"""Brown-spot area (%) of banana photos, the Python version of the ImageJ threshold method.

    python experiments/browning_area.py photos/ --out data/E5/browning.csv --preview

Take the photos on a plain white or blue background with a fixed light and camera
height. For each image:
    banana pixels = saturated enough and not background-bright (HSV)
    brown pixels  = banana pixels with hue 5-40 deg and low brightness
    brown_pct     = brown pixels / banana pixels x 100
Check the thresholds on a few photos with --preview (writes *_mask.png next to the CSV)
and adjust them before measuring every photo; then keep them fixed for the whole study.
"""
from __future__ import annotations

import argparse
import sys
from pathlib import Path

import numpy as np
import pandas as pd
from PIL import Image

sys.path.insert(0, str(Path(__file__).resolve().parents[1]))

SATURATION_MIN: float = 0.25   # below this it is background or glare
BROWN_HUE: tuple[float, float] = (5.0, 40.0)   # degrees
BROWN_VALUE_MAX: float = 0.55  # brown is darker than yellow peel


def masks(path: Path) -> tuple[np.ndarray, np.ndarray]:
    hsv = np.asarray(Image.open(path).convert("RGB").convert("HSV"), dtype=float) / 255.0
    hue, saturation, value = hsv[..., 0] * 360, hsv[..., 1], hsv[..., 2]
    banana = (saturation > SATURATION_MIN) & (hue < 90)  # yellow, green-yellow and brown peel
    brown = banana & (hue >= BROWN_HUE[0]) & (hue <= BROWN_HUE[1]) & (value < BROWN_VALUE_MAX)
    banana |= (value < 0.25) & (saturation > 0.1)  # very dark rotten patches still count as banana
    brown |= (value < 0.25) & (saturation > 0.1)
    return banana, brown


def main() -> None:
    parser = argparse.ArgumentParser(description=__doc__, formatter_class=argparse.RawDescriptionHelpFormatter)
    parser.add_argument("folder", type=Path)
    parser.add_argument("--out", type=Path, default=Path("browning.csv"))
    parser.add_argument("--preview", action="store_true")
    arguments = parser.parse_args()
    arguments.out.parent.mkdir(parents=True, exist_ok=True)
    rows = []
    for path in sorted(p for p in arguments.folder.iterdir() if p.suffix.lower() in (".jpg", ".jpeg", ".png")):
        banana, brown = masks(path)
        percent = 100 * brown.sum() / max(banana.sum(), 1)
        rows.append({"photo": path.name, "banana_pixels": int(banana.sum()), "brown_pct": round(percent, 1)})
        if arguments.preview:
            preview = np.zeros(banana.shape + (3,), dtype=np.uint8)
            preview[banana] = (240, 210, 60)
            preview[brown] = (110, 60, 20)
            Image.fromarray(preview).save(arguments.out.parent / f"{path.stem}_mask.png")
    table = pd.DataFrame(rows)
    table.to_csv(arguments.out, index=False)
    print(table.to_string(index=False))


if __name__ == "__main__":
    main()