:
@@ -26,71 +28,46 @@ def wait_for_files_active(files):
                 raise Exception(f"File {file.name} failed to process.")
     print("All files are active.")
 
-def process_inputs(image, audio, pdf, video, user_prompt):
-    """Processes the inputs and generates a response using the Gemini API."""
-    files = []
+def process_inputs(image, video, user_prompt):
+    """Processes inputs and returns the Gemini response."""
+    parts = []
+    
     if image:
-        files.append(upload_to_gemini(image, mime_type="image/png"))
-    if audio:
-        files.append(upload_to_gemini(audio, mime_type="audio/mp4"))
-    if pdf:
-        files.append(upload_to_gemini(pdf.name, mime_type="application/pdf"))
+        parts.append(upload_to_gemini(image, mime_type="image/png"))
     if video:
-        files.append(upload_to_gemini(video, mime_type="video/mp4"))
+        parts.append(upload_to_gemini(video, mime_type="video/mp4"))
 
-    if files:
-        wait_for_files_active(files)
+    # Add the user prompt as a Part object
+    if user_prompt:
+        parts.append(Part(text=user_prompt))
 
-    # Initialize chat session with the model
-    chat_session = model.start_chat(
-        history=[
-            {
-                "role": "user",
-                "parts": [{"fileData": {"fileUri": file.uri, "mimeType": file.mime_type}} for file in files] + [{"text": user_prompt}],
-            }
-        ]
-    )
+    # Wait for files to process
+    wait_for_files_active([p for p in parts if isinstance(p, Part) and p.uri])
 
-    # Send the user's message and get a response
-    response = chat_session.send_message(user_prompt)
+    # Create a chat session and send the message
+    model = genai.GenerativeModel("gemini-2.0-flash-exp")
+    chat_session = model.start_chat()
+    response = chat_session.send_message(parts)
     return response.text
 
-# Model configuration
-generation_config = {
-    "temperature": 1,
-    "top_p": 0.95,
-    "top_k": 40,
-    "max_output_tokens": 8192,
-    "response_mime_type": "text/plain",
-}
-
-# Initialize the Gemini model
-model = genai.GenerativeModel(
-    model_name="gemini-2.0-flash-exp",
-    generation_config=generation_config,
-)
-
-# Gradio interface setup
+# Gradio interface
 with gr.Blocks() as demo:
-    gr.Markdown("## Multimodal Chat with Gemini AI")
+    gr.Markdown("## Multimodal Chat with Gemini API")
 
     with gr.Row():
-        image_input = gr.Image(label="Upload an Image")
-        audio_input = gr.Audio(label="Upload an Audio File")
-    with gr.Row():
-        pdf_input = gr.File(label="Upload a PDF")
-        video_input = gr.Video(label="Upload a Video File")
-
-    user_prompt = gr.Textbox(label="Your Prompt", placeholder="Enter your question or command here...")
-
+        video_input = gr.Video(label="Upload a Video", type="filepath")
+        image_input = gr.Image(label="Upload an Image", type="filepath")
+    
+    user_prompt = gr.Textbox(label="Your Prompt", placeholder="Enter your question here...")
+    
     submit_button = gr.Button("Submit")
-    output = gr.Textbox(label="Gemini AI Response", interactive=False)
+    output = gr.Textbox(label="Gemini Response", interactive=False)
 
     submit_button.click(
         fn=process_inputs,
-        inputs=[image_input, audio_input, pdf_input, video_input, user_prompt],
+        inputs=[image_input, video_input, user_prompt],
         outputs=output
     )
 
-# Launch the Gradio app
+# Run the app
 demo.launch()