moonlight_window.py

from PIL import Image, ImageDraw, ImageFont
import numpy as np
import os


# config
pix_per_char = 300

font_support_dict = {
    "simhei": "simhei.ttf",
    "Deng": "Deng.ttf",
    "simsun": "simsun.ttc",
    "simkai": "simkai.ttf",
}
bg_color = (255, 255, 255)      # white
ans_color = (0, 0, 0)           # black
char_color = (0, 0, 255)        # blue
split_color = (127, 127, 127)   # gray

os.makedirs("output", exist_ok=True)


# while True:
def moonlight_window(chars: str, font: str):
    if len(chars) >= 10:
        return f"Fail! {chars} is too long!"
        
    # default font: simhei
    font = ImageFont.truetype(font_support_dict.get(
            font, 
            font_support_dict["simhei"]
        ), pix_per_char)  # Font type

    char_num = len(chars)
    width, height = char_num * pix_per_char, pix_per_char
    image = Image.new('RGB', (width, height), ans_color)
    draw = ImageDraw.Draw(image)


    # get answer image step 1: write characters on answer color
    for idx, char in enumerate(chars):
        position = (idx * pix_per_char, 0)
        draw.text(position, char, font=font, fill=char_color)

    # get answer image step 2: make the characters bi-valued (preperation for flood)
    img_arr = np.array(image)
    mask = (img_arr==char_color).astype(np.uint8).sum(axis=-1)==3
    mask3 = np.dstack([mask.astype(np.uint8)]*3)
    img_arr_dyed = np.array(char_color).astype(np.uint8).reshape((1, 1, 3)) * mask3
    image = Image.fromarray(img_arr_dyed)
    draw = ImageDraw.Draw(image)

    # get answer image step 3: dye background with flood
    ImageDraw.floodfill(image, (0, 0), value=bg_color)

    # get answer image step 4: split characters from each other
    for idx in range(1, char_num):
        xy = (idx * pix_per_char, 0, idx * pix_per_char, pix_per_char)
        draw.line(xy, fill=split_color)
    
    # get answer image step 5: draw four sides of the image
    for xy in [
        (0,             0,                  0,              height - 1),
        (width - 1,     0,                  width - 1,      height - 1),
        (0,             0,                  width - 1,      0         ),
        (0,             height - 1,         width - 1,      height - 1),
    ]:
        draw.line(xy, fill=split_color)
    image.save(f"output/{chars}_A.png")


    # get question image step 1: dye background
    img_arr = np.array(image)
    mask = (img_arr!=ans_color).astype(np.uint8).sum(axis=-1)!=0
    mask3 = np.dstack([mask.astype(np.uint8)]*3)
    img_arr_dyed = np.array(bg_color).astype(np.uint8).reshape((1, 1, 3)) * mask3
    image = Image.fromarray(img_arr_dyed)
    draw = ImageDraw.Draw(image)

    # get question image step 2: split characters from each other
    for idx in range(1, char_num):
        xy = (idx * pix_per_char, 0, idx * pix_per_char, pix_per_char)
        draw.line(xy, fill=split_color)
    
    # get question image step 3: draw four sides of the image
    for xy in [
        (0,             0,                  0,              height - 1),
        (width - 1,     0,                  width - 1,      height - 1),
        (0,             0,                  width - 1,      0         ),
        (0,             height - 1,         width - 1,      height - 1),
    ]:
        draw.line(xy, fill=split_color)
    image.save(f"output/{chars}_Q.png")

    return "Success!"
    # return f"Success! Please refer to output/{chars}_Q.png and output/{chars}_A.png"