import youtube_dl
import subprocess
from PIL import Image
import pytesseract
import cv2
import numpy as np


VIDEO_URL = 'https://www.youtube.com/watch?v=hHW1oY26kxQ'


def get_media_url(ytb_url):
    ydl = youtube_dl.YoutubeDL()
    video = ydl.extract_info(ytb_url, download=False)
    return video['formats'][-1]['url']


def get_white_mask(img, lower=np.array([150, 150, 150]), upper=np.array([255, 255, 255])):
    return cv2.inRange(img, lower, upper)


def save_text_screen(media_url):
    subprocess.call(f'ffmpeg -i "{media_url}" -vframes 1 -y tmp_screen.png', shell=True)

    image = Image.open('tmp_screen.png')
    w, h = image.size
    image.crop((0, 0, w, h-1000)).save('tmp_text_screen.png')


media_url = get_media_url(VIDEO_URL)
save_text_screen(media_url)

cv2.imwrite('tmp_mask_text_screen.png', get_white_mask(cv2.imread('tmp_text_screen.png')))

result = pytesseract.image_to_string(Image.open('tmp_mask_text_screen.png'), lang='eng')
print(result)