Get contents of netflix genres

get_genre_contents_nfx.py

from json import dumps
from re import compile as recompile
from requests import session as resession

ANIME_CATEGORIES = {
    'Anime': 7424,
    'Anime Action': 2653,
    'Anime Comedies': 9302,
    'Anime Dramas': 452,
    'Anime Fantasy': 11146,
    'Anime Features': 3063,
    'Anime Horror': 10695,
    'Anime Sci-Fi': 2729,
    'Anime Series': 6721
}

PAGE_ITEMS = [
    'authURL',
    'BUILD_IDENTIFIER',
    'ICHNAEA_ROOT',
    'API_ROOT',
    'API_BASE_URL',
    'gpsModel'
]

BASE_URL = 'https://www.netflix.com'

URLS = {
    'shakti': '/pathEvaluator',
    'login': '/login',
    'watch': '/watch'
}

def init_session():
    session = resession()
    session.headers.update({
        'User-Agent': 'Mozilla/5.0 (Macintosh; Intel Mac OS X 10_12_3) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/57.0.2987.133 Safari/537.36',
        'Accept-Encoding': 'gzip'
    })
    return session

def login(session, email, password):
    page = session.get(url=BASE_URL + URLS.get('login'))
    user_data = get_token_data(content=page.text, items=PAGE_ITEMS)
    login_payload = {
        'email': email,
        'password': password,
        'rememberMe': 'true',
        'flow': 'websiteSignUp',
        'mode': 'login',
        'action': 'loginAction',
        'withFields': 'email,password,rememberMe,nextPage',
        'authURL': user_data.get('authURL'),
        'nextPage': ''
    }

    # perform the login
    login_response = session.post(url=BASE_URL + URLS.get('login'), data=login_payload)
    return (session, get_token_data(content=login_response.text, items=PAGE_ITEMS))

def get_token_data(content, items):
    account_info = {}
    react_context = recompile('reactContext(.*);').findall(content)
    for item in items:
        match = recompile(
            '"' + item + '":"(.+?)"').findall(react_context[0])
        if len(match) > 0:
            account_info.update({item: match[0].decode('string_escape')})
    return account_info

def get_category_contents(session, categories, user_data):
    paths = []
    for category in categories:
        paths.append(['genres', categories.get(category), 'su', {'from': 0, 'to': 100}, ['summary', 'title']])
        paths.append(['genres', categories.get(category), 'su', {'from': 0, 'to': 100}, 'boxarts', '_1280x720', 'jpg'])

    payload = {
        'paths': paths,
        'authURL': user_data.get('authURL')
    }

    parameters = {
        'model': user_data.get('gpsModel')
    }

    headers = {
        'Content-Type': 'application/json',
        'Accept': 'application/json, text/javascript, */*',
    }

    api_url = user_data.get('API_ROOT') + user_data.get('API_BASE_URL') + '/' + user_data.get('BUILD_IDENTIFIER') + URLS.get('shakti')
    return session.post(url=api_url, data=dumps(payload), params=parameters, headers=headers)

def process_response(response, categories):
    _ret = []
    _raw_data = response.json()

    # video data
    videos = _raw_data.get('value').get('videos')
    for video_id in videos:
        _video = videos.get(video_id)
        _ret.append(dict(
            id=video_id,            
            title=_video.get('title'),
            type=_video.get('summary').get('type'),
            url=BASE_URL + URLS.get('watch') + '/' +  video_id,
            boxart=_video.get('boxarts', {}).get('_1280x720', {}).get('jpg', {}).get('url')
        ))
    return _ret

# RUN IT
EMAIL = '[email protected]'
PASSWORD = 'My_Secret_Netflix_Pass'

session, user_data = login(session=init_session(), email=EMAIL, password=PASSWORD)
response = get_category_contents(session=session, categories=ANIME_CATEGORIES, user_data=user_data)
print process_response(response=response, categories=ANIME_CATEGORIES)