app.py

import gradio as gr
import torch
import torch.nn as nn
from torchvision import models, transforms
from PIL import Image
import numpy as np

# =======================================
# CONFIGURATION
# =======================================

DEVICE = torch.device("cpu")  # Hugging Face Free Tier = CPU only
CLASS_NAMES = ["Non-Glaucoma", "Glaucoma"]
MODEL_PATH = "model_fold_0.pth"

# =======================================
# LOAD MODEL
# =======================================

try:
    model = models.resnet18(weights=None)
    model.fc = nn.Linear(model.fc.in_features, 2)

    state_dict = torch.load(MODEL_PATH, map_location=DEVICE)
    model.load_state_dict(state_dict)

    model.to(DEVICE)
    model.eval()

except Exception as e:
    raise RuntimeError(f"Failed to load model from {MODEL_PATH}\nError: {str(e)}")

# =======================================
# IMAGE PREPROCESSING
# =======================================

transform = transforms.Compose([
    transforms.Resize((256, 256)),
    transforms.CenterCrop(224),
    transforms.ToTensor(),
    transforms.Normalize(
        mean=[0.485, 0.456, 0.406],
        std=[0.229, 0.224, 0.225]
    )
])

# =======================================
# PREDICTION FUNCTION
# =======================================

def predict_fundus(image):
    if image is None:
        return "Please upload a retinal fundus image to begin.", None

    try:
        img_pil = Image.fromarray(image).convert("RGB")
        img_tensor = transform(img_pil).unsqueeze(0).to(DEVICE)

        with torch.no_grad():
            output = model(img_tensor)
            probs = torch.softmax(output, dim=1)[0].cpu().numpy()

        pred_idx = int(np.argmax(probs))
        confidence = float(probs[pred_idx])
        label = CLASS_NAMES[pred_idx]

        result_text = f"""
### Analysis Result

**Prediction:** {label}  
**Confidence:** {confidence:.1%}

**Non-Glaucoma Probability:** {probs[0]:.1%}  
**Glaucoma Probability:** {probs[1]:.1%}

---

⚠ This tool is for research and educational purposes only.  
It must not be used for clinical diagnosis or medical decision-making.
        """.strip()

        img_display = np.array(img_pil.resize((400, 400)))

        return result_text, img_display

    except Exception as e:
        return f"Error during analysis: {str(e)}", None


# =======================================
# PROFESSIONAL HIGH-CONTRAST CSS
# =======================================

custom_css = """
body {
    font-family: 'Segoe UI', sans-serif;
    background: #ffffff;
    color: #111827;
}

.gradio-container {
    max-width: 1100px !important;
    margin: auto;
}

h1 {
    color: #1e3a8a !important;
    font-weight: 700 !important;
    text-align: center;
}

h3 {
    color: #1f2937 !important;
    font-weight: 600 !important;
}

.markdown {
    color: #111827 !important;
}

.upload-zone {
    border: 2px dashed #64748b;
    border-radius: 12px;
    padding: 20px;
    background: white;
}

.result-panel {
    background: white;
    border-radius: 12px;
    box-shadow: 0 4px 15px rgba(0,0,0,0.08);
    padding: 24px;
    min-height: 380px;
}

.note {
    font-size: 0.95em;
    color: #374151;
    margin-top: 16px;
}
"""

# =======================================
# GRADIO INTERFACE
# =======================================

with gr.Blocks(theme=gr.themes.Default(), css=custom_css) as demo:

    gr.Markdown("""
    # Glaucoma Screening – Fundus Image Analysis

    Upload a retinal fundus photograph to receive an AI-based probability assessment.
    """)

    with gr.Row(equal_height=True):

        with gr.Column(scale=5):
            gr.Markdown("### Upload Fundus Image")

            input_image = gr.Image(
                type="numpy",
                label="",
                elem_classes=["upload-zone"],
                height=480,
                image_mode="RGB"
            )

            analyze_btn = gr.Button("Analyze Image", variant="primary")

        with gr.Column(scale=5):
            gr.Markdown("### Analysis Result")

            output_text = gr.Markdown(
                value="Upload an image and click Analyze to begin.",
                elem_classes=["result-panel"]
            )

            output_image = gr.Image(
                label="Uploaded Image (Resized)",
                type="numpy",
                height=400,
                interactive=False
            )

    gr.Markdown("""
    <div class="note">
    <strong>Important:</strong> This is an experimental AI model trained on limited data.
    Results should be interpreted cautiously and verified by a qualified ophthalmologist.
    </div>
    """, elem_classes=["note"])

    analyze_btn.click(
        fn=predict_fundus,
        inputs=input_image,
        outputs=[output_text, output_image]
    )

# =======================================
# LAUNCH (HF Compatible)
# =======================================

demo.launch()