import gradio as gr
from typing import Any, Dict
from urllib.request import urlopen, Request
from io import BytesIO
from pathlib import Path
from PIL import Image
from functools import lru_cache
from moderators import AutoModerator


_MODEL_CACHE: Dict[str, Any] = {}

EXAMPLE_ITEMS = [
    ("https://assets.clevelandclinic.org/transform/LargeFeatureImage/cd71f4bd-81d4-45d8-a450-74df78e4477a/Apples-184940975-770x533-1_jpg", "viddexa/nsfw-detection-2-mini"),
    ("https://encrypted-tbn0.gstatic.com/images?q=tbn:ANd9GcSbRwt56NYsiHwrT8oS-igzgeEzp7p3Jbe2dw&s", "viddexa/nsfw-detection-2-mini"),
    ("https://img.freepik.com/premium-photo/portrait-beautiful-young-woman_1048944-5548042.jpg", "viddexa/nsfw-detection-2-mini"),
]

CUSTOM_CSS = """
.header {
    text-align: center;
    padding: 2rem 1rem;
    background: linear-gradient(135deg, #14b8a6 0%, #0d9488 100%);
    color: white;
    border-radius: 8px;
    margin-bottom: 2rem;
}
.header h1 { margin: 0; font-size: 2.5em; }
.header p { margin: 0.5rem 0; opacity: 0.9; }
.header a { color: white; text-decoration: underline; }

/* Hide built-in Gradio footer */
footer {
    display: none !important;
}

/* Custom footer styling */
.custom-footer {
    text-align: center;
    padding: 1rem;
    color: #888;
    margin-top: 2rem;
}
"""

EXAMPLES_DIR = Path("examples")
EXAMPLES_DIR.mkdir(exist_ok=True)


@lru_cache(maxsize=32)
def download_image(url: str) -> Image.Image:
    """Download and return PIL Image from URL."""
    req = Request(url, headers={"User-Agent": "viddexa-gradio-demo/1.0"})
    with urlopen(req, timeout=20) as resp:
        return Image.open(BytesIO(resp.read())).convert("RGB")


def prepare_examples():
    """Download and prepare example images for display."""
    examples = []
    for i, (url, model) in enumerate(EXAMPLE_ITEMS):
        img_path = EXAMPLES_DIR / f"example_{i}.jpg"
        if not img_path.exists():
            try:
                img = download_image(url)
                img.save(img_path, "JPEG", quality=95)
            except Exception as e:
                print(f"Warning: Could not download example {i}: {e}")
                continue
        examples.append([str(img_path), model])
    return examples


def load_model(model_id: str, token: str | None = None) -> Any:
    """Load model with caching."""
    if model_id not in _MODEL_CACHE:
        _MODEL_CACHE[model_id] = AutoModerator.from_pretrained(model_id, token=token, use_fast=True)
    return _MODEL_CACHE[model_id]


def analyze(image_path: str | None, image_url: str | None, model_id: str, token: str | None = None):
    """Run inference and return classification scores."""
    if not image_path and not image_url:
        raise gr.Error("Provide an image or URL")
    
    img = Image.open(image_path).convert("RGB") if image_path else download_image(image_url)
    model = load_model(model_id, token)
    results = model(img)
    
    classifications = results[0].classifications if hasattr(results[0], "classifications") else results[0]["classifications"]
    return {str(k): float(v) for k, v in (classifications.items() if isinstance(classifications, dict) else [(c["label"], c["score"]) for c in classifications])}