app.py

import gradio as gr
from openai import OpenAI
import os
from PIL import Image
import tempfile

# Initialize the OpenAI client with OpenRouter
client = OpenAI(
    base_url="https://openrouter.ai/api/v1",
    api_key="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",
)

def process_input(text_input, image_input):
    messages = [{"role": "user", "content": []}]

    # Add text input if provided
    if text_input:
        messages[0]["content"].append({
            "type": "text",
            "text": text_input
        })

    # Add image input if provided
    if image_input is not None:
        # Save the image to a temporary file
        with tempfile.NamedTemporaryFile(delete=False, suffix=".jpg") as temp_file:
            image_path = temp_file.name
            if isinstance(image_input, str):  # If it's already a file path
                Image.open(image_input).save(image_path)
            else:  # If it's an image object from Gradio
                image_input.save(image_path)

        # Encode and add the image
        import base64
        with open(image_path, "rb") as image_file:
            base64_image = base64.b64encode(image_file.read()).decode('utf-8')

        messages[0]["content"].append({
            "type": "image_url",
            "image_url": {
                "url": f"data:image/jpeg;base64,{base64_image}"
            }
        })

        # Clean up temporary file
        os.unlink(image_path)

    # If neither text nor image is provided
    if not text_input and image_input is None:
        return "Please provide either text, an image, or both."

    try:
        # Call the API
        completion = client.chat.completions.create(
            extra_body={},
            model="meta-llama/llama-4-maverick:free",
            messages=messages
        )

        return completion.choices[0].message.content
    except Exception as e:
        return f"Error: {str(e)}"

# Create the Gradio interface
with gr.Blocks() as demo:
    gr.Markdown("# LLaMA 4 Maverick AI Assistant")
    gr.Markdown("Enter text, upload an image, or both to get a response from the AI.")

    with gr.Row():
        with gr.Column():
            text_input = gr.Textbox(label="Text Input", placeholder="Type your question or prompt here...")
            image_input = gr.Image(label="Image Input (Optional)", type="pil")
            submit_btn = gr.Button("Submit")

        with gr.Column():
            output = gr.Textbox(label="AI Response", lines=10)

    submit_btn.click(
        fn=process_input,
        inputs=[text_input, image_input],
        outputs=output
    )

# Launch the app
if __name__ == "__main__":
    demo.launch()