app.py

import os
import gradio as gr
import requests
from PIL import Image

# Set your API keys (replace with your actual keys or use environment variables)
nvidia_api_key = os.getenv("Vision")  # NVIDIA API Key
imagebb_api_key = os.getenv("ImageAPI")  # Imgbb API Key

# NVIDIA API Endpoint
invoke_url = "https://ai.api.nvidia.com/v1/gr/meta/llama-3.2-90b-vision-instruct/chat/completions"

def upload_image_to_imgbb(image_path):
    """Uploads an image to ImgBB and returns the URL."""
    url = f"https://api.imgbb.com/1/upload?key={imagebb_api_key}"
    with open(image_path, "rb") as image_file:
        response = requests.post(url, files={"image": image_file})
    if response.status_code == 200:
        return response.json()["data"]["url"]
    else:
        raise ValueError(f"Image upload failed: {response.json()}")

def analyze_image(image, instruction):
    """Analyzes the image using NVIDIA’s Llama 3.2 Vision Instruct model based on the provided instruction."""
    try:
        # Save image locally
        image_path = "uploaded_image.png"
        image.save(image_path)

        # Upload image to ImgBB and get the URL
        image_url = upload_image_to_imgbb(image_path)

        # NVIDIA API Request
        headers = {
            "Authorization": f"Bearer {nvidia_api_key}",
            "Accept": "application/json"
        }
        payload = {
            "model": "meta/llama-3.2-90b-vision-instruct",
            "messages": [
                {
                    "role": "user",
                    "content": [
                        {"type": "text", "text": instruction},
                        {"type": "image_url", "image_url": {"url": image_url}}
                    ]
                }
            ],
            "max_tokens": 512,
            "temperature": 0.1,
            "top_p": 0.1
        }

        response = requests.post(invoke_url, headers=headers, json=payload)
        response_data = response.json()

        # Extract the response
        if "choices" in response_data:
            return response_data["choices"][0]["message"]["content"]
        else:
            return f"Error in response: {response_data}"

    except Exception as e:
        return f"Error: {str(e)}"

# Gradio interface
iface = gr.Interface(
    fn=analyze_image,
    inputs=[
        gr.Image(type="pil", label="Upload Image"),
        gr.Textbox(label="Instruction", placeholder="Enter your analysis instruction here.", lines=2)
    ],
    outputs="text",
    title="Deep Image Analysis using LLM",
    description=(
        "Upload an image and provide instructions to analyze it using Llama 3.2 90B Vision. "
        "You can upload and analyze multiple pictures, but one at a time."
    ),
    live=False,
)

# Launch the app
iface.launch()