import argparse

import pandas as pd
import skimage.color
import skimage.io


def get_pixel_values(im, pixel_table, white_obj, threshold, offset=[0, 0]):
    data = skimage.io.imread(im)
    if len(data.shape) == 3 and data.shape[-1] > 1:
        data = skimage.color.rgb2grey(data)
    x = []
    y = []
    img_height = data.shape[0]
    img_width = data.shape[1]
    for j in range(img_width):
        for i in range(img_height):
            if not white_obj:
                if data[i, j] <= threshold:
                    x.append(i + offset[0])
                    y.append(j + offset[1])
            elif data[i, j] >= threshold:
                x.append(i + offset[0])
                y.append(j + offset[1])

    df = pd.DataFrame()
    df['x'] = x
    df['y'] = y
    df.to_csv(pixel_table, sep="\t", index=False)


if __name__ == "__main__":
    parser = argparse.ArgumentParser(description="Create a csv table with Coordinates of the ROI")
    parser.add_argument("im", help="Paste path to out.png (output created by transformation)")
    parser.add_argument("pixel_table", help="Paste path to file in which list with all pixles > threshold should be saved")
    parser.add_argument("--white_obj", dest="white_obj", default=False, help="If set objects in image are white otherwise black", action="store_true")
    parser.add_argument("--threshold", dest="threshold", default=0.5, help="Enter desired threshold value", type=float)

    args = parser.parse_args()
    get_pixel_values(args.im, args.pixel_table, args.white_obj, args.threshold)