nick_web.py

import json
import os
from re import search

import requests
import youtube_dl
from bs4 import BeautifulSoup

subtitle_type = "ttml"  # ttml | vtt | both
domain = ""
output_dir = ""


class Episode(object):
    def __init__(self, MGID):
        self._download_item(self.get_data(MGID), self.save_information())

    def get_data(self, mgid):
        data = requests.get(
            "https://media.mtvnservices.com/pmt/e1/access/index.html?uri=mgid:arc:episode:nick.intl:{}&configtype=edge".format(
                mgid), headers={"Referer": link}).json()
        self.title = data["feed"]["title"].replace("/", " ")
        self.description = data["feed"]['description']
        self.image_url = data["feed"]["image"]["url"]
        self.airDate = data["feed"]["items"][0]['airDate']
        self.duration = data["feed"]["items"][0]['duration']
        self.show_title = data["feed"]["items"][0]['title']
        self.episode_number = data["feed"]["items"][0]['group']["category"]['episodeN'][-2:]
        self.season_number = data["feed"]["items"][0]['group']["category"]['seasonN']
        if len(self.season_number) < 2:
            self.season_number = "0" + self.season_number
        self.linearPubDate = data["feed"]["items"][0]['group']["category"]['linearPubDate']
        self.language = data["feed"]["items"][0]['group']["category"]['language']
        self.country = data["feed"]["items"][0]['group']["category"]['countryName']

        return str(data["feed"]["items"][0]["group"]["content"]).replace("&device={device}",
                                                                         "") + "&format=json&acceptMethods=hls"

    def save_information(self):
        dirname = os.path.join(output_dir, self.show_title + " " + self.country, "Season " + self.season_number)
        if not os.path.isdir(dirname):
            os.makedirs(dirname)
        title = "{show} - S{season}E{episode} - {title}".format(show=self.show_title, season=self.season_number,
                                                                episode=self.episode_number, title=self.title)
        return os.path.join(dirname, title)

    @staticmethod
    def _download_item(url, output):
        item = requests.get(url, params={
            "acceptMethods": "hls",
            "format": "json",
        }).json()["package"]["video"]["item"][0]
        try:
            src = item["rendition"][-1]["src"]
        except KeyError:
            print("error while getting episode")
            return
        if "transcript" in item:
            if subtitle_type == "both":
                for f in range(2):
                    sub_type = ["ttml", "vtt"]
                    subtitles = next(i for i in item["transcript"][0]["typographic"] if i["format"] == sub_type[int(f)])
                    with open(f"{output}." + sub_type[f], "w", encoding="utf-8") as file:
                        file.write(requests.get(subtitles["src"]).text)
            else:
                subtitles = next(i for i in item["transcript"][0]["typographic"] if i["format"] == subtitle_type)
                with open(f"{output}." + subtitle_type, "w", encoding="utf-8") as file:
                    file.write(requests.get(subtitles["src"]).text)

        ydl_opts = {'format': 'best',
                    "outtmpl": output + ".%(ext)s",
                    "fixup": "detect_or_warn",
                    "hls_prefer_native": True,
                    }
        with youtube_dl.YoutubeDL(ydl_opts) as ydl:
            ydl.download([src])


def get_MGID():
    soup = BeautifulSoup(requests.get(link).text, "html.parser")
    return json.loads(soup.find(type="application/ld+json").contents[0])["@id"]


def Season(season_link):
    episodes = []
    soup = BeautifulSoup(requests.get(season_link).text, "html.parser")

    for episode in soup.find_all(class_="item full-ep css-q2f74n-Wrapper e19yuxbf0"):
        # print(episode.contents[0]["href"])
        episodes.append("https://" + domain + episode.contents[0]["href"])
    return episodes


Input_links = []
in_link = input("Video or Series:")
Input_links.append(in_link)

for input_link in Input_links:

    domain = search("(?:[a-z0-9](?:[a-z0-9-]{0,61}[a-z0-9])?.)+[a-z0-9][a-z0-9-]{0,61}[a-z0-9]", input_link).group(0)

    # check input list
    if "show" in input_link:
        for link in Season(input_link):
            Episode(get_MGID())
    else:
        Episode(get_MGID())