"""Standalone inference for the Tiny Vial CNN Hugging Face release."""

from __future__ import annotations

import argparse
from pathlib import Path

import numpy as np
from PIL import Image, ImageDraw
import torch
from safetensors.torch import load_file
from torch import nn


IMAGE_SIZE = 64


class TinyVialDetector(nn.Module):
    def __init__(self) -> None:
        super().__init__()
        self.features = nn.Sequential(
            nn.Conv2d(3, 12, 3, stride=2, padding=1), nn.ReLU(),
            nn.Conv2d(12, 24, 3, stride=2, padding=1), nn.ReLU(),
            nn.Conv2d(24, 32, 3, stride=2, padding=1), nn.ReLU(),
        )
        self.head = nn.Sequential(
            nn.Flatten(),
            nn.Linear(32 * 8 * 8, 64), nn.ReLU(),
            nn.Linear(64, 4), nn.Sigmoid(),
        )

    def forward(self, image: torch.Tensor) -> torch.Tensor:
        return self.head(self.features(image))


def predict(image_path: Path, weights_path: Path) -> tuple[int, int, int, int]:
    model = TinyVialDetector()
    model.load_state_dict(load_file(weights_path, device="cpu"))
    model.eval()
    with Image.open(image_path) as source:
        width, height = source.size
        resized = source.convert("RGB").resize((IMAGE_SIZE, IMAGE_SIZE), Image.Resampling.BILINEAR)
    array = np.asarray(resized, dtype=np.float32).copy()
    tensor = torch.from_numpy(array).permute(2, 0, 1).unsqueeze(0) / 255.0
    with torch.inference_mode():
        cx, cy, box_width, box_height = model(tensor)[0].tolist()
    x0 = max(0, min(width - 1, round((cx - box_width / 2) * width)))
    y0 = max(0, min(height - 1, round((cy - box_height / 2) * height)))
    x1 = max(x0, min(width - 1, round((cx + box_width / 2) * width) - 1))
    y1 = max(y0, min(height - 1, round((cy + box_height / 2) * height) - 1))
    return x0, y0, x1, y1


def main() -> None:
    parser = argparse.ArgumentParser()
    parser.add_argument("image", type=Path)
    parser.add_argument("--weights", type=Path, default=Path("model.safetensors"))
    parser.add_argument("--output", type=Path, default=Path("prediction.png"))
    args = parser.parse_args()
    bbox = predict(args.image, args.weights)
    image = Image.open(args.image).convert("RGB")
    ImageDraw.Draw(image).rectangle(bbox, outline="#76b900", width=2)
    image.save(args.output)
    print({"bbox_2d_tight": list(bbox), "output": str(args.output)})


if __name__ == "__main__":
    main()