utils.py

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


def sort_chars(char_list, font, language):
    if language == "chinese":
        char_width, char_height = font.getsize("制")
    elif language == "korean":
        char_width, char_height = font.getsize("ㅊ")
    elif language == "japanese":
        char_width, char_height = font.getsize("あ")
    elif language in ["english", "german", "french", "spanish", "italian", "portuguese", "polish"]:
        char_width, char_height = font.getsize("A")
    elif language == "russian":
        char_width, char_height = font.getsize("A")
    num_chars = min(len(char_list), 100)
    out_width = char_width * len(char_list)
    out_height = char_height
    out_image = Image.new("L", (out_width, out_height), 255)
    draw = ImageDraw.Draw(out_image)
    draw.text((0, 0), char_list, fill=0, font=font)
    cropped_image = ImageOps.invert(out_image).getbbox()
    out_image = out_image.crop(cropped_image)
    brightness = [np.mean(np.array(out_image)[:, 10 * i:10 * (i + 1)]) for i in range(len(char_list))]
    char_list = list(char_list)
    zipped_lists = zip(brightness, char_list)
    zipped_lists = sorted(zipped_lists)
    result = ""
    counter = 0
    incremental_step = (zipped_lists[-1][0] - zipped_lists[0][0]) / num_chars
    current_value = zipped_lists[0][0]
    for value, char in zipped_lists:
        if value >= current_value:
            result += char
            counter += 1
            current_value += incremental_step
        if counter == num_chars:
            break
    if result[-1] != zipped_lists[-1][1]:
        result += zipped_lists[-1][1]
    return result


def get_data(language, mode):
    if language == "general":
        from alphabets import GENERAL as character
        font = ImageFont.truetype("fonts/DejaVuSansMono-Bold.ttf", size=20)
        sample_character = "A"
        scale = 2
    elif language == "english":
        from alphabets import ENGLISH as character
        font = ImageFont.truetype("fonts/DejaVuSansMono-Bold.ttf", size=20)
        sample_character = "A"
        scale = 2
    elif language == "german":
        from alphabets import GERMAN as character
        font = ImageFont.truetype("fonts/DejaVuSansMono-Bold.ttf", size=20)
        sample_character = "A"
        scale = 2
    elif language == "french":
        from alphabets import FRENCH as character
        font = ImageFont.truetype("fonts/DejaVuSansMono-Bold.ttf", size=20)
        sample_character = "A"
        scale = 2
    elif language == "italian":
        from alphabets import ITALIAN as character
        font = ImageFont.truetype("fonts/DejaVuSansMono-Bold.ttf", size=20)
        sample_character = "A"
        scale = 2
    elif language == "polish":
        from alphabets import POLISH as character
        font = ImageFont.truetype("fonts/DejaVuSansMono-Bold.ttf", size=20)
        sample_character = "A"
        scale = 2
    elif language == "portuguese":
        from alphabets import PORTUGUESE as character
        font = ImageFont.truetype("fonts/DejaVuSansMono-Bold.ttf", size=20)
        sample_character = "A"
        scale = 2
    elif language == "spanish":
        from alphabets import SPANISH as character
        font = ImageFont.truetype("fonts/DejaVuSansMono-Bold.ttf", size=20)
        sample_character = "A"
        scale = 2
    elif language == "russian":
        from alphabets import RUSSIAN as character
        font = ImageFont.truetype("fonts/DejaVuSansMono-Bold.ttf", size=20)
        sample_character = "Ш"
        scale = 2
    elif language == "chinese":
        from alphabets import CHINESE as character
        font = ImageFont.truetype("fonts/simsun.ttc", size=10)
        sample_character = "制"
        scale = 1
    elif language == "korean":
        from alphabets import KOREAN as character
        font = ImageFont.truetype("fonts/arial-unicode.ttf", size=10)
        sample_character = "ㅊ"
        scale = 1
    elif language == "japanese":
        from alphabets import JAPANESE as character
        font = ImageFont.truetype("fonts/arial-unicode.ttf", size=10)
        sample_character = "お"
        scale = 1
    else:
        print("Invalid language")
        return None, None, None, None
    try:
        if len(character) > 1:
            char_list = character[mode]
        else:
            char_list = character["standard"]
    except:
        print("Invalid mode for {}".format(language))
        return None, None, None, None
    if language != "general":
        char_list = sort_chars(char_list, font, language)

    return char_list, font, sample_character, scale