Create app.py

app.py

@@ -0,0 +1,80 @@
+import os
+import time
+import gradio as gr
+import google.generativeai as genai
+from google.generativeai.types import Part
+
+# Configure Gemini API key
+genai.configure(api_key=os.environ["GEMINI_API_KEY"])
+
+def upload_to_gemini(filepath, mime_type):
+    """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 Part.from_uri(file.uri, mime_type=mime_type)
+
+def wait_for_files_active(files):
+    """Waits until all uploaded files are processed and active."""
+    print("Waiting for files to become active...")
+    for file in files:
+        while True:
+            status = genai.get_file(file.name)
+            if status.state.name == "ACTIVE":
+                break
+            elif status.state.name == "PROCESSING":
+                time.sleep(5)
+            else:
+                raise Exception(f"File {file.name} failed to process.")
+    print("All files are active.")
+
+def process_inputs(image, video, user_prompt):
+    """Processes the inputs and generates a response using Gemini API."""
+    parts = []
+
+    # Convert files to Part objects with mime types
+    if image:
+        image_part = upload_to_gemini(image, mime_type="image/png")
+        parts.append(image_part)
+
+    if video:
+        video_part = upload_to_gemini(video, mime_type="video/mp4")
+        parts.append(video_part)
+
+    # Add user prompt as a text Part
+    if user_prompt:
+        text_part = Part(text=user_prompt)
+        parts.append(text_part)
+
+    # Create the `contents` list
+    contents = parts
+
+    # Ensure files are processed
+    wait_for_files_active([p for p in parts if hasattr(p, "uri")])
+
+    # Initialize Gemini model and generate content
+    model = genai.GenerativeModel("gemini-2.0-flash-exp")
+    response = model.generate_content(contents=contents)
+
+    return response.text
+
+# Gradio Interface Setup
+with gr.Blocks() as demo:
+    gr.Markdown("## Multimodal Chat with Gemini API")
+
+    with gr.Row():
+        video_input = gr.Video(label="Upload a Video")
+        image_input = gr.Image(label="Upload an Image")
+
+    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)
+
+    submit_button.click(
+        fn=process_inputs,
+        inputs=[image_input, video_input, user_prompt],
+        outputs=output
+    )
+
+# Run the Gradio App
+demo.launch()