import gradio as gr
import json
import torch
from torch import nn
from torchvision import models, transforms
from huggingface_hub import hf_hub_download
from PIL import Image
import requests
import os
from io import BytesIO

# Define the number of classes
num_classes = 2

# Download model from Hugging Face
def download_model():
    model_path = hf_hub_download(repo_id="jays009/Restnet50", filename="pytorch_model.bin")
    return model_path

# Load the model from Hugging Face
def load_model(model_path):
    model = models.resnet50(pretrained=False)
    model.fc = nn.Linear(model.fc.in_features, num_classes)
    model.load_state_dict(torch.load(model_path, map_location=torch.device("cpu")))
    model.eval()
    return model

# Download the model and load it
model_path = download_model()
model = load_model(model_path)

# Define the transformation for the input image
transform = transforms.Compose([
    transforms.Resize(256),
    transforms.CenterCrop(224),
    transforms.ToTensor(),
    transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225]),
])

# Function to predict from image content
def predict_from_image(image):
    # Ensure the image is a PIL Image
    if not isinstance(image, Image.Image):
        raise ValueError("Invalid image format received. Please provide a valid image.")

    # Apply transformations
    image_tensor = transform(image).unsqueeze(0)

    # Predict
    with torch.no_grad():
        outputs = model(image_tensor)
        predicted_class = torch.argmax(outputs, dim=1).item()

    # Interpret the result
    if predicted_class == 0:
        return {"result": "The photo is of fall army worm with problem ID 126."}
    elif predicted_class == 1:
        return {"result": "The photo is of a healthy maize image."}
    else:
        return {"error": "Unexpected class prediction."}

# Function to handle image from URL or file path
def predict_from_url_or_path(url=None, path=None):
    try:
        # If URL is provided, fetch and process image
        if url:
            response = requests.get(url)
            response.raise_for_status()  # Ensure the request was successful
            image = Image.open(BytesIO(response.content))
            return predict_from_image(image)
        
        # If path is provided, open the image from the local path
        elif path:
            if not os.path.exists(path):
                return {"error": f"File not found at {path}"}
            image = Image.open(path)
            return predict_from_image(image)
        else:
            return {"error": "No valid input provided."}
    except Exception as e:
        return {"error": f"Failed to process the input: {str(e)}"}

# Gradio interface
iface = gr.Interface(
    fn=lambda url, path: predict_from_url_or_path(url=url, path=path),
    inputs=[
        gr.Textbox(label="Enter Image URL", placeholder="Provide a valid image URL (optional)", optional=True),
        gr.Textbox(label="Or Enter Local Image Path", placeholder="Provide the local image path (optional)", optional=True),
    ],
    outputs=gr.JSON(label="Prediction Result"),
    live=True,
    title="Maize Anomaly Detection",
    description="Provide either an image URL or a local file path to detect anomalies in maize crops.",
)

# Launch the interface
iface.launch(share=True, show_error=True)